TURUN YLIOPISTON JULKAISUJA
ANNALES UNIVERSITATIS TURKUENSIS
SARJA - SER. B OSA - TOM. 274
HUMANIORA
REASONS FOR MOVE:
A study on trends and reasons of internal migration
With particular interest in Estonia 1989 – 2000
Mare Ainsaar
TURUN YLIOPISTO
2004
2
From the Department of Social Policy
University of Turku
Turku, Finland
Supervised by
Professor Olli Kangas, PhD
Department of Social Policy
University of Turku
Turku, Finland
and
Ismo Söderling, Ph.D
Population Research Institute
Helsinki, Finland
Reviewed by
Elli Heikkilä, PhD
Institute of Migration
Turku, Finland
and
Tiit Tammmaru, PhD
Institute of Geography
University of Tartu
Tartu, Estonia
Dissertation opponent
Professor Olli Kultalahti, PhD
Department of Regional Studies and Environmental Policy
University of Tampere
ISBN 951-29-2658-X
ISSN 0082-6987
Tartu University Press
Cover: Photo by Viktor Salmre from the collection of Estonian National Museum
2644:949
Design: Tiia Ilus
Cover design: Aita Linnas
3
TO KADI AND HANNES
4
5
Content
INTRODUCTION ........................................................................................
7
1. URBANISATION AND MIGRATION REVERSE IN
DEVELOPED COUNTRIES
Introduction and definitions ......................................................................
1.1. Macro level factors of migration ..........................................................
1.2. Macro level theories of settlement development ................................
1.3. Empirical studies of migration reverse ................................................
1.4. The case of Estonia .............................................................................
1.4.1. Urbanisation from 1900–1983..................................................
1.4.2. Migration reverse 1983–2000 ...................................................
Conclusions: General trends of population placement ...............................
19
22
25
29
36
38
42
62
2. WHY PEOPLE MOVE
Introduction ...............................................................................................
2.1. Different groups of reasons and well-being approach in migration
research ...............................................................................................
2.2. Data ....................................................................................................
2.3. Prevailing reasons for migration and deconcentration ..........................
2.4. Age and gender ....................................................................................
2.5. Human capital and self-actualisation ...................................................
2.6. Housing ...............................................................................................
2.7. Economic resources .............................................................................
2.8. Social environment ..............................................................................
2.9. Distance and information ....................................................................
Conclusions: Well-being factors in migration ..............................................
72
78
80
92
100
104
108
112
117
122
3. MIGRATION AS BEHAVIOUR
Introduction ...............................................................................................
3.1. On the search for a model ...................................................................
3.2. Appraisal and coping strategies ............................................................
3.3. Decision making ..................................................................................
3.4. Individual differences ..........................................................................
131
132
138
143
147
71
6
3.5. Interviews with people ........................................................................
3.5.1. Method......................................................................................
3.5.2. Methodological results...............................................................
3.5.3. Appraisal, needs, triggering reasons and other reasons...............
3.5.4. Search for a new place...............................................................
3.5.5. Evaluation and choice ................................................................
Conclusions: Future of behavioural theory .................................................
149
149
151
151
157
163
165
4. REGIONAL POVERTY AND FREEDOM TO MOVE
Introduction – rationality for equality ........................................................
4.1. Regional equality and equity ................................................................
4.2. Regional versus individual poverty .......................................................
4.3. Data and method .................................................................................
4.4. Regional poverty in Estonia .................................................................
4.5. Income and migration..........................................................................
Conclusions: Did it exist, freedom to move? ..............................................
173
176
178
179
183
186
190
CONCLUSIONS ........................................................................................... 195
REFERENCES ............................................................................................... 201
7
Introduction
Why was this book written?
Different theories have different opinions about the impact of migration on the
regional development. Many researchers see migration as an essential determinant in
economic, urban and demographic processes, source of growth and spatial distribution of socio-economic opportunities or “a phenomenon capable of resolving and
balancing out the conflicts between various subsystems, most particularly between
the economic and socio-political” (Tervamäki 1987). At the same time migration is
not always flexible enough to react to regional disparities. The reason is hidden in
lack of freedom of movement.
This book has two objectives: to contribute to the development of migration
theory and to present reliable empirical data on the territorial movement of people
in Estonia during the last decade of the 20th century — the period that has caused
the most confusion among researcher so far. Both objectives hold enough challenges
for migration researchers. The ten uncertain years in Estonian migration research
have brought about many speculations, stray inquiries and contradictory conclusions
relating to migration, development of the population and the affecting factors.
Many questions were in the air as I started writing this book: Does the human
population move towards more concentrated or homogenous settlement system?
What are the leading determinants of migration and regional development? Can we
predict individual migration behaviour? Are migration models different for different
groups? Is poverty a barrier for migration freedom in Estonia? This book tries to
answer these questions.
One could ask also: why to study the migration of people at all? If we exclude the
plain interest in the development of human society, migration trends definitely tell
us something about the lifestyles of people, territorial connections in the society and
eventually also about the distribution of resources. Because of involvement of human migration to almost all spheres of human life we can argue, that migration flows
can be important signs about general development of society.
Migration and social justice
Ideas of social justice have a geographical dimension. Despite different understandings
about how geographical justice or justice on space should be defined and measured,
the regional approach to well-being is particularly important because space itself
seems to have an influence on people’s abilities and equality. Migration is one primary mechanism used to achieve equilibrium in society (Morrill 1970). The fortune
8
or misfortune of being a citizen of region with low well-being remains important
social problem. Migration and public policies have the power to change regional disparities. Still, this book doesn’t analyse regional equality generally, but concentrates
its attention into questions of migration as a result and cause of regional disparities.
We assume in this book that voluntary migration is a movement towards higher personal well-being. In that way migration fulfils the role of voting with the feet (Fidlay
and Rogerson 1993) and has an equalising effect on regional supply-demand mismatches. Migration can be a source of increasing personal well-being as well as a
source of inequality due to its selective openness to various groups. Beside of being a
tool of access for resources in society, freedom for migration itself is a resource. In
the contexts of poverty appraoch we can argue that the lack of freedom to move can
be labelled as “mobility poverty”. In migration, like in society generally, stands the
notion that “under circumstances of grinding poverty, people have little freedom for
manoeuvre” (Goodin et al 1999). Thereby poverty doesn’t need to appear in form of
economic shortages, but can exist also in a form of deficit of human or social capital
and freedom to move.
Voluntary migration requires always at least some degree of freedom of movement. Freedom is considered to be an ability to decide, to choose; at the same time
it can be the absence of necessities, coercions or constraints. Often distinction between negative (from) and positive (ability to) freedoms is made. The both types of
freedom are appropriate in migration studies. Here and later in this book we employ
the concept of positive freedom to move. It is based on argument that indepentently
from negative freedom, not to move, exists positive freedom to move. Reasons causing migration restrictions are marked in migration studies as barriers and lack of
them can be interpreted in terms of negative freedom. Migration barriers are additional sources of inequality, as people will have different access to resources in society. All reasons affecting migration can become also barriers of movement. Therefore, migration is an essential social indicator of positive freedom to improve one’s
well-being.
What is migration?
There are plenty of different definitions of migration, guided by the diversity of
approaches. The most simple and common approach to migration sees it as a permanent change of residence with crossing some administrative borders. Still, border
crossing is a rather statistical view on migration and in a more scientific approach we
could take into account all changes of living place. However, the concept of presence
seems to change over the time. The change of permanent home, used most commonly by researchers, might not be valid nowadays anymore to describe human destinations in space. Time spent in different areas, or access to services can be even
more appropriate. For example, Mulder and Hooimeijer (1999) regard also travel,
relocations of workplace and consumption as spatial mobility. According to this approach migration can be defined as multiple relocation.
When considering the goals of behaviour, migration can be viewed as a tool in
achieving some objectives in life careers (Mulder 1993). The notion that migration is
9
an instrumental behaviour in the course of life leads to certain expectations about
migration dynamics within and between cohorts. Cohort differences in migration
can appear as a consequence of changes in the timing of events in one or more parallel careers. According to rational approaches voluntary migration takes place when
the perceived relative advantage of the new location exceeds the costs of leaving
previous daily activity place. The guiding approach of migration in this book is
that migration is an adaptation decision to the lack of well-being, performed in
a situation of restricted information.
Economists see migration as a process of relocating labour force and equalising
economic needs with demand. Migration is influenced by several micro and macro
level factors. Often people move to a place with better prospects for living, working,
studying or raising children. Thus the migratory flows could be regarded as indicators of social welfare. Surveys have shown that migration is selective in nature, offering better opportunities and greater freedom of choice to more active people with
better human, social and economic capital. In addition, several regional factors like
territorial differences of demographic, socio-economic and social mobility conditions shape migration. Therefore, migration trends indicate indirectly the mismatches
of individual and regional level well-being.
Content of the book
Migration is an outcome of interaction between environment, individual needs and
subjective decision-making process. Three main levels of analyses can be distinguished
in migration research: (1) Macro level (classical push and pull factors, barriers, intervening opportunities, territorial differences, etc); (2) Aggregated individual data level
(demographic groups, groups by economic status, etc); (3) Individual level (behavioural approach) (Figure 1).
This book presents different level approaches to migration between the same
covers. The unifying element of different chapters is the empirical time-space scope,
namely migration processes in Estonia in 1990s. As all migration can be described as
an interaction between population and environment, individual level analyses concentrate mainly on behaviour of people, aggregated individual level analyMacro level
ses on behaviour of groups of people
regional push or pull forces
and macro level analyses on differences
on space which can be incentives for
migration. If the macro level approach
Aggregated individual data
tries to find the spatial regulators of migration patterns, then micro approach
explains differences in migration by investigating the background and reasons
Individual behaviour
of individual migration decisions. The
theoretical model determines the nature of data used in analyses. Thus, the Figure 1 Different levels of migration
psychological or sociological approach research
10
requires individual level data, while the classical economic or demographic approach
uses primarily combined regional indicators. Macro and micro level studies are in
essence complementary, and when used simultaneously, provide the best overview
of the factors that affect migration. The author of this book is convinced that behavioural or psychological approaches hold better opportunities for explaining migration
in all its diversity. On the third chapter this model is developed and checked with
some individual level empirical data. Basic elements from behavioural approach have
influenced the text in all chapters of this book.
While in developed and less developed countries the reasons and mechanisms of
internal migration can be rather different we focus our interest only on economically
more advanced countries. Another great divide is that between voluntary and forced
migration. Here we have restricted our main interest to voluntary migration.
The first chapter describes changes in Estonia on the framework of other developed
countries and analyses the background factors of migration on macro level. The aim
of the first chapter is to give an overview of general trends of urbanisation in the
developed world and reasons of growth and decline of urban population. Second half
of the first chapter follows the changes in one country — in Estonia. The aim of the
country study is to provide comprehensive analyses throughout different periods of
migration history in one country. Still the main emphasis is on period 1990–2000, as
most confusing and less studied period. Chapter comes to the conclusion that, although usually the settlement of population has been seen as outcome of production
forces, this may no longer be the case in situation where the means of transportation
have improved and people have more freedom of movement.
The second chapter analyses different classifications of reasons of migration and
seeks the impact mechanism of the main factors of migration on aggregated individual level. It tries to find answers to the questions why and how migration flows
are created and how are different social groups involved to migration. These answers
are important in order to analyse different levels of freedom of migration. Every
factor having an impact to migration is a potential barrier of migration in other situation. Three different approaches — well-being approach, theory of needs and classical migration reasons are used to elaborate comprehensive classification of migration components. Latter, migration streams in Estonia and other countries are analysed according to this classification. Classification consists of seven factors — demographic composition, human capital, dwelling, income, social resources, environment and distance. Discussion about the main reason of migration at certain time
periods is developed.
The behavioural explanations in migration decision-making are the least studied area
in migration research. The aim of the third chapter is to develop a behavioural model
of migration decision-making and to evaluate its validity with empirical data.
The fourth chapter analyses the freedom to move on regional level. Regional poverty
is studied as indicator of equality or inequality in society. We assume that an inequality of outcome is more legitimate if people have more equal opportunities to make
11
choices and less legitimate in a situation of missing choices. This principle is valid
also for territorial mobility. The fourth chapter investigates links between regional
and individual poverty and freedom to move in Estonia in the end of 1990ies. Our
particular interest is on barriers for movement, caused by regional poverty.
In that way four chapters of book try to find answers to questions what have
happened, why have happened, how have happened.
Estonia as a case
Migration in Estonia has been the main source of original empirical data in this book.
Estonia has been chosen on the one hand because the author is most closely connected with it and feels bold enough to say something about this country but, on the
other hand Estonia is also a country with exactly the appropriate area and fine settlement structure in order to study internal migration. The small population and rather
simple settlement system of Estonia suites well for migration studies. The settlement system consists of one main centre — capital with inhabitants of 1/3 from the
total population, three bigger towns on the 200 km radius from capital, second order
county capitals, smaller towns and rural settlements (Figure 2). This plain settlement system makes an influence of system to processes transparent and allows the
country use as a model. Only the lack of genuinely long distances can be regarded as
the disadvantage. The maximum length of internal migration in Estonia can reach
about 350 km. Still the “shortage of distance” has started probably to affect migration more at the end of 1990s, when improving transportation, communication and
information opportunities have reduced the subjective distances even more.
Several studies (see Tammaru 2001b, Golledge and Stimson 1990) have argued
that socialist countries experienced different settlement development than capitalist countries and that here the main forces behind spatial development were different from those in Western countries because of different economical set-up. Estonia
is an interesting because has an experience of both economic systems from shortterm history. From 1945–1990 Estonia belonged to the group of developed countries and the subgroup of socialist countries. 1989–2000 is often seen as a transformation decade from centrally planned to market economy in Estonia. In the beginning of 1990s all classical features of a centrally planned economy were characteristic for Estonia. Essential changes in economy were obvious in the middle of 1990s
and by the end of the ten-year-period all market forces had taken their classical
forms.
At the same time Estonia has passed different stages of urban-rural development
during recent thirty years, which makes it perfect case for analyses of migration
turnaround. Up to 1983 mainly urbanisation processes prevailed in Estonia, later
migration outflows from bigger towns has marked deconcentration processes. Such a
turnaround points in history are the best periods for researchers to analyses reasons
of process.
12
Tallinn
Narva
Tartu
Pärnu
0-10
10-20
20-240
240-2610
Figure 2 Population density in 2000, urban settlements and main roads in Estonia
Throughout the book the division into nine types, where possible, is employed in
classifying Estonian local municipalities: four urban (capital, county centres and
Narva, satellite towns, other towns) and five rural types (the close to the capital and
county centres, the close hinterland of smaller towns, rural municipalities with essential roads, rural municipalities with railway connection and peripheral rural municipalities). Classification is based on the functional — hierarchical distribution of
the settlements (see Olsson 1965, Morrill 1970). The main principle of this division
lays on assumption that settlements with different population size and functions
have different roles and prospects for success. Increasing number of functions creates better position in “settlement market”. The numerated distribution used in this
book corresponds to the hierarchy of municipalities. In defining the municipalities
into groups their highest possible position has been taken into account. Every municipality can belong only into one group and, in case it bears characteristic features
of several groups, it has been attributed a higher position. On the whole, municipalities have been divided into nine groups. Several times proposals were made to cut
down the number of groups during the writing of this book. Municipalities with a
railway passing through there turned out to be the least popular. Laura Saurama, one
of the early reviewers of the manuscript, made even a nice suggestion that the whole
book could be a life story of different municipalities. Unfortunately, this book did
not become a story about seven brothers or a story about Snow White and the seven
dwarfs, which would have undoubtedly possessed a greater literary appeal. After
some discussion, the communes with railways retained their position, because the
earlier studies (Tammaru 2001a) have shown that railway has been a significant engine of development for municipalities, and the final division into nine classes was
13
set up. Figure 3 shows the number of settlements in different groups and the total
population living in this type of settlements in 2000.
population (thousands)
450
69
400
350
250
40
30
150
16
100
50
50
41
34
200
70
60
48
300
80
10
16
12
20
10
1
0
0
capital
county
centres
satellite
towns
other
towns
rural,
close to
big towns
rural,
close to
small
towns
total population in all settlements in 2000 (thousands)
with
rural with
essential railway
roads
rural
perifery
number of local units in this group
Figure 3 Number of settlements and total population in different settlement types
in 2000
Table 1 Permanent population by types of settlement (2000 Population census)
>250,000
urban
Number of municipalities
Total population
1
400,378
50,000–
250,000
urban
2
169,849
10,000–
50,000
urban
11
233,262
<10,000
urban
Rural
28
119,722
205
446,841
Along with the hierarchical distribution also the more common distribution into
urban-rural municipalities and distribution according to the population size has been
employed in this book. Table 1 displays the number of municipalities and the total
size of population according to latter division.
Original empirical data in the book
Description of migration up to 1990s in Estonia is based mainly on population statistics. Starting from 1990s the quantity of migration statistics increased but the reliability of data declined. This has caused the increased attention on sociological research. Analysis of the migration has undoubtedly been rendered more difficult by
the lack of reliable registers and long-term timeseries. One of the most reliable data
sources about migration in 1989–2000 is the population census from 2000. Simultaneously with statistics sociological surveys are used to describe the period from 1990
to 2000.
14
Altogether five different data sources are used in order to analyse migration patterns
in Estonia in the 1990s. All databases in this book can be divided into two groups: (a)
individual level data (b) macro level data. Table 2 gives an overview of sources and
types of data. None of them is a totally perfect data source and all have their own
limitations. When combining the results from different sources, we can still get a
rather trustworthy picture. Conclusions about migration processes can be made only
if the results are supported by several independent data sources.
Table 2 Main databases and their content used in this book
Data base
Local
Governments
Data Base (B)
Living
Conditions
Survey 1999
(A)
Migrant
Survey 1997
(A)
Qualitative
survey (A)
Population
Census 2000
(A, B)
Period for
migration
analyses
1989–2000
Content
Remarks
Demographical, social, economic, geo- Administrative
graphical indicators of 247 municipali- databases, 2000 poputies.
lation census data,
author’s compilations.
Used in chapters 1
and 4.
Before 1983 Consists of information about last Used in chapter 2.
1983–1988 change of residence of respondents Limitations: Limited
1989–1999 and future migration plans. 2453 res- analytical capacity of
pondents reported about the change of migration
reasons.
residence in Estonia, 628 of them du- Possible
memory
ring 1989–1999, 447 between 1983– problems (see Auriat
1999. Weighted sample is used
1991)
1989–1997 Mail survey among those, who registe- Used in chapter 2.
red a change of residence in population Low response rate
register in 1995. 489 cases of the (45%)
change of residence
1995
Behavioural approach of migration
1989– 2000 The territory of Estonia was divided
between 42 cities and 205 rural
municipalities. Taking into account the
changes of names, types and borders of
administrative units, the 1989 Census
results have been recalculated for
comparability of data according to the
distribution and borders of administrative units in 2000 (2000 population
…2001)
Chapter 3
Used in chapters 1, 2
and 4.
No reasons of migration were asked
There is no general consensus about names of spatial migration processes (see Chapter 1). In this book deconcentration is defined as an outflow of population from
larger towns into less densely populated areas, suburbanisation is an outflow of population mainly into the nearest hinterland of bigger towns and counterurbanisation is
a migration to the more distant areas, if not defined otherwise.
15
Acknowledgements
The research has been completed with support from Tyoverkko network of Finnish
Academy of Science, Oxford University Scholarship and SIMO. Part of the research
has been financed by Estonian Science Foundation.
This book is, in a way, a milestone in the migration research history of my life.
Therefore, I would like to express my greatest appreciation to my first scholarly
teacher professor Ann Marksoo from the University of Tartu, whose influence once
raised in me the interest in migration issues. Professor Olli Kangas and Ismo Söderling
from the Department of Social Policy of the University of Turku have helped me
during the several Finnish periods in both practical and scientific questions. I’m grateful for Olli Kangas for his time and numerous helpful comments about different
versions of manuscript. Comments made by him had pushed me to jump higher and
higher and in the end of the day, I’m quite happy about it. Laura Saurama’s invoking
and encouraging remarks about the manuscripts had improved it a lot. Dear all from
the Department of Social Policy at Turku University, your friendliness and warmth
have made it easy to work in “North” and I would suggest it as the best place for
writing books also to everybody else. I thank also Marika Kivilaid, Riina Leinbock and
other employees of the Statistical Office of Estonia, the Ministry of Social Affairs
and the Tax Board of Estonia for their kind and professional help in data exchange. I
render thanks also to Ülle Marksoo from the Ministry of Social Affaires for the offer
to work with Living Conditions Survey. And finally, thanks to Leho who has acted at
times as the first reader, technical consultant, and supported both practically and
morally.
Liina-Mai Tooding, Takis Venetoklis have been the good counsellors in the questions of statistical data analyses and Judit Stömpl in qualitative research methods.
Technical help was provided by Greta Tischler (first versions of maps), Monika Tenno;
Tuuli Semevski, Sirje Ainsaar and Alar Helstein (proof-reading). I appreciate also
comments made by Elli Heikkilä and Tiit Tammaru as first official pre-reviewers.
And finally, comforts of different ferry firms sailing between Tallinn and Helsinki
and Finnish Railway have added several pages to this book.
16
17
Urbanisation and migration reverse
in developed countries
1
18
19
1. Urbanisation and migration reverse
in developed countries
Introduction and definitions
This chapter aims to give an overview of the general urbanisation trends in developed
countries and analyse macro level reasons for the growth or decrease in urban population. The main emphasis is on the role of migration in the population change. We
also give an overview of trends and background factors of urban-rural migration based
on previous research.
The change of the whole population is shaped by different demographic processes. In this chapter, we will mainly analyse the role of internal migration in the
urbanisation and deconcentration processes. Since there is a wide diversity of urbanisation and deconcentration concepts, we start from an overview of different
options and notions that are used in the book. The first part of the chapter discusses
macro level factors and theories of migration in order to create a general theoretical
framework. Our focus is on migration turnaround, which has occurred in many countries. The chapter gives an overview of previous studies on migration reverse and the
main factors behind it. The subchapter 1.4 describes the change of urban-rural population in Estonia and the role of migration in this process. The main emphasis is on
the period of 1990-2000, as the most confusing and the least studied period. The
municipality level internal net migration data is compared with economic and geographical background indicators. In the empirical analyses of migration trends in
1990-2000, we consider Estonia as a closed country model since it allows for a more
transparent observation of changes1.
Definitions
There is no consensus about the definitions of urbanisation and stages of population
deconcentration (see Berry 1976, Frey 1987, 1988b, Champion 1989a, 2001,
Johnston 2000, Tammaru 2001a). The least controversial one seems to be the definition for urbanisation, but even here, several approaches can be applied. The most
commonly used definition for urbanisation is an increase of urban population. None1
A overview of the influence of natural change and international migration in Estonia during the
period of 1990–2000 is presented by Tammaru et al (2003).
20
theless, urbanisation can be measured by different characteristics such as net migration gain, the area of land being developed for urban use, changing life styles, etc.
Fielding (1982) defines urbanisation as a process, which can be described as a positive relationship between migration gain and settlement size. Accordingly, urbanisation takes place if settlements with a bigger population grow more due to migration
than smaller ones. Still, even within this framework the results can be rather different depending on the definitions of settlement units. Several classifications of units
can be used: administrative units of local authorities, metropolitan areas (urban core
with at least 100,000 inhabitants, and surrounding communities), labour areas, commuting areas, etc. (see Champion 2002).
Definitions of deconcentration are even more controversial. In this book, we will
make a distinction between three terms: deconcentration, suburbanisation and
counterurbanisation. The terms refer to population outflow from more densely
populated areas (towns), but the borderline between suburbanisation and
counterurbanisation remains ambiguous. In the USA, where the phenomenon was
first observed, counterurbanisation was interpreted as population deconcentration
away from large (especially the largest and oldest) urban settlements. However,
there are different views on the extent of counterurbanisation. In his earlier works,
Champion (1989a) defined counterurbanisation simply as redistribution within the
same metropolitan area. Ford (1999) specified it as migration to the adjacent and
more distant rural areas. Similarly, Champion (1992:469) stated that
counterurbanisation can be characterised as “a growth rate of rural areas in favour of
more peripheral regions, smaller settlements, or cities of intermediate scale”. Berry
(1976) believed that counterurbanisation is redistribution along urban hierarchy
rather than across broad geographic regions. Frey (1987) employs a somewhat unified approach, suggesting that counterurbanisation is more like a process of population shift from large cities and more heavily urbanised regions to less densely populated areas down the urban hierarchy. According to Fielding (1982), the best indicator of counterurbanisation is negative correlation between the size of settlement and
all net migration. And finally, according to the theory of differential urbanisation
(Geyer and Kontuly 1993), the beginning of deconcentration is marked with the
decline of primate city functional area net migration. Within this framework,
counterurbanisation is the latest stadium in the process of deconcentration. The
latter is described by the diminishing population of primate and intermediate city
areas due to migration. Once again, the definitions of deconcentration can be combined with two different indicators: considering just net migration or all population
change. However, almost all approaches described here define deconcentration as
population distribution in favour of settlement units on a lower hierarchical level.
The definition of suburbanisation is as unclear as counterurbanisation. The dictionary refers to suburbanisation as “denoting the process by which cities expand
peripherally, initially by out-migration of population and economic activity from dense
urban cores, to less dense contiguous settlements” (Suburbanization 1998). However, some researches do not consider suburbanisation as a part of deconcentration
but rather like a geographical diffusion of concentration (see Champion 2001). In
addition to the most common criteria for suburbanisation like total population growth
or positive net migration of the surroundings of bigger towns, plenty of other
21
Net migration
suburbanisation measurement criteria can be employed. Suburbanisation is sometimes interpreted as a geographical expansion of towns and urban living style (Halliday
and Coombes 1995). Fisher (2003:554) presented the idea that “the key to differentiating suburbanisation and counterurbanisation lies in measuring the behaviour
and motivation of migrants”. Therefore, the search of a more rural way of life, for
instance, can be one characteristic of counterurbanisation that does not depend on
the distance of migration.
In this book, deconcentration is defined as an outflow of population from
larger towns into less densely populated areas (sometimes also called
deurbanisation, desurbanisation, or disurbanisation). Deconcentration includes
suburbanisation and counterurbanisation. Suburbanisation is an outflow of population mainly to the nearest hinterland of bigger towns (also called spill over, or seen
as a continuous growth of towns over their borders), and counterurbanisation is migration to the more distant areas (Figure 1.1). Consequently, the distance of migration from larger towns is an essential differentiator between suburbanisation and
counterurbanisation. Distinction by distance from the main centre is especially suitable for Estonia, which is a country with 1/4 of the classical full settlement system
(Figure 2).
Decon
cen
tration
Count
er urba
nisat io
n
Suburbanisation
Urban positive net migration
Town
Distance from centre
Figure 1.1 Definitions of migration processes and distance from main centre(s)
as primary concept used in this book
Figure 1.1 illustrates the definitions of suburbanisation, deconcentration and
counterurbanisation as net migration gain in different distances from centres. Although urban and rural administrative borders are essential benchmarks in definitions of spatial development, the spatial processes are not limited with borderlines
and might cross them. For example, it might be difficult to make a clear difference
between territorial spill over effect of towns and growing popularity of rural life.
People changing their place of residence beyond suburbs and becoming a part of rural
population can still retain basic functional ties with the city and urban lifestyle, and
can be considered more alike to urban population. Moreover, different countries use
different measures to define “urban”. Although the formal statistical approach is
22
widely used because of data availability, it has several functional shortcomings. Despite the fact that there are several options to distinguish between urban and rural
life (life style, employment structure, density indicators), we will only apply the
administrative approach. Alternative definitions are not employed in our empirical
analyses due to limitations to comparative data in this book.
1.1 Macro level factors of migration
Decreasing mobility seems to be a universal phenomenon in Europe. Many researchers have reported the decline of mobility in the last 20 years in different countries
(Poulain 1996, Gonzalez and Puebla 1996, Cairncross 1997, Rogers and Rajbhandary
1997, Lin 1999, Coll and Stillwell 1999, Illés 2000, Herm 1999) and changes in the
timing of migration behaviour (Mulder and Manting 1993, Baccaini and Courgeau
1996). The changing migration intensity could give an explanation to the driving
forces behind migration. The features with a possible influence on the outcome of
migration can be classified under six broad categories.
The administrative statistical factor. Analyses of migration by administrative
units are highly dependent on the classification rules and the size of units. Bigger
units will decrease and smaller units increase statistically the number of migrants. In
Estonia, for example, an average municipality size was quite similar to the municipalities in Germany or Denmark in 1990s, but these municipalities were bigger than
in many Central European countries and much smaller than in other Nordic countries (Appendix 1.1). As a result, similar actual migration intensity produces statistically lower mobility rates in, for example, Nordic countries and higher rates in countries with smaller statistical units.
The selectivity of migration. Selectivity was one of the first qualities of migration to be discovered. Selectivity by age is probably the best known “migration rule”,
but migration is selective also by other characteristics: wealth, civil status, human
capital, etc. A more detailed overview of possible selection mechanisms of migration
is offered in the second chapter of this book. Migration selectivity indicators might
vary in time and space. Some of those rules are more and some less persistent in
time. For example, White (1989) compared migration in the 19th and 20th century
France and found that while in the past, migration was an activity of the poor, it had
now become an activity of the wealthy2 . All factors influencing migration selectivity
are potential sources of inequality if they become barriers for the freedom to move.
Demographic potential. The population size in the old and new places of residence was among the first indicators used in classical migration models in order to
predict the migration capacity between two settlements. According to the gravity
model, the number of migrants between two settlements is proportional to the population size of these settlements and inversely proportional to their distance from
each other. The influence of population size can be interpreted as a joint result of the
This difference can reveal the changed benefits from staying and moving, or a changed strategy
of successful life career.
2
23
demographic potential, diffusion of information, and the probability of selection
between different opportunities. However, the total amount of population is not the
best indicator for predicting migration, because migration is selective by age. The age
structure of population is an important determinant of a region’s migration potential. The volume of migration should increase (assuming that other important factors
remain unchanged) when a more sizeable generation reaches the critical age (Plane
1993, Rogers and Sweeney 1998 and others). Therefore, the age structure of population is an important determinant of the migration potential of a region. In all developed societies, people in the age of 18–30 make the majority of moves, and the
amount of people in a particular age influences the intensity of overall migration
flows (see also Chapter 2.4).
Regional differences and space. Starting with Stouffer (1940), most macro level
migration equations assume that large regional disparities encourage migration and
that regional equalisation decreases movement. Migration is closely related to distance and space. Distance determines the type, reasons and nature of migration.
Regional imbalances of economic development are often seen as a main source of
migration, although not the only important regional factor and not necessarily over
long distances (Green et al 1986). Still, the influence of any regional disparity is
seldom univalent. Regional differences in income are suitable for demonstrating the
interaction between different factors. In order to understand the link between wealth
and mobility, demographic and social indicators of groups should be considered. For
example, Plane and Rogerson (1991) proposed an idea of eightfold relation to illustrate mutual influence between three factors — migration, demographic and economic development (Figure 1.2). The Figure shows a link between the proportion
and mobility of the cohort attending the job market in the age of 20–24. Mobility
decreases when a large generation enters the labour market and it becomes minimal
when the last members of the cohort seek a job (B). When a smaller cohort follows
a large generation, migration activity increases and reaches its peak at the generation
change (C–D). This theoretical generalisation took into consideration not only the
inertia of economic demand but also the influence of demographic cycles and economic perspectives. It is found that
migration propensities of young Mobility
adults tend to be lower for large coC
horts due to competitive labour
markets and depressed job opporLarge cohort
tunities (Pandit 1997). In concluA
sion, macro level economic migraD
tion streams are composed of
Smaller cohort
mainly two factors — demographic
and economic development.
B
Commuting and other alternatives for migration. Many
Share of 20/24 year old people in population
studies mention commuting and
Figure 1.2 Mobility and the proportion of
the development of communicapeople aged 20–24 in population
tion facilities as alternatives for miSource: Plane 1993
gration (see Zelinsky 1971,
24
Dziewoñski and Korcelli 1981, Mulder and Hooimeijer 1999). Poulain (1996) finds
that the only good explanation for changes in migration is based on “continuous
improvement in transportation and communications, which extends the daily journey to work and also the higher level of social welfare which has reduced the need to
change residence in order to take another job”. Many papers analyse the role and
directions of commuting (see Green 1995, Green and Meyer 1997, Fuguitt 1991,
Cross 1990, Borgegård et al 1995, Illés 2000, Martin 2001, Gottlieb and Lentnek
2001). Surveys report an increase over time in the amount of commuters, in the
distance of the journey and in the commuting area (Barr 1997, Wood and Carter
2000, Tammaru 2001c). One factor encouraging commuting can be increasing ownership of second homes (Johnson and Salt 1992). The spreading of commuting opportunities leads to greater freedom in the selection of the places of residence and
employment. People become more independent from distances and have more freedom to choose a place of living (see different references by Findlay and Rogerson
1993). Accordingly, the ability of people to fulfil their individual well-being preferences also rises. But still, commuting requires extra resources. Stockdale et al (2000),
for example, have found that commuters are generally more affluent people. It leads
to the hypothesis that permanent commuting might be a privilege of more advanced
and higher income societies. Nevertheless, it would be useful to distinguish between
commuting as a lifestyle and commuting as a temporary inevitable period before
permanent migration.
Freedom to move. One explanation to different mobility rates of people in different life stages, countries or regions can be the differing levels of freedom to react
accordingly to the push and pull signals from the environment or personal needs.
From macro reasons, a person’s freedom to move seems to be influenced primarily
by the level of economic development. It is a well-reported phenomenon that economical upraise coincides with greater freedom and increased ability to change the
place of residence (Bengtsson and Johanson 1995, Borgegård et al 1995, Kauppinen
et al 1997). Migration intensity is the lowest at times of economic decline and depression. The level of economic development is especially important for economically motivated people. However, the freedom of movement does not depend only
on economic aspects. Migration freedom can be influenced by the availability of
information, resources, mental readiness, or even belief in success. The impact of
different factors to the freedom to move is discussed more thoroughly in chapter
two.
Different determinants (administrative classifications, factors determining the
selectivity of migration, demographic potential, regional differences, alternatives for migration, and freedom to move) have different degrees of impact on
migration. The overall migration pattern forms as a result of mutual influence
between all these factors. Some of those factors are contradicting by character,
some supplement each other. At different times different reasons play key roles
in the migration processes because of the change of reasons for regional disparities and the changing importance of well-being factors.
25
1.2 Macro level theories of settlement development
The majority of regional development theories see
the development of regional space from the point
Town
Town
of view of economic development. Economic space
models conclusively see the settlement system as
dependent on production rules (see also Elliot and
Town
Town
Perry 1996) and do not pay particular attention to
Cen tra l
the role of migration. Among these theories, spato wn
tial theories by von Thynen (1895), Weber (1909),
Lösch (1954), and Christaller (1933) are historiTow n
Town
cally the most important. All those models are
based on the assumption that the space is plain and
the interaction between production rules, market Figure 1.3 Hierarchical concentric model of production, after
prices, and distance creates spatial structures. All
Christaller and Lösch
the afore-mentioned models predict the development of a hierarchical concentric space production
structure (Figure 1.3). None of these approaches analyses the role of migration, but
rather they describe the more general outcome of human production activities on
space. Although during the time when these models emerged, the population was
closely linked to production areas, we can still assume that the authors claimed to
describe development of the settlement system generally.
Modern approaches use quite the same economic arguments in predicting regional changes. In von Böventer’s view (1970), the three main driving factors behind
regional economic development are (1) the external and internal economies or
invisibilities, (2) the demand for land, and (3) transport costs. Von Böventer believes
that the first two factors tend to encourage spatial concentration and “centrifugal
and centripetal forces interplay via the level of transport costs” (325). He found that
a change in assumptions concerning the relative strength of these factors can lead to
many different spatial structures. One extreme is the total geographic concentration
of production when agglomeration economies dominate. The second is a complete
dispersion of all production on a homogeneous plain when linear homogenous production function applies for all economic activities.
One of the latest approaches to the development of settlement is the spatialgrowth model. The model is based on the concepts of W. Rostow and E. Taaffe and is
the spatial expression of Rostow’s stages of economic growth3 . According to this
model, there are three stages of spatial development, which describe the growth and
3
Rostow (1978) saw economic growth as occurring in five stages. Initially, technology is primitive
and social structures are rigid and hierarchical. Production per capita is low and change is rare. On
the second stage, possibly because of outside stimuli, investments rise, the infrastructure starts to
develop, and there is growth in agricultural and industrial sectors. The next stage, ‘take-off ‘, is
based on these preconditions. This is a short period of time during which the economy and society
are transformed. Investments and savings rise, and new industries grow in primary and manufacturing sectors. The growth gives rise to the ‘drive to maturity’. Industrial development now diversifies, imports fall, and investment is still high. The final stage, the ‘age of high mass consumption’, is
reached as consumer goods are of increasing importance, real income rises, and the welfare state is
established” (Spatial-growth model ... 1997).
26
decline of settlements and ports: their production functions, domestic and international market ties. Another interesting theory from the viewpoint of migration and
regional development is the endogenous growth theory. The theory is based on the
assertion that the accumulation of physical and human capital causes the spill over
effect. As for regional development, it would be well suited to explain the phenomenon of deconcentration. Using the endogenous growth theory, Durlauf et al (1996)
see four main factors of economic growth: (1) the accumulation investments, (2) the
increase in the quality of labour force, (3) the relocation of resources from low to
high productivity sectors, and (4) the technological change. However, the authors
claim that, according to their results, these factors influence income per capita rather
than the growth rate.
Despite the obvious connection between migration and economic factors, there
are only a few economic theories referring directly to the role of migration. Modern economic theories see the employment market as the main determinant of regional development (Kaldaru and Päll 2003). For example, Fischer (1998) analysed
why and when migration matters in regional economic development and growth. He
believes that the impact of migration depends on the way different markets and
economies differ from each other. Migration can be seen as a relief to regional trade
shortcomings, which helps avoid economic shocks, speeds up the growth convergence process and redistributes people from technologically disadvantaged to advanced regions. Finally, migration can determine the distribution of potential gains
from economic integration.
Based on the above mentioned wide-ranging macro level economic factors, a variety of analyses on the reasons and driving forces of migration have been produced.
The majority of them use the migration gravity model as a starting point, according
to which migration is dependent on the settlement size and distance between settlements. Sometimes, approaches based on the migration gravity model are unified
under the common name of spatial-interaction theory (see Spatial-interaction…
1997). Still, these equations are used mostly as purely empirically descriptive tools
without any theoretical claims and therefore cannot be considered as theories.
Some macro level approaches developed within the framework of migration research could also belong to this group of spatial development models. Mabogunje’s
(1970) system model sees the development of the environment in the framework of
interactions within a rural-urban system, based on information flows sent by migrants and consequences of changes. According to this approach, regional changes
are an outcome of different responses from the external environment, the urban
system, and the stimuli. The model itself allows for quite various interpretations and
adds another important feature — information — to the factors considered in urbanrural development. Although this model has mostly been used for analysing urbanisation, it does not state clearly whether the direction of general regional development is towards urbanisation or not. According to Zelinsky’s (1971) mobility transition model, migration streams change according to the development stage of a
region or a nation. Geographical mobility increases with development until the point
where transport will be sophisticated enough and migration may be replaced by commuting, or when communication facilities reduce the need to human mobility.
Zelinsky’s model presents a relationship between the general level of development,
the amount of different migration streams, and the level of communication facilities.
27
In addition to spatial development models, several descriptions with plausible regional development stages have been developed (Table 1.1). They all see the general direction from frequent thinly distributed settlements towards concentration
and, finally, return to a more evenly distributed type of settlement. In the latter
stage, the mixture of lifestyles, disappearance of clear differences between urban
and rural, and counterurbanisation are predicted.
Table 1.1 Stages of urban development according to different authors
Gibbs 1963
Formation of towns and
start of their growth.
Urban population is a
minority. Rural
population grows quicker
than urban.
Towns grow quicker than
rural settlements, but
rural population in
absolute terms will not
decrease
Growth of towns
continues. Rural
population diminishes
because of ruralurban
migration
Peak of concentration.
Accelerated growth of
metropolis. Absolute
decline of population in
small towns and rural
settlements. Formation of
agglomerations.
slowdown of the growth
of town centres. Growth
of population on the
hinterland of towns.
Distribution of people
begins to balance.
Hall and Hay 1980
Ioffe 1987
Equal
settlement.
The
concentration
points are determined by transportation
Zaiontskovskaia 1985
Autonomic
development of rural
and urban areas until
the
demographic
deficit.
Industrial revolution
and concentration of
people to the big
towns
Accelerated growth Concentration of rural
of core regions, de- settlers to towns.
population of areas Decay of frontiers.
between agglomerations.
Concentration
continues, spread of
concentration areas.
Suburbanisation.
Growth of areas near
towns. Concentration
around towns. Change
of values. High efficiency in all socialeconomic spheres
Deconcentration,
Agglomerations. Urareas between agglo- ban and rural popumerations are filled lation is integrated.
Close contacts between rural and urban
areas, intensive, short
production and other
connections.
Accelerated growth
of small scenic rural
areas. Stagnation of
old centres.
Deconcentration
spread from urbanisation centres to the
whole
territory.
Stagnation and depopulation of big
towns.
Spread of agglomerations and accelerated growth of
population on their
periphery.
28
There are three main views on the future rural-urban development: deconcentration
as a major shift in population redistribution (see Zelinsky 1971, Frey 1995, Long and
Nucci 1997, Wardwell 1977, Johnston and Beale 1994, Champion 1992),
deconcentration as a mere temporary exception in a long process of concentration
(Champion 1989b, Fielding 1993, Frey 1987, Frey 1988a, Johnston and Beale 1994),
and a mixed approach.
Forecasts for urban development
Many researchers (Shumway and Davis 1996, Beale and Fuguitt 1990, Johnson and
Beale 1994) believe that the outmigration of people from bigger towns might be a
continuous process. This belief is based on the assumptions of increasing freedom
and the desire of people to select a scenic living place, prevailing preference of living
places out of metropolitan areas, smaller dependence on the location of the job or
lower competition in sparsely populated regions (Borgegård et al 1995). Others see
the continuous deconcentration merely as a backwash effect within diffusion theory
(Chapman 1979). Deconcentration after a stage of intense concentration would also
fit well with the endogenous growth theory.
The idea of a continuing concentration of population is mainly based on the
assumption of the rising importance of space for the geographical organisation of
society. According to the concentration approach, deconcentration is rather an exception than a continuous trend. The redistribution reversal is explained by the coinfluence of unique demographic and economic circumstances (Frey and Speare
1992). It is believed that continuing concentration is a result of advantages of densely
populated regions — better accessibility to the resources and decreased transportation costs. The United Nations vision of the future (World ….2001) also supports
the notion of selective but continuous concentration growth. The authors of the
report found that during the last ten years population has increased more in urban
areas, and this tendency will probably continue in the future. But the increase in the
numbers of urban population will be unequal, depending on the development level:
urban population grows quicker in less developed areas.
According to the cyclical model (Hall and Hay 1980, Berry 1988, Champion
1992, Geyer and Kontuly 1993, Long and Nucci 1997, Kontuly and Geyer 2003),
deconcentration is regarded as a phase in continuous cycles of counterurbanisation
and urbanisation. Champion (1992:475) proposes an explanation of three types of
factors co-influencing the final outcome on a certain territory: factors that facilitate
deconcentration for a period of time, factors that tend to cause greater concentration, and factors that may become effective at different times, depending on the
circumstances. These three sets together can produce considerable variations on
space. There is also an alternative for concentrated dispersion — shrinking of old
centres and concentration of non-metropolitan area population (Richardson 1980).
Analysing the relationship between migration and FIRE (finance, information, real
estate) workplaces, Richardson proposed it as another possibility for the future development.
29
1.3 Empirical studies of migration reverse
Evidence from different countries shows that at a certain level of development of
society, the urbanisation processes will be replaced with deconcentration. However,
there is lack of unanimous understanding of the reasons behind migration turnaround and large volumes of literature have been dedicated to the essence and trends
of urbanisation and counterurbanisation (Berry 1976, Berry and Dahman 1977,
Champion 1987, 1989b, 1992). The following chapter gives a brief overview of same
explanations of migration turnaround. Drastic changes serve as a good opportunity
to study the essence of migration processes more closely.
Although there have been some reports about earlier signs of urban reversal of
concentration in London and America (Korcelli 1984), the counterurbanisation process became well known and was recognised in the USA in the 1970s. After 1970, the
process accelerated. According to Hall and Hay (1980:87), “cores virtually ceased to
grow and with continuing losses from the non-metropolitan areas — the rings actually accounted for more than the entire net growth of the population”. However,
these generalisations included rather substantial variations among individual countries (Hall and Hay 1980). It was found that counterurbanisation was related to the
hierarchy of towns. As the correlation between the ranks of hierarchy and the rate of
population growth was negative, the main contributors to counterurbanisation were
found to be the towns in higher hierarchical order. Before the end of the 1970s, the
first signs of population turnaround had already shown in some bigger American
towns. At the same time, a considerable slowdown of rural inflow was registered in
the United Kingdom, where only remote rural areas still grew in population (Champion1981, 1987).
A comparative study by Gordon (1978) on metropolitan cores, ring and hinterland, and non-metropolitan areas in 18 countries revealed three different groups
according to the differences in population growth in 1950–1970:
1) Counties where the city size was positively associated with the growth of its
hinterland area, but negatively with its own growth rate;
2) Deconcentration at the level of both urban regions and their cores (UK, Netherlands, Switzerland);
3) An increasing concentration on all spatial levels (Spain, Italy, Finland, Japan)
(see Korcelli 1984).
Korcelli (1984) analysed different surveys carried out during 1970s and summarised
it as follows: Spatial deconcentration of settlement was a widespread phenomenon
but its forms varied among individual countries and regions (359). Several authors
proposed an administrative explanation of regional changes, stating that
counterurbanisation was the expansion of metropolitan communities beyond standard metropolitan boundaries. Consequently, counterurbanisation was more specifically a growth of urban areas beyond the administrative boundaries, as people moving out preserved strong ties with urban areas (see also Wardwell 1977, Hall and Hay
1980, Korcelli 1984 ). The growth of urban areas was selective. For example, Johnson
and Salt (1992) argued that during counterurbanisation, smaller towns showed a
more rapid growth of employment than the countryside. However, people still often
preserved ties with towns despite living in the countryside.
30
In many countries, deconcentration seemed to be the most obvious in the 1970s
(Champion 1989a, 1989b). In early 1980s, contrary to the rest of the decade,
deconcentration became less dominant. In some countries, primarily in the USA,
the urbanisation processes accelerated and reurbanisation appeared (Ogden and Hall
2000). In many other regions, however, the processes of deconcentration were still
quite obvious (Table 1.2). Generally, the situation in the 1980s was much more
complicated and diverse than in the 1970s, although it needs to be stressed that
different administrative circumstances complicate clear comparisons. For example,
Dahms (1995) found that the apparent slowdown of counterurbanisation or even its
reversal was often the result of reclassification of former rural areas to urban areas.
In the 1990s, the situation remained diverse. The latest European-wide research
project (Rees and Kupiszewski 1999) reported different trends for different countries in the 1990s, similarly to the previous decades. There are controversial results
about migration trends from the United States. Elliot and Perry (1996) analysed
differences among metropolitan areas in the USA in 1965–1990 and reported clear
contrasts. 85% of all net gains of established metropolitan territories derived from
the exchange with central counties, regardless of the geographic subsector. This led
the authors to the conclusion that “recent patterns of metropolitan dispersion may
in fact reflect an alternative form of suburbaniation rather than extended
Table 1.2 Urbanisation (+) and deconcentration (–) in some countries during
the 1970s–1990s
Country
Australia
Austria
Belgium
Canada
Czech
Denmark
Estonia
Finland
France
Hungary
Ireland
Italy
Norway
Poland
Portugal
Spain
Sweden
The
Netherlands
UK
USA
West Germany
1970s
1980s
1990s
Urban % Urbanisation, Urban % Urbanisation, Urban % Urbanisation,
1970
deconscen1980
deconscen1990
deconscentration
tration
tration
85.2
85.8
85.1
67.5
+, 67.2
+
67.0
ND
94.3
95.4
+, 96.5
ND
75.7
75.7
+
76.6
ND
52.0
+
74.6
ND
74.8
79.7
83.7
84.8
ND
64.9
+
69.7
71.1
50.3
+
59.8
+,61.4
+, 71.0
73.3
74.0
ND
48.5
+
56.9
ND
62.0
51.7
+
55.3
56.9
ND
64.3
66.6
66.7
65.4
+
70.5
+
72.0
+
52.3
+
57.9
60.7
25.9
+
29.4
+
46.7
66.0
+
72.8
75.4
81.1
+,83.1
+
83.1
ND
86.1
88.4
88.7
88.5
73.6
-
-
ND — no data, Source: Appendix 1.2
88.8
73.7
-
+
-
89.1
75.2
85.3
-
31
deconcentration from existing suburban areas” (508). Analysing the USA population
census data from 1990–2000, Lopez and Hynes (2003) concluded that the results
indicated deconcentration. The sprawl index showed shifts towards a more equal
population distribution.
In the 1990s, the majority of European countries still had some signs of
deconcentration. Based on previous references to the direction of internal migration
and general population development from the 1970s to the 1990s, we can classify
countries into three groups according to their deconcentration and concentration
parameters (Table 1.2):
1. Countries where deconcentration has been replaced by reurbanisation (Austria, Belgium, Finland, USA, Canada)
2. Countries with a tendency to deconcentration during the last decades (Estonia, Germany, Poland, Italy, Australia, Czech Republic, Hungary, Italy, Poland,
Spain, Portugal, the Netherlands, UK, Western Germany)
3. Countries with no clear signs of deconcentration (Norway)
We might expect that the minimum level of urban population which will create
preconditions to deconcentration can be around 70%. However, it is obvious that the
beginning of deconcentration does not depend solely on the level of urbanisation.
For example, Norway as a country with the most persistent urbanisation tendencies
has a rather high level of urbanisation, but Ireland displayed some signs of
deconcentration already with the urban population level of 55% (Table 1.2).
Reasons for migration reverse
The process of deconcentration has been quite fragmented and several questions
about the background and impact of these changes are still unclear. Many researchers see migration as an essential determinant in economic, urban and demographic
processes (Richardson 1973, Peltonen 1982, Tervamäki 1987, Goldstein et al 1997)
and almost all agree that migration is influenced by some kind of regional well-being
differences. In the light of this, Vining, Jr (1985) presented interesting findings,
which linked the possibility of migration reversal with general wealth indicators. He
concludes that the reduction of growth of the core regions conforms to five conditions: (1) national economy passes the threshold of approximately $ 4000 in gross
domestic product per capita, (2) severe economic regression, (3) limits to the growth
of agglomerations, (4) administrative and political monopoly, (5) changes in technology. Still, there is no consensus about reasons or general stages of migration reversal.
Moreover, conclusions are often based on assumptions rather than empirical evidence because of lack of comparative data.
In a research conducted at the University of Michigan (Frey 1988a), a hypothesis
that would explain the halting of the development of city cores was tested by empirical studies in 13 developed countries in 1970–1980. The results showed that a
part of the data supported the hypothesis that the main reason for deconcentration
is of social nature. The process is based on such conditions of the housing market
that allows people to achieve all possible objectives, including socio-ecological ones.
Changes in the production structure, improvement of the living standard and the
32
development of transportation will end a situation where production determines the
location of producers and consumers. This model was especially appropriate in the
case of the USA, Canada, the UK, the Netherlands, Austria, and Italy. Similarly,
many other authors (Long 1988b, Wilson 1988) have seen the reason for
deconcentration mainly in greater freedom of choosing the residential area and diminishing influence of production forces. Several researchers have found that the
change in migration streams in the 1970s was a change in behaviour combined with
a change in housing preferences (Berry 1988).
It is accorded to the resident-consumer and asserts that longstanding preferences for lower
density locations are becoming less constrained by institutional and technological barriers.
Changes in the industrial structure, a rising standard of living and technical improvements
in communication and production are leading away from a situation where both producer
and consumer space is dictated by production constraints (Frey 1988a:597).
It was found that in some other European countries (Belgium, Germany, Switzerland, Denmark) the main reasons for regional reorganisation were outdated functions of the economy, and migration flows depending mainly on the restructuring
of the economy (Frey 1988a).
The majority of researches stress the mutual influence of several factors. Champion (1989c) found that most of the reasons for deconcentration have been relatively similar — economical, social and technological changes. Later, Champion
(1992) stated that the reasons for deconcentration and reurbanisation are quite similar but opposite in character. Vining and Kontuly (1978) analysed counterurbanisation
in 18 different countries and found that, despite there being no precise explanation
for migration turnaround, the main reasons seemed to be economic restructuring
and the influence of government policies. Dahms and McComb (1999) gave an overview of the reasons for counterurbanisation in the 1990s and listed the following
factors: attraction of environment, improved communication, state policy, unique
economic and demographical conditions, service industries and restructuring of production, rural amenity environment, and push factors from the cities. Others
(Marksoo 1988, Stockdale et al 2000, Heikkilä 2003) assumed that the change of
rural environment (infrastructure) itself made the migration change possible.
Johnston (2000) listed both the attractions and disadvantages of metropolitans
for employers and households as factors of counterurbanisation: attractiveness, land
use, transport, labour force, relocation of production and the human dimension,
lower costs, life quality preferred by the elderly.
After analysing previous research, Findlay and Rogerson (1993) found that probably all mentioned factors, which are listed in different surveys as causes of migration, could be seen as one single factor under the common term “life quality”.
Basically, they argued that migration is triggered by the influence of personal preferences on the factors that are considered to be the most important. Accordingly, the
reasons related to the quality of life could simultaneously explain both out and in
migration of a region. At the same time, different population groups have different
reasons for migration. People with differing lifestyles have unequal possibilities in
the same environment.
33
Analyses of the reasons for decocentration and urbanisation (Table 1.3) during 1970s–
1990s show the overwhelming dominance of lifestyle-related preferences and
housing reasons among the arguments used as explanations of migration. “Lifestyle” factors are often related to property, property prices, push forces out of towns
because of stress, life quality preferences of different age groups and improving transportation opportunities. The second group by frequency are economic reasons. It is
plausible that different circumstances prevail in different countries. However, the
attempt to rank reasons does not rule out the hypothesis about the coexistence of
several factors.
Despite the fact that Johnson and Fuguitt (2000) found after analysing data from
the USA between 1950 and 1995 that continuity in age-specific trends had endured
through good and the bad times, we assume that the demographic potential of different groups is one of the factors that influence the general migration picture through
prevailing preferences and places of destination. In order to analyse migration processes we should bear in mind that different social groups have different sensitivity to various stimuli and needs. Therefore, migration can be dependent on the
demographic potential of different population groups, their level of freedom to move
and specific pull and push factors in the society specific to these population groups.
For example many surveys report that the leading group among out-of-town
outmigrants are retired people (Beale 1977, White 1990, Johnson and Salt 1992,
Halliday and Coombest 1995, Shumway and Davis 1996, Coll and Stillwell 1999,
Kok 1999, Rees and Kupiszewski 1999) and older people in the working age (Frey
1989, White 1990, Cross 1990). The other group, often mentioned as urban
outmigrants, consists of young families with children and people aged 25–34 (Fielding 1982, Borgegård et al 1995). Even during the period of reurbanisation there is
evidence that during the arrival of young people to towns, older (Poulain 1996,
Baccaini and Puman 1996) and wealthier people (Illeris 1996, Gonzalez et al 1996)
tend to leave towns. Reurbanisation is also partly led by an increased number of
immigrants (Long and Nucci 1997, Bontje 2001), by retired people (Shumway and
Davis 1996, Dahms and McComb 1999), and by the growing number of single households (Bontje 2001).
34
Table 1.3 Reasons for deconcentration and reurbanisation in the 1970s–1990s
1970s
Economic
changes
Changes in
services,
communication
Government
policy, public
sector
Lifestyle,
preferences
Demographic
Long and Nucci 1997; economic restructuring, location flexibility of many
industries
Keeble 1989; small firms, change of production from manufactured based
to service based. New technologies give greater location flexibility and
manages has been able to consider non-economic issues in their location
decision making — such as proximity to attractive physical and social
environment
Frost and Spence 1981
Champion 1992; growth of employment in particular localised industries
like mining, defence, tourism, effect of economic recession on rural-urban
and return migration
Kelley and Williamson 1984; economic forces pulling migrants into the cities
Long and Nucci 1997; improvement of rural infrastructure,
transformation of economy towards services that could be performed in
variety of locations
Johnson and Fuguitt 2000; diminished friction of distance, communication, transport
Champion and Illeris 1990; improvement of transportation and communication
Fielding 1982, 1987; labour
Dean 1987; labour relocation
Champion 1992; improvement in transport and communications
technology and education, health and other infrastructure in rural areas
Borgegård et al 1995; growth of jobs in the public sector, planning policy,
increased access to private car and state subsidy for commuting, attitudes
Vining, D. R. Jr 1985; government policy
Champion 1992; availability of government subsidies for rural activities,
spatial government policies
Heikkilä 2003; "village activities": infrastructure, housing plots
Borgegård et al 1995; increased access to private transport, attitudes;
Bolton and Chalkley 1989; quality of life, property availability and prices
acceptable income
Champion 1989c; lifestyle, preferences
Champion and Illeris 1990; increased preference for owner occupied
houses, growth of tourism and outdoor recreation
Fidlay and Rogerson 1993; Changes in domestic sphere, crime, stress,
congestion, erosion of family unit
Mitchelson and Fisher 1987; commuting
Green et al 1986; lifestyle
Champion 1992; commuting, social problems in large cities, change in
residential preferences of working-age people and entrepreneurs, changes
in age structure and household size and composition
Champion 1992; reduction of stock of potential out-migrants living in
rural areas, acceleration of retirement migration
35
Table 1.3 Reasons for deconcentration and reurbanisation in the 1970s–1990s
(continuation)
1980s
Economic
changes
Changes in
services,
communication
Government
policy, public
sector
Lifestyle,
preferences
Demographic
Reijo and Valkonen 1993; industrial jobs
Bontje 2001; economic downturn, increase of immigration, growing
number of single households
Long and Nucci 1997; loss of cost saving advantages of rural areas in
international business, bank loans investments in commercial real estate
Frey 1987; organisation of production
Frey and Speare 1992; restructuring of towns because of better econ
development, industrial and employment structure
Coll and Stillwell 1999; industrial restructuring
Findlay and White 1986; new distribution of labour
Champion 1992; lower job turnover, contraction in primary production,
less manufacturing investment, reduced rates of house building and greater
difficulties of retirees in selling their metropolitan homes
Rayer and Brown 2001; employment growth, family income, sustenance
differentiation
Heikkilä 2003; improvement of employment opportunities in the cities
Kok 1999; services
Boyle 1995; military migration
Gordijn and Eichperger 1996; government housing policies
Champion 1992; change in government policy, including general
expenditure cutbacks as well as switches in spatially targeted policies
towards inner-city regeneration
Dahms and McComb 1999; way of life
Walmsley et al 1998; physical environment and social infrastructure,
imago
Halliday and Coombest 1995; retirement and way of life, removal of the
constraints of employment
Hautamäki 1984; greater value to quality of life
Cross 1990; commuting, increased private car ownership, less restrictions
in location decisions
Parr 1987; Commuting and more desirable residential location
White 1990; housing as main reason of couterurbanisation in France
Rayer and Brown 2001; home ownership
Bontje 2001; increase of immigration, growing number of single
households
Long and Nucci 1997; immigration
Smailes 1996; demographic potential and concentration
Yoshitaka 1999; Demographic factors linked with labour-market
restructuring
Champion 1992; ageing of the 1960s baby boom into the city-loving
young adults of the 1980s.
36
Table 1.3 Reasons for deconcentration and reurbanisation in the 1970s–1990s
(continuation)
1990s
Economic
changes
Changes in
services,
communication
Government
policy, public
sector
Lifestyle,
preferences
Illés 2000; housing and labour market
Shumway and Davis 1996; manufacturing
Poulain 1996; better economic conditions encourage couterurbanisation
Johnson and Salt 1992; more rapid growth of employment in the smaller
towns and cities than in the true countryside
Rayer and Brown 2001; employment growth, family income, sustenance
differencation
Bonaguidi and Abrami 1996; good social services, high amenities and low
cost of living
Poulain 1996; improvement in transportation and communications
Reijo and Valkonen 1993; services
Lanaspa et al 2003; from one side congestion and crime as push force of
biggest towns and same services, culture amenities at the smaller towns as
pull factor
Borgegård et al 1995; size of public sector and public expenditure will
determine the possibility to remain in remote areas
Wiessner 1999; tenure building, investments, land prices, town renewal
Gonzalez et al 1996; public investment
Wiessner 1999; increase of individual car traffic
Kok 1999; attractive and cheaper housing opportunities in rural areas
Illeris 1996; housing
Gonzalez and Puebla 1996; quality of life, accommodation, supply and
demand of housing
Gordijn and Eichperger 1996; quality and prices of housing, insecurity
Deconcentration, related to government housing policies
Johnson and Salt 1992; the growing importance of second homes
Rayer and Brown 2001; home ownership
1.4 The case of Estonia
This subchapter will look at the urban-rural development of the settlement system
in Estonia in order to describe changes and analyse the underlying reasons. The description of trends up to 1989 is mostly based on previous studies. There are several
comprehensive studies about population development in Estonia in the 20th century.
Historians and statisticians (Pullat, 1978, Karjahärm 1992, Reiman 1927, Tomberg
1930) have studied the earlier periods of the century. Since 1959 the migration of
population and general settlement development has been the research object of geographers: prof A. Marksoo (see Kurs and Toots 2000), Laas (1978, 1987), Ainsaar
(1997a), Tammaru (2001a). Chapter 1.4 provides a more profound analysis of the
latest period 1989–2000. In order to describe the changes from 1989 to 2000 the
most recent population census data are analysed. A combination of general migration
flows with macro-level economic data enables an analysis of the changes in the light
of general regional development.
37
Table 1.4 Stages of population development in Estonia 1920–2000
Period and Natural
economic
increase
order
International
migration
1920–1938
market
economy
1945–1960
socialist
1960–1983
socialist
1983–1990
socialist
Positive
Starting from 1924 Urbanisation
outmigration
Positive
Positive, promoting Fast urbanisation. From the beginning of the
urbanisation
1950 decline of positive urban net migration
Positive
Positive, promoting Urbanisation and concentration of settlement
urbanisation
on all levels of urban-rural hierarchy
1990–2000
Emerging
market
economy
Internal migration
Positive
Positive, promoting Weak positive rural net migration. Negative
urbanisation
internal net migration in higher rank towns
(more than 100,000 inhabitants). Slow
decrease of overall migration capacity
Negative Negative, promoting Controversial results about internal migration.
depopulation
of Census data from 1989-2000 show outtowns
migration from bigger towns
Table 1.1 presents a summary of the general stages of demographic trends in the
20th-century Estonia. During that period the country experienced periods with positive and negative natural increase, growth and decline of towns and transformations
from one economic system to another.
From 1940–1989 the socialist economic and political order ruled in Estonia. For
half a century Estonia belonged to the Soviet Union, although in the status of a
Soviet Republic. Therefore it might be important to highlight shortly some distinctive features of this period as well as the major differences between the socialist
system and market economy in order to analyse the impact of society on migration
trends. The first peculiarity was a lack of big gaps in peoples’ incomes in socialist
countries. Salaries were much less dependent on the level of education or profession
than in countries with market economy. Therefore, we can expect the migration to
be less dependent on social-professional or income status. Secondly, as market prices
for several goods (land, rents) were missing, there were no market economy based
regulation mechanisms. The official housing policy was meant to avoid, as far as
possible, social differences between regions (Musil 1993). Regional development in
socialist countries was much more influenced by central planning as well as various
administrative methods in settlement strategies, than in other regions. As a result of
artificial lack of unemployment and lack of large income differences, other factors
and personal preferences (e.g. environment) could play a more important role in
internal migration. And finally, in a situation of low rents and a continuous demand
for dwellings, the availability of housing was often the most important and deficit
value, especially for younger age groups. Also, an available flat often played the primary role in job selection.
38
1.4.1 Urbanisation from 1900–1983
A faster growth of towns was obvious in Estonia already at the end of the 19th
century. The total number of towns did not rise, but the increase occurred mainly
because of migration (Laas 1987). In the early 20th century inflow of rural people to
towns was a continuous trend. The growth of towns was promoted by specialisation
as well as by the development of industry and transportation.
80
70
%of urban population
60
50
40
30
20
10
0
1897
1913
1922
1934
1941
1950 19551959
1965 1970 19751979
19851989
1995 2000
Figure 1.4 Urbanisation in Estonia, 20th century
In the 1920s and 1930s, which was a period of negative natural growth for towns,
country-to-town migration was the only process avoiding urban depopulation and
providing support for the development of urban settlement. The main reasons for
urbanisation were a growing need for industrial workers in towns and the surplus of
labour in the countryside. In north-eastern Estonia several relatively large industrial
settlements sprang up in the areas of oil-shale mining. At first, however, those settlements were not granted official town rights and so their inhabitants continued to be
counted along with rural population. Although the rural-urban population changes
were still relatively slow, higher fertility in the countryside could not balance the
losses caused by outmigration to towns and so the rural population diminished.
Urbanisation attained its maximum rate in the period following World War II
(Figure 1.4). The process was considerably influenced by immigration from other
Table 1.5 Sources of growth of urban population in Estonia (Average change per
year)
Administrative change
Natural increase
Internal migration
International migration
Total change
1946–1959
2220
3610
6390
8060
20280
1959–1991
150
5610
3170
4550
13480
1989–2000
–2600
–2400
–1220
–11040
–17260
Source: 1946–1991 calculated from Tammaru (2001a,c), internal migration 1989–2000 Local Municipality Database
39
Table 1.6 Sources of growth of rural population in Estonia (Average change per
year)
1946–1959
–2220
–280
–6670
3060
–5830
Administrative change
Natural increase
Internal migration
International migration
Total change
1959–1991
–150
–300
–3330
1360
–2270
1989–2000
2600
–1180
1220
–2550
90
Source: 1946–1991 calculated from Tammaru (2001a, c), internal migration 1989–2000 Local
Municipality Database
Soviet Republics. The joint effect of administrative reforms, natural population dynamics and intensive country-to-town migration resulted in a dramatic slump in rural population (Table 1.6), contrasting with a simultaneous urban growth (Table 1.5)
in 1946–1959. In 1959 56% of the Estonian population lived in towns.
Up to the 1960s the main source of urbanisation was international migration,
while later natural growth became more important (Table 1.5). Also in the 1960s the
rate of urban growth at the expense of rural depopulation remained continuously
high (Figure 1.5). Immigration from other soviet republics was the source of growth
for both urban and rural population, but the majority of net migrants stayed in towns.
As the internal natural growth (births-deaths) for towns decreased, the role of migration became more important for urban development, although the absolute numbers of urban net migration started diminishing during the 1960s (Figure 1.6). Still,
the urban settlement was not homogeneous. International immigrants settled mainly
in the capital and the north-eastern industrial towns, thus increasing the growth rate
of those towns.
20
thousands
15
10
5
-5
Figure 1.5 Internal urban net migration in Estonia (Statistics)
Source: 1946–1990 Tammaru (2001a), 1991–2000 ESA
2000
1997
1994
1991
1988
1985
1982
1979
1976
1973
1970
1967
1964
1961
1958
1955
1952
1949
1946
0
thousands
40
110
100
90
80
70
60
50
40
30
20
10
0
-10
1956-60
-20
-30
-40
-50
-60
-70
-80
-90
-100
1961-65
1966-70
external mig
1971-75
1976-80
internal mig
1981-85
1986-90
1991-95
1996-99
natural inc
Figure 1.6 Components of urban growth 1956–1999 in Estonia
30
20
thousands
10
0
1956-60
1961-65
1966-70
1971-75
1976-80
1981-85
1986-90
1991-95
1996-99
-10
-20
-30
-40
external mig
internal mig
natural inc
Figure 1.7 Components of rural growth in Estonia in 1956–1999
Rural population was shaped most of all by internal migration (Figure 1.7). Continuous urbanisation was stimulated by push forces originating from the nationalisation
of agriculture as well as by several measures of central economic planning.
The deepening concentration of rural settlement was supported by the official
policy of division of all settlements into those with a future and those without one.
That division was applied as a basis for regulation of building and social planning. As
a result of this policy construction in settlements “without a future” was stopped
and small schools were closed, which in turn operated as a push factor for
outmigration. Rural population lost mainly younger and more educated people: first
girls in search of a better job or education, then young men. Due to gender selectivity
of rural outmigration the demographic difference between town and country was
enhanced even more. The selective nature of the migration caused a considerable
drop in the quality of rural labour and rural population in general. Lack of specialists
41
was particularly acute in agriculture, although the share of people involved in agriculture was more than 9% and the number of workplaces in agriculture diminished
more quickly than all population in rural areas (see Tammaru 2001a, p 155).
By the beginning of the 1970s the rate of concentration reached its maximum in
Estonia, having acquired some features characteristic of a high phase of urbanisation
(Kümmel 1986), notably, a slowdown in the relative growth of urban population and
a concentration of most people in major urban centres. Although the demographic
resources for town development were exhausted, urbanisation still continued with
some inertia. More than 1/3 of the growth of urban population resulted from international migration (Figure 1.6). The territorial preferences of immigrants played an
important role in urbanisation. Most of them settled in the capital and the northeastern industrial region and increased the share of city-dwellers. Smaller towns,
however, started losing inhabitants in the 1970s (Marksoo 1974).
The internal migration directions of the 1970s also gave evidence of arriving at
the peak of urbanisation (Table 1.7). The growth of urban population brought about
an increase in the role of town-to-town migration and a decrease in the migration
from one rural area to another. A slight fall in country-to-town migration and a rise in
the opposite direction accompanied those processes (Marksoo 1987).
Table 1.7 Directions of internal migration (% of the total numbers of internal
migrants)
Direction
Urban-urban
Rural-urban
Urban-rural
Rural-rural
All
1959–64
33.5
29.1
14.6
22.8
100
1965–69
35.9
27.1
17.3
19.7
100
1970–76
37.9
26.1
18.5
17.5
100
1986–87
30.8
25.2
25.4
18.6
100
1989–2000
27
23
30
20
100
Source: 1959–76 Marksoo (1987), 1986–1987 Ainsaar (1990), 1989–2000 Population census data
New tendencies of slowdown in the urban growth rate and of suburbanisation
emerged in the mid-1970s and early 1980s (Kümmel 1986, 1987, Marksoo 1988).
Simultaneously with a sharp decline in the urban internal net migration (Figure 1.8)
a decrease occurred in concentration and in the differentiation of regional population growth.
A comparison of the urban/rural population growth statistics of 1977–1984 enabled
Pragi (1988) to divided Estonia into the following three regions:
1. Western Estonia (the West-Estonian islands, Haapsalu county and the western
parts of Harju and Pärnu counties) was a rather stable region with a few stagnant rural population areas;
2. Southeastern Estonia and the coast of Lake Peipsi (the entire counties of Põlva
and Võru, a major part of Tartu county, the eastern parts of Valga and Jõgeva
counties) appeared a peripheral territory with a medium level of urbanisation
and a rather unstable rural population. The countryside of this area was suffering from depopulation almost everywhere, while in some areas the process
was extremely rapid;
42
3. In the rest, comprising a major part of Estonia — settlement differentiation
continued, but slowly and without causing considerable loss in rural population. Nearly everywhere the absolute numbers of population showed an upward trend.
In the beginning of the 1980s internal migration depended mainly on jobs, living
space, and the profile and location of educational institutions. According to A. Lõo
(1987), who studied agricultural workers in 1981, the two main conditions mentioned as primarily favouring the settling of labour were “a good flat” and “a decent
job”. A sophisticated service sector was considered much less important. Ü. Marksoo
(1989) suggested additional reasons for migration. Notably, she found that rural-tourban migration of that time was caused not by shortcomings in the service infrastructure of rural areas, as by difficult working conditions, strained relations in the
labour collective, and family reasons.
1.4.2 Migration reverse 1983–2000
In the late 1980s, large areas in the Soviet Union were still dominated by concentration tendencies. Urbanisation processes showed some signs of stabilisation and change
only in Estonia and Lithuania. This shows that the deconcentration started from the
most urbanised regions. Several additional factors have been used to explain the
deconcentration, including, e.g. the implementation of certain complex
socio-economic measures and an exhaustion of the demographic potential of the
rural areas (see Ainsaar 1990, Ioffe 1987).
Figure 1.8 shows that the demographic potential of different age groups cannot
be the only explanation for the deconcentration trends. The proportion of 25–35year-old people does not correspond at all with the fluctuations in urban net migration. Share of people aged 55 or more fits better but can only partially explain the
migration turnaround.
Studies by Marksoo (1988) show that the migration turnaround started from
Tallinn and Tartu — the towns on the highest level of urban settlement hierarchy in
Estonia, while half of the overall growth occurred in the hinterland of Tallinn. Later,
on all levels of rural hierarchy, migration streams to towns diminished and Tallinn
and the urban conglomeration of the North-East of Estonia ceded part of their inhabitants to the nearest hinterland.
The proximity of all the deconcentration to major centres have been the reason
why some authors has interpreted it as hidden urbanisation (Kuddo 1988). Still with
time the area with positive net migration in rural regions enlarged. Urban outflow
spread from closer hinterland of Tallinn to other regions. Up to 1982 the growth of
rural population was limited to Harjumaa. By 1985, however, three new counties
were added to the rural growth area. The increase in rural population was not exclusively due to closeness to larger urban centres, as rural growth was observed on a
much wider territory (Ainsaar 1990). In nearly all villages, still affected by depopulation, the drain was considerably lower than in the 1970s (Marksoo 1988). This was
-2
10
-4
5
-6
0
urban netmigration
share of >55
age group as % from total population
2000
1998
1996
1994
1992
1990
1988
1986
15
1984
0
1982
20
1980
2
1978
25
1976
4
1974
30
1972
6
1970
netmigration in thousands
43
share of 25-35
Figure 1.8 Urban net migration and the share of different age groups from total
population (Migration data: 1946–1990 Tammaru (2001a), 1991–2000 ESA)
mainly the result of the diminishing number of leavers from the countryside rather
than an increasing flow of inhabitants from towns to villages. At the same time, the
flow of rural population into county centres and small towns still continued, but on
average at a 2–3 times lower level than before. In the mid-1980s migration flows
between the countryside and county centres as well as between the countryside and
other small towns were balanced (Marksoo 1988). Only the centres of state and
collective farms continued as points of concentration (Marksoo, Ü. 1989).
Migration data from 1986–1987 revealed that the growth of towns still occurred
at the expense of young women who started their studies and pensioners. The main
motives for moving to towns were family reasons, lack of professional work in the
countryside, or conflicts in the previous place. A case study conducted in three municipalities in the hinterland of the capital (Kuusalu, Haljala, Vihula municipalities,
Ainsaar 1990) showed that people left villages and went to towns mainly because of
studies, but after graduation they usually returned (especially young men). The main
reasons to stay in towns were the lack of a professional job, or family reasons. Among
urban-to-rural migrants very often it was the husband who moved to the countryside
in search of a job while the rest of the family stayed in town. In many cases the main
motive to move to the country was the lack of living space in towns. Most of the
newcomers did not work in their speciality in the countryside, doing agricultural jobs
instead and they lived not far from the centre. Their future plans for the place of
living seemed to be more dependent on the opinion of wives. A higher readiness to
stay in the countryside was revealed among a bit older (aged 30) families with children. Altogether, net migration was largely dependent on the economical well-being
of collective farms. Villages with weaker economic indicators were still losing population.
44
Marksoo (1988) believes that although the changes were due to bilateral measures of
town and country, the decisive role in the changes of rural-urban migration patterns
belonged to the labour quotas enforced in most of the major Estonian towns. Those
quotas affected first and foremost internal migration from country to town. Another
important factor was the economic and social level of the rural areas. Certain improvement had occurred in girls’ admission to vocational country schools. Some
growth could also be observed in the number of service sector jobs as well as women’s employment in such jobs in the country. New jobs and professions for women
were created in the countryside (see Kuddo 1988, Lõo 1987), which encouraged
young women to remain in rural areas. According to Kuddo (1988) the most likely
reasons for the growing popularity of country among young graduates was based on a
better availability of housing and higher income prospects than in town. Many young
families left towns and tried to settle in the countryside, because of the lack of flats
in towns and the growth of the prestige of country life.
As a result of the above described changes in migration, the age-gender structure
of population in rural units improved by the late 1980s: the share of old people
decreased, and the gender rate became more balanced. But it must be emphasised
that the positive changes occurred only in some regions, which were more favourable
and economically better developed.
The period 1989–2000
After the political changes of 1990 new tendencies pervaded the whole society. The
positive natural increase of the population was replaced with a negative one as a
result of the falling birth rate, while the negative international migration influenced
most of all the population of towns. Gradual change from planned economy to market economy shook all spheres of life. A sharp drop of the GNP was accompanied
with a rise in social inequalities and overall social insecurity (see Deacon 2000). One
of the first signs of economic change was the rise of accommodation prices in 1991.
In migration statistics it was reflected by a sharp rise of rural inmigration in 1992
(Figure 1.8). However, part of the unnaturally high rise of the numbers of urban-torural movers was the result of a fictitious registration of living place because of economic reasons. Hoping to avoid the growth of communal taxes some family members were officially registered as inhabitants of the countryside where communal
taxes were lower.
Also other signs of growing social and regional inequalities were revealed in
the beginning of 1990. A rise of unemployment (Eamets 1994, Pettai and Sõstra
1997) and changes in the housing market (Ainsaar et al 1996) changed the overall
environment of migration. At the same time social stratification grew quickly, so that
by the end of 2000 Estonia had become a country with the highest income differences among economically more advanced countries (Saar 2003). Because of the
growing economic disparities we could expect an increase of economic reasons in
migration during 1989–2000. Figures 1.9 and 1.10 reflect the economic and housing
indicators by settlement type. The income and unemployment rates maintained a
45
5000
6
4000
5
Income level
3500
3000
4
2500
3
2000
1500
2
1000
1
500
Registered unemployment 1996
Registered unemployment 2000
Income 1996
Rural periphery
Rural with
railway
Rural with large
roads
Rural around
smaller towns
Rural around
county centres
Other towns
Satellite towns
County centres
0
Capital
0
Level of registered unemployment in Sept
7
4500
Income 2000
Figure 1.9 Average level on incomes and unemployment4 by settlement types in
1996 and 2000 (Local Municipality Database)
45
40
35
30
25
20
15
10
5
0
Capital
County
centres
Satellite
towns
Other
towns
Rural
close to
county
centres
New dwelling built 1991-2000 per inhabitant
Rural
close to
smaller
towns
Rural with Rural with
Rural
large
railway periphery
roads
share of 15-34-year-old (1989)
Figure 1.10 New dwellings per inhabitant, built in 1991–2000 (Average by settlement types, Population census data from Municipality Database)
4
There can be up to two times differences in registered and survey based unemployment data in Estonia.
Although the survey data have higher reliability, it provides reliable unemployment data only for large
regions. All analyses in this book are based on local municipality level data. Although official data do not
reveal all of the unemployment, we can assume that the gap between official and non-official data is
quite similar in all municipalities, which still makes it possible to analyse registered unemployment as a
useful indicator of differences.
46
similar shape of distribution by hierarchical settlement types throughout the years.
General economic situation was better in the top of the settlement hierarchy. In
1991–2000 the rural municipalities close to bigger towns enjoyed the quickest rise
of new dwelling space per inhabitant. Housing conditions were among the most
important values in internal migration during the Soviet period. As a hypothesis it
was assumed that it is likely that at least some of its importance has been replaced
with economic values in the market based economy.
According to the classical rules of migration, a rise in regional differences will lead
to growing migration flows, in case of missing barriers. As a result of growing regional
differences a rise of residential mobility can be expected. However, this did not
happen in Estonia. On the contrary, according to official statistics, the volume of
internal migration did not rise and was relatively low as compared with the previous
decades (Figure 1.11). The overall registered migration intensity stayed much lower
than in the 1980s. Although it can be a sign of migration barriers or other alternatives
for migration, at least part of it can be explained with changes in migration registration. Also irregular migration rate is obvious evidence about the influence of administrative changes in migration registration. Analysing changes in the age structure of
migrants in time Herm (1999) found essential differences in the age structure and
numbers of urban arrivals. The number of arrivals to rural areas changed comparatively less, but the decrease was still noticeable. There can be different reasons for
the low migration rate: non-registration of moves, diminished demographic potential, migration barriers, increasing commuting, or a diminished need for moves and
regional differences in well-being. Although the demographic potential of people
with the highest migration risk diminished slightly (Figure 1.11) it cannot be the
only explanation. A sharp statistical drop of migration activity had occurred already
before the decline of the share of the15–34-year-old people in the society.
45
40
rate /%
35
30
25
20
internal migration rate
2000
1998
1996
1994
1992
1990
1988
1986
1984
1982
1980
1978
1976
1974
1972
1970
15
% of 15-34 in population
Figure 1.11 Internal migration rate and share of 15–34-year-olds in total population (Author’s calculations from statistics)
47
A quite plausible explanation for the internal migration decline, or at least some part
of it, is non-registration of moves. In the light of a decline in the overall reliability of
migration statistics in the 1990s, it was the migration registration of young, around
15–30-year-old persons that became the most doubtful (Herm 1999, Tammaru and
Sjöberg 1999). In earlier years the students who changed their place of residence
temporarily were “signed out”5 of their place of residence and then signed in at their
place of study as temporary residents. Since the Estonian passports came to the
issued with only the permanent place of residence noted down, most of the students
who had changed the place of residence for the study period were excluded from the
migration data. Age specific migration rates revealed that the decrease of internal
migration had taken place mainly on account of the decrease of non-registered changes
of the place of residence of 15–34-year-old males and 15–29-year-old females (Herm
1996). In addition, the migration registration activity of students was dependent on
changes of social security benefit entitlement rights and other benefits offered by
local municipalities in the late 1990s.
Another plausible explanation for the migration decline might be increasing commuting, as a substitute for classical migration. A survey conducted by Tammaru
(2001c) revealed that the number of commuters had increased by nearly 40% since
the end of the 1980s in Estonia. A majority of the commuters still moved within the
came county and only 3% crossed the county borders. The age structure of the commuters suggests, that commuting may explain only part of the migration decrease.
An average commuter was older than an average person generally and they probably
commuted in order to have a job at all, rather than to have a better job (Tammaru
2001c). According to population census data of 2000 16% of those who changed
their place of residence in 1989–2000 remained commuters, while only 6% of nonmigrants practised commuting.
Internal migration of 1989–2000 according
to the population census of 2000
Lack of an explicit requirement for registration of residence and fictitious migration
registrations out of economic reasons (lower communal taxes, better schooling opportunities for children) is the reason why we will use only 2000 year population
census data to analyse the regional trends of internal migration from 1989–2000 in a
more detailed way. To get a picture of population shifts during those eleven years,
two population censuses from 1989 and 2000 are compared. Census data are undoubtedly the most reliable source for internal migration studies. Evaluation of the
reliability of census data are presented, e.g. by Tamm (2001) and Tamm (2003).
As during the census of 2000 only the present place of permanent residence and
the residence in 1989 were recorded. The differences revealed are not reverberations of overall migration, but represent the total net migration during 11 years. The
“Signed out” — a reference to a population registering system. Persons reported that they are
leaving their present place of residence and moving somewhere else.
5
48
respondents were asked if they were living at the same place as in 1989. If the
answer was “no”, the question about the previous place of living was asked. Change
of the name of the residence place is not considered as a change of residence and all
population data are recalculated according to municipality borders in 2000.
Because of the specific character of population census only the movement of
those people who had not emigrated and were alive at the end of the period (year
2000) was recorded. Population census data also included information about people
who had immigrated during that period, but this information will be not used for
analysis. In order to analyse internal migration without external influences Estonia
will be regarded as a closed system.
Migration trends were analysed along the geographical hierarchy of settlements.
Mainly two types of classification were used. The first was based on a rural-urban
typology and the size of urban municipalities. The second, more sophisticated classification lies on the urban-rural type and settlement function indicators (see Introduction).
General trends of internal migration
Population census data revealed that between 1989 and 2000 the total population of
the country diminished by 12.5 %. The drop of population was the greatest in towns
with a large non-Estonian population, and in small towns. A total increase of population took place only in the hinterlands of bigger towns (Kivilaid et al 2002). The
same is true for internal migration (Figure 1.12). International migration was a source
of total population loss in all settlement types and natural increase was slightly positive only in the hinterland of the capital (Tammaru et al 2003).
Altogether 199,000 persons, or 14.5 %, did not live in the same municipality in
2000 and in 1989. A comparison of men and women revealed higher migration activity of women. Also previous surveys conducted in Estonia have shown that women
have a higher migration activity (Marksoo 1987, Ainsaar 1994) and shifted life cycles, i.e. women’s migration intensity peaks earlier than men’s (Ainsaar 1994, Katus
et al 1999).
Remarkable was the migration intensity difference between titular and non-titular ethnic groups (Figure 1.13). Lower migration intensity of non-Estonians can be
indicative of some kind of barriers. Generally the internal migration age-curve of
Estonians followed the traditional shape of migration age structure with a slight rise
in women of older ages groups.
Altogether internal migration played a rather unsubstantial role in the urban population change (Figure 1.14) during 1989–2000. The main determinants of population change were international migration and natural increase (Table 1.5). Internal
migration was more important only in rural areas, where it supported population
growth, while natural growth and international migration diminished it (Table 1.6).
For the period 1989–2000 had witnessed most remarkable administrative transformations of urban settlements into rural ones.
49
Siserände jaotus
<-15
-14-0
0-10
>11
Figure 1.12 Net internal migration rate 1989–2000 (Population census)
50
45
40
35
% of group
30
25
20
15
10
5
0
<15
15-19
20-24
25-29
30-34
m Est
35-39
40-44
45-49
w Est
50-54
55-59
mothers
60-64
65-69
70-74
75-79
80-84
85+
wothers
Figure 1.13 Age specific migration rates by gender and ethnic group (mESt =
Estonian men, wEst = Estonian women, m other = men of other ethnic groups, w
other = women of other ethnic groups, Population census data)
50
capital
50 000-250 000 10 000-50 000
urban
urban
<10 000 urban
rural
per 1000 inhabitants
10
0
-10
-20
-30
-40
-50
netmigration
total change
Figure 1.14 Total population change rate and net internal migration rate in towns
with different population size and rural areas in 1989–2000 (Population census)
Figure 1.15 shows a crude rate of in — and outmigrants at the end of the period
according to the size and type of settlement. Crude net migration draws a quite even
picture of the loss of population due to internal migration in all urban settlement
types, and the growth of population due to migration in rural municipalities, typical
for deconcentration. The net migration loss was most significant in small urban settlements. Overall migration intensity was higher in rural areas. However the correlation between the size of settlement types and migration intensity is at least partly
the statistical result because of smaller territory of the municipalities.
Altogether the share of movers from-town-to-rural areas increased and fromcountry-to-town movers from all internal migrants decreased compared with previous periods (Table 1.7). The remarkable rise of rural-rural migration can be explained,
at least partially, with substantial changes in the administrative setup of settlements.
60
per 1000 inhabitants
50
40
30
20
10
0
-10
capital
50 000-250 000 10 000-50 000
urban
urban
outmigration
inmigration
<10 000 urban
rural
netmigration
Figure 1.15 Internal migration rate by settlement type in towns with different
population size and rural areas 1989–2000 (Population census)
51
Figure 1.16 shows that the influence of internal migration was not uniform in all
urban and rural areas, while the differences were rather significant. In the overall
group of towns, that generally lost population due to internal migration, satellite
towns were an exception. Together with rural settlements in the hinterland of towns,
these towns have grown most remarkably due to internal migration starting from
1989, and the rural periphery has shrunk. Internal immigration has been most salient
during 1989–2000 in the hinterland of county centres and in satellite towns.
Outmigration shows slightly lesser differences. Absolute (Figure 1.17) indicators
80
70
per 1000 inhabitants
60
50
40
30
20
10
outmigration
inmigration
Rural
periphery
Rural with
railway
Rural with
roads
Hinterland of
smaller
towns
Hinterland of
county
centres
Other towns
Satellite
towns
-20
County
centres
-10
Capital
0
netmigration
Figure 1.16 Internal migration rate by settlement type 1989–2000 (Population
census)
70000
60000
50000
40000
30000
20000
Inmigration
Outmigration
Rural
periphery
Rural with
railway
Rural with
roads
Hinterland of
smaller towns
Hinterland of
county
centres
Other towns
-20000
Satellite
towns
-10000
Capital
0
County centres
10000
Net migration
Figure 1.17 Internal migration in absolute numbers by settlement type 1989–
2000 (Population census)
52
reveal that the capital, county centres, and the hinterland of towns prevailed in the
migration picture with reference to absolute numbers. The county centres, their
closer rural hinterland, and capital also had a more salient absolute net migration.
Figure 1.18 shows the place of residence, in 1989, of people who moved between
two population censuses, and their final destination in 2000. We are particularly
interested in the sources of growth of two settlement types, which grew most because of internal migration between 1989–2000 — satellite towns and the closer
rural hinterland of big towns. Although these data do not suffice to draw final conclusions about the sources of net migration growth, it seems that the remarkable
growth of satellite towns and the hinterland of the county centres occurred because
of outmigration from towns.
rural periphery
rural with railway
living place in 2000
rural with roads
rural around small towns
rural around big towns
satellite towns
other towns exp satellite
county centres
capital
0
50
100
150
200
250
300
350
probability per 1000 inhabitants in 1989
urban other county
urban same county
rural same county
rural other county
Figure 1.18 Origin of people who moved in 1989–2000 (Probabilities per 1000
inhabitants, Population census)
In order to test the homogeneity of settlement types we classified all settlements
into four groups according to their net internal migration (positive or negative) and
the total change (positive or negative) (see Tables 1.8 and 1.9). It occurs that during
1989–2000 the total population grew only in 14 municipalities out of 247. The net
internal migration growth was observed in 111 municipalities. More than half of the
municipalities belonged to the group where both internal migration and total population growth were negative. Municipalities with a total positive growth and a positive internal migration were very rare — 3% (Table 1.8). Distribution according to
settlement size offers a rather homogeneous picture of towns as declining municipalities of which only less than one third experienced a positive net internal migration (Table 1.8). Rural municipalities also lost their population, but half of them had
a positive net internal migration.
53
Table 1.8 Municipalities by population size in 1989 and internal migration combined with total population growth between 1989–2000 (Number and % of all
settlements belonging to this group, Local Municipality Database)
Municipalities by internal
migration and total
population growth
>250 000
urban
50 000–
250 000
urban
10 000–
50 000
urban
<10 000
urban
Rural
Total
+internalm,+ total growth
–
–
–
–
+internalm, – total growth
–
– internalm,+ total growth
–
1
25.0%
–
4
30.8%
–
– internalm ,– total growth
1
100.0%
3
75.0%
9
69.2%
6
25.0%
2
8.3%
16
66.7%
7
3.4%
93
45.4%
5
2.4%
100
48.8%
7
2.8%
104
42.1%
7
2.8%
129
52.2%
100.0
100.0
100.0
100.0
100.0
100.0
Total
Table 1.9 Municipalities by type in 2000 and internal migration combined with
total population growth between 1989–2000 (Number and % of all settlements
belonging to this group, Local Municipality Database)
Internal
migration and
total population
growth
Capital
County
centres
Satellite
towns
Other
towns
Rural
close to
county
centres
Rural
close to
smaller
towns
Rural
with
large
roads
Rural
with
railway
Rural
periphery
Total
+internalmig
+ total growth
–
–
–
–
6
17.1%
–
1
2.4%
–
–
7
2.8%
+internalmig
– total growth
– internalmig
+ total growth
– internalmig
– total growth
Total
–
2
12.5%
2
12.5%
12
75.0%
100.0
5
50.0%
–
5
31.2
–
26
74.3%
–
11
68.8%
100.0
3
8.6%
100.0
17
41.5%
1
2.4%
22
53.7%
100.0
5
41.7%
–
5
50.0%
100.0
25
53.2%
2
4.3%
20
42.5%
100.0
19
27.5%
2
2.9%
48
69.6%
100.0
104
42.1%
7
2.9%
129
52.2%
100.0
–
1
100.0%
100.0
7
58.3%
100.0
Once again, larger diversities will pop up if we analyse groups of municipalities by
functional ranks. Table 1.9 shows internal differences within functional municipality
groups. Although almost all towns belonged to the group with declining total population and the majority also to the negative internal migration group, the group of
satellite towns was an exception. Half of satellite towns had grown because of internal migration. Almost all rural municipalities that grew and had positive internal
migration were situated in the hinterland of bigger towns. Positive internal migration
was characteristic of the majority of the “close to town” type settlements. 91% of
the municipalities in the hinterland of bigger towns had positive internal migration.
The opposite margin among rural settlements was the periphery. Internal negative
net migration was reported in 72% of municipalities in rural periphery. In sum, according to the criteria of total population growth and internal migration, the rural
hinterland of county centres was the most homogeneous settlement type.
54
Population census data showed that internal migration caused a decline of population at the top of the settlement hierarchy, a rather prosperous situation in the
middle, and a declining population situation towards periphery, which is indicative of suburbanisation rather than counterurbanisation.
Figure 1.19 presents the net internal migration growth according to the distance
from capital and county centres. A great loss of population due to migration was
visible in all centres. The population has grown because of migration both in the
closer hinterland of the capital and of the county centres. This is an obvious sign of
suburbanisation. There has been some growth also in more remote municipalities,
but this growth was more occasional and smaller in numbers.
2500
1500
500
-500
010
-1500
30
50
70
90
110
130
150
170
190
210
230
250
-2500
-3500
-4500
-5500
km
from county centre
from capital
Figure 1.19 Net internal migration according to the average distance (km) from
the capital and county centres 1989–2000 (Local municipality database from Population census data)
Demography of internal migration
The age structure of migrants reveals that the capital (Figure 1.20) has grown at the
expense of 15–29-year-olds. More young families with children and people older
than 30 left the capital than arrived there during 1989–2000. Satellite towns have
won population in all age groups, but especially in young families (Figure 1.22).
Quite similar to the migration age structure of satellite towns is the growth of the
rural hinterland of bigger towns (Figure 1.23). County centres and other towns, on
the other hand, have been the source of outmigration in almost all age groups (Figure
1.21). Only the group of 15–24-year-old women (female students) has increased
because of migration during the last tens years in county centres. Interesting is the
migration structure of more remote rural areas: despite the loss of up to 30-year-old
people, older immigrants aged between 45 and 75 have arrived in the countryside
(Figure 1.24).
55
Macro level explanations
Assuming that the general regional development on municipality level has been rather
linear in 1989–2000 (Figure 1.9), and the population change of recent years has a
little greater impact on net migration, data from 1995–2000 were used as macro
level background indicators to analyse migration changes. Many previous studies show
that population development depends on the functions and geographical position of
settlement (about Estonia see Tammaru 2001a). Capital and other centres often
fulfil the role of economic growth centres. According to this presumption we could
expect that distance from centres can have some influence on the overall spatial
processes. Economic well-being data will reflect the influence of economic push-pull
forces and new housing will be related to change of living conditions and environment.
In order to analyse the determinants of internal migration, correlation analyses of
crude net migration rate6 , total net migration, and crude total change were carried
out with municipality background indicators (new dwelling units by sq m built in
1991–2000, municipality budget per inhabitant, unemployment ratio in 1995–1996,
population size in 2000, distance from the county centre, distance from the capital,
inhabitants per 1 sq km) and population indicators (average income of an employed
person in 1996, poverty assistance per person in 1997, share of commuters among
20–65-year-old population).
Bivariate correlation analysis was used mainly as the first step of analyses, because
of probable danger of multicollinearity between variables. As it was conceivable that
in different settlement groups different forces can shape the migration situation, the
correlations were computed for each settlement type separately as well (Table 1.10).
Net migration rate can be interpreted as an indicator of the general attractiveness of a place. Crude internal net migration had strong positive correlations with
population change rate, new housing built during the period, average incomes, share
of commuters and strong negative correlations with unemployment ratio, poverty
assistance per inhabitant and municipality resources by inhabitant. Distance from
capital and municipality population size showed weaker but also significant correlations. During 1989–2000, municipalities with less unemployment and poverty, fewer
municipality resources per inhabitant, with more new residential space built per
capita, bigger individual incomes and less distant from the capital, rather than the
opposite, experienced a bigger growth of population due to internal migration. Distance from the county centre and population density did not have any statistically
significant correlations with the internal migration rate. Characteristically for
suburbanisation, a strong correlation between the net migration rate and the share
of commuters emerged. It indicates that a big proportion of migrants remain commuters after change of residence. Analyses of population census data showed that
this is particularly true in closer hinterland of bigger towns. Almost every fourth
person who moved to a rural municipality close to bigger towns was a commuter.
6
A crude rate is used here because of the particularity of data — only the Population census
internal migration sample. At the same time our aim was to standardise the net migration according
to settlement population size, in order to evaluate the real impact of net migration to settlement.
56
6000
6000
men leaving
women leaving
5000
men leaving
women leaving
5000
men arrived
men arrived
women arrived
4000
women arrived
4000
men
3000
men
3000
women
2000
2000
1000
1000
0
women
0
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
10
85
-1000
-1000
-2000
-2000
Figure 1.20 Internal migration in the
capital by gender and age 1989–2000
(Population census)
women leaving
men arrived
800
women arrived
men
600
20
25
30
35
40
45
50
55
60
65
women
75
80
85
men leaving
women leaving
2500
men arrived
women arrived
2000
men
women
1500
400
70
Figure 1.21 Internal migration in
county centres by gender and age
1989–2000 (Population census)
3000
men leaving
1000
15
1000
200
500
0
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
0
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
-500
-200
Figure 1.22 Internal migration in satellite towns by gender and age 1989–
2000 (Population census)
men leaving
8000
women leaving
men arrived
6000
women arrived
men
4000
women
2000
0
10
-2000
-4000
-6000
Figure 1.23 Internal migration in rural hinterland of bigger towns by gender and age 1989–2000 (Population census)
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
Figure 1.24 Internal migration in rural areas except close to bigger towns,
by gender and age 1989–2000 (Population census)
0.14
0.42*
–0.05
0.13
0.43
–0.01
Other town, N = 116
Rural close to county centre, N = 34
Rural close to smaller town, N = 47
Rural with large road, N = 41
Rural with railway, N = 12
Rural periphery, N = 69
0.13
0.14
0.13
–0.08
0.46**
–0.24
–0.08
–0.30
0.00
–0.10
0.02
** Significant on level 0.01, * Significant on level 0.05
0.24
Satellite town, N = 10
Crude net migration rate, N = 247
0.58*
0.03
0.15
Net internal migration, N = 247
County centre, N = 16
0.14*
Crude population change rate, N = 247
0.21
–0.06
–0.09
–0.15
–0.53**
–0.38
–0.47
–0.24
–0.13*
–0.16*
–0.10
0.00
0.02
–0.15
–0.19
0.26
0.19
–0.03
0.02
–0.13
–0.07
–0.21**
–0.26*
0.12
–0.36*
0.12
–0.53**
–0.10
–0.04
–0.43
–0.26**
–0.21**
–0.28**
–0.13
0.07
–0.13
–0.11
–0.29
0.19
0.21
–0.55*
–0.18**
–0.17**
–0.17**
0.12
0.30
0.45**
0.12
0.66**
0.11
0.06
–0.29
0.23**
0.12
0.19**
–0.01
–0.62
–0.09
–0.01
–0.18
–0.42
–0.20
–0.38
–0.23
–0.23**
–0.19**
0.35**
0.20
0.15
0.44**
0.71**
–0.11
0.95**
–0.09
0.42**
0.42**
0.45**
–0.25*
0.85**
0.36*
–0.59**
0.61**
–0.08
0.54
0.03
0.46**
0.44**
0.43**
Poverty
Share of
Average MuniciDistance Unemplo
InhaNew
assistance
Distance
commuPopupality
income of
from
yment
bitants
dwelling
resources
per
from
ters from
lation
a person
county
rate
per 1 km
1991–
per inh
person
capital
popusize 2000
1996 1996–2000 2000
centre 1995–96
1989
1997
lation
0.61**
0.81**
0.67**
0.63**
0.86**
–0.15
0.84**
–0.55*
0.64**
0.59**
1.00
Crude
population
change
rate
Table 1.10 Correlation of crude netmigration rate, net internal migration, crude population change with municipality level background data (Spearman correlation coefficients)
57
58
Among the group of migrants from smaller towns to closer hinterland this share was
19%. These numbers are even higher for working age population, because now the
reference group included children and old people as well. Still the relationship between commuting and net migration was not similar in all settlement types. Firstly it
was important only in rural municipalities and secondly in the rural hinterland of
smaller towns, and in peripheral municipalities commuting turned out to have a
negative correlation with net migration. It means that in those types of municipalities, areas with higher net migration had less commuters than others. Altogether it
refers to different mechanisms of settlement functioning on different levels.
Correlations of net internal migration indicators were similar to the migration
rate correlations, but with some minor exceptions. A disappearance of correlation
with incomes was the most significant change in case of net migration data. Quite
unexpected was the absence of a significant correlation with distance from county
centres, which is visible on the map (Figure 1.12). The reason might be the fluctuating essence of this distance. At the time when the closer hinterland of county centres enjoyed a clear growth rate, 50 km from the centres experienced a decline and
rather distant areas, again, a rise.
In different settlement categories, different factors seemed to play an important
role. New housing stock, share of commuters and unemployment were related
with the internal migration rate most frequently, while the other indicators had occasional importance. Net migration changes in rural hinterland of bigger towns was
best described by background indicators that were quite similar to the overall predictors of the country level. However, the small number of statistically significant
correlations within settlement types might be a sign of either a high unity within
group (missing differences) or a diversity of relationships.
In order to examine possible coinfluens of the internal migration determinants
on macro level we applied a binary logistic regression model with all background
indicators as independent variables and positive/negative migration as the dependent one. After the exclusion of insignificant predictors, the final statistically significant models were achieved (Table 1.11). The models show that five indicators have
had an influence on migration: new dwelling, incomes, population density, distance
from capital and unemployment in some groups of settlement. Different factors
seem to have influenced migration in different settlement types.
A model for the whole country showed that, the more new living areas were
built, the less densely populated was the municipality, and the higher were incomes,
the higher was the probability for expected positive migration. The all-country model
worked better as a predictor of negative net migration in municipalities than of positive net migration.
A separate view of different urban groups did not give any clear results for models. Distance from the capital was the only important variable for all urban internal
migration growth in general, but the model was quite weak. For all rural areas new
dwelling, income and population density formed a significant model. Still, within
different rural settlement types, the predictors of growth vary. The most interesting
was the model for peripheral areas where distance from the capital had a positive
impact on migration, while unemployment had a strong negative influence.
59
Table 1.11 Models for positive/negative net migration by settlement types (Binary logistic regression, best models for positive/negative net migration, Exp (B)
negative net migration =0, Local Municipality Database)
Type of municipalities
Arguments
New dwelling 1991–2000
Average incomes 1996
Unemployment rate 1995–96
Distance from Tallinn
Inhabitants per 1 sq km 1989
Nagelkerke R
All
municipalities
All
urban
All
rural
Rural close
to smaller
town
Rural with
large roads
Rural
periphery
1.008
1.001
–
–
0.998
0.341
–
–
–
0.985
–
0.193
1.008
1.002
–
–
0.997
0.349
1.014
–
–
–
–
0.437
–
1.005
–
–
–
0.370
1.009
–
0.376
1.026
–
0.401
Discussion of migration trends in Estonia 1989–2000
Internal migration data from 1989–2000 proved, that processes of deconcentration,
which started in the beginning of the 1980s, did not stop with political-economic
changes7 . On the contrary, the period from 1989 up to 2000 showed a continuation
of depopulation of bigger towns because of internal migration. In the last section of
the chapter some factors influencing migration were analysed. It was not possible to
test scientifically the impact of change of the economic system on suburbanisation,
because of the limited time series of economic municipality level data. Still a remarkable correlation between migration trends and new housing was observed, which
hints indirectly to a connection between real estate market and suburbanisation. The
second plausible explanation for outmigration from bigger towns can be higher price
level push forces, which might have more influence on older people. The inflow of
people to the countryside is sometimes explained with the effect of privatization in
the 1990s. In the course of restitution many urban people got back their land or
country house, which could inspire them to move to the countryside. Nevertheless,
the macro level data used in this chapter cannot either verify or deny this hypothesis.
It was found that those who left towns were more inclined to choose new residence places in municipalities with a higher average income. This result might still
suffer from self-dependency, because in a situation where movers are people with
higher incomes (see Chapter 2.2), their arrival itself will cause an average increase of
income in their region of destination. Still the fact about the interrelationship of
income and net migration remains valid, although income level cannot be interpreted
clearly as a reason of immigration. The situation is different in peripheral municipalities, where a correlation between unemployment and migration was found. It is a
much clearer sign about the influence of economic forces on migration.
7
Using different approaches of measuring urban units, we can get different results. About Estonia
see also, Tammaru et al (2003).
60
Analysis of different settlement types lead us to the conclusion that it is relevant to
treat settlement groups separately, because the results refer to different migration
forces working in different settlement types. Altogether satellite towns and closer
hinterland of bigger centres were the most popular settlement types for internal
migrants. Municipalities with railway, however clearly lost some of their previous
advantages.
Internal migration data from 1989–2000 showed large differences in migration
intensity by ethnic groups. We did not analyse this phenomenon more deeply in this
chapter, but the reason could be in some kind of cultural or information barriers and
borders of traditional settlement areas, mentioned also in previous studies (Ainsaar
1995).
0-15%
15-25%
> 25%
0-15%
15-25%
> 25%
Figure 1.25 Share of commuters of the 20–65-year-old population in 2000 (Population census)
Positive net migration area and commuting area (Figure 1.25) revealed a remarkable
coincidence in the late 1990s in Estonia. We can interpret the outcome as indicative
of specialization of space according to production and non-production functions.
Additional data would be necessary to test if the process of spreading commuting is
voluntary of forced because of concentration of workplaces. Indirectly the voluntary
character of outmigration from towns suggests the prevalence of voluntary movement. It is also remarkable that the commuting area follows mainly the network of
highways. It points to the fact that commuting is related to the freedom of selecting
one’s living place and better possibilities to achieve well-being. Still some surveys
have shown that commuting might be an outcome of lack of options rather than
61
voluntary selection of lifestyle (Jolkkonen et al 2003). This analysis did not give a
clear answer to those questions.
From the point of view of regional development, it is important to note that the
migration turnaround started from bigger towns. Consequently the reasons of migration turnaround should also be related to bigger towns. Price level, lack of dwelling,
availability of short distance jobs, time spent on transportation, infrastructure —
these can all be factors which influence town-to-rural migration. The Estonian data
showed mainly tendencies characteristic of suburbanisation — most people moved
not too far from centres, to areas with cheaper land, and a considerable amount of
new dwelling. Still, macro-level data processing enables only preliminary conclusions
as to the reasons of migration change. The reasons for turnaround are studied in
more detailed in Chapter 2.
The analysis of the migration flows allows of the assumption that in Estonia
deconcentration was driven by three factors in the main: (1) the spread of commuting decreased in-migration to towns, (2) the increase of housing opportunities outside towns supported out-migration from larger towns and, finally, (3) towns hold a
demographic potential necessary for outmigration.
62
Conclusions: General trends of population placement
In Chapter 1 we investigated the general rules of population placement in developed
countries on macro level. Different macro level determinants — administrative reasons, factors determining the selectivity of migration, demographic potential, regional differences, alternatives of migration, and freedom to move — have a different impact on migration. The overall migration pattern forms as a result of mutual
influence between all these factors.
There is no well-formed theory that could predict the regional development of
settlement. Instead, the role of the missing theory has often been fulfilled by descriptive lists of changes in time. Many theoretical approaches see the historical trend
of urban-rural development like a track on which society is passing stages of concentration and deconcentration. Still the majority of theories, covering regional development relate the changes with economic productive forces. However, productive
forces might not be the only explanative power, in a situation of diminishing economic and other costs of transportation and an increasing freedom of people to choose
their action spaces. The last decades allow making the assumption that in many
countries migration has been replaced by commuting and the people have become
more independent of the location of their place of work.
Literature shows an overwhelming dominance of lifestyle-related preferences and
housing reasons among the motives used to explain migration turnaround. The second group by frequency are economic reasons. The main forces that work against the
concentration of population can be partial replacement of migration by commuting,
improvement in the means of transport and communication, restructuring of production and a relative reduction of communication costs, and people’s deliberation
from work (decrease of the proportion of working people in the whole population, as
well as the freedom of people relative to the location of their place of work). It is
plausible that in different regions different circumstances prevail.
In different countries deconcentration processes emerged at different times, but
most countries have passed through this stage during some period of their development. Consequently, the apparent phenomenon that in some countries there has
been no suburbanisation would require a more detailed analysis. One possible explanation could be that a more flexible policy of city boundaries “has eaten” the
suburbanisation in these countries, so that the actual growth of the cities is smoothed
out. Another explanation could be related to the distinctive character of city environments (population density, milieu).
Chapter presents more detailed data about migration turnaround in the case of
Estonia. All urban-rural development of Estonia in the 20th century can be divided
into two periods: urbanisation up to 1983 and deconcentration from 1983–2000.
Internal migration has had more impact on rural population. A deeper insight into
the internal migration processes of Estonia during 1989–2000 revealed a continuous
outflow of people from towns into rural areas. This direction of migration persisted
despite the economic, political and demographic changes since 1983. Altogether,
the share of migrants from-town-to-rural areas showed an increase, while from-country-to-town movers decreased as compared with previous periods.
63
Population census data from the period 1989–2000 showed clear depopulation of
bigger towns and an inflow of people into the nearest rural hinterland of the towns.
Satellite towns and closer hinterland of bigger centres were the most popular settlement types among internal migrants. Municipalities with railway suffered clear loss
of the privileges they had previously. The capital and the county centres experienced
the greatest decline in absolute numbers of net migration. In relative terms, the
greatest changes took place in the hinterland of county centres and in satellite towns,
which gained 20–30 additional persons per 1000 inhabitants during this period. Large
numbers of people settled from county centres into the closer hinterland of towns
and from the capital reaching the distance of up to 60–70 km from the centres. The
main age group supporting suburbanisation were men and women of 25 and older in
the closer rural hinterland of bigger towns.
New housing was the most significant macro level indicator of positive net internal migration. Different factors seem to have influenced migration in different settlement types. We can conclude that the process of internal migration between 1989
and 2000 can be labelled mainly as suburbanisation, with rural economic selectivity.
As a methodological outcome of the Estonian data it can be mentioned that the
classification of settlements according to their geographic hierarchy provided a more
informative picture about the differences of internal migration between towns and
rural areas, than just using groups according to the size of settlements.
64
Appendix 1.1 Size of municipalities in some European countries in 1990
Country
Austria
Belgium
Bulgaria
Czech.Rep.
Denmark
Estonia
Finland
France
Germany
Italy
Hungary
Netherlands
Poland
Slovakia
Sweden
Total surface
area km2
83 857
30 518
110 912
78 864
43 093
45 227
338 145
551 602
357 041
301 277
93 033
41 864
312 683
49 036
449 964
Average surface
area of local
authorities km2
40
52
432
13
157
178
669
15
176
37
30
53
127
17
62
Average population
per local authority
3 340
16 960
35 000
13 730
18 760
5 930
10 870
1 580
4 925
7 130
3 340
23 200
15 560
1 850
4 930
% of municipalities with
population < 5000
inhabitants
91
17
8
95 (1991)
7
85
49
95
83
73
91
11
28
95
3
Source: The Size of Municipalities, Efficiency and Citizen Participation (1995), Local and Regional
Authorities. Europe No 56.
65
Appendix 1.2 Urbanisation and deconcentration processes from 1970–2000 in
developed countries
1970s
1980s
Urbanisation
Czechoslovakia – (Korcelli
1984)
Estonia – (Kümmel 1986,
Marksoo 1987, Tammaru
2001)
Finland – (Frey 1988a)
slowdown of concentration
in core regions (Vining and
Pallone 1982), number of
communes
with
neg
migration
decreased
(Tervamäki 1987)
GDR – (Frey 1988a)
Hungary – (Frey 1988a)
Ireland – (Champion 1992)
Italy – (Champion 1992)
Japan – (Korcelli 1984,
Frey 1988a)
New-Zealand
–
(Frey
1988a)
Norway
–
(Champion
1992)
Poland – (Parysek and
Kotus 1995, Frey 1988a)
Portugal
– (Champion
1992)
Spain – (Champion 1989)
Deconcentration
Australia – (Champion 1989)
Belgium – (Korcelli 1984,
Champion 1992)
Canada – (Korcelli 1984,
Frey 1988a)
Denmark – (Illeris 1980,
Frey 1988a, Champion 1992)
France – (Champion 1992)
FRG – (Korcelli 1984,
Champion 1992, 1989)
Italy – (Frey 1988a, Bonaguidi and Abrami 1996,
Champion 1989)
Netherlands – (Frey 1988a,
Champion 1992)
Sweden – (Ahnström 1980,
Illeris 1980, Korcelli 1984,
Frey 1988a, Champion 1992)
Switzerland (Illeris 1980,
Champion 1992)
UK – (Illeris 1980, Korcelli
1984, Frey 1988a, Champion
1989)
USA – (Berry and Dahman
1977, Long and Nucci 1997,
Korcelli 1984, Elliot and
Perry 1996)
Other
Austria
–
urbanisation
(Champion 1992), deconcentration (Frey 1988a)
Sweden – outmigration from
bigger towns, but still
concentration
(Borgegård et al 1995)
Netherlands – outmigration
to the most remote villages at
the time of concentration
(Gordijn and Eichperger
1996)
Austria – (Champion 1992)
Canada – (Keddie and
Joseph 1991)
Belgium – (Frey 1988a)
FRG – (Frey 1988a)
Norway
–
(Champion
1992)
Portugal
– (Champion
1992)
Sweden
– (Champion
1992)
USA – (Frey and Speare
1992)
Australia – (Dahms and
McComb 1999)
Belgium – (Champion 1992)
Denmark – (Frey 1988a,
Champion 1992)
Estonia – (Marksoo 1988)
France – (Champion 1992)
Germany
–
(Champion
1989)
Ireland – (King et al 1996,
Champion 1992)
Italy – (Champion 1992)
Netherlands – (Gordijn and
Eichperger 1996)
Poland – 1980–1985 (Parysek and Kotus 1995)
Sweden – (Frey 1988a)
Switzerland – (Frey 1988a)
Spain – 1981–1999 5 largest
cities lose population (Lanaspa et al 2003)
UK – (Frey 1988a, Stillwell
et al 1996, Dahms and
McComb 1999)
Finland – several reports support urbanisation (Reijo and
Valkonen 1993, Tervamäki
1987, Champion 1992)
Heikkilä (2003) show loss of
urban population and growth
of rural areas because of
migration
Sweden – positive net migration in urban areas in the
early 1980s, later deconcentration, but less numerous,
deconcentration led by families with children, and aged
> 45 (Borgegård et al 1995)
Poland – 1984 –1994 loss of
population in biggest towns
to the surroundings, migration gain from smaller settlements to 50,000– 500,000
towns of (Kupiszewski et al
1998)
Netherlands – 1980–1984
deconcentration, 1984–1988
concentration
(Champion
1992)
USA – outmigration from
metropolitans (Long and
Nucci 1997)
66
Appendix 1.2 Urbanisation and deconcentration processes from 1970-2000 in
developed countries (continuation)
1990s
Finland – (Aro 1997,
Laakso 1998, Tervo 2001)
Hungary – urbanisation
with suburbanisation (Kok
1999)
Poland – (Kok 1999, Rees
and Kupiszewski 1999)
Norway – (Rees and Kupiszewski 1999)
Russia – (Treivish et al1
1999)
Romania – (Rees and Kupiszewski 1999)
Australia – suburbanisation
(Fisher 2003, Walmsley et al
1998)
Czech R. – (Pavlik and
Kucera 2002)
Hungary – suburbanisation
(Döv ényi et al 1998, Illés
2000)
Italy – (Bonaguidi and
Abrami 1996)
Poland – suburbanisation
(Kupiszewski et al 1998)
Netherlands
–
suburbanisation in a smaller
amount (Gordijn and Eichperger 1996, Rees and Kupiszewski 1999)
Spain – 1981–1999 5 largest
cities lose the population (Lanaspa et al 2003)
UK – (Stillwell et al 1996,
Rees and Kupiszewski 1999)
USA – (Long and Nucci
1997, Johnston and Beale
1994, Nucci and Long 1995,
Fuguitt and Beale 1996,
Shumway and Davis 1996,
Lopez and Hynes 2003)
Estonia – reports about of
concentration
(Rees and
Kupiszewski
1999)
and
deconcentration
(Ainsaar
1999)
Finland – 1989–1992 increase of provincial centres
with 10,000–30,000 inhabitants, migrants moved out
from 10 largest cities. 1994
increase of largest towns
(Kauppinen et al 1997)
Italy, Portugal, Germany,
Czech Republic – change of
settlement system from urbanisation to deconcentration
(Rees and Kupiszewski 1999)
67
Appendix 1.3 Typology of local government units (Status according to the administrative division of 2000)
Tallinn
County centres,
cities
Haapsalu, Jõgeva, Jõhvi, Kohtla-Järve, Kuressaare, Kärdla, Narva, Paide, Põlva,
Pärnu, Rakvere, Rapla, Tartu, Valga, Viljandi, Võru
Satellite towns
Elva, Kehra, Keila, Kiviõli, Maardu, Narva-Jõesuu, Püssi, Saue, Sindi, Türi
Other towns
Kallaste, Kilingi-Nõmme, Kunda, Loksa, Mõisaküla, Mustvee, Paldiski, Põltsamaa,
Räpina, Sillamäe, Suure-Jaani, Tamsalu, Tapa, Tõrva, Võhma
Hinterland of
Tallinn, county
centres and cities
Audru, Haaslava, Harku, Jõelähtme, Jõgeva, Jõhvi, Kaarma, Karula, Kiili, Kohtla,
Luunja, Paide, Põlva, Pühalepa, Pärsti, Rae, Rakvere, Rapla, Ridala, Saarepeedi,
Saku, Saue, Sauga, Suure-Jaani, Tartu, Tõlliste, Türi, Tähtvere, Vaivara, Viimsi,
Viiratsi, Vinni, Võru, Väätsa, Ülenurme
Hinterland of small
cities*
Abja, Aegviidu, Alatskivi, Anija, Antsla, Halinga, Helme, Iisaku, Imavere, JärvaJaani, Järvakandi, Karksi, Kasepää, Kehtna, Keila, Koeru, Kohila alev — town,
Kohila, Kohtla-Nõmme, Konguta, Kõo, Kose, Lehtse, Lohusuu, Loksa, Lüganuse,
Maidla, Märjamaa alev — town, Märjamaa, Nõo, Orissare, Otepää, Paikuse,
Põltsamaa, Rõngu, Räpina, Saarde, Saksi, Sonda, Suure-Jaani, Tamsalu, Toila,
Tootsi, Vastseliina, Viru-Nigula, Väike-Maarja, Vändra alev — town, Värska
Rural with roads
Albu, Ambla, Are, Aseri, Avinurme, Haljala, Hummuli, Häädemeeste, Kadrina,
Kambja, Kanepi, Kernu, Koigi, Kõue, Kuusalu, Laeva, Lihula, Misso, Mäetaguse,
Oru, Õru, Padise, Puka, Puurmani, Risti, Roosna-Alliku, Rägavere, Saare,
Sõmerpalu, Sõmeru, Surju, Tabivere, Taebla, Tahkuranna, Tori, Torma, Tudulinna,
Valgjärve, Vasalemma, Vigala, Vändra
Rural with railway
Kabala, Nissi, Oisu, Olustvere, Orava, Palamuse, Palupera, Raasiku, Rakke,
Sangaste, Vastse-Kuuste, Veriora
Rural periphery
Ahja, Alajõe, Avanduse, Emmaste, Haanja, Halliste, Hanila, Illuka, Juuru, Kaisma,
Kaiu, Kareda, Kihelkonna, Kihnu, Kolga-Jaani, Kõlleste, Koonga, Kõpu, Kõrgessaare,
Kullamaa, Käina, Kärla, Käru, Laekvere, Laheda, Laimjala, Lasva, Lavassaare, Leisi,
Loodna, Lümanda, Martna, Meeksi, Meremäe, Mikitamäe, Mõniste, Mooste, Muhu,
Mustjala, Mäksa, Noarootsi, Nõva, Paistu, Pajusi, Pala, Peipsiääre, Pihtla, Piirissaare,
Põdrala, Puhja, Pöide, Raikküla, Rannu, Rõuge, Ruhnu, Salme, Taheva, Tali,
Tarvastu, Torgu, Tõstamaa, Urvaste, Valjala, Vara, Varbla, Varstu, Vastsemõisa,
Vihula, Võnnu, Vormsi
* Small cities — other cities and cities without municipal status (Abja-Paluoja, Antsla, Karksi-Nuia, Otepää)
70
69
Why people move
2
70
71
2 Why people move
Introduction
Migration research is more than 100 years old. After Ravenstein’s (1885) migration
rules, tens of different migration indicators have been used as explanatory factors.
Those indicators have had different names and coverage, and there have been many
attempts to classify the reasons for migration into different groups according to their
function. Although it may seem that migration studies have been quite well covered
by comprehensive research, several question have been left open. Why are some
reasons important and others not? Do people take into account one or several reasons? What determines the importance of one or another reason at a different time?
Why sometimes economic and sometimes lifestyle reasons prevail? How are different reasons connected with each other? Why are some people more mobile and
others more settled? And, finally, a question that has been repeatedly asked in migration studies: “Why do people move?” This chapter will touch upon all these questions.
Conditions affecting migration are often understood as “reasons” for migration.
Different authors use different classifications of the reasons and see their role from
different angles. The aim of this chapter is to find the impact mechanism of the main
factors of migration. These factors will be grouped according to the theory of needs,
well-being and traditions of migration research. The chapter includes an overview of
the influence of seven groups of factors on internal migration — population, selfactualisation, housing, economic conditions, social environment, natural environment
and distance. The chapter seeks to answer the following question: “Which factors
determine the mobility and barriers of migration in the society at a certain moment?” Part 2.1 is a theoretical overview of different approaches and grouping strategies of the reasons for migration. The rest of the chapter analyses the influence of
those reasons as barriers or sources of freedom to move. All sub-sections include
empirical data from Estonia. The subsequent growth of geographical differences in
economic well-being (see Chapter 1.4.2) would allow us to expect changes in the
reasons for migration, primarily the increasing importance of economic reasons which
started in the 1990s. Since migration turnaround in Estonia took place in 1983, we
will compare different reasons for migration before and after 1983 (i.e., during urbanisation and suburbanisation periods), where it is methodologically possible. Empirical analyses of data have two aims: (1) to give an overview of the motives of
migration in the conditions of migration turnaround, (2) to analyse the reasons for
migration in different population groups and in different directions of migration.
Since migration data are often controversial and might have a low level of reliability,
we will draw a final conclusion only if different sources of data give similar results.
72
2.1 Different groups of reasons and
well-being approach in migration research
Over the history of migration research, a variety of reasons have been used to analyse
migration patterns. There have been several attempts to unify motivational reasons
for change of residence under common groups and to give them some explanative
power. Theory of push and pull factors (Bogue 1969) is probably the most frequently used approach in migration research. Although there are some reports that
this approach may not always be suitable (Skinner et al 1983, Moon 1995), the basic
idea of pulling and pushing forces still remains meaningful. Some researchers believe
that there is even a symmetrical relationship between the factors that motivate people to leave a permanent residence and those that attract them to another specific
locus (Moon 1995, Shuval 1982). In 1940 Stouffer introduced “intervening opportunities” as one group of alternative pulling forces. A later classification used by
Simmons (1984) divided the reasons for migration into immediate and secondary
determinants. According to this approach, immediate determinants are the forces
that directly motivate migration (wages, job availability, education opportunities).
Secondary determinants are contextual forces and development strategies that differentiate wages, jobs, and schools in different rural and urban areas. This approach
is rather similar to Chamberlain’s (1988) findings about positive, negative and global
factors. He came to the conclusion that the existence of a second-order factor of
general well-being receives strong empirical support.
The classification developed by Clark and Onaka (1983), however, is somewhat
different. They divided the reasons of household relocation into three subgroups:
forced relocation, voluntary adjustment (housing — space, quality, design, cost, tenure change; neighbourhood — physical environment, social environment, public services, workplace, shopping, school, family, friends), and induced reasons. According
to this classification, induced reasons are changes in employment (job change, retirement) and life cycle (household formation, changes in marital status, change in household size). Without agreeing completely with the whole classification by Clark and
Onaka, it seems fruitful to distinguish induced reasons from other reasons. By nature, induced reasons are very similar to trigger forces. Wiseman (1980) used the
term “trigger issues” for induced reasons. According to him, trigger issues could be
life-cycle change, lifestyle change, age related losses, change of personal resources,
community ties.
The reasons for migration are seen from a new angle in Mulder’s (1993) theoretical framework, in which the author presents them as a tool to achieve the goals that
arise from parallel careers in the life course. According to this approach, individual preference, resources, and constraints influence the extent to which certain
events and circumstances in parallel careers lead to migration. Mulder considers the
household, housing, educational and occupational careers. The career providing the
goal that is to be achieved by moving, is designated a triggering career. Mulder also
sees a twofold influence of social context: (1) it has a direct influence on people’s
opportunities to realise the wish to migrate; (2) it influences individual resources,
constraints, and preferences. Individual resources and constraints are shaped by eco-
73
nomic circumstances and the social security system in the context of social acceptability of certain preferences.
Migration is a result of interaction between the environment and an individual.
Since migration is human spatial behaviour, the forces behind migration should be
similar to behaviour in general. The grouping of the reasons for migration should be
based on common behavioural reasons, in accordance with the theory of needs. The
grounds of the theory of needs lay in Maslow’s (1954) work. According to the main
argument of this theory all people have certain needs which can grouped hierarchically according to their priority. Needs are motivators of human behaviour. Later,
plenty of works have shown that the hierarchy of needs is not necessarily valid and
found other subjective aspects of Maslow’s theory. For example, Doyal and Gough
(1991) give an overview of different interpretations of human needs and argue that
needs are socially constructed and thus lack an objective standardised base. Needs
are dependent on culture. As many other researchers, Douglas and Cough distinguished between basic needs (nutrition, housing, non-hazardous environment, appropriate health care, security in childhood, significant primary relationships, physical) and intermediate needs. They argued that basic needs could endorse neither an
absolute minimum standard nor a culturally relative one. Needs change in time and
are different for different people.
The second useful approach applicable in migration studies is wellbeing approach.
While approaches of need concentrate more on necessities that are essential for
human functioning, the well-being approach treats different needs on a more equal
basis. However, well-being approach still takes into account all necessities important
for life (Ringen 1995, Filkins et al 2000, Christakopoulou et al 2001) and is therefore closely related to the theory of needs. One possible difference between those
two approaches can be that the theory of needs puts greater emphasis on objective
well-being, whereas well-being is used to a large extent to measure subjective evaluation and satisfaction. Subjective approach is also very appropriate in migration studies (Schulze et al 1963, Fernandez and Dillman 1979, Basu 1992, Liao 2001, Theodori
2001).
A distinction between positive and negative components of well-being could also
be interesting for migration research. Positive well-being implies satisfaction, whereas
negative well-being can be interpreted as ill-being (Chamberlain 1988:584). Despite
the fact that positive and negative approaches to well-being are mainly used in the
affective dimensions (600), we argue that in migration research it can well be interpreted in the framework of push and pull forces. Perceived positive well-being suits
to explain pull factors, or in some occasions obstacles, and negative well-being is
related to push factors. After analysing large amounts of literature, Chamberlain
(1988:586) concludes that negative and positive affects can be independent: “separate dimensions of well-being are not just opposite ends of bipolar continuum. Those
things which make people happy are not the same which make them unhappy”. In
traditional migration studies it means that push factors are not necessarily antipodes
of pull factors.
In order to classify the reasons for migration into groups we use Migrant Survey
(the data is described in Chapter 2.2). The survey fits our research well because it
provides a comprehensive list of pull and push factors, which were active in the
74
migration process. All migrants filled the stratified questionnaire about their decision-making criteria and process: evaluation of conditions considered as push and
pull factors and the importance of those factors for them1 .
A factor analysis of a combination of push reasons (reasons for leaving), pull reasons (reasons for choosing the place of residence) and the age of migrants produced
eleven combinations of migration incentives (Appendix 2.2):
1. The pull factor of larger centres, which was also related to the level of services
in these centres. This described the largest amount of variances, which is
16%.
2. Employment (related to income and having a job).
3. Reasons related to environment and security, including neighbours.
4. Housing.
5. Household career reasons (moving to parents’ or relatives’ home, closeness of
friends and relatives, starting a life together with the spouse, moves related to
well-known places).
6. Self-actualisation. The younger generation of migrants: push factors associated mostly with studies, starting a job, moving away from parents.
7. Distance (between job and home, condition of roads).
8. Change in the household structure (changes in family composition that were
mostly associated with changes in marital status, i.e., moving away from or to
the spouse).
9. Social capital (problems related to children, school, language, dissatisfaction
with a job).
10. Elderly people’s desire for the change of environment.
11. Separation.
In all groups, except the first (1) and the last (11), both pull and push factors were
represented. The first group included only pull factors and the last group only push
factors. This indicates that although pull and push factors may be asymmetrical they
usually appear together.
Table 2.1 summarises migration incentives from different approaches under seven
main groups: population, housing, economic resources, social resources, natural environment and distance. This division will be used later in the book in order to analyse
empirical data. According to our hypothesis, factors often used in migration research,
like ethnicity and gender, do not influence migration by themselves. Instead, the
differentiation mechanism of the results of migration works through the indicators
described in Table 2.1. However, since a great deal of literature on migration discusses the connections between migration and gender, we examine it separately later.
Principal factors influencing migration are often the same on the macro and individual levels, although different labels or names have been used in different approaches. Table 2.1 also reveals the analogies between the theory of well-being and
the theory of needs and their correspondence to the classical reasons for migration.
Two last columns of the table display a link between the most commonly used migration data and the theories of well-being and needs. Migration factors coincide to a
1
Table 2.4 reflects the basic structure of the questionnaire.
Housing market
1
income,
Climate
Distance, knowledge of place or Time, security
lack of information, cost of
movement
Preferences, health needs
Needs
Need to maintain control,
security
Physical environment for
basic needs
Love and belonging, need
for
affiliation,
safety,
intimacy
Self-actualisation as a tool
for achieving other needs,
physiological needs
Self-actualisation, feeling
of being important, need
for achievement
Need for personal space
Social
contacts, All individual needs
networks, social capital
Well-being approach
Human capital status,
power, influence, comfort, stimulation
Housing
Quantity and quality of
housing
Employment, income
Wealth, income and
expenses, availability of
services
Because of other family member, Quantity and quality of
Need for closeness, need for social contacts, social
separation, social and security capital.
needs, services etc.
According to Maslow’s classification - motivation towards attaining one’s intellectual potential
Distance
Employment,
unemployment
Housing
Social relations, civil status, social services, cultural services, security, comigration, distance from
relatives etc
Better climate, natural
environment,
urban
environment, etc
Distance,
communi- Information, perception
cation facilities
of distance, costs of
movement
Regional differences in
employment,
salaries,
poverty, investments
Social
Availability of services,
environment and number of relatives,
capital
informants
etc,
communication,
comigrants
Natural
Climate, natural or artienvironment
ficial environment
Economic
resources
Housing
Macro level
Migration research
Individual level
Most commonly used wording of
reasons
Population
Population age structure Age, life stage changes, Changes in civil status (marriage,
sub-groups and size
gender, preferences
etc) and household composition,
need
for
different
tenure,
preferences, etc
Self-actualisation1 Number of school places, Career possibilities, social Studies, promotion possibility,
job opportunities
status, studies
recreation possibilities
Factor
Table 2.1 Macro level and individual level factors and explanations used in migration research, well-being approach and the
theory of needs
75
76
great extent with indicators used in the theories of well-being and needs. A logical
connection between the indicators in the research on needs, well-being and migration implies that it could be possible to unify approaches under the theory that is
based on needs. The possibilities of meeting one’s needs and their actual realisation
are often monitored in the research on well-being. In that content, voluntary migration is a type of behaviour that helps people meet their needs better and achieve the
condition of well-being. Some difficulties in linking theories of migration and theories of needs might arise from the specific character of the individual data on needs
and macro level migration data.
Perceived
well-being factors
in the
environment,
neutral, push, pull,
and triggering
factors
Individual
needs, change of
needs or
perceived
environment
An idea about
migration
Type of migrants,
motivation and
freedom to move
Gainscost +barriers =
benefit
Response
1. Migration
2. Change of
environment
3. Adjustment of
needs
Figure 2.1 A simplified model of the factors for migration
There are at least five important stages in migration that have the greatest impact on
the whole process: (1) individual needs, (2) perceived differences in environment,
(3) response to misbalance between needs and environment, (4) genesis of migration intention, (5) and individual characteristics of persons involved in this process.
Figure 2.1 reflects a simple model of this system. Interestingly, only a few factors of
all potential well-being factors become triggering, push and pull agencies. One possible explanation to the selectivity of triggering reasons can be the lack of perceived
differences in environment. Some conditions, which are similar to all regions, might
remain neutral from the point of view of migration. Most of the literature on migration does strive towards determining clear differences between push, pull, and triggering factors. In this work, triggering factors are defined as the factors that instigate
migration desire because of changes in environment or needs. Push and pull factors,
on the other hand, are permanent factors pleasant or unpleasant that could also
become factors influencing migration.
The present settlement situation can be described as a moment of equilibrium,
in which the needs, environment, and personal resources are in a balanced situation. Any occasion that changes this balance could become a triggering reason
for migration.
77
Research on the individual level of migration shows that people tend to behave differently in similar situations. Migration could be seen as a way to achieve greater
personal satisfaction or the well-being of one’s family. Households have different
inclination towards migration and differing degrees of freedom in making decisions
about migration (Golledge and Stimson 1990), which might be the reason why households with a similar socio-economic status tend to respond similarly to external and
internal stimuli conditions of migration (Hartshorn 1992). There is a good reason to
believe that the pattern of behavioural reasons is a summary of many individual
characteristics, and that its roots could be explained through the theories of needs
and behaviour. The type of a migrant (determined by various individual features
such as life stage, education, experience, etc) is closely linked to personal needs
changing in each life stage, and to personal resources. Migration can be interpreted as
an attempt to meet the needs through moving to another place. One feature characteristic of migration is that only those requirements, the fulfilling of that is dependent on regional differences, can be met by moving to another place. If a person is
already considering the idea of moving, he or she will make a decision about possible
responses. We can assume that different people use different strategies. According
to the rational model, people will subjectively take into account possible gains, costs
and barriers, and then make the decision. However, changing the environment by
migration is only one possible solution from several coping options. There are many
other coping strategies for the situations where needs and perceived environmental
opportunities do not match. This explains why people with good migration reasons
do not always change their place of residence. However, in this chapter we will not
give a detailed analysis of all possible responses and their mechanisms, but will concentrate on a more general description of migration processes.
The other reason for not changing a place of residence, besides lack of needs, can
be migration barriers. Many factors influencing migration can be perceived as both
stimuli and barriers. Migration decision is made within the normative structure of
social limitations, alternatives and barriers restrict the scope of individual choice
(Gang and Stuart 1999), that is, freedom to move. Similarly to triggers, the risk of
mobility barriers might be distributed over the life span in a systematic way (Foss
1984). Still, there are some relatively universal barriers. Some of the most common
individual barriers are economic resources needed for migration, information, and
distance. Nieminen (1983), for example, found that the main problems related to
migration in Finland in 1977–1978 were finding suitable housing (29%), job (24%),
and leaving friends behind (24%), all of which can be treated as migration barriers.
In addition to external barriers, an individual decision threshold (see Morrison
1972) might become an internal barrier and determine the possible response to objective migration factors. Table 2.2 illustrates mutual relations between different
barriers and incentives pertaining to migration (costs and benefits). The approach is
based on the bounded rationality model of behaviour. According to this model, the
reasons for migration will be formulated only when the costs or benefits of staying or
moving rise above the decision threshold. But even in that case migration cannot take
place before the barriers of migration have been overcome. This rather plain interrelationship can explain several unsolved phenomena in migration research: firstly, why
people react to similar migration stimuli with a different probability of migration;
78
Table 2.2 Response to the need of migration and the combination of decision
threshold and balance between costs and benefits of migration
Low decision
threshold
High decision
threshold
Benefits > costs of migration
Migration, frequent movers
Costs > benefits of migration
Dissatisfaction, other coping strategies
Wish to move, but less frequent Keeping the same situation, dissatismovers or non-movers
faction, psychological coping
secondly, why we find people in different stages of migration (for example, wishing
to migrate but not doing so because of barriers, too high cost or too low benefits;
people with good reasons for migration remaining settled because of high decision
threshold or people continuously on the move).
The following chapters analyse possible incentives and obstacles in the migration
of different groups of people. Chapters 2.4 — 2.9 analyse migration patterns according to factors presented in Table 2.1. For that purpose, references on previous studies
and original empirical data from Estonia are used.
2.2 Data
Two main sources of data are analysed in order to investigate the reasons of migration
during the 1990s in Estonian — Migrant Survey and Living Conditions Survey.
Migrant Survey was designed and conducted with the specific purpose to analyse the formation of the migration decision and the reasons for migration. It was
conducted as a retrospective mail survey by the author in 1997. A random sample
was formed from the Population Register data from those people who had registered
their change of residence from one municipality to another in 1995. 1500 questionnaires (230 in Russian language) were sent to the new addresses. There was also a
request on the envelopes to return the response letter even if the persons to whom
the questionnaire was addressed did not live at that address. The researchers received 668 answers (60 in Russian). The results showed that 55.6% of the respondents had changed their place of residence from one municipality to another, 23.9%
had only registered their migration in 1995, but lived at the registered destination
already before, 13.1% moved within the same municipality, and 7.2 % registered
their move, but did not move in reality. 179 responses from the people who had not
changed their place of residence or had done it before 1989, were omitted from the
subsequent analysis. Consequently, only those responses were considered for further
analyses that indicated a change of residence in 1989–1997. Altogether, 489 questionnaires with reports about movement in 1989–1997 remained for the analyses.
The majority of the changes of residence (63%) occurred in 1995, 20% in 1989–
1994, and the rest in 1996–1997. Of all movements, 81 were rural-urban, 137 urban-urban, 94 urban-rural, and 124 rural-rural. A later comparison with population
census data (Table 1.8) revealed only 8% overrepresentation among rural-to-rural
movers and 8% under representation among urban-to-rural migrants. However, differences in direction can be partly explained by different time scopes, as census data
79
covers the period of 1989–2000 and survey data only the beginning of the period.
The Migrant Survey questionnaire posed questions on the satisfaction of the respondent with the previous and present living environments, amenities and household income, as well as on the triggering reason of migration, range of possible pulling
reasons, factors affecting the selection of the new living place, and barriers. As an
addition to structured question blocks, the respondents were asked to specify the
triggering reasons in an open question and the answers were later coded into six
groups (Appendix 2.1). The influence of possible pull and push forces was examined
with the list of structured questions. The structured questions were also grouped
into classes of reasons (Appendix 2.1). Figure 2.2 gives an overview of the general
structure of the questions concerning migration used in the survey.
Satisfaction in origin areas
Triggering
reasons
Push forces
Barriers
Pull forces
Satisfaction in destination areas
Figure 2.2 Composition of the Migrant Survey
The weakness of the survey was a low response rate (45%). In part, this could be the
fault of the low reliability of the addresses in the Population Register, because
registrating the place of residence was basically voluntary. However, the age curve of
the respondents (Appendix 2.3) was quite similar to the migration curve as indicated by the population census (Figure 1.13), which allows us to assess the responses
as representative by age. Also, the representation of respondents in the main directions of migration is rather good, which allows to see this survey as a trustworthy
source.
The second source of data used in this chapter is the Living Conditions Survey
from 1999. The survey offers an opportunity to compare reasons of change of residence before 1983 with the later period in Estonian migration research history. The
survey was conducted by Estonian Statistical Office, Ministry of Social Affairs and
the University of Tartu in 1999. The survey contains only a limited range of questions on migration history — the time of the last migration, previous living place, the
reason for migration, and future migration plans. All retrospective data were collected only regarding the last move. Assuming that some people tend to change the
place of residence more often than others, the movement of the more frequent
movers is probably underestimated in our results. The other disadvantage of the
Living Conditions Survey was a rather modest set of reasons for migration. Only four
types of reasons — purchase of accommodation, change of the job or getting a new
job, studies and family reasons — were identified in the survey. The initial database
was weighed against the last population census results (see Oja and Tiit 2002). From
4,726 respondents, 70% had an experience of migration in their lives, 628 of them
had moved in Estonia between 1989–1999, and 447 between 1983–1988. For the
80
analysis, all migration history of migrants was divided into three periods: before 1983,
1983–1988, 1989–1999. 1989–1999 coincides best with the last census data and is
of particular interest to our study. We will also look separately at the period of 1983–
1988, because 1983 was the turning point in rural-urban migration and we can expect similar trends from the later period. Thus, we use the period before 1983 as a
reference stage.
The main differences between the two surveys are the time span covered by the
research and the way of posing questions on migration. In the Living Conditions
Survey the respondents were asked to specify only one reason from a very limited
range (employment, studies, housing and family) but Migrant Survey gives a much
richer range of explanations. Different clusters of reasons for migration in the two
surveys do not allow an entirely synchronised comparison of migration motives, but
they afford to give a rather broad overview of the general motives. Appendix 2.1
gives an overview of the principles of classification of the reasons for migration that
are used in this chapter, if not indicated otherwise. The reasons for migration are
classified into six groups — household career, housing, economic, self-actualisation,
social environment, natural environment, and accessibility reasons. Several migration factors could have a number of explanatory meanings. To get the best classification of the reasons for migration, the classification was adjusted during the work
process according to the results of the factor analyses (see page 74). As seen in Table
2.4, all clusters of reasons were not surveyed symmetrically in the Migrant Survey.
Therefore, the values of reasons for migration were calculated as an average of different reasons belonging to this group.
Living Conditions Survey gives a general overview of migration in different periods and the role of different population groups in it. Migration Survey allows a more
detailed analysis of data by groups.
2.3 Prevailing reasons for migration and
deconcentration
The reasons for migration reveal the secrets of moving. The reasons are the guiding
forces, the real background behind the forces that trigger, pull and push people from
one place to another. Without reasons, migration would cease to exist. Therefore,
the motivation for migration not only reveals why people move, but it can also tell
what people long for, what they want to achieve, and who they would like to become. The choice of moving or not moving is simultaneously influenced by many
different factors. Migration is prompted by triggering reasons, but other factors relating to people’s needs and well-being are probably also considered when thinking
about changing the place of living. A change in the balance of the forces that urge
people to move or stay is required in order for a person to move. This change can be
brought about by changes in the personal status or in the environment.
Different reasons for migration prevail in different development stages of society.
It has been argued, for example, that, at a time when there is a large supply of labour
market and a rapid expansion of production, the most important factors are the
presence of job opportunities and the distribution of production facilities
81
(Stankuniene and Sipavitsiene 1989). This leads to an impression that migration
depends mainly on the distribution of production facilities and, although the factor
of living conditions also influences the situation, it remains of secondary importance.
In a situation where labour demand exceeds the demographic potential the situation
changes significantly. Behavioural motivation becomes more important. There are
more opportunities to choose the place of living and working, the factor of living
conditions and subjective factors play an increased role, and there is more freedom
to realise personal interests.
The differences between supply and demand can be easily observed in different
directions of migration. For instance, it has been found that the intensity of moving
to cities is determined by the development of the city — the extent of investments
and available housing, functional diversity, services, job opportunities, and differences in the state of living (Moissejenko and Mukomel 1982, Rybakovsky 1987,
Nieminen 1983 et al). In case of outmigration from the cities, the decisive factors
are the demographic structure of the city and availability of housing (Moissejenko
and Mukomel 1982, Nieminen 1983).
Several authors have expressed the idea that simultaneously with the development of society, the increase of general wealth, and spreading of communication
networks, the influence of distance decreases, and people’s subjective preferences
and pulling factors be especially important (Jagijelski 1980, Cadwallader 1989). Subjective preferences might become more decisive in the case of short distances, e.g.,
moving within the city (Cadwallader 1989). Findlay and Rogerson (1993) explain a
tendency towards freedom of choice with greater flexibility of employment — an
increasingly higher proportion of population is becoming more independent of fixed
office locations and this gives people more freedom in the choice of dwelling.
Still, if we would like to make forecasts about the dominating reasons in society we assume that this reason can be identified as the most deficient factor that can
be improved by migrating. Perhaps this hypothesis might for example find some
support from a survey, which indicated that people in Eastern and Northern Finland
tended to attach more importance to work in the migration process, and people in
the South to housing (see Söderling 1983).
Prevailing reasons in Estonia
Many studies show that the motives of migration differ, as do the groups of migrants
themselves, depending on the direction of movement. A trend of internal migration
from larger towns into the countryside, which started in 1983, prevailed throughout
the 1990s in Estonia. However, moving to the towns remained popular among
younger population groups. Previous studies have shown that in Estonia, the dominant reasons for internal migration into urban areas have been family-related motives
(Ainsaar 1990), employment (Ainsaar 1990, Tammaru and Sjöberg 1999), and living
conditions (Ainsaar 1999). Migration to rural areas was primarily caused by the desire to improve living conditions (Lõo 1987, Tammaru and Sjöberg 1999, Ainsaar
1999), enhance salary opportunities (Lõo 1987) or, in the case of specialists, a suitable job (Kuddo 1988). Until 1990, migration into rural areas was also undoubtedly
82
influenced by the state policy of compulsory job placement upon graduation from
the university. Generally, the motives leading to migration were studied rather superficially during the 1990s.
Knowledge about the general development of society and migration trends allowed us to suggest some hypotheses about migration in Estonia during the 1990s.
We predicted that migration caused by economic reasons might have increased, different migration motives prevail in rural-urban and urban-rural direction and finally
that suburbanisation has been accompanied by a change in the reasons for migration.
Table 2.3 Reasons for migration in different time periods (%, Living Condition
Survey)
1989–1999
1983–1988
<1983
Total
Apartment
23.6
16.4
Employment
24.4
34.3
Studies
6.9
4.7
Family
34.6
37.4
Other
10.4
7.2
Total
100 (N = 627)
100 (N = 446)
10.1
14.7
31.5
30.2
10.5
8.5
40.9
38.6
6.7
7.7
100 (N =1367)
100 (N = 2440)
The Living Conditions Survey revealed remarkable changes in the reasons for
migration during the 1990s (Table 2.3). The most essential reason was the rising
number of moves because of housing in the period of 1989–1999. Family reasons
maintained their priority and employment related migration declined, in contrast to
our expectations. Moves related to studies formed a rather minor part of all moves.
Figure 2.3 depicts a ranking of trigger factors and all factors taken into account in
the process of migration in the beginning of the 1990s according to the Migrant
Aveage importance of factor
60
50
40
30
20
10
0
-10
-20
household
career
housing
economic
all reasons
social
environment
selfactualisation
natural
environment
triggering
Figure 2.3 Triggering reasons compared to all other reasons taken into account
in migration (Average * 100: –1 — did not take into account, 0 — it was not important
to me, 1 — took into account to a certain extent, 2 — it was important, Migrant Survey)
83
Table 2.4 Push and pull factors of migration as studied in the Migrant Survey
(% of people who took the respective factor into account in their migration decision)
Reasons for leavingPush factors
Economic
Absence of professional job
Unemployment
To work after graduation
Low income
Other work related reasons
Conditions (amenities, price)
To have a bigger flat, house
To have a smaller flat, house
End of rent contract
Changes in family
Break up with partner
Emancipation from parents
Other family members
Neighbours, workmates
Because of children
Away from relatives
Ethnic composition
25
25
12
32
20
55
41
4
8
35
16
23
27
10
30
11
17
Self
actualisation
Studies
11
Environment
Natural environment
Change of living environment
Job’s distance from home
34
62
18
Housing
Household
career
Social
environment,
services
Accessibility
Influenced a decision about new place
of residence. Pull factors
43 Possibility to get a job
45 Better income
65 Better living conditions
23
21
15
13
30
37
42
19
15
11
31
38
44
Cohabitation
To parents’ home
To relatives’ dwelling
Neighbours
Level of schools
Closer to familiars
Security of closer environment
Phone connection availability
Level of local health care
Studies
Cultural and social life
Recreation possibilities
Environment
26
17
40
53
31
Job’s distance from home
Roads
Distance from centres
Knowledge of place
Availability of consumer goods
Survey. The results of the two surveys were quite similar. The Migrant Survey also
showed that the dominant triggering factors were related to household career and
family life. Housing and economic reasons remained less important.
The results would be somewhat different if, instead of triggering reasons, all factors involved in migration decision-making would be considered. These are the factors that influenced the migration decision, but were not necessarily the most important factors that caused the move. The results shown on Figure 2.3 represent the
average values of all assessments by the respondents. The value of an assessment
could vary from a factor being firmly taken into account to a definite refusal to take
this factor into account. In the latter case the assessment was coded as negative. If
the factor did not have any importance whatsoever for the person, it received a value
of 0 and did not influence the average value2 . On the average, people claimed that
2
This coding is used also later in analyses of the Migrant Survey
84
the main factor they took into account when choosing a place of residence was the
environment, housing was the next in importance, and economic factors were third.
The Migrant Survey allows us to take a closer look at the specific division of
reasons according to the push and pull forces (Table 2.4). Although this table only
includes the responses of the people who considered the respective factor important, instead of average values, the method still gave similar results to the previous
ones. Reasons related to the living environment were the main factors causing migration (62%). The second group of push factors was formed by several factors related
to housing, next came factors related to the changes in household composition and
employment status. The factors that were most often considered when choosing a
new place of residence, or pull factors, were better living conditions (65%), good
reputation of the location (53%), income (45%), environment (44%), employment
opportunities (43%), security (42%), and distance from a centre (40%). Consequently, most of the desired criteria of a new place of residence were characteristic
of a larger centre or its close suburbs.
Reasons for deconcentration
Family and employment have been the primary reasons for migation throughout the
years in Estonia. Still, there were variations of reasons in different directions of migration (Table 2.5). Comparison of urban-rural and rural-urban directions reveal that
dwelling has been much more important reason in moving from towns to rural areas
and job and studies prevailed in urban arrival, then in other directions. There was
comparatively less differentiation in family moves.
There were also important age-motive differences in both migration directions
(Figures 2.4 and 2.5). Firstly, it was obvious that the rate for urban in-movers was
generally much higher than for rural immigrants, because of their large numbers.
Urban in-movers were younger and moved in younger ages more because of studies
and family reasons. A particularity of migration into rural areas was a bigger share of
200
perthousand
thousands migrants
per
migrants
180
160
140
Other
120
Family
100
Studies
80
60
Job
Apartment
40
20
0
<15
15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59
>59
Figure 2.4 Age and reasons for rural-to-urban migration 1983–1999 (Living Conditions Survey)
85
80
per thousands migrants
70
Other
60
Family
Studies
50
Job
40
Apartment
30
20
10
0
<15
15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59
>59
Figure 2.5 Age and reasons for urban-to-rural migration 1983–1999 (Living Conditions Survey)
rural 1983-1999
other towns 1983-1999
capital1983-1999
rural <1983
other towns <1983
capital <1983
0%
10%
20%
30%
Apartment
40%
Job
50%
Studies
60%
70%
Family
80%
90%
100%
Other
Figure 2.6 Reasons for in-migration, time periods and destination type (Living
Conditions Survey)
rural 1983-1999
other towns 1983-1999
capital1983-1999
rural <1983
other towns <1983
capital <1983
0%
10%
20%
30%
Apartment
40%
Job
50%
Studies
60%
70%
Family
80%
90%
100%
Other
Figure 2.7 Reasons for iout-migration, time periods and destination type (Living
Conditions Survey)
86
Table 2.5 Migration by direction and reasons for migration (%, Living Conditions
Survey)
Urban-rural
Urban-urban
Rural-urban
Rural-rural
Total
Apartment
Job
Studies
Family
Other
Total
23.3
8.6
5.6
19.4
11.8
25.3
28.7
31.3
27.8
28.9
0.7
10.0
12.6
0.8
7.7
37.8
43.1
42.4
43.7
42.3
12.8
9.4
8.0
8.3
9.2
100 N=407
100 N=1095
100 N=1017
100 N=661
100 N=3180
Reasons of positive or negative net
migration
housing migrants in ages 20–34. These variances were also statistically significant
(Kruskal-Wallis test).
Comparing the reasons for in-migration before and after 1983 we can see the
increase of dwelling-related moves both in to rural and urban areas (Figure 2.6).
Moving to rural areas has always been more closely related to housing than moving to
towns. Another important change was the increase of employment-related migration into the capital, but not into other towns. On the whole, the biggest shifts took
place in migration related to the capital.
There were also clear differences in the reasons for leaving by settlement types
before and after 1983 (Figure 2.7). The most remarkable change in out-migration
was the more than double increase of share of housing migrants among people who
left the capital in 1983–1999. The share of housing migration increased in all areas,
but this increase was less significant outside the capital. Other reasons for migration
had changed less.
As a result, in 1983–1999 rural areas gained people mainly because of non-economic
reasons such as dwelling and family, and towns because of economic reasons (Figure
2.8). The capital attracted people also as an education centre. On the whole, economic factors did not essentially affect the migration balance and, therefore, did not
influence the proportion of rural-urban population.
130
110
90
70
50
30
10
-10
-30
-50
-70
other
Family
Studies
Job
Apartment
capital 1983-1999
other towns 19831999
rural 1983-1999
Figure 2.8 Net migration by reasons and settlement types 1983–1999 (Living Conditions Survey)
87
<1983rural-urban
1983-1988r-u
1989-1999r-u
<1983urban-rural
1983-1988u-r
1989-1999u-r
0%
10%
20%
30%
Appartment
40%
50%
Change of job
60%
Studies
70%
80%
Family
90%
100%
Other
Figure 2.9 Reasons of urban-rural migration in different time periods (Living
Conditions Survey)
Figure 2.9 shows the reasons for migration in different directions of moves. We are
mainly interested in the motivation of the migration from rural areas to towns and
from towns to rural areas. A comparison of the two time periods reveals that the
most remarkable changes had occurred in the urban-rural migration. Migration from
towns into rural areas because of housing increased over the course of years. However, employment-related migration into rural areas had been constantly decreasing.
The reasons for rural-urban migration had remained quite unchanged, with minor
variations. The most important change in the movement from rural areas into towns
was the decreased importance of migration related to studies. If before 1983 the
proportion of migrants moving from rural areas to towns because of dwelling was
very small, the doubling of this small proportion appeared as a fairly large change.
Figures 2.10 and 2.11 reveal age specific migration changes before and after 1983
in net migration. Net migration before 1983 had a clearer and more powerful direction towards towns. Data before 1983 show that towns grew mainly because of
family and employment related reasons at an expense of less than 24-year-old movers. The fact that the column of over 50-year-old migrants is almost non-existant
might be due to partial underrepresentation because of the sample and survey
method. Respondents who belonged to this age group in 1983 or earlier were too old
in 1999 to be totally represented in the sample. The results from later than 1983
were more scattered. At first, the survey data revealed a smaller net migration in
numbers. Towns lost people mainly because of housing reasons in all age groups and
gained people because of job opportunities. Although changes occurred in all age
groups, the most remarkable changes took place among people over 25, who left
towns because of housing and all other reasons apart from jobs. To people over 50,
migration gain analyses by reasons gave very similar results to the younger generation. The only difference was a smaller net migration.
The Migrant Survey data allow us to consider more thoroughly the factors that
influence moving. Figure 2.12 shows different migration directions and all reasons
88
600
Reasons of net migration
500
other
family
400
studies
job
300
apartment
200
100
0
<25
25-49
50-64
-100
Age of migrants
Resons of posi tive or negative net migration
Figure 2.10 Urban net migration by age and reasons before 1983 (Living Conditions Survey)
other
30
family
20
studies
job
10
apartment
0
-10
<25
25-49
50-64
-20
-30
-40
-50
Age of migrants
Figure 2.11 Urban net migration by age and reasons 1983–1999 (Living Conditions Survey)
taken into account in the migration decision. Positive values show important factors
and negative values indicate the aspects that were not considered. The different
heights of the columns reflect that people moving from rural areas to towns consider
more aspects important than people moving in other directions. Figure reveals that
people tend to consider more factors in urban-rural and rural-urban migration, since
it involves a change of the environment. Movement from towns to rural areas seems
to be fairly similar to rural-rural and town-town migration, except for a greater role
of the environment and a smaller role of economic factors. In all of these three cases,
distance, self-actualisation and social environment had no special significance. Ruraltown migration was different from other types, mainly because of prevalence of
economic and self-actualisation reasons in that direction.
89
3
Average importance of factors
2.5
2
distance
natural environment
1.5
social environment
self-actualisation
1
economic
0.5
housing
life career
0
.
-0.5
-1
town-town
town-rural
rural-rural
rural-town
Change of satisfaction after migration
Figure 2.12 All migration reasons taken into account (–1 — did not take into
account, 0 — it was not important, 1 — took into account to a certain extent, 2 — it
was important, Migrant Survey)
2.5
2
1.5
1
0.5
0
-0.5
-1
town-town
town-rural
rural-rural
neighbours
generally housing
possibilities for children
work generally
services
region generally
rural-town
possiblity to influence local life
Figure 2.13 Migration directions and differences of satisfaction at the place of
destination compared to previous residence (2 — very satisfied, 1 — rather satisfied, –1 — not very satisfied, –2 — not at all satisfied, 0 — it is not important to me,
Migrant Survey)
90
Figure 2.13 reflects the most remarkable changes in satisfaction resulting from migration. After summarising the changes, it appears that on the average, the increase
in satisfaction was the highest among the rural-urban movers. They were most satisfied with services and possibilities for children. Urban-rural movers complained about
fewer opportunities for children and employment, but were satisfied with services.
The least change can be detected in the satisfaction of the town-town movers, but
on the average, it had increased with respect to all mentioned aspects of life. Consequently, it appeared that people, irrespective of their directions and factors of migration, were all in all more satisfied after moving than before.
50
45
% of all respondents
40
35
30
large extent
25
little
20
15
10
5
0
leaving
people who I
know
price of
dwelling
place of
dwelling
size of
dwelling
expenses
find a job
Figure 2.14 Difficulties in migration (Migrant Survey)
The Migrant Survey data also provide information about migration barriers. The
biggest difficulty was to find a dwelling with an appropriate price (Figure 2.14).
Almost half of the migrants experienced difficulties resulting from leaving friends or
relatives behind, the cost of migration, or some dwelling problems.
Discussion of data from Estonia
Remarkable changes in the reasons for migration occurred during the 1990s. The
main change was the increased share of housing-related moves and a relative decrease of employment-related reasons in migration. Consequently, the presumed
increase in employment-related migration in the 1990s did not find confirmation.
There are several possible explanations to this.
The prevalence of housing factors over employment-related migration could partly
prove some earlier results (Ainsaar 1999, Bonifacy et al 1997), according to which
migration in developed societies is less economy-related. In the case of Estonia, the
small share of employment-related migration could result from short distances of
migration. However, this does not explain the change in employment-related migration over time, because the territory has remained the same. One possible explanation could be the replacement of employment-related migration by commuting mi-
91
gration, which did grow remarkably in the last years (Tammaru 2001c). The influence of commuting was most remarkable in the hinterland of larger towns. The
Population Census data also showed that if 16% of all migrants remained commuters
in 2000, the share of communters among those who moved to the closer hinterland
of bigger towns was 32.5%. Another explanation could be that the volume of employment-related migration has not decreased, but the share of other reasons has
increased, but this is not a plausible interpretation in the light of decreasing mobility.
The rapid increase of housing-related reasons compared to earlier periods could be
explained by limited possibilities of changing one’s living conditions before the 1990s.
Namely, people had fewer opportunities to find an accommodation on their own.
Usually it was the employer who provided the employees with flats (see Tammaru
2001a). Therefore, it could be assumed that, even if the possibility of getting a flat
was the main motivation in migration, the primary objective still was a job as a means
of acquiring other benefits. The total volume of housing construction decreased in
the 1990s, compared to the 1980s (Kõre and Hendrikson 2000), but people themselves became more involved in improving their living conditions.
In Chapter 2.1 we developed ‘the most deficient migration factor’ approach.
According to this approach, the dominant reasons for migration are the most deficient factors in society, and these factors could be improved by moving. The dominant reasons in Estonia in 1989–1999 were family-related reasons, which, according
to our results, were not always directly related to housing. Due to lack of data on
family-related regional deficiencies, it is difficult to combine this result with the
most deficient migration factor approach. However, the increase of housing reasons
can be explained by the coincidence of differences in regional opportunities and
needs of housing.
The 1990s in Estonian migration history are also marked by the continuance of
suburbanisation. It is characteristic that the migration turnaround is mostly associated with those larger towns where the process began. Therefore, the reasons for
suburbanisation must be related to the environment characteristic of larger towns. A
survey in Estonia proved that the biggest changes in the reasons for migration were
related to in- and out-migration to the capital — increased in-migration because of
employment accompanied by increased out-migration because of housing. Between
1983 and 1999, rural areas gained people mainly because of non-economic reasons
such as dwelling and family, and towns because of economic reasons. On the whole,
economic factors did not influence the migration balance and, therefore, did not
affect the proportions of migration in rural and urban areas.
Less than a half (30–46%) of all movers experienced difficulties in migration.
Even though we do not have any comparative data, we could estimate that these
were rather moderate difficulties. The relatively low barriers could be the result of
the small size of the country or the self-selection of migrants. Self-selection is a
process in which the people who ponder the actual migration decision are the same
groups who already have better abilities for overcoming the barriers.
92
2.4 Age and gender
Age and life stages
Plane (1993) has classified all age-related migration studies into three groups: demographic potential studies, migration intensity studies and studies that deal with the
change of reasons for migrating in the same age groups. The following chapter will
give a general overview of all these approaches.
One of the most popular topics in migration studies has been regional demographic potential affecting the number of migrants. It is commonly thought that the
increase of demographic potential is accompanied by an increase in the migration
volume and demographic pressure, but there are also some differing viewpoints (see
Kelley and Williamson 1984). Cohort studies are popular in research on the change
of migration age indicators over time, i.e. the mutual influence of different generations and factors. For example, Rogerson (1987) and Long (1988a) found that the
size of cohorts might influence the timing of migration. Members of small cohorts
tend to move earlier in their life cycle than members of large cohorts. Lastly, from
the group of studies that deal with age-related differences in migration behaviour,
Mulder’s (1993) keywords are “propensity change” and “compositional change”.
Compositional change is caused by a relative change in the proportion of mobile and
less mobile groups in the population (increased number of single mobile people).
Propensity change is shaped by changes in the inclination to move (growth of migration among single people because of improved housing conditions). Analyses of individual life course characteristics in the Netherlands revealed that propensity change
was more important than compositional change despite that considerable influence
from individual characteristics was found.
A great number of studies done on migration report age differences in migration
intensity throughout the life stages. There is an almost universal age-related migration curve that seems to be valid in many countries (Warnes 1992, see also Figure
1.13). In many countries also the direction of migration is often dependent on age.
For example, there seems to be a trend in 1990s in Europe that young families tend
to leave urban centres and agglomerations and move to suburban and rural areas
(Rees and Kupiszewski 1999).
Empirical surveys report a sharp increase in migration between ages 15–25, which
is related to multiple coincidences of life events, some related to migration (Willekens
1987, Ainsaar 1994). Such transitional periods are typically concentrated around
adolescence and early adulthood. Migration peak is explained by several co-occurring events. Some events, like marriage and migration, tend to occur during a comparatively short period of time (see Nieminen 1983, Grundy and Fox 1985) and
they are called synchronized events (Wagner 1989). Another view is to see the key in
interpreting migration intensity in household careers and stages or life courses3
(Clark and Onaka 1983, Mueser et al 1988, Nijkamp et al 1993). Life course specific
migration can be explained with the fact that resources and constraints are also specific to careers. Careers providing individual resources and constraints have also been
3
Lewis (1982) suggests to use instead of “life stage” age free term “life-course”
93
called conditioning careers (Mulder 1993). She argues that the changes in parallel
careers are not evenly distributed over the life course and, therefore, neither is migration. However, not only triggers but also barriers of mobility are distributed over
the life span in a systematic way, which might cause migration differences (Foss
1984). The results of age selectivity are as follows: relatively high mobility of younger
children moving with their parents, very settled schoolchildren, peak of migration
after graduation from school, and later gradual decline of mobility. Accordingly, the
most plausible explanation for the migration peak in younger ages seems to be a lack
of essential migration barriers in terms of binding obligations and costs, as well as a
large number of unmet needs. Higher mobility of younger people has also been explained by the life span mobility investment theory (Clark 1982). According to this
theory, a future prospect of the migrant is an important determinant of the critical
difference between improvements and costs that would motivate an actual move. It
has been found that human capital is more transferable in the early stages of life than
in later ones. Familiarity with surroundings and occupation can be interpreted as
non-transferable investments and, therefore, they represent a barrier. Young people
move more often because they have fewer ties to their family, work, and profession.
This hypothesis supported by the findings of Mann (1973) who found that young
adults were in all cases more favourably disposed towards any kind of move that
improved their housing position. He explained the phenomenon of higher mobility
of the youth by the fact that young people often live with their parents or in rented
apartments which is not very binding.
From the peak of migration intensity until retirement migration can be better
understood within the theory of bounded rationality. Most important factors influencing migration are reasons related to the housing market and occupation (Warnes
1992). The financial possibilities and interests of the people change with age. Several
authors (Morgan and Robb 1981, Liao 2001) found that financial motivation of migration decreases rapidly with the increase of age. Permanent employment becomes
a binding factor and decreases migration. Also an accumulation of different types of
capital increases the costs of movement. For example, Gallaway (1969) found that
with the increase of age the need for financial compensation of both objective and
subjective losses increased as well. Consequently, migration is more costly for elderly people. But there can be different explanations as well. Campell et al (1976)
found that older individuals report lower levels of happiness but higher levels of
satisfaction than younger people. It may indicate that they are also generally more
satisfied with the present situation because of different adaptation strategies, regardless of lower happiness with the situation.
A new increase in migration intensity comes with retirement, although it is not as
intense as it is when people are in their 20s and 30s. The conditions of the labour and
housing market become less important for migration, but they still exert a certain
influence (Warnes 1992). A great interest in the research of migration of elderly
people during the last decade (see King et al 1998) can be explained by the role of
older generations in suburbanisation (Ford and Champion 2000), contraurbanisation
(Berry and Dahman 1977), and reurbanisation (Law and Warnes 1976) processes.
For years, mainly the general migration trends of young people have been studied,
but with the increased proportion of older generations in society, it is more often the
elderly who determine the overall netmigration balance (Shumway and Davis 1996).
94
Research shows that older people have a tendency to move to smaller places
(Borgegård et al 1995) and they are more likely to cover long distances (Rogers
1992). Aged people seem to lay more weight on non-economic factors connected
with the quality of life (Rannikko 1986, Koivukangas 1986, Johnson and Salt 1992,
Liao 2001) or services. For example, Syrjänen (1982:216) found that elderly single
people or elderly single couples without children moved to the built-up areas in
order to be nearer to the shops, social welfare, and health services they need.
The rise of the migration curve in a relatively old age has been associated with the
end of the active old age, worsening health condition, and widowhood that forces
the elderly to move to their families or into care institutions (Warnes 1992, Glaser
and Grundy 1998) in order to receive some help. For example, Corden and Wright
(1993) found that the initial resolution of older people about whether to move into
a new home was influenced by the services and the health and housing conditions
available for the elderly.
Age related migration in Estonia
Age and gender sensitive migration curve for the time period 1989–2000 is presented in Chapter 1.4.2 (Figure 1.13). The age-related differences in migration are,
to a great extent, conditioned by the life stage and age selective propensity of barriers and needs in life span. It could be assumed that people in different age groups
experience different pull and push forces that incite them to move. We believe that
pull, push and well-being factors change with age because there are changes in the
values, needs and opportunities related to migration. Below we look at the age specific data from Estonia from the 1990s. All reasons for migration according to age
groups of migrants and only the triggering reasons are presented.
Figure 2.15 shows the age-related changes in the groups of reasons taken into
account in migration and measured in the Migrant Survey. The general trend seems
to be the decrease of an importance of different factors with age. Some factors
changed their positions as migration reasons during lifetime, like self-actualisation
and social environment, others retained importance throughout the life span. Selfactualisation was important for younger people, but not important at all for the
elderly, at least in the context how it was measured in the Migrant Survey (Appendix
2.1). Social environment as a factor of migration was also important only in the early
adulthood (25–34-years).
While the detailed picture of the Migrant Survey on the triggering factors that
influence migration should be viewed primarily as a collection of hypotheses due to
relatively small amount of cases (Figure 2.16), the age-related curves of reasons for
migration of the Living Conditions Survey present a fairly reliable picture of the
leading triggering reasons of migration in the 1990s (Figure 2.17). The comparison
of the triggering reasons of the two surveys gave basically the same outcome even
though different measuring instruments were used which confirms the reliability of
the results. Both surveys showed the growth of family or social capital and job-related reasons in the age group of 15–24. Social capital related reasons were the main
cause of the migration peak around the age of 20. Also, education was more important in younger ages.
95
Figures 2.18 and 2.19 depict some individual reasons of migration that changed the
most with age. Figure 2.18 shows the reasons the importance of which increased
with age and figure 2.19 shows reasons that were more important for younger people. The assessment of the reputation of the place, closeness to friends and relatives,
parents’ home, better income, road conditions, neighbours, phone connection, medical services, and social life did not change with age. The importance of natural environment increased, while factors related to possibilities to organise free time, the
level of schools and housing became less important with age.
Table 2.6 Triggering reasons for migration and factors taken into account in migration process (Linear regression, only statistically significant B, Migrant Survey)
Triggering reason = dependent variable
Possibilities taken into
account in migration
process
Household career
Housing
Economic
Self-actualisation
Natural environment
Social environment
Household career
Housing
Economic
0.467
–0.100
–0.173
–0.157
–0.170
0.234
–0.152
0.303
–0.185
–
–
–
-0.113
–
0.514
–
–
–
The age-related curves changed significantly in the turning points of life stages. For
example, the role of the distance between home and workplace generally increased
with age, but its importance disappeared around retirement age. The opposite tendency could be detected with respect to natural environment becoming more important in retirement age. Several rapid changes in values took place in the group of
25–34-year-old people (there was a change in the average importance of the level of
schools, closeness to bigger centres, and security).
It was assumed that people whose movement has been caused by different triggering reasons consider different factors in choosing their new place of residence.
The Migrant Survey allowed dividing people into groups according to their main
triggering reason and thus, measure the influence of different push and pull factors
of migration. The changes in peoples’ life stages were coded in the Migrant Survey as
household career reasons (Appendix 2.1). Regression analysis indicates (Table 2.6)
that people who moved because of household career reasons, took into consideration
the improvement of household career opportunities and the social environment4 . All
other factors were clearly not important at all in the whole migration process because of household career. Somewhat unexpected was, for example, the under-consideration of housing factors that could in principle be a part of the decision to move
due to household career. These results indicate that household career and housing
needs do not necessarily coincide.
4
Regression is used here in order to include possible simultaneous influence of several factors.
96
100%
Important and unimportant factors in
migration decision
80%
34%
60%
31%
40%
48%
50%
64%
33%
35%
54%
41%
74%
60%
20%
64%
30%
46%
37%
27%
41%
24%
distance
25%
34%
30%
42%
75%
35%
8%
20%
39%
natural environment
social environment
self-actualisation
economic
housing
0%
life career
-20%
-40%
15
20
25
30
35
40
45
50
55
Age of migrants
Figure 2.15 All reasons taken into account in migration and age of migrants
(Average for age group: -1 - did not take into account, 0- it was not important to me,
1- took into account to a certain extent, 2- it was important, Migrant Survey)
other
120
environment
100
social capital
studies
80
job
housing
60
40
20
0
0-14
15-19
20-24
25-29
30-34
35-39
40-44
45-59
45-
Figure 2.16 Age and triggering reasons of migration (Migrant Survey)
other
350
family
300
studies
250
job
apartment
200
150
100
50
0
<15
15-19
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
>59
Figure 2.17 age and reasons for changing place of residence in 1983–1999 (Living Conditions Survey)
97
1.5
Impo rtan ce
1
0.5
0
-0.5
-14
15-24
25-34
35-44
45-54
55-65
-1
Age of migrants
closeness to bigger centres
better living conditions
free time
level of schools
Figure 2.18 Migration factors that become more important with age in migration (Average: –1 — did not take into account, 0 — it was not important to me, 1—
took into account to a certain extent, 2 — it was important, Migrant Survey)
0,8
0,6
Importance
0,4
0,2
0
-0,2
-14
-14
15-24
25-34
35-44
45-54
55-65
-0,4
-0,6
Age of migrants
security
natural environment
Figure 2.19 Migration factors that are more important in younger ages in migration (Average: –1 — did not take into account, 0 — it was not important to me, 1—
took into account to a certain extent, 2 — it was important, Migrant Survey)
Gender
Selectivity or the lack of selectivity according to gender in migration is connected to
a wider cultural background and it can vary over time, geographical scope, and different migration types. We argue that gender itself is not a factor that determines
migration, but its influence becomes apparent through cultural attitudes, norms, and
conventions. De Jong (2000) found in a study about Thailand that intentions, expectations, norms, and gender roles were the key elements in the “black box” of making
the decision for migration. In connection with the norms, local labour market
98
features also shape migration outcomes according to gender. For example, in Latin
America, there is a predominance of women in migration streams into cities, which
is due to a great demand for female domestic servants in urban areas. At the same
time, in South Asia and Africa, women are more likely to stay at home to run farms
and businesses, which means that men predominate in both rural-urban and ruralrural migratory streams (Simmons 1984). In empirical research, we found both: the
results that confirm the prevailing role of men in migration (Bach and Smith 1977,
Marksoo 1987, Stark 1988, Mulder 1993, Goldstein et al 1997) and the results that
state the opposite (Naukkarinen 1969, Nilsson 2001).
Several studies (Naukkarinen 1969, Mulder and Wagner 1993, Ainsaar 1994)
point out that the migration activity of women and men varies over the life span. The
migration activity of men is usually postponed. It could be explained by the shift in
life stages. Mulder and Wagner (1993) found that age specific differences were small
for long distance moves, but more visible for short distance (< 50 km) moves. Still
in case of both distances, women passed the peak of moves earlier than men. The
main explanation for the earlier migration peak of women could be the difference in
marrying age. For example, Mulder (1991) using the data from the Netherlands
found that the age and sex differences in migration almost disappeared when marriage synchronization effects were taken into account.
Foss (1984) believes that women have traditionally stronger ties between the
life-cycle, educational and occupational life spans. It is also reported in many studies
that women and men tend to have different reasons for migration. Simmons (1984)
summarises several studies and concludes that, for men the main reasons for migration are the current wages and unemployment, but for women the reasons for migration also include the “marriage market” in addition to the labour market opportunities. For example Naukkarinen (1969), after studying internal migration in Northern
Finland, found that migration of women was dependent upon the business cycle, but
the influence of this cycle was relatively small compared to men. The weaker connection of the migration of women to economic circumstances could also be a result
of the specific characteristics of women’s jobs. Halfacree (1995) came to the same
conclusions about the reasons of women’s lesser dependence on the labour market.
He used different sociological theories and empirical results and found that women
tend to be concentrated around jobs that are geographically relatively ubiquitous
and, therefore, easier to obtain.
Nilsson (2001) on the other hand, using the Swedish population register and the
migration database of Umeå University from 1985–1995, found that men and women
also behave differently. Women constituted a more homogenous group according to
their behaviour while men were more influenced by their place of birth and education. An interesting finding was that migration had a negative effect on the income of
women with children, whereas it had a positive effect for women without children
and in all groups of men. Similar results were obtained by LeClere and McLaughlin
(1997).
Presuming that stress can be one cause of migration, it might prove fruitful to
explain the differences in the migration of men and women (reasons, intensity) by
factors causing the most stress. Namely, research has shown that men and women
have different receptivity towards stress caused by life events (Conger et al 1993).
99
Men are more open to professional and financial events, whereas women are more
receptive to stress induced by family members (but not by friends).
Gender differences in Estonia
Earlier migration studies in Estonia have shown that reasons for migration can be
selective according to gender. Marksoo (1987) gives a thorough overview of migration differences between men and women in Estonia in 1959–1976. The study indicates that the migration of women had higher efficiency than the migration of men.
Although women were proportionally not particularly predominant in migration,
their relocation through migration was much effective than the relocation of men.
By preferring large towns, women were the main initiators of urbanisation and the
more active side when the family moved into town. Living in a family decreased the
territorial mobility of women and women took more into consideration the needs of
other family members and they were more often than men in the role of passive comovers. Finally, when choosing a workplace, women gave less consideration to salary
than men and took more into account the working conditions.
The study of the life course of the cohort who graduated from high school in
1983 found that the only reason for the higher than usual migration activity among
young men was the military service. Regarding all other reasons, the migration of
young women in ages 18–28 was two to three times more frequent (Ainsaar 1994).
On the other hand, the comparison of the life events of two generations of Estonians, disregarding the military moves, showed that gender did not influence migration
activity (Ainsaar 2002b). The data of the Living Conditions Survey from 1999 showed
that men tend to be more settled during their life than women (Table 2.7). Similar
results were obtained from Population census data in 1989–2000 (Figure 1.13).
Table 2.7 People who have changed their place of residence according to gender
(%, Living Conditions Survey)
Male
Female
Total
Settled
33.7
26.6
29.8
Migrant
66.3
73.4
70.2
Total
100 N= 2135
100 N= 2591
100 N= 4726
The migration of men and women was influenced by different reasons. Although the
pattern of reasons changed over time men moved more often because of employment, while women moved more because of family reasons (Table 2.8). In time,
the differences in employment-related reasons have decreased, but the gap in the
family reasons has remained the same in 1989–1999.
One explanation for different reasons for migration among men and women could
be the differences in values. We can assume that, since men and women move
because of different triggering reasons, they might also have different values and
factors they take into account in migration. Indeed, analysis of the data from
Migrant Survey, taking into account all reasons for moving, revealed differences
100
Table 2.8 Gender and reasons of migration in different time periods (%, Living
Conditions Survey)
Men <1983
Women <1983
Men1983–1988
Women1983 –1988
Men 1989–1999
Women 198 9–1999
Total men
Total women
Dwelling
9.9
10.2
18.3
14.6
24.7
22.4
13
11
Job
34.8
29.4
37.6
31.3
25.5
23.6
33
26
Studies
8.4
12.2
1.8
7.3
6.7
6.9
7
8
Family
38.5
42.7
33.8
40.8
30.4
38.6
37
47
Other
8.4
5.5
8.5
6.0
12.7
8.5
10
8
Total
100
100
100
100
100
100
100
100
between men and women in the evaluation of the household career and housing
factors in migration decision. Women attributed higher value than men to the household career as well as to the housing factors during the migration process 5.
Accordingly, Estonian data from the 1990s somewhat supported previous
research on variations between male and female migration concerning family
reasons and explained them with possible value differences.
2.5 Human capital and self-actualisation
Human capital has traditionally been measured by education, training or experience
(Becker 1993, Lin 2001). This chapter gives an overview of empirical and theoretical
findings about migration incentives and barriers related to human capital. In the
analysis of Estonian data, three components of human capital, i.e. education, health
and previous migration experience, are used.
Many migration studies imply that the personal characteristics of people influence migration. One possible explanation is that the statistical differences between
people are caused by the selective nature of migration and people with certain characteristics are more opened to changes. Bogin (1988), for example, found that those
who migrated were always taller than those who stayed. There have been some studies (Walmsley 1982) that support the assumption that introverts and extroverts have
different preferences for the place of residence, but this aspect is still relatively
poorly researched and questionable from the viewpoint of migration. Morrison (1972)
started with the notion that migrants are thought to be more ambitious and achievement-oriented than non-migrants. The author was seeking evidence for migrating
individuals having certain psychological characteristics, but he did not find any. The
author believed that the explanation could be in the age and occupation of people.
People in some occupations and younger people tend to be more achievementoriented and mobile. Even characteristics such as intelligence have been employed
5
Household career features included also the family.
101
to explain migration differences (see Mangalam and Morgan 1968). The majority of
literature about the influence of human capital on migration covers education and
previous migration experience.
Education is an important migration factor mainly through its connection to the
socio-economic status of the migrant (Charney 1993). Several studies indicate that
the migration prospect of people with less education might be more restricted (see
Hartshone 1992, Jolkkonen and Kilpeläinen 2002). However, Linnemann and Graves
(1983) found that people with better education tend to be more settled when their
migration is not related to the change of jobs. Newbold (1998), who studied the
influence of education on the out-migration from California in 1990, did not find
any selectivity of education. However, people with poor education moved into states
that were somewhat closer. Results revealed no significant or systematic selectivity
among migrants with better education, while self-selection was more important
among the poorly educated non-migrants. However the self-selectivity by education
was not verified. The perceived difference in wages reduced the likelihood of outmigration among the poorly educated, but increased the likelihood of migration
among the more highly educated.
Mulder’s and Manting’s (1993) findings refer to the more complicated migration
strategy of more highly educated people. They found that the young and highly
educated choose a more flexible strategy (nest leavers), related more to individualistic attitudes. But it was more the combination of being young and highly educated
than the two determinants separately.
Human capital provides people with better prospects for changing their place of
residence. Better education and skills enable more effective searching and information processing and better achievement of well-being. However, education could
also have an adverse effect: the geographic options of people with a specific education can be much more limited than the options of the people with common professions.
Experience
It is possible to classify people into groups, based on the intensity of movement (see
Plane 1993): chronic movers who often and constantly change their place of residence, individuals with a great degree of mobility and non-movers. The same grouping might also be valid for households (Hartshorn 1992). According to their influence upon the household decision to move, migrants have often been divided into
active or innovative (Lloyd and Dicken 1972), and passive groups — e.g., the family
members of an active individual. The traditional explanation of this classification is
the domination of the needs of one family member, but a similar classification would
also apply for greater receptiveness and responsiveness of some people. Hägerstrand
(1957) identified as active migrants those who were constantly seeking for a suitable
location that would help them to achieve financial well-being, and as passive
migrants those who followed the impulses from their circle of acquaintances, i.e.
mainly from those who had already moved successfully. White and Woods (1980)
divide people into innovators versus traditionalists. They believe that sometimes
even the traditionalists can move when the environment starts to change too rapidly.
In order to explain the differences in migration activities, several researchers (Brown
102
and Moore 1970, Morrison 1971) use the term of “decision threshold” known from
the behavioural approach of migration research. Individuals or a household with a
low decision threshold are more mobile than those with a higher decision threshold.
The studies on American youth commencing their careers (Black 1983) clearly
demonstrated that migration depends on the previous migration experience. In time,
the results of the search process improve, which also helps to explain why more
experienced people are better migrants (Weibull 1978, Clark and Smith 1982). Previous migration experience and the intensity of migration could be connected to the
learning process, better knowledge or skills and, through this, also to the decrease of
risks and time loss.
Human capital might produce contradictory impact to migration. There is some
evidence in research on migration differences being conditioned by the nature of
different people. Analysing individual human capital differences has not been the
mainstream of migration research and more empirical data is needed before any final
conclusions can be drawn. Therefore, in the following we will turn our main attention to the analysis of human capital, the settledness-mobility of the people, and
migration distance. Within the framework of human capital, we will consider the
following indicators: education, health, and previous experience of moving.
Human capital and migration in Estonia
Several earlier migration studies suggest that migration might be selective according
to education (Ainsaar 2002a, 2002b). Neither the data of the Living Conditions
Survey (settled or migrant) or of the Migrant Survey (the number of places of residence) confirmed the increase or decrease of migration activity as a result of education. Since age affects both health and education, this relation was studied separately
in different age groups. But even this analysis of the Living Conditions Survey data
did not show any correlation between education and settledness. People with different levels of education were no more active nor passive in migration. However,
the analysis of the Living Conditions Survey data according to education groups
showed that people with higher education differed from the rest of the groups in
their reasons for moving. They migrated more often than others because of workrelated and housing reasons during 1989–1999 (Table 2.9), but since the number of
Table 2.9 Education and reasons of migration among internal migrants 1989–
1999 (% of educational group, Living Conditions Survey)
Attained level of
education
Basic or less
Secondary
Higher
Apartment
Job
Studies
Family
Other
Total
21.3
22.0
27.3
18.9
23.9
28.8
3.3
9.1
5.6
39.3
35.0
30.8
17.2
9.7
7.6
100 N= 122
100 N= 309
100 N= 198
103
migrants whose data was analysed was fairly small, this could be interpreted as a
preliminary hypothesis.
In the Living Conditions Survey, we also tested the earlier results according to
which people with better grades tended to move into larger settlements. Migration
into certain environment can be explained by different values (Ainsaar 2002b) and
opportunities. The Living Conditions Survey used only the data of attained level of
education in 1999 and it was not possible to analyse the level of education at the year
of migration. The survey did not reveal any significant changes in migration directions of people with different education (Table 2.10). People with lower education
had played a dominant role in rural-rural migration in all time periods. The percentage of people with higher education was higher in urban-rural and urban-urban
migration.
Table 2.10 Attained education by 1999 and the direction of last migration in
three time periods (%, Living Conditions Survey)
1989–1999
Urban-rural
urban-urban
rural-urban
rural-rural
total
1983–1988
urban-rural
urban-urban
rural-urban
Rural-rural
total
< 1983
Urban-rural
urban-urban
rural-urban
Rural-rural
total
Basic or less
15.6
11.7
23.1
30.1
19.4
Basic or less
26.5
13.0
19.0
36.4
22.8
Basic or less
28.8
19.0
42.6
54.4
35.2
Secondary
46.4
45.2
54.6
48.7
48.2
Secondary
37.3
40.0
40.5
46.9
41.4
Secondary
36.0
37.3
33.0
33.6
34.9
Higher
38.0
43.1
22.3
21.2
32.3
Higher
36.3
47.0
40.5
16.8
35.8
Higher
35.2
43.8
24.4
11.9
29.9
100 (N=179)
100 (N=188)
100 (N=130)
100 (N=156)
100 (N=653)
100 (N=102)
100 (N=185)
100 (N=126)
100 (N=143)
100 (N=556)
100.0%(N=125)
100.0%(N=722)
100.0%(N=751)
100.0%(N=360)
100.0%(N=1958)
Finally the Migrant Survey examined the relationship between education and the
distance of migration, satisfaction after migration, options and the time spent on
selection of the new place of residence. The only statistically important correlation
(sig 0.005) turned out to be the one between education and the number of dwellings
available. It indicated that people with better education had more options and
they moved more often to places they had not been before than migrants with
lower education.
104
Health
Based on the Life Conditions Survey, the self-assessment of health was compared
between people who had moved and those who had not. The analysis of the whole
sample showed a surprising correlation between migration and the assessment of
one’s own health: people who had never moved claimed that their health was much
better. However, this was, above all, a result of the different age structure of settled
and non-settled people. The correlation between migration and health condition
disappeared after this relation was analysed separately in different age groups.
Experience
It could be assumed that previous migration experience is a benefit when it comes to
the next move because the acquired experience enables to decrease the psychological cost of moving or improve the quality of the decision. Migrant Survey allowed to
examine the correlation between the number of moves after the age of 18, the time
spent on the last move, options for migration and satisfaction with it. The correlation
analysis revealed that the only important connection was with the number of
options. Those who had more experiences with migration had more choices. This
might indicate that frequent movers have better search strategies, but it might also
be a result of other factors.
2.6 Housing
Housing is an aspect that is considered in all migration events. In this subchapter, we
will study when it becomes the most important factor or the main reason for changing place of residence and how it influences the overall migration decision.
Both qualitative and quantitative characteristics of the future place of residence
are considered in migration: size, quality, design, cost, and ownership (Clark and
Onaka 1983, McCarthy1976, Millington 2000). It is found that housing migration
dominates in shorter migration distances and in case of longer distances, the migration decision is based on other factors so that housing remains only one of them
(Forrest and Murie 1990). Suburbanisation has been seen as a typical result of short
distance housing migration (White 1990).
Often the financial possibilities of the family and feelings towards the location of
the dwelling play a decisive role in making the choice of housing (Clark 1976, see
also barriers in Chapter 2.3). Therefore, the migration because of housing reasons is
dependent on income and the housing market and can be influenced, besides other
factors, for example, by the availability of loans. Also, the form of ownership is
important. As a general rule, tenants and lessees move more often than owners
(Mulder and Manting 1993). Consequently, the characteristics of housing both in
destination and original areas influence the territorial migration activity and may be
barriers to movement or its supporting factors (see Myers et al 1997, Knox and
Pinch 2000). For example, Clark et al (2003:158) found that “in economically
vibrant housing markets with rapidly growing population and high level of new construction of dwellings, households move more than in housing markets with a more
stable population and lower levels of new constructions”. Empirical research proves
105
that new housing might aid in-migration of all age groups (Millington 2000). In areas
with old housing market, on the other hand, the shortage of available living places
could become an essential barrier for newcomers. For example, Lyons and Simister
(2000) observed that in-migrants were less successful on London housing market
than on the labour market.
Selection of housing is a selection of life style. Scheider
and Kasper (2003) conceptualise the choice of housing location in the framework of life
style as phenomenon interrelated with social structure and
daily mobility in the context of Figure 2.20 Interaction between choice of housing, life
space-time structures (Figure style, social structure and daily mobility
2.20). Life stages and housing Source: Scheider and Kasper (2003)
needs are probably one of the
most studied aspects of migration (Rossi 1980, Clark and Onaka 1983,
McCarty 1976, Borgegård et al 1995). Clark et al (2003) develop an idea of a housing career meaning that each move and each new dwelling is a step in the career, a
step closer to the house that meets best the needs and aspirations of the household
at the moment. The household reaches equilibrium, but may fall out of equilibrium
when there is a change in the family structure. They found that most changes in
housing stages show an upward trend by criteria of price/quality. Low income can be
an essential barrier to housing career. A close connection was found between housing, family status and income. A study in Estonia (Käärik et al 2001) suggests that
when families with children got into financial difficulties, the first thing they sacrificed was their housing conditions. Also, a survey conducted in Helsinki in 1981
showed that household characteristics of intra-city migrants were more important
determinants of residential mobility than area characteristics (Valkonen and Martelin
1986). The differences in the probability of moving between the high and lowincome areas were opposite to what was assumed. The explanation was that if households in low-price areas sold their apartments, they could only afford larger apartments in the same low-price areas. Households living in the high-price areas where
the price of apartments was higher had more freedom in their decision to move
which may explain the higher mobility in “better” areas.
All empirical research analysed here showed that housing migration should have a
strong connection to life stages, financial possibilities and housing market. The availability of appropriate housing and the price can be essential barriers to moving.
Housing migration in Estonia
Macro-level analyses showed that there was a strong correlation between new housing and net migration in 1989–2000 (Chapter 1.4.2). Individual survey data revealed
that the reasons related to dwelling were among the most important ones in all age
106
groups and altogether were placed among the first three in the ranking of migration
reasons in the 1990s (Figures 2.3, 2.8).
Housing acts equally as push and pull factor and influences all age groups,
although the dwelling needs may vary. It is interesting to note that the average
satisfaction of people after the move, i.e., migration outcome, was also greatest
with respect to the dwelling-related factors (Figure 2.21). As expected, the people
who moved for housing reasons primarily considered housing conditions in their
decision-making (Table 2.6). They devoted significantly less attention to employment, income, and household career opportunities.
We can assume that opportunities to get a suitable dwelling might be related to
financial resources. An analysis was made on whether wealth had any significant
impact on the chances of finding a new dwelling. The Migrant Survey data showed
that household income per capita did not show a statistically important correlation
with the time spent on the whole migration process. However, the wealthier households had more options for choosing (sign 0.048). It is possible that the benefits of
wealth will be reflected not only in the search process but also in the outcome.
Correlation analysis showed that wealthier people were more satisfied with the
comforts of their dwelling (0.002), but this did not apply for other housing related
factors.
In Migrant Survey, the term “environment” involved mainly two indicators: “natural environment” and “living environment”6 . A factor analysis of all circumstances
considered in the migration process (see page 74 and Appendix 2.2) showed that
0.6
0.5
0.4
0.3
0.2
roads
possiblity to influence
local life
work generally
relations at work
services
neighbours
income
region generally
security
-0.2
possibilities for
children
-0.1
comfortable of
housing
0
generally housing
0.1
size of tenure
satisfaction after - before migration
0.7
Figure 2.21 Increase in satisfaction with different aspects of well-being after
migration (Differences between average satisfaction in previous and present living
place, –2 not at all satisfied, 2 very satisfied, Migrant Survey)
The definition of the living environment was not specified in a more detailed way. The present
wording and subjective interpretation comes directly from the survey. The analyses showed that
“living environment” might be a mixture of natural environment and living conditions.
6
107
living conditions were often taken into account simultaneously with environment
factors. This is also a reason for analysing “environment” reasons together with the
housing ones, because it was difficult to distinguish between closer and more apparent living environment in this survey. The Migrant Survey clearly revealed the difference of “environment” reasons in different directions of migration. Only 7.7% of the
rural outmigrants claimed that environmental conditions were an important reason
for migration. Among people moving out of the capital, the same indicator was 26%.
Environment and security were the only factors considered in choosing a new dwelling the rural pulling force of which was stronger than the urban pulling force (Table
2.5). The importance of the environment as a pull factor increased with age (Figure
2.16).
60
% from all who desire to leave
50
40
30
20
10
0
capital
town (>50,000
pop.)
family
work
other urban
residence
rural
studies
Figure 2.22 Place of residence and reasons for the desire to leave (Living Conditions Survey)
A summarised overview of the correlation between environment and migration can
be attained from the analysis of migration desires of people living in different settlements (Figure 2.22). Unsatisfactory living conditions posed the greatest problem for
the residents of the capital in 1999. The survey revealed that almost 1/3 of the
potential migrants from the capital wanted to move to rural settlements in the surrounding county. At the same time, there was also a reverse pattern — people in
rural settlements wanted to move into towns and people in smaller towns wanted to
migrate into the capital or some other larger town. The main attractions of the capital were family ties, employment opportunities and better possibilities for studying.
The desire to move to other towns was mostly related to the hope of finding a job,
and people moving to rural areas wanted to improve their living conditions or to
move in with their parents (Ainsaar 2002a).
108
2.7 Economic resources
Income, economic resources and social status tend to be interrelated. Economically
motivated people act according to the perceived highest economic outcome and
economic migration is especially important in longer distances (Olsson 1965). Usually, the movements that are related to jobs or economic indicators can be viewed
according to the rational model of maximizing benefits. Several studies refer to the
evidence that migration is economically beneficial for the migrants: Greenwood
(1975, 1985) reported that migrants in the USA had bigger income than nonmigrants, Skrede (1977) found a greater return on educational investment among
the more mobile people in Norway compared to the less mobile ones. The benefits
of migration may also exist in the form of economic opportunities (Stinner and van
Loon 1992).
At the same time there are several studies that do not explicitly support the
results about the benefits of migration (Westerlund 1995) or claim that the benefits might be a result of probable self-selection (Natosteen and Zimmer 1982). For
example, Smits (2001) found, studying married men and women in the Netherlands, that when people had moved over longer distances, their earnings were higher
than the earnings of the people who had not moved or had covered a shorter
distance. One of the explanations was that migrants might be a favourable selfselected group. When this favourable self-selection was taken into account, male and
female migrants turned out to earn significantly less than their non-migrating counterparts. The other explanation could be that the migrants avoid even greater poverty and lastly, migration strategies can have different aims, for example, risk minimisation instead of profit maximisation (Brown et al 1998, Simmons 19847 ).
However, despite the fact that the availability of jobs can direct labour migration,
the relationship is not always straightforward (Zaiontskovskaia 1985). A potential
migrant who looks for better job conditions might have several other options besides
migration: for example, vertical or horizontal mobility within the same region or
commuting. Horizontal and vertical moves in labour market can occur simultaneously with migration, especially when the migration is caused by shortages in structural job markets. However, some empirical studies show that most people tend to
continue to work in the same sector where they had their previous job (Hart 1973).
Lots of literature discusses the connections between developing service markets
(Reijo and Valkonen 1993, Karjalainen 1989, Bengtsson and Johanson 1995, Newbold
1998) and the influence of large corporations on migration (see Sell 1990). Some
regions do not attract people due to the general availability of jobs, but rather
because of the presence of available jobs of a particular type. These regions are
more favourable for certain activities and therefore, attract also specific types of
migrants. Sometimes such successful regions have been called “escalator regions”.
The spheres that make regions economically successful change over time and are
dependent on the overall development of the society. Occupational difference in
migration can be caused by the fact that different socio-economic groups migrate
in response to different stimuli: manual and unskilled workers for higher wages or
7
Both use the example of family migration.
109
professional workers for promotion opportunities (Hart 1973, White and Woods
1980). There seem to be different migration push and pull forces for different groups
of economic migrants.
On the individual level, several studies suggest that people with a higher socioeconomic status have more freedom of choice in making a decision to migrate
(Valkonen and Martelin 1986). Consequently, the status, economic possibilities, and
education together influence migration through freedoms and barriers (Figure 2.23).
The economic stimuli are especially influential in case of economically motivated
individuals or groups of people who value income and try to gain economic profit.
The economic factors might influence people from younger age groups up to retirement (Grundy and Fox 1985, Warnes 1992) and the socio-economic differences
shape the behaviour of people even within the same age group (Grundy and Fox
1985).
Education
Migration
Income
Social status
Freedom,
barriers
Non-migration
Figure 2.23 Interaction between social status, income, education and migration
Mulder and Hooimeijer (1999) reported that people with higher incomes and higher
education tend to be more attached to their labour market careers than others. They
also have reasons to move more frequently and over longer distances, because highly
paid specialized jobs are geographically more dispersed than lower status jobs. Higher
mobility of higher status migrants has been explained by a greater number of freedoms
and possibilities (Berry and Dahman 1977, Warnes 1986, Hartshone 1992, Liao
2001). It seems to be essentially true in case of longer migration distances when the
workplace is changed at the same time as the living place (Johnson et al 1974,
Linnemann and Graves 1983, Long 1988b, Fielding 1989, Walker et al 1992). The
results of the research on the connections between migration and social status are
similar. Many studies in developed countries show that members of the middle class
tend to move more often than their co-citizens with a lower status. On the one hand,
social relations of the middle class provide its members with better information
about the existing alternatives and decrease the influence of the distance; on the
other hand, the specific characteristics of their work might cause them to move over
longer distances. Exceptions are specialists whose work requires constant maintenance of a client base — doctors, architects, lawyers, etc. (Ladinsky 1967).
110
One possible explanation of lower migration intensity among the less specialised workers may be better opportunities to increase their earnings through occupational mobility within present living area. For the lower status groups, migration is
also more expensive: they have less information about available alternatives and the
search for alternatives demands more time and money (Jones 1990). In addition to
migration being more expensive for people with less income, the distance of migration is restricted by the lack of information about suitable job opportunities (Jones
1990) and also, the information might be dependent on the contacts and previous
migration experience (Amrhein 1985). The search for alternatives is money and
time consuming and, therefore, the decision for migration must be made in a relatively insecure situation (Cadwallader 1989). At the same time, the less educated
might not need to move as far to remain in the same occupational sector.
Economic migration in Estonia
On the grounds of the increasing importance of work and income as the main assurance of access to other social benefits and regional differences in income we can pose
a preliminary hypothesis on the rise of the importance of employment-related
migration in Estonia during 1990s. Despite the expected leading role of employment-related migration, the results of the surveys conducted in the 1990s showed
that economic reasons appeared to be only on the second or third place in the hierarchy of reasons (Figure 2.3) and their position did not rise during the 1990s (Table
2.3). The Migrant Survey confirmed the results of Living Condition Survey (Table
2.8) about different involvement of men and women in economic migration. The
correlation between migration and work as a reason for migration was stronger for
men than for women. Nearly 44% of the men based their migration decision on
employment-related reasons, while women considered employment factors in 36%
of the cases. Employment-related factors were more significant for people with
vocational and post-secondary vocational education and it was also more difficult for
them to find employment in their respective professions.
The analysis of factors taken into account in the assessment of potential destination places (Table 2.6) showed that people who moved because of economic triggering reasons were persistently taking the economic conditions (employment, income)
into consideration. The circumstances connected to household career were, on the
contrary, considered not important at all.
According to the Migrant Survey, income was the most influential push factor of
different employment-related reasons (Table 2.4). 1/3 of all migrants found low
income to be an essential reason for migration, ¼ migrated because of unemployment or absence of a professional position. Yet, the detailed analyses proved that not
all of those who complained about low income reported it as a factor they took into
consideration in selecting a new place. 43–45% of migration decisions were influenced by the expected level of income in the new location or the possibility to get a
job.
111
% of people who have changed the
residence 1989-1999
The analysis of the reasons for migration among wealthier and poorer people
revealed as the only significant difference the fact that the wealthiest people moved
more because of housing-related reasons. All other income groups had similar migration reasons. While the average percentage of reasons related to the improvement of
housing in the Migrant Survey sample was 15%, the wealthiest quartile reported
housing as their reason of migration in 23% of the cases.
We might expect that people
25
with higher income tend to migrate
more often than others. This phe20
nomenon can be explained mainly
with the fact that they have greater
15
freedom to move and a wider range
of options. For the lower status
10
groups, migration is more expensive
and the search for alternatives might
5
require more resources. Indeed, the
Living Conditions Survey showed
0
that people who were financially
wealthy
nor rich nor poor
close to poverty
poor
better off had changed their place
of residence more often during the Figure 2.24 Change of residence and selflast ten years (Figure 2.24), the assessment of wealth (Living Conditions
same result was supported by the Survey)
Population Census data. Although
no questions about income were asked in the census, people belonging to the higher
employment status groups were more mobile during 1989–2000 compared to the
lower status groups. 19.7% of the people with higher positions (ISCO classification
groups 1–3) had changed their place of residence, while the same number for people
in lower positions was 15.7%.
The Migrant Survey revealed that the wealthier people moved over longer
distances (sign 0.005) and they had more options to choose from. The Living Conditions Survey also showed that the wealthier had more plans to relocate in the future
(sign 0.018).
Socio-economic status influences the degree of freedom of migration and its
barriers. Most empirical migration studies have shown that people with higher
economic status move more often and over longer distances, whereas people
with poor income are less mobile due to bigger barriers, such as information
barriers, and the range of local opportunities. Analyses of Estonian data showed
similar tendencies about the positive correlation between wealth and mobility
during 1990s.
112
2.8 Social environment
In this section the role of social capital and social services in migration process are
discussed. Social capital and services supplement each other in providing the framework for social security. In this chapter, they will be referred to as social environment. Social environment consists of a range of different factors: social capital, services, attachment (psychological memories), security. This chapter summarises briefly
the empirical research related to migration and analyses empirical data from Estonia
in the 1990s.
Social environment is an essential factor of overall well-being. Myhren (1998)
studied well-being, using an unstructured interview in two housing areas in Norway,
and specified four main factors influencing environmental well-being: aesthetic (special architecture, pleasant environment, green space), functional (presence of communication, shops, schools), social relations and individual and personal factors (roots,
memories which create associations to previous places of residence). Three last factors belong to the group of social environment. The importance of all these factors
might change with age and social position. Although the Migrant Survey data did not
reveal significant differences in the evaluation of social environment by different age
groups in Estonia, it has been found previously that the importance of social environment varies by groups and might be acquired from different sources. For example,
personal contacts in the neighbourhood are more important to the young, elderly
and poor people, and the members of the middle class are more orientated towards
social contacts that are based on the workplace and holiday locations than to closer
living environment (Hartshorne 1992).
Social capital is linked to the network of social environment and can also be
interpreted as a part of social security. When analysing the wide range of different
definitions of social capital, Johnston and Percy-Smith (2003) found that the most
common feature for social capital was trust. Social capital is formed by family relations, friendship, neighbours and other social ties. The benefits of migration may also
be hidden in the transferability of human capital into social capital and vice versa.
Coleman (1988) and Burt (1992) point out that the more social capital one has, the
better he or she can use the investment in human capital. Social capital adds to,
rather than replaces the human capital in managers’ income attainment (Burts 1992).
Human capital produces social capital, but the effect is not very strong. According to
research, social capital exhibits a strong effect on managers’ income and generates
high income mainly through a network that is based on strong ties. Sanderfur and
Laumann (1998) see the benefits of social capital in information, influence, control,
and social solidarity. Rook (1984) has postulated that the availability of social support from others may reduce threats to well-being. For example, a study by Oh
(2003) showed that tree main channels of social bonds important to the urban elderly were friendship, social cohesion and trust, and informal social control. The study
still proved that social bonds did not influence mobility intentions directly but
through improved residential satisfaction. On the other hand, emotional involvement between the significant others may also create a bind, which reshapes individual behaviour and becomes a triggering factor in migration. An outcome of social
capital is not necessarily only positive. Social capital has all features characteristic to
113
other types of capital — from freedom on one side to responsibilities and burden on
the other. Social capital can become a barrier if migration strives to break up important relations or a triggering reason if migration decision is lead by the motive to
establish closer social relations. Therefore, it is appropriate to view social capital as
property (see also Putnam 1993, Lin 2001).
Although there is no general agreement about the agents of social capital (Johnston
and Percy-Smith 2003), we argue that family relations are the most widespread
sources of social capital. Consequently, the studies of migration and social capital
should include marriage and family analyses. Indeed, when compared to other social
ties in the context of household career, marriage as an event in life has been studied
most thoroughly. It is correct in the context of life event analyses, but simultaneously people belonging to the household will be agents of social capital.
Marriage has been seen as a settling factor in a longer perspective (Kawabe and
Liaw 1992, Warnes 1992, Mulder and Manting 1993, Mulder 1993, Newbold 1998),
but it must be mentioned that this result can be an effect of age and other factors
(housing). Schofield’s (1990) research among English teachers showed that civil status factors (married, single) influenced teachers’ ability to engage in career related
migration, frequency and direction of migration. Mulder and Wagner (1993)
reviewed a number of studies about migration, marriage, and fertility career. They
argued that the common finding that married people move less than the unmarried,
is not true at short distances (less than 50 km), if marriage is considered as an event
which influences the probability of another event. The explanation was that married
people were more committed to the dwelling and environment, since they had to
account for (at least) two people’s careers, obligations, and other types of local ties.
Mulder (1993) also found that single people move more often than cohabitants and
married people. Many of their moves were, however, related to the event of forming
a union (marriage and migration synchronization). Interestingly, there were large
differences between cohabitants and married people. Cohabiting people appeared to
be much closer to single people than spouses in their migration behaviour.
Migration behaviour of married people seems to be a combination of risk minimisation and simultaneous efforts to meet the needs of several household members.
Mulder and Hooimeijer (1999) found that the presence of the second earner in
couples and families resulted in a strong negative influence on long-distance moves.
They supported the idea that restrictions on migration arise from having to take into
account more than one daily activity. Mincer (1978) and Newbold (1998) claimed
that marital status and family composition affect the utility associated with migration, because family size influences the costs associated with moving. Families must
consider the changes in the income of its members, and the structure of the social
environment (school, workplaces). According to a study of Chang et al (2003) about
economic and social status in household decision-making in Taipei, mobility decision
varied according to households’ social and economic organisation. The higher a
person’s economic and social status in the household was, the more his or her needs
were counted in the migration decision. Snaith (1990) found that the mobility
decision of dual career households is complex and can produce a range of possible
outcomes. Quality analyses in dual career households showed that often the decision
was made according to the needs of the household member with a less common
profession.
114
There is also ample empirical evidence concerning the connections between marriage and the distance of migration, but these analyses usually do not explain how
these connections work. Sandefur and Scott (1981) and Warnes (1992) found that
marriage affects migration distance; long distance migration was more affected by
marital status than short distance: married people tended to move over longer
distances. On the other hand, Mulder (1991) observed that, generally, marriage related movement itself belongs more often to the short distance migration category.
Law and Warnes (1982) argue that married couples are more likely to be innovative
in the choice of destination by moving to remote areas. However, such a result could
also be influenced by third factors, such as house prices or natural environment
indicators.
Emotional attachment is a part of social capital. Several studies explore the phenomenon of community emotional attachment (Swanson et al 1979, DeJong and
Fawcett 1981, Goudy 1982, Bell 1992, Beggs et al 1996). Surveys show that the
longer one lives in the same place, the higher the probability of remaining there is,
because of ties and habit (e.g., Morrison 1971, Sandefur and Scott 1981, Nieminen
1983, Courgeau 1985, Hartshorne 1992). Knox and Pinch (2000) called this principle “cumulative interia” which is usually explained in terms of emotional attachment
that develops towards the dwelling and immediate neighbourhood, and the reluctance to sever increasingly strong and complex social networks in favour of the
unknown quantity of the pattern of daily life elsewhere. DaVanzo (1978), for example, emphasised that the connection between unemployment and migration willingness were clearly visible on recent migrants, whereas people who had not moved for
a longer period of time did not always respond to unemployment with migration.
The explanation could be the influence of community attachment. For example,
Theodori (2001) used data collected in a general population survey from a random
sample of individuals in four communities in Pennsylvania to analyze the effect of
community satisfaction on self-assessed individual wellbeing. Substantial support
was found to the argument that satisfaction with the community and attachment to
the community were associated independently and positively with individual wellbeing. Greater community satisfaction and attachment resulted in higher levels of
perceived well-being. As migration is related to well-being, greater well-being can
also lead to lower emigration numbers.
Migration and social environment in Estonia
Factors related to family have been the dominating reasons for migration throughout
the years in Estonia (Table 2.3). From the different push factors of the social environment, needs of the family members (30% needs of the children, 27% needs of
other family members, Table 2.4) were the prevailing reasons for migration. The
main pull factors were the perceived security of the surrounding environment (42%)
and distance from friends and relatives.
According to the Living Conditions Survey there were more individuals among
married people who had moved at least once in life than among single people. The
same result was reported by Tammaru and Sjöberg (1999), who explained it with
115
the fact that there are many single households among the elderly who do not move
any more. The Living Conditions Survey examined all registered movements during
the lifetime, which, therefore, were not dependent on certain life periods. However,
the result might not indicate the greater freedom of movement of households, but
the influence of household formation as a migration event.
Table 2.15 People with and without children by their triggering reasons (% of the
whole group, Migrant Survey)
Household
career
Housing
Economic
No children
With children
24
34
12
17
10
16
11
7
15
7
9
4
19
15
100 N=185
100 N= 322
Total
30
15
14
8
10
6
17
100 N= 507
Social Natural
Self
actuali- environ- environ- Other
ment
ment
sation
Total
We can expect considerable differences in the migration behaviour of different groups
of people depending on their social capital. Four different household types were
compared in the Migrant Survey: single, living with a partner, with children, without
children. Cohabiting people may also have been with children. There were no differences in the triggering reasons between the people who lived with a partner and
single people, but the differences emerged between people with and without children (Table 2.15). People with children moved more because of household related
reasons. There were also differences concerning other reasons, but due to the small
size of the sample these results were statistically not completely reliable.
Table 2.16 Factors taken into account in migration by household composition
(Average values: –1 — did not take into account, 0 — it was not important, 1— took
into account to a certain extent, 2 — it was important, Migrant Survey)
Household
career
Housing
Economic
Selfactualisation
Social
environment
Natural
environment
Distance
Single
Partner
0.08*
0.31*
0.43
0.49
0.28
0.29
0.04
-0.07
-0.12
-0.13
0.39*
0.64*
-0.01
-0.08
No children
With children
0.09*
0.28*
0.35*
0.54*
0.24
0.32
-0.03
-0.02
-0.22*
-0.06*
0.56
0.53
-0.09
-0.02
Type of
household
* Statistically significant differences (0.05)
When analysing all factors considered in the migration process, instead of using
triggering factors alone, the results of the analysis may differ. Table 2.16 shows the
differences according to household types with all factors considered in migration.
Compared to families without children, those with children assigned statistically
more importance to household and housing factors and were more inclined to
116
consider social environment. People who migrated with a partner considered household and natural environment factors more often than single people. Comparing the
different groups we can argue that having children was the most influential migration process factor.
It could be assumed that social ties (and resulting obligations) govern the process
of migration, because the needs of several family members must be considered
simultaneously, which could complicate the moving process. Migrant Survey indicated that married people took more time to search for a new dwelling than single
people (Table 2.17), and families without children tended to move over longer distances.
Table 2.17 Household status, time lag between getting the idea of moving and
the change of residence, and distance (Migrant Survey)
Type of household
Status
Single, divorced, widow
Married, cohabiting
No children
With children
Time from idea to act
(in months)
Mean
N
9.1
188
11.2
254
11.5
154
9.2
269
Total
10
Average time for search of
new destination (in months)
Mean
N
3.0*
176
4.5*
216
3.8
139
3.8
239
4
Distance (km)
Mean
97
104
122*
90*
N
197
274
161
289
101
* Statistically significant differences (0.05)
Social ties can be a positive as well as negative capital and, therefore, we examined
the influence of partnership and the number of children on the difficulties arising
during relocation (Table 2.18). It appeared that children often complicate the process, but the presence of a partner tends to reduce the difficulties; this could also
illustrate the different character and impact of the social capital. Children complicate the change of residence, but having a partner will make it easier.
Table 2.18 Difficulties in migration and household composition (Only statistically important models, linear regression, Migrant Survey)
Children
Partner
Price of
housing
0.130
–0.102
Appropriate
place
0.103
–
Appropriate
size
0.115
–
Find a job
Leaving familiars
0.102
–
–
–0.127
In conclusion, it can be said that social resources tend to decrease the risks of
migration, but also increase the cost of migration. Estonian empirical data also
revealed different influence of social capital on freedom to move.
117
2.9 Distance and information
Some research state that visual and
personal contacts are the main
source of information that define
the distance (Lloyd and Dicken
1972, Chapman 1979) and perceived distance can be quite different from metric distance. Studies
show that most people have a relakm from capital
tively vague idea of their surrounding metric world (Cadwallader
students
retired
employeers
1989) and as the possibilities of virtual information channels advance, Figure 2.25 Migration distance from the
it has been argued that the physical capital among rural-capital-rural migrants
distance is about to lose its role at all by social groups (Source: Ainsaar 1990)
(Margolis 1977). Therefore, it is
theoretically more justified to use the concepts of mental and economic distance
(time and cost of travel) in migration studies, but the metric distance is still a more
often applied concept because of access to data.
Distance itself does not affect migration directly (except as a factor of regional
differences), but it influences migration through implicit costs and information of
moving. Therefore, distance should be viewed as a migration barrier that is connected to the decrease of information flows and the increase of the cost of migration. Longer distance rises expenses for information, transportation, psychological
costs (Carlstein et al 1978, Bengtsson 1981, Borgegård and Håkansson 1995). However, the function between distance and information is not always linear, because of
uneven spread of information over distance.
An interesting phenomenon related to the distance of migration is the selectivity
of the reasons for migration by distance. Many authors have found that migration
over long distances is more often instigated by work-related motivation (Morrison
1972, Gleave and Cordey-Hayes 1977, Tervamäki 1987, White 1990, Stillwell 1991,
Bonaguidi and Abrami 1996) or educational motivation (Mulder and Hooimeijer
1999), while migration over shorter distances occurs more often because of dwelling
reasons. Mulder (1993:212) explains the distance-reason relationship with locationspecific capital:
35
30
% of group
25
20
15
10
5
0
0
25
50
75
100
125
150
175
200
225
In the period during people live in a particular place, they build up ‘capital’ in the sense of
investments in a dwelling, a position in the labour market, and all sorts of local ties. They
only leave their present location if they expect to improve their situation elsewhere. Moves
therefore tend to take place in the area within daily reach, unless there is a particular reason
for educational or occupational career.
Several authors (Olsson 1965, Stillwell 1991) have noticed the decrease of migration distance with age. Survey results from Estonia (Ainsaar 1990) concerning return
migrants revealed that the general rule that migration volume decreases with
distance was applicable only to people in the working age. The migration volume of
118
students and retired people even rose slightly with the increase of distance (Figure
2.25). The average migration distance in a certain time period and region is influenced by a proportion of people in specific age groups and with specific migration
motives from all movers (Warnes 1992, Stillwell 1991).
Information is a decisive factor in all stages of the migration process. Information determines the final choice, but on the other hand, it is already crucial in the
development of migration willingness, because dissatisfaction with the present condition is formed in comparison with familiar options. Information limitations might
become barriers. Lack of information causes insecurity and restricts migration,
whereas familiarity with the place supports migration. The amount of information
could also explain the tendency to prefer physically and psychologically close locations (home region). The similar effect of information availability might apply to
larger centres with a wider spread of both private and public information (Lloyd and
Dicken 1972).
In addition to the amount of information, the cost of information acquisition
might influence migration. The rational school of migration studies presumes that
the act of migration takes place only on the condition that the gained or expected
benefit exceeds the costs of migration. The amount and available sources of information might vary to a great extent in different individuals (groups) with different
characteristics (Herzog and Schlottmann 1984, Herzog et al 1993). For instance,
Herzog et al (1993) found that in the 1970s, first time movers in the USA made a
much bigger effort to gain information and, therefore, entered the labour market
much better informed than the recurring migrants. However, their costs for migration process were also higher. One strategy to lower the cost of moving is to use
cheap information. For example, using similar information channels is one reason
why people from particular communities tend to follow the same migration paths
(Morrill 1970). Finally, people have different aspirations on sufficient information
for migration. Cahill (1994) studied the process of change of dwellings, and found
that better educated and wealthier people were more interested in information and
also used it more. The amount of information and the ability to process it might also
explain the tendency that people with previous migration experience move more
often. Pred’s (1967) initial information matrixes already consisted of two arrows:
the amount and quality of information and personal skill to use the available information. Arsdol (1986) found that experienced movers were more orientated to
future mobility and were better able to actualise a moving plan and choice than
people who had not moved in the past.
The role of distance and information in Estonia
Distance influences the flows of information and the cost of migration, and, as a
result, the volume of migration decreases with distance. Correlation analysis
between all factors considered in migration and the distance of migration showed
that economic factors increased with distance, whilst housing as the reason for migration dominated in shorter distances. The same result is illustrated on Figure 2.26,
which presents the triggering reasons, and on Figure 2.27, which shows all considered migration factors.
119
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
0-30
31-60
social relations
61-90
studies
91-120
job
housing
121-150
environment
151-200
other
Figure 2.26 Triggering reasons and migration distance km (Migrant Survey)
0.8
0.7
Importance
0.6
0.5
0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
0-30
31-60
61-90
91-120
121-150
151-200
>200
Distance km
household career
housing
economic
self-actualisation
social environment
distance
Figure 2.27 Distance and average importance of migration factors taken into
account in migration decision (–1 — did not take into account, 0 — it was not
important, 1 — took into account to a certain extent, 2 — it was important, Migrant
Survey)
While the average moving distance of those who moved with the purpose of
improving their living conditions was around 61 km, those who moved because
of social relations or self-actualisation reasons covered twice as long distances.
(Table 2.19). People who moved for self-actualisation reasons also needed the longest time period between getting the idea of migration and realising it. The people
who moved because of housing took the longest time to find a particular destination.
120
Table 2.19 Triggering reason, migration distance, time from idea to migration,
realisation time of migration, and the number of alternatives (Means and total
cases, Migrant Survey)
Triggering reasons
Distance (km)
Household career
Housing
Economic
Self-actualisation
Social environment
Natural environment
Total
103.9* (145)
61.2* (75)
105.2* (65)
112.9* (36)
155.4* (47)
71.3* (29)
100.3 (478)
Time of pondering
(months)
7.9* (135)
13.1* (72)
7.1* (63)
15.8* (34)
11.4* (44)
7.1* (29)
10.2 (448)
Time of search
(months)
3.5* (116)
5.5* (59)
3.6 (60)
2.8 (32)
3.8 (38)
3.7 (27)
3.8 (397)
* Significant differences on level 0.05 with at least some groups (independent samples T-test)
The quickest movers were the people moving mainly because of household changes,
environmental or economic reasons. Almost all movers, irrespective of the reasons
for moving, had an average of 1.5 alternative options and there were no differences
between the groups with different triggering reasons.
The hypothesis about a link between distance, the number of migration options
and time for migration decision, using the the Migrant Survey data was tested. No
correlation could be detected between migration distance and the number of
options, familiarity of the place, and the time of search for a suitable location. It is
possible that the personal factors played a more important role. Therefore, the
correlation between distance and various personal features like wealth, education
and household composition was examined. The only statistically significant result
was that wealthier households (sign 0.00) and households without children (sign
0.10) tend to cover longer distances and people with a partner took more time to
search than single people (Table 2.17).
Similarly, the average distance of migration was not different in different migration directions. The rural-urban migration was the shortest with respect to both
Table 2.20 Time and distance by migration directions (Migrant Survey)
Direction of
migration
Town-town
Town-rural
Rural-rural
Rural-town
Total
Time for the search
(months)
Average time from
idea to act (months)
Mean
3.6
4.2
4.2
3.5
3.9
Mean
7.7*
15.8*
7.6*
9.9
9.9
N
118
74
97
70
359
N
123
89
114
75
401
Average distance of
migration (km)
Mean
95.9
96.6
96.9
82.9
93.9
N
133
91
122
80
426
* Significant differences on level 0.05 with at least some direction (independent samples T-test)
121
Table 2.21 Distance and difficulties in migration (Pearson correlation, Migrant
Survey)
Dwelling with appropriate price
Dwelling in appropriate place
Dwelling with appropriate size
To find a job
Financial cost
Leaving close people
Distance
Dwelling
with
appropriate
price
–0.13**
–0.04
–0.04
0.19**
–0.01
0.22**
1
0.59**
0.49**
0.07
0.31**
–0.01
Dwelling in
appropriate
place
1
0.61**
0.09
0.26**
–0.05
Dwelling
with
appropriate
size
1
0.12*
0.30**
–0.05
To
find
a job
Financial expenses
1
0.24
0.22
1
0.16
** Significant correlation at the 0.01 level (2-tailed), * at the 0.05 level
duration and distance, however, this difference was not statistically significant
(Table 2.20). The only statistically meaningful result was that people who moved
from towns to rural areas took the most time to ponder their decision — more than
a year on the average.
However, it appeared that distance increases difficulties that people experience
in migration. Distance complicated the search for a new job and made parting from
friends and relatives harder (Table 2.21). Yet finding a dwelling for a suitable price
was more difficult for those who move over shorter distances. The explanation might
be that since most people who moved over short distances were interested in
improving their living conditions, they were also more fastidious when it came to
choosing a new dwelling. The location of the dwelling, its size and financial costs of
the move did not have correlations with the distance of migration.
122
Conclusions: Well-being factors in migration
The influence of population, social and economic conditions, housing, natural environment and distance can function as push (ill-being), pull (well-being), trigger
(arousal) factors, or barriers (obstacles, registering features). The factors discussed
in migration research are quite compatible with the well-being and needs theory
framework. However, different factors have different explanatory power. Some
factors have important influence on migration on their own (human capital, social
capital, economic conditions, household composition, housing market, distance),
others combine the features of several merged factors (gender, nationality). The
influence of all these features is strongly intertwined.
In this chapter, a hypothesis was suggested that the dominating reason for migration is the part of the well-being, which under current circumstances is the most
deficient, and which can be changed by moving. Earlier studies allow us to make the
assumption that people consider several factors even when they change their place
of living due to one triggering reason, because they try to develop several parallel
careers and, usually, many needs have to be met at least on the minimal level. Why
are some people more mobile and others less so? The answer lies in the relationship
between the macro features of external environment (environmental differences,
space, information) and individual characteristics.
The analysis of the reasons for migration in Estonia allowed us to claim that
in the 1990s, the prevalent triggering reasons for migration were factors related to
the family. Employment and housing belonged to the second group of important
reasons. If all factors were taken into account in the migration decision, the most
influential factors in migration turned out to be environment and housing. Both
results demonstrated the diminishing influence of production forces on peoples’
choice of the living place in the stage of suburbanisation.
Detailed conclusions
Concerning the reasons for migration:
• The most remarkable change in the reasons for migration has been the
increase of housing-related migration in 1989–1999.
• Migration because of economic reasons decreased, but there were major differences between different settlement types.
• On the whole, the biggest changes in 1989–1999 took place in migration
related to the capital. On the background of a general decrease of employment-related migration, employment-related migration increased among inmigrants into the capital. Among the people leaving the capital the percentage
of those who left because of housing-related reasons doubled. Consequently,
the total increase of migration related to the capital due to in-migration was
primarily a result of study migration and, and to a lesser extent, employmentrelated migration.
• The importance of economic factors increased over longer distances and housing as the reason of migration was more important in shorter distances. Social
relations as the main reasons for migration prevailed in all migration distances.
123
Concerning the groups of migrants:
The selectivity of migration by age is generally known and it can be associated with a
person’s life stages and values. The analyses once again proved the importance of age
as one of the key factors of migration.
• The analysis of the reasons of migration indicated their selectivity by age.
• Self-actualisation was important for young people, but it was not at all important for the elderly. The importance of the natural environment increased
with age. Household career seemed to be important only during certain
crucial periods of lifestage changes.
• Although in 1989–1999, men moved more for employment-related reasons
and women for family-related reasons, there were certain changes taking place
over time. The differences in employment-related reasons between the genders decreased, but family-related migration was still more typical for women
than for men.
• The impact of education on migration behaviour could not be detected.
• Other observed features of human capital — such as previous experience of
migration and health condition — were also not confirmed as influential
factors of migration in Estonia.
• Social capital influences migration in various ways. Children complicate the
change of residence, but having a partner makes it easier. People with children
made more moves because of their household career.
• Economic capital supports mobility. Wealthier people were more mobile. The
wealthiest moved most because of housing-related reasons. The reasons for
migration in all other income groups were very similar.
• The vast majority of migrants were more satisfied with different well-being
aspects than before migration. On average, the increase in satisfaction was the
highest among the rural-to-urban movers.
Concerning migration behaviour:
• Usually, migration was simultaneously influenced by both pull and push
factors. Only moving towards settlements with better infrastructure (pull),
the desire of elderly people to move into different environment (push), and
need of privacy (push) were clearly distinguished as one-sided forces.
• People who moved because of different triggering reasons considered different factors when choosing a new dwelling. Consequently, the hypothesis about
a relatively narrow channelling of triggering reasons and factors considered in
moving found confirmation.
• Migration distance in Estonia had no correlation with the duration of information search or the number of alternatives in migration process, but it was
related to different triggering reasons for migration.
• More factors were considered in urban-rural and rural-urban migration, i.e.,
when a change of environment took place, than in simple urban-urban or
rural-rural migration.
• The main barrier of freedom of migration was the price of dwelling.
124
Methodological results:
• When migration was measured by triggering reasons alone, it produced different results from measuring migration by including all considered reasons.
• 11 types of reasons were specified as the major components of migration. The
classification was based on a combination of detailed pull, push and age data.
It is possible to use the classification for developing future survey questionnaires.
• Major differences in the characteristics of the people who had different reasons for moving indicated that the groups who move for different reasons
should be analysed individually in the same way as different age groups.
• The comparison of macro and individual analyses confirmed the importance
of housing as one leading migration factor in Estonia in the 1990s. Consequently, the data of the two levels were complementary. Individual data gave a
more thorough picture of different choices and behaviour of different groups,
but they were quite insensitive to regional differences. The macro-level data
enabled getting an overview.
Group of reasons
Migrant Survey
Migrant Survey
Push forces
Pull forces
1.Household career. Changes in Emancipation (separation from parents) To stay together with spouse
life stages
Changes in composition of family (birth
of a child, separation, marriage)
Separation from spouse
2.Self-actualisation. Motives
Studies
Local, cultural and social life
directed to increasing human
Other possibilities for rest and self-development
capital and giving more
possibilities
3.Housing. Motives related
To have a bigger living place
Possibility to improve living conditions
directly to desire to change
To have smaller living place
tenure
To change living conditions (amenities,
price)
End of rent contract
4.Economic. Employment,
Lack of professional job
Possibility to get a job
income
Lack of job
Better income
Low income
Dissatisfaction with job because of
other reasons
Attendance to the first job after studies
Because of work or studies of another Closeness to parents, friends, acquaintance
5.Social environment.
Household members, social
family member
Back to parents’ home
relations, social services
Environment to grow up for children Possibility to stay with relatives
and their education
Ethnic composition of area
Contacts
with
co-workers
and Security
neighbours
Neighbours
Because of language and cultural Phone connection possibilities
environment
Level of local health care
Level of schools
Wish to change living environment
Natural environment
6.Natural environment.
Environmental problems
7.Accessibility. Distance
Distance from home to job
Distance from job to home
Distance to bigger centres
Knowledge of the place
Possibility to buy food and consumer goods
Quality of roads
Appendix 2.1 Classification of reasons for migration in different surveys
Change of job or getting
a new job
Job
Environ-ment and infrastructure
Altruism, co-migration, Family reasons
need for closer social
relations, need for social
separation
Purchase of accommodation
Housing
Studies, preferen-ces, Studies
self-develop-ment
Migrant Survey
Living Conditions
Triggering
Survey
Changes in composition Family reasons
of house-hold
125
126
Appendix 2.2 Migration factors (Migrant Study Rotated Component Matrix)
1
2
3
4
5
6
7
8
9
10
Factors taken into account in selection (pull factors)
Possibility to get a job 0.23 0.77 0.00 –0.08 0.14 0.16 0.02 –0.01
0.03 –0.02
Possibility to improve living conditions 0.10 –0.07 0.18 0.77 0.01 0.10 –0.02 -0.06
0.03 –0.07
Closeness to relatives 0.22 0.05 0.10 –0.01 0.75 –0.04 –0.00 –0.01
0.08
0.04
Parents’ home –0.03 –0.04 0.06 –0.11 0.75 –0.12 –0.04 –0.10
0.08 –0.21
To relatives –0.05 0.00 0.00 0.04 0.72 –0.04 0.12 0.02 –0.22
0.15
Together with spouse –0.09 0.06 0.12 –0.06 –0.17 0.26 0.02 0.69 –0.03 –0.10
Environment –0.12 –0.10 0.79 0.07 0.05 –0.11 –0.06 0.01 –0.03
0.05
Ethnic composition 0.17 –0.13 0.63 –0.06 0.13 0.03 0.22 0.03 –0.04
0.08
Security 0.09 –0.12 0.73 0.18 0.04 –0.08 –0.03 –0.08
0.04 –0.12
Social life 0.62 0.14 0.15 –0.13 0.22 0.21 –0.04 –0.08
0.18
0.31
Recreation possibilities 0.57 0.20 0.09 –0.08 0.17 0.35 –0.06 –0.09
0.19
0.27
Better income 0.18 0.78 –0.03 –0.06 0.07 0.22 –0.03 –0.00
0.10
0.08
Distance from job 0.16 0.20 0.09 0.00 0.05 0.08 0.77 0.05
0.04 –0.02
Quality of roads 0.59 0.09 0.14 0.03 –0.10 –0.09 0.33 0.13
0.10 –0.10
Neighbours 0.32 0.01 0.44 0.00 0.08 –0.03 0.24 0.05
0.13 –0.31
Phone connection 0.69 –0.00 0.09 0.11 –0.03 –0.05 0.02 0.12
0.06 –0.27
Accessibility of health care 0.81 0.13 0.02 –0.00 0.05 0.06 0.04 –0.04
0.07
0.05
Possibility to buy common goods 0.74 0.20 0.05 0.02 0.09 0.05 0.19 –0.00 –0.00
0.03
Quality of schools 0.50 0.23 –0.08 0.04 0.09 0.12 –0.03 –0.05
0.51 –0.06
Closeness to the bigger centres
Knowledge of the place
0.67 0.24 –0.04
0.27 0.12 0.33
Age
0.00 0.06 0.10
Reason of change of residence (push factors)
Lack of professional job opportunites 0.21 0.66 –0.18
Lack of job 0.08 0.78 –0.10
Distance between home and job 0.10 0.24 –0.04
Low income 0.16 0.67 –0.04
Dissatisfaction with job 0.01 0.52 –0.07
To get a bigger living space –0.12 –0.07 0.16
To get a smaller living space 0.18 0.00 –0.05
To change living conditions 0.12 –0.00 0.02
To change a living environment 0.07 0.20 0.43
Studies
First job
Family members
Emancipation
Separation from a spouse
Separation from relatives
Because of children
Interpersonal relations with others
Changes in family composition
Environment conditions
Language problems
0.17
0.08
0.17
0.08
0.07
0.02
0.27
0.11
0.07
0.05
0.23
0.00
0.13
0.36
0.16
–0.00
–0.08
0.23
0.00
–0.22
–0.00
–0.04
–0.10
–0.04
–0.04
0.12
–0.13
0.02
–0.05
0.29
–0.08
0.64
0.21
0.08
0.04
–0.22
0.08 0.15 0.06 0.00
0.06
0.52 0.13 –0.03 0.08
0.11
0.00 –0.58 0.02 –0.06 –0.22
11
0.12
0.04
–0.05
–0.11
0.14
–0.03
0.10
0.00
0.02
0.09
0.06
0.06
0.13
–0.14
–0.16
–0.12
0.08
0.06
0.12
0.17 –0.01
0.04 –0.09
0.37
0.05
–0.11
–0.06
0.06
0.00
–0.05
0.79
0.01
0.79
0.31
–0.05
–0.07
0.02
–0.02
0.10
–0.06
0.02
–0.03
0.20
0.00 0.31 –0.03 –0.07
0.16 –0.14
–0.12 0.17 –0.00 –0.04 –0.04 –0.15
0.12 0.83 –0.01
0.05
0.07
0.02
0.00 0.10 –0.12
0.18 –0.04 –0.04
–0.05 0.15 –0.06
0.41
0.10
0.02
–0.00 –0.01 0.07
0.04 –0.05
0.01
–0.09 0.16 0.05 –0.02 –0.02
0.05
–0.06 0.11 –0.16
0.06
0.20 –0.03
0.12 –0.04 0.01
0.01
0.46 –0.02
–0.00
–0.07
0.02
0.14
–0.24
0.02
0.21
–0.00
0.02
0.26
0.03
–0.08
0.00
0.02
–0.12
0.17
–0.06
0.04
–0.07
0.01
0.06
–0.13
0.68 0.09 –0.05
0.04
0.08 –0.13
0.74 0.07 0.00 –0.03
0.14
0.04
0.07 0.21 0.09
0.21 –0.18
0.28
0.53 0.08 0.38 –0.15 –0.14
0.27
–0.29 0.09 0.60
0.03
0.16
0.00
–0.05 0.08 0.04
0.04 –0.00
0.86
–0.06 –0.01 0.06
0.68
0.14
0.11
0.08 0.17 –0.00
0.53 –0.04 –0.10
–0.02 –0.02 0.83
0.03 –0.02
0.05
0.01 –0.07 –0.06
0.05
0.33 –0.07
0.07 0.23 0.01
0.34
0.50 –0.12
Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.
a Rotation converged in 11 iterations.
127
Appendix 2.3 Age and gender of respondents of the Migrant Survey
80
70
men
number
60
women
50
40
30
20
10
0
15
20
25
30
35
40
45
50
age
Appendix 2.4 Previous and present place of residence, respondents of the
Migrant Survey
Capital
"Real movers" previous
"Real movers" present
All respondents previous
All respondents present
60
29
128
77
County
centres +
Narva
49
60
105
112
Satellite
towns
Other
towns
5
10
12
16
7
22
17
38
Hinterland of Other
bigger towns country
26
38
47
68
100
88
176
173
Total
247
247
484
484
Appendix 2.5 Age differences in factors taken into account when choosing a new
dwelling (The results of the T-test in relation to the mean values 90% significant
level: –1 definitely not, 0 not important for us, 1 — a certain extent, 2 — very
important, Migrant Survey)
Possibility to get a job
Better living conditions
Closer to known people
To parents’ home
To relatives’ space
Together with spouse
Environment
Ethnical composition
Security
Social life
Free time
Better income
Distance to job
Roads
Neighbours
Phone connection
Medical care
Possibility to obtain goods
Level of schools
Closer to centre
Knowledge of the place
–14
.396
.071
.406
.966
.539
.002
.218
.007
.032
.522
.932
.347
.000
.993
.775
.177
.885
.465
.216
.117
.815
15–24
.199
.354
.540
.558
.483
.004
.401
.071
.671
.900
.426
.039
.000
.264
.678
.351
.768
.675
.336
.818
.884
25–34
.513
.146
.453
.867
.209
.762
.396
.837
.250
.283
.242
.267
.000
.168
.561
.126
.139
.368
.001
.031
.160
35–44
.242
.035
.887
.397
.835
.090
.082
.264
.788
.427
.244
.766
.000
.874
.665
.376
.251
.066
.178
.867
.180
45–54
.691
.050
.478
.725
.057
.431
.218
.499
.235
.432
.127
.425
.537
.606
.236
.192
.475
.803
.000
.294
.590
55–64
.071
.467
.542
.630
.021
.000
.401
.344
.325
.842
.032
.109
.000
.724
.559
.438
.666
.630
.001
.170
.868
128
129
Migration as behaviour
3
130
131
3 Migration as behaviour
Introduction
The behavioural approach is the least studied and youngest area in migration
research. By the 1970s, it was clear that previous approaches to migration had
exhausted their theoretical resources. Migration models were not valid for all types
of migration flows; they were applicable only to specific types of societies, and could
explain only “normal” events (Zelinsky 1983). Behavioural approach to migration
was born in conjunction with the development of behavioural sciences. The very
first publications in the field that have retained their importance were papers by
Wolpert (1965), Brown and Moore (1970).
Unlike macro level approaches, the behavioural approach strives to explain individual or household migration by analysing different stages in migration decision
process. The motivational factors and limitations of behaviour are important determinants of the outcome. More than other theories, the behavioural approach to
migration emphasises the importance of subjective perceptions of people. It is based
on the assumption that the behaviour of people is based more on a perceived world
and less on the objective real world (Walmsley et al 1998, Lazarus and Folkman
1984, Cadwallader 1989), and a subjective interpretation of the cause is also the
reason of behaviour (see de Bruijn 1992). Consequently, a large number of migration
studies that analyse the movement of people from the point of view of subjective
reasons could already be considered as partially behavioural approaches. However,
because of the relative youth of the behavioural approach and significant labour
input required for data acquisition in this framework, behavioural migration studies
constitute a minority in all migration research. Yet the behavioural approach has
proven itself (Jones 1978, Cadwallader 1989) and shows a good potential for developing a deeper understanding of migration.
After the birth of the approach in the 1970s, there have been surprisingly few
attempts to expand the theories or test their validity against empirical data. Still,
some mainly theoretical writings (Roseman 1971, Molho 1986, Cadwallader 1989,
Moon 1995) discuss several elements of the whole process. However, these theories
have seldom been examined in the context of empirical data. The aim of this chapter
is to construct a principal behavioural model of migration and to evaluate its structure using empirical data from Estonia. In order to develop this theory, facts known
from migration research are used alongside with psychological theories. The main
emphasis will be on the mental stages and individual strategies of behaviour.
132
3.1 On the search for a model
In the second chapter it was assumed that there are at least four important factors in
migration that have the greatest impact on the whole process: (1) people’s individual
characteristics and needs, (2) perceived differences in the environment, (3) the emergence of the idea of migrating, (4) response. But what are the steps of behaviour
within these larger stages of migration behaviour? Different authors have identified
different stages of the process. Some behavioural element can be found already in
the theory of intervening opportunities by Lee (1969)1 . The following is a summary
of a number of Lee’s theses with some additional interpretations. Lee proposed that:
1. Different factors affect different people in different ways. Accordingly, we can
distinguish groups of people who respond to the same general sets of factors with
a similar strategy. Perception of different local factors is a more important determinant of migration than the actual situation. Thus, personal intelligence, sensitivities, number of contacts, sources of information, and awareness of conditions
elsewhere enter into the evaluation of the situation at origin.
2. Persons living in an area have an immediate and often long-term acquaintance
with the area and they are usually able to make considered and unhurried judgements. This is not necessarily true of the factors associated with the area of destination.
3. While migration may result from a comparison of factors at origin and destination, a simple calculus of +s and –s does not determine the act of migration. The
balance in favour of the move must be enough to overcome natural inertia (distance, laws). The same set of obstacles affects different people in different ways.
4. Migrants responding primarily to the plus factors at destination tend to be positively selected. They preserve their opportunities elsewhere and can weigh the
advantages and disadvantages at origin and destination. Degree of positive selection increases with the barriers.2 Migrants responding primarily to the minus
factors at origin tend to be negatively selected. There could be a danger that
minus factors seem so overwhelming to the whole group in large regions that they
may not be selected at all. The group might not perceive the poorness of their
situation, or they might consider their opportunities no better elsewhere and so
just wait for the improvement of the situation.
5. Personal factors also affect the individual threshold to migrate. Some of them are
more constant throughout time, some change with life cycle. In addition, there
are personalities who are resistant to change — change of residence — and there
are personalities who welcome change for the sake of change. Consideration of
personal differences could also prove useful in finding answers to the question,
why some people are firmly settled in one place, despite their needs. A person
who has migrated once and has broken the bonds that tie him to the place in
which he has spent his childhood is more likely to migrate again than the person
who has never migrated before.
6. And finally the decision to migrate is never completely rational and for many
people the rational component is much less important than the irrational.
1
2
These statements were posed by Lee (1969) as hypothesis.
This idea is repeated also by Myrdal (1957).
133
Lee’s (1969) ideas already reflected the behavioural notions of + and — forces, of
different reaction of different people to these forces, and of the barriers of movement. The only lacking element was the explanation why these processes take
exactly this form. Brown and Moore (1970) in collaboration with John Silk, James
Simmons and Julian Wolbert, saw migration as a process of adjustment where change
of residence served as a tool to achieve a higher satisfaction of desires and needs. The
receiver of information was “seen as an active element who samples available information channels with varying intensity to serve his own interest” (8). Brown and
Moore differentiated between two main phases in the migration process.3 The first
phase consisted of the development of discontent or stress resulting from interaction between internal and external forces. The stress can be relieved by adapting the
needs, changing the environment, or moving into another environment. The second,
a broader phase of migration process consisted of the decision to relocate. Levin
(1951) divided all intentional actions broadly into three phases: the first phase was a
motivational process, the second an act of choice, and the third the action itself.
Golledge and Stimson (1997), like many other geographers (see Boyle et al 1998),
start their process modelling from the point of decision about migration (Table 3.1).
They identified three phases of the process: the decision to migrate, decision about
where to move, and the final selection of destination. The decision-making is
affected by different propensities to move and different degrees of freedom, utility
of the place, preferences, choice, and cognitive framework that can depend on the
time of making the decision and on the needs of the person. Chapman (1979) proposed a four-stage model of decision-making: (1) goal formation, which is related to
stimulus, stress. The goal can be economic (economic man), social, political, or consist of multiple goals; (2) Acquiring information — search, experience; (3) evaluation of choices in a situation of certainty or uncertainty. Evaluation may involve formal testing of alternatives by means of analytical procedures or it may be based on a
value judgement of the decision maker; (4) Decision and positive or negative feedback. After this stage, the circle might start all over again.
Loyd and Dicken (1972) developed a spatial model of location selection for
businesses, which seems rather relevant also for humans. The model starts with the
perception of stress and stress tolerance. The second phase is an attempt to make
adjustments at the existing location. In case of maladjustment, requirements for a
new location are specified (establishing the level of aspirations). Then follows a search
process according to existing spatial knowledge, as well as preliminary filtering, search
for suitable location within acceptable area, evaluation of alternative locations, comparison of satisfactory alternatives, and finally modification of the level of aspirations
if the search process has failed.
To sum it up, one could assume that the behavioural model of migration would
have much in common with the general models of human behaviour. From the variety of general psychological theories, the theory of planned behaviour or theory of
trying could be best suited for analysing human migration. A planned type of action
requires that the actor has a reason for action and considers it realistic to initiate the
steps needed for action (Clark 1984, Bagozzi 1992, Abelson and Levi 1998). The
theory of planned behaviour is complementary to the theory of reasoned action with
3
The paper applies only to intra-urban migration.
134
an added component — perceived behavioural control. Perceived behavioural control is defined as “the persons” belief as to how easy or difficult performance of the
behaviour is likely to be” (Abelson and Levi 1998:457) and it is influenced by thoughts
about personal deficiencies or external obstacles that might possibly thwart the performance of an act (Ajzen and Madden 1986:181). The main difference between the
theory of trying and other theories is that this theory allows viewing action as a
sequence of attempts to achieve the final performance, “in which the attempts
involve physical as well as mental efforts following the formation of an intention to
try” (Abelson and Levi 1998:183). This theory also introduces the notion of a selfregulatory process. Self-regulatory processes mean “the monitoring appraisal, and
coping activities that translate attitudes into intentions, subjective norms into intentions and intentions into actions leading to goal attainment. These self-regulatory
processes involve connate, emotional, social, and volitional sub-processes”. All
together, the self-regulatory processes are an appraisal that leads to emotional reaction and coping responses.
There is reason to believe that people use different models of behaviour in solving
different problems (see also Gold 1980, Reggiani 1991). Similarly, different strategies could be used in different stages of the formation of the migration decision. The
most important question within these frameworks is — on which principles are the
decisions of the people based? Often various rational approaches have been developed in order to explain human behaviour4 . Of course, it might be principally wrong
to apply a single model to all people. Abelson and Levi (1998:234) criticize such
standardisation with an argument, which states that “... whether humans are rational
is reminiscent of arguments such as whether humans are basically selfish or whether
human nature is intrinsically aggressive”. There are also some problems concerning
the definition of rationality. Elster (1993) emphasizes the question of indeterminacy, i.e., situations in which the rational choice is not well defined (genuine uncertainty). The author discusses the question whether preferences can be said to be
rational (time preference, attitudes to risk, regret), and concludes that construction
of rationality is partly discovery and partly decision. However, there seem to be
some similar elements in migration and rational behaviour. For example, the presence of a specific goal has often been considered as a characteristic of migration
(Golledge and Stimson 1990:258, Hartshorne 1992) as well as of rational behaviour.
Interpretation of migration is very similar to the ideology of the rational model of
behaviour, since many researchers (Abelson and Levi 1998:233, Goldthorpe 1998)
find that rationality is defined precisely by the goal-directed behaviour. A precondition of a rational model of behaviour is that people behave in a way that is profitable,
making choices that they believe will improve their well-being. Several migration
researchers (Willekens 1987, Golledge and Stimson 1990, Mulder 1993) emphasise
that migration is not an end in itself, but rather a means of attaining something:
wealth, status, comfort, activity, autonomy, affiliation, morality, or well-being. Even
the definition of the term “reason”, often used in migration studies, is based on the
presumption that it is a rational motivation of action (see Colman 2001). According
to Mulder (1993), migration is an adjustment mechanism for meeting the needs
4
Even more often the rationality is used as explanative argument of behaviour in biology.
Act itself
Place preference and
utility choice
Re-evaluation and
Decision
restructuring of search
process
New search and evaluation, restructuring
of search procedure
Decision action. Feedback
Evaluation of choices
Decision to stay or
move, different degrees
of freedom
Altering of aspirations
Adjustment of needs,
change of environment, relocation,
Act of choice
Goal (stimulus, stress)
Chapman (1979)
Acquire of information
Collective needs,
continuos stimulus,
stress
Motivational process
Golledge & Stimson
(1997)
Aspirations
Simultaneous search
and evaluation
Brown & Moore
(1970)
Levin (1951)
Table 3.1 Steps of migration decision according to different authors
Evaluation of alternatives
Attitudes and adjustment.
Perceived behavioural control
Intention to migrate, wording
of aspiration
Selectiv search and evaluation
Change of equilibrium between
needs resources and environment, appraisal
Migration model used in this
book
Modification of aspiration level Decision, act adjustment new
equilibrium
Evaluation of alternative locations
Comparison of satisfactory
alternatives
Attempt to adjust at existing
location
Specify location requirement
Search process
Preliminary filtering
Search for suitable location
within acceptable area
Stress, tolerance
Loyd & Dicken (1972)
135
136
arising from parallel careers of household, family, education, housing, and occupation. But if the same goals can be achieved without relocation, people could use
other, less costly solutions.
Yet the studies of rationality have had their ups and downs. The first set of
rational theories of behaviour was based on the assumption that actors have perfect
knowledge and ability to behave. However, such constructs were not confirmed and
there emerged some doubts concerning the rational theories (Schoemaker 1982,
Hogarth and Reder 1986). Later it was understood that actors are still rational, but
they tend to use subjective rationality. “Actors may hold beliefs…. and pursue of
actions, for which they have “good reasons” in the circumstances in which they find
themselves, even though they may fall short of the standards of rationality….”
(Abelson and Levi 1998:171). A new interpretation of rationality was offered by
Simon (1957, 1982), who developed a model of bounded rationality. According to
this approach, “behaviour may appear to be economically or spatially irrational, but it
merely reflects the outcomes of variations in individual ability to cope with and store
information that is fragmented and incomplete while operating under severe time
constraints” (Simon 1957:8). Gradually the simple homo economicus approach was
replaced by a much more complicated concept. New terminology was employed in
the discussion concerning human behaviour: low-risk behaviour, minimum regret,
satisfaction, bounded rationality, risky and uncertain behaviour. It appeared that
people could base their decision-making on quite different factors. Still there is no
agreed concept on the character of humans as decision makers. Some authors conclude that human decision maker has been variously seen as a corrigible rationalist, a
bounded rationalist, an error-prone intuitive scientist5 , a slave to motivational forces,
or the butt of faulty normative needs (Abelson and Levi 1998:233, see also Leyens
and Dardenne 1996). The latter seems to be the most interesting of these different
approaches. It also enables the widest application of the rational model. According
to the butt of faulty normative model, human reasoning cannot be faulty: it sets its
own standards and, therefore, always allows constructing human behaviour as
rational. There are clear rational elements in migration behaviour: for example, often
a solution is processed in a manner that minimises efforts and provides satisfying
solution, which is good enough in certain circumstances (Brown and Moore 1970,
Mellers et al 1998) in order to rationalise resources per action.
Important restrictions for the application of the theory of rationality are the limits of rationality. Sabini (1992:685) concludes that there are at least four limits of
rational action. Different studies show that people can be expected to be less
rational in their intentions and behaviour
1. When habits oppose reason. (At the same time, a habit can also be a rational way
of behaviour, habits save time and resources.)
2. When emotions are involved.
3. When people do not believe that they are capable of doing what reason dictates.
4. Because of moral reasons.
5
Intuitive scientist — similar to bounded rationalist, but boundedness comes not from the impracticality of spending the time and effort to make truly optimal judgements but from genuine failure
to appreciate normatively appropriate strategies (Abelson and Levi 1998:233)
137
All these reasons for deviation from rationality are also relevant for migration behaviour.
Despite the limits, many authors (Goldthorpe 1998, Boudon 1994, De Bruijn
1992, Hecter and Kanazawa 1997) believe that a rational theory could still be the
best theory for sociology. For example, the rational theory of human capital has been
successfully applied to migration (see Cadwallader 1989:499). Speare (1971) demonstrates using the Taiwan migration study the strong explanatory power of the costbenefit indicators. Nevertheless, the data from Taiwan also proved that, despite the
good results with regard to rationality, people might have only vague idea of real
costs and benefits.
A model constructed on the basis of psychological and behavioural theories of
migration (Figure 3.1) shows that needs, well-being, learned coping strategies, perception and information are the central attributes of the whole migration model that
influence all stages of migration. All those attributes depend on human, social and
economic resources. The initial settled situation could be characterized as equilibrium between human needs, resources, information and perceived environment. Disturbance of the equilibrium leads to the appraisal of the situation and
development of adjustment strategies. Adjustment strategies can be regionally local
(repairing the flat, new job in neighbourhood, adjustment of needs) or non-local
Equilibrium
between
environment,
personal needs,
resources
Decision about
migration or return
to the stage of all
adjustment
strategies
Evaluation of
alternatives and
present and future
place of residence
Appraisal - change of
equilibrium between
environment,
personal needs,
personal resources,
attitudes or coping
strategies
Individual needs,
resources (inc. skills
of adjustment,
attitudes)
Search and
modification of
search
Abandonment of
local coping and
intention to migrate.
Wording of
aspirations
Decision about
search
Coping strategies:
local and nonlocal.
Information search.
Perceived
behavioural control
Coping without
migration
Figure 3.1 Behavioural migration model
138
(migration). Most often studied “reason of migration” in that context is a well-being
factor, which is expected to change through moving. We can assume that in a situation when a person or household has decided to search for possibilities for changing
their place of residence, they already formulate the expected outcome or specific
aspirations. The subsequent search procedures are presumably concentrated on this
factor. After some potential places of residence have been found, they will be
assessed and compared to each other as well as to the possibilities in the present
dwelling. Presumably, a preliminary selection takes place already during the search
process, and only such cases will be accepted for final consideration in which the
factor causing migration is on a satisfactory level (e.g., when movement is caused by
the need to find a job, new places of residence are chosen only from among those
places where there is a high probability of finding a job). After comparing the alternatives, the decision to migrate is made or abandoned. In the second case, people
return to the stage of considering alternative adjustment strategies. One possibility
could be temporary abandoning of migration intention, i.e., by considering other
coping strategies. Through this, a new (temporary) equilibrium is achieved until new
changes appear in equilibrium.
Although one might think that the described model is suitable only for describing
voluntary migration, it is also valid for enforced migration and for different distances
and types of migration. In the framework of this approach, enforcement can be
interpreted as a need to follow certain rules. In milder cases of enforced migration
(e.g., co-migration), the enforcement can even be interpreted as part of the needs of
the person (need for social contacts, need for social approval). In the following sections, this model will be elaborated in more detail and results of inductive-deductive
interviews will be presented. Theoretical modelling of migration behaviour will be
presented in three thematic groups: appraisal and coping strategies, choice, individual differences.
3.2 Appraisal and coping strategies
This section analyses the following migration process episodes and concepts: needs,
well-being, attitudes, appraisal, stress, adjustment without change of residence.
The model starts from the situation of equilibrium of needs, resources and perceived options. Changes in individual needs or environment perception lead to a
change in personal well-being, which within migration aggregated individual data
level analysis (Chapter 2) is called a well-being reason. Brown and Moore (1970)
based their approach on external and internal forces as the foundations of the behavioural chain of migration. Needs and resources can be labelled as internal forces and
perceived options as external forces. Change of needs could be seen as a factor that
explains many migration reasons related to life stages, whereas age variations explain
varying degrees of migration intensity. People in certain ages and life stages experience greater transformation of needs and freedom to act according to needs. These
changes are reflected in higher mobility, because people have more “reasons” to
139
change the place of residence. Sometimes a distinction has been made between
inner (self esteem, social anxiety, self-rated health, congruence between expectations and achievement) and outer (satisfaction with friends, residential satisfaction,
perceived quality of time use) dimensions of subjective well-being (Lawton 1983,
Lawton et al 1983). In migration, both factors seem to have important influences. As
the needs (see Wetherly 1996) and their significance for people are very diverse,
there are numerous combinations of different needs and satisfactions that could
evoke the idea of migration. The majority of migration research is based on questionnaire-type information gathering. In that style of survey, the needs could be indirectly measured through statements about value judgements. Several earlier migration studies have confirmed the influence of values on migration (Kephart 2003 et
al). For example, studies of migration behaviour of cohorts have confirmed that
career orientation and high regard for self-actualization influence migration (Ainsaar
2003). A survey conducted in Estonia (Ainsaar 1997b) proved that incorporation of
values into migration equations increases the explanatory power of migration models. The value expectancy model developed by De Joung and Fawcett (1981) enables combining the influence of regional differences and personal needs to explain
migration forces. According to the model, the strength of migration intentions follows from the value attached to the outcome of the relocation multiplied by the
expectancy that relocation will lead to the desired outcome. Among the values
attached to relocation is the decrease in residential stress or, on the other hand,
increase in residential satisfaction. This model could also be called “cognitive calculus”, because “The value expectancy model assumes that people will usually behave
in a forward looking positive way, making choices that they believe will maximise
their well-being” (Golledge and Stimson 1990:258).
Our model is based on the assumption that the whole process is triggered by
changes in the state of equilibrium. Consequently, the process starts with changes
and perception of these changes. Perception is a process in which a person becomes
aware of receiving information (Colman 1992). In migration, the process of perception is mostly conscious perception, as persons must formulate more or less clear
goals of activity. Abelson and Levi (1998) define problem recognition as realisation
that a discrepancy exists between what is and what should be. This fits with
Moon’s (1995) distinction between four levels of value settings: the level of ideas,
the level of expectations, tolerance level, and the lowest threshold of intolerance.
According to Moon’s interpretation, values between lowest threshold and general
tolerance level create reaction. Readiness to migrate depends on the gap between
the real situation and the ideal. As indicated by our model, the changes in internal or
external factors are always accompanied by some individual reaction. Several
authors use the term action threshold to label the need for stimuli required for
action (Mintzberg et al 1976, Abelson and Levi 1998), others use the term tolerance threshold or stress tolerance (Brown and Moore 1970). Moon (1995) calls
zone between tolerance level and the lowest threshold of intolerance ‘reactionary
zone’. However, during this period people will probably use some other type of
adjustment mechanism (internal), because every important change affects the state
of a person’s well-being and certain adjustment efforts are required to achieve a new
equilibrium.
140
Miller and Starr (1967) describe two possible ways leading to problem perception:
(1) a challenging event may prompt the person to notice that a problem exist —
triggering reasons in migration approach; (2) individuals are observing certain diagnostic signs that indicate departure from equilibrium. Mintzberg et al (1976) propose that the challenging event might often consist of a set of stimuli rather than a
single stimulus, and their influence is cumulative. The cumulative character of
migration stimuli can be illustrated with results by Ni Laoire (2000), who found that
migration reasons could also derive from the individual’s past biography, not only
from the present situation and the final reasons could also be explained by the influence of a gradual accumulation of factors.
Our current knowledge about human behaviour enables making the assumption
that decisions made by the individuals might be influenced both by the past and by
the future (Foss 1984:53, Luuk and Tulviste 2002). It might take a long time before
the decision to migrate is made. Over this long period, several strategies are
employed to adjust to or modify the factors that influence the life situation, and the
conditions necessary for making the move are arranged. The migration decision is
not identical to the actual move and the vocabulary of motives will often be related
to the last link in a long causal chain. The final drop in that process can be the
so-called triggering reason in migration behaviour.
Several authors (Clark and Cadwallader 1973, Brown and Moore 1970, Chapman
1979) assume that the whole process is triggered by stress6 , although sometimes a
precise definition of stress is not specified. According to the stress approach, moving
is not a standard solution — people essentially tend towards settling and relocation
would require a significant amount of pressure (stress). In our model, the whole
process is triggered by the disturbance of equilibrium. This approach assumes that in
principle, people are mobile, but movement could be inhibited by various barriers.
Therefore, sufficiently high degree of motivation or freedom to move would be
required to overcome the barriers. According to the last approach, people constantly
reassess their needs and options and are always in principle ready to migrate (see
Foss 1984). Yet this starting point needs further investigation. For example, Speare
(1971) found that many non-migrants appear to have never given any serious consideration to the idea of moving anywhere, which could also mean that they used other
coping strategies. On the other hand, migration research has identified several functions of pull forces, which are not necessarily related to stress.
Psychological literature uses the term appraisal to designate the phase of personal activation following the emergence of the reason of migration and its perception. Appraisal is the evaluation of internal or situational conditions as they apply to
one’s well-being (Abelson and Levi 1998). The whole process is led by desires and
desires are fundamental psychological determinants of intentions7 . Appraisal and
“Stress — psychological and physical strain or tension generated by physical, emotional, social,
economic, or occupational circumstances, events or experiences that are difficult to manage or
endure” (Colman 2001:711)
7
Here we bear in mind only volitive, not appetitive desires. “A volitive desire, like an attitude
towards an act, is based on reasons, but unlike attitude it implies a motivational commitment.
Common words expressing volitive desires are “want”, “wish”, would like and possibly “covet”
(Abelson and Levi 1998:185)
6
141
intentions are expected to emerge if the evaluations are strong enough, but according to the authors, this part is weakly handled in theories (Abelson and Levi
1998:186). Lazarus (1991) assumed that appraisal processes of internal and
situational conditions lead to emotional responses and these, in turn, to coping
activities. Even two appraisal stages can be identified: primary and secondary. In
primary appraisal one assesses:
1. The motivational relevance of the conditions leading to the appraisal;
2. The motivational congruence, or extent to which the conditions thwart or
facilitate achievement of one’s goals;
3. One’s ego involvement.
Secondary appraisal addresses the resources or options for coping with internal or
situational conditions:
1. Attribution to oneself or to another of credit or blame for any harm or benefit;
2. Self-efficiency with regard to acting on the situational conditions;
3. Self-efficiency with regard to regulating one’s own internal states;
4. Expectations of forces operating beyond one’s control (Lazarus 1991:187).
Coping strategies, intention of migration
Lazarus (1991) differentiates between two general types of coping: problem focused
and emotion focused. Problem focused coping consists of efforts to overcome or
reduce the effect of an undesirable situation (changing the physical situation, breaking off relationship, or persuading someone to do something to remove an external
threat). Emotion focused coping refers to cognitive strategies to master, reduce or
tolerate an undesirable situation. Such strategies might entail denial, avoidance of
thinking about an appraisal, or re-conceptualising the source of dissonance or its
meaning. Brown and Moore (1970) distinguished between three possible adjustment strategies: adjustment of needs, restructuring the environment and relocation. In the classification of Lazarus (1991), relocation belongs to the group of
problem-focused coping. Our prime interest lies in the question, how and when
people get the idea to improve their well-being through migration. As migration
always involves some expenditure of resources, one could assume that before deciding for migration people consider local coping strategies. A decision to migrate is
made only if the expected benefit is greater than the accompanying costs and the
same outcome could not be achieved without migration. Supposedly, the migration
decision “matures” through three phases and in each phase an individual must
answer the following questions:
1. What are my well-being prospects in present location?
2. What would be my well-being prospects elsewhere?
3. Which is the more beneficial option for coping?
Moon (1995) considers the most important factor in making the migration decision
the question, how well the person values his or her moorings. The conceptualisations
of mooring are subjective and it includes the following segments:
1. The level of mooring (anchoring, cohesion);
2. The fact that the issue is personally important;
142
3. The (in)ability to escape from conflict and, on the other hand, the level of
freedom to manipulate social and cultural factors.
According to Moon (1995:514) attitudes depend on a “degree of optimism toward
place habitat”. Optimism raises tolerance. Migration is a process that is expected to
bring benefit. This benefit is an expected and imagined benefit (Moon 1995, Newbold
1998), which in the case of limited information depends on the image of certain
regions. For example, in the case of growing regions, regional prospects can be overoptimistic, the impression of the opportunities that they offer might be easily overestimated. Declining areas, on the other hand, often have an unfavourable image that
aggregates their problems (Chapman 1979).
When choosing relocation or when considering its feasibility, a person presumably
starts to collect information about suitable options. Yet it is also possible that the
decision to migrate is initiated by some new information about a potential place of
residence (change of perceived options). Apparently in many cases, the decisionmaking takes place simultaneously with the search for alternative possibilities. Constructing a model of migration, the first question would be: is the decision to leave
current place of residence made before or simultaneously with the choice of a new
dwelling? Several authors (Valkonen and Martelin 1986, Brown and Moore 1970)
propose that the decision to leave and the choice of dwelling are separate decisions,
although psychological models of behaviour could quite naturally picture also a
correlation between the decisions.
Information is important at all stages and the changes in information reshape the
whole situation. People acquire information from various channels. Particularly important in migration studies are the spatial differences in information processing,
which will be analysed in more detail later in this book.
After migration has been pre-selected as one possible adjustment strategy, this
situation could be called migration intention. This stage is often described in migration research as willingness to relocate. In addition to other individual characteristics
(see 3.5), attitudes have been identified as an important factor in the development
of migration intention (Triandis 1971). According to the theories of planned behaviour, attitudes are influenced by subjective norms and perceived behavioural control, and are a source of intentions (Ajzen 1991). Some authors (Ajzen and Fishbein
1980, Fredericks and Dossett 1983) have found that some attitudes affect behaviour directly and others indirectly through intentions (Bentler and Speckart 1979,
Liska 1984, Tallman et al 1993). The results of Pagozzi and Yi (1989) show that both
intentions and attitudes influence behaviour. When intentions were well formed,
they completely mediated the effect of attitudes on behaviour, in keeping with the
theory of reasoned action. When intentions were poorly formed, the mediating role
of intentions was reduced and attitudes had a direct effect on behaviour. De Jong
(2000) found in the case of the Thai sample a correlation between intentions to
move and actual migration. Intentions were a more significant predictor of permanent migration than of temporary migration. However, Abelson and Levi (1998)
believe that a favourable attitude is not sufficient for stimulating an intention and
that something more is needed to perform a motivational role.
143
The missing motivational link in the attitude-intention relationship seems to be related to
the subjective experience of desiring to perform an action. To desire to do something implies a motivational commitment to do it, if we assume that a person believes he or she can
do it (Abelson and Levi 1998:184).
MacCrimmon and Taylor (1976) repeat the same idea and claim that for the next
steps of action individuals must also be motivated to resolve the problem and must
perceive that there exist necessary abilities and resources to do so. In migration
studies, the term “barriers” is often used to mark conditions that could inhibit the
process of migration. In place of barriers, in behavioural model, we use the concept
of resources. Resources can be economic or emotional resources, as well as human
capital, social capital or information. We assume that resources and barriers are two
sides of the same phenomenon and we treat barriers as lack of resources to overcome
obstacles. Therefore, in order to start the action, people should not perceive the
barriers as being too strong or the resources as being too scarce.
3.3 Decision making
This subchapter describes, how people specify their location requirements, develop
aspirations, process information, select between alternatives, make a judgement, and
finally make a decision.
When choosing between many alternative coping strategies, people already need to
apply some rules of choice. In principle, the same mechanisms will be used at the
next stages of behavioural circle of migration (decision about information search,
decision about move or stay, decision about destination). As the general principles of
decision-making are relatively similar, we will discuss them together. The whole
procedure of making a decision is determined by goals, information search, particularities of information processing, judgement and choice principles.
We assume that migration is a chain of conscious and goal-oriented actions. When
deciding about entering the search process, people probably already set up some
aspirations that could be corrected and modified along the way as necessary (Hollis
and Nell 1975). Summarizing some earlier works, Brown and Boore (1970) thought
that regional aspirations are shaped by three different factors: the importance of
each criterion, the subjective probability of finding a vacancy of a specified grade
class for a given criterion dimension, and desirability of that criterion. We have
reason to believe that aspirations are related to the original reasons of migration, i.e.,
to internal or external factors. According to the needs theory, a person must simultaneously consider several of his/her needs and seek at least a minimum level of satisfaction and, consequently, take into account several different criteria. Ni Laoire
(2000) has obtained some empirical results to prove this. De Bruijn (1992) adds to
the theory that the level of satisfaction should pertain to some behavioural option for
which the actor will settle. Aspirations are also factors that determine the manner
of information search and processing. Abelson and Levi (1998:274) believe that
144
“decision makers generally treat these evaluative dimensions as non-commensurable” and there can be two possible rules of data processing in that situation:
a) by alternative — each alternative in the choice set is processed and evaluated
as a whole according to level of aspirations and then the choice is made based
on these evaluations;
b) by attribute — all alternatives are compared on a single attribute, then on
another, etc. Decision-making, therefore, requires ranking the attributes in
the order of their importance.
Feasibility of the aspiration often determines the duration and complexity of the
whole search process. The results of a migration survey in Finland 1977–1978
showed that the idea of migrating came to one-fifth of the migrants less than one
month before the actual move, the final decision to move was made by over one-half
of the migrants less than one month before moving (Nieminen 1983). Data from
Estonia (Table 2.19) demonstrated that the average time from the intention to a
migration decision was around 10 months and average active search period lasted 4
months. This reveals that the act of migration is sometimes performed very quickly
after the principal decision and sometimes it takes more time. Here several factors
could play a part: distance, barriers, freedom to migrate, housing market situation
etc. Whereas in short distances the active search might take less time than the decision-making itself (Hartshorn 1992), migration over longer distances requires more
time and information. The decrease of physical distances leads to a decrease of all
cost components: economic resources, cultural differences, social costs and finally
time to acquire information (see Amundsen 1985, Maier 1990, Nordvik 2001). It is
found that the efficiency of the used information channel depends on the distance.
Amount and quality of information can significantly alter the whole process of
decision-making, including the formation of attitudes. Consequently, the amount of
available information is an important factor in migration decision-making. For example, Boudon (1994) claims that situational limits on knowledge influence decisionmaking more than simply failures in information processing. Availability of information also determines the range of risk behaviour. It has been assumed (DaVanzo
1981, McAuley and Nutty 1985) that the perceived level of risk is an important
factor in migration decision. DaVanzo (1981) proposed that migration could altogether be related to the willingness to take risks; this explains why younger and more
adventurous people have a higher rate of migration. However, McAuley and Nutty
(1985) did not find a clear confirmation for this hypothesis. The explanation may lie
in different risk perception by different people, or differences of individual and household data. Structuring and restructuring of information takes place through the whole
period of problem-solving (Kikas 2002). There can be several limitations on information. Most analysed in migration decision-making are limitations caused by space and
space perception (Wolpert 1965, Woods 1982, Cadvallader 1989). Accuracy of the
information depends on the distance and information channels. The space for which
the information is possessed is called perceived space and by the nature of contacts
it can be divided into an active space, where information is gained through observation or daily visits, and an indirect contact space where information is gained from
other sources (Brown and Moore 1970). Personal experience tends to play a more
important role over shorter distances (Hartshorn 1992). Nieminen (1983) found
145
Number of evaluations necessary
to reach a choice decision
that advance information most accurately corresponded to reality when people
migrated on active space. Therefore, it could be assumed that people who move over
shorter distances possess more information and more experiences with respect to
this region. In addition, information could come from indirect sources. Constrictions of the indirect space have been seen as a factor limiting especially the migration
of groups with lower social status (Jones 1990), which is the reason, why relocating
over longer distances is more risky and difficult for this group.
Because of importance of information in migration process, it is quite expected
that people are more likely to select large destinations rather than smaller ones
(Stillwell 1991). Larger places have the advantage of public image and information
availability so that they can more easily become migration destinations. The cause of
this advantage lies in the nature of information processing. Additionally, close proximity to many other places might result in high relative accessibility (Louviere 1984).
There are different opinions concerning the question, whether judgement and
choice are two sequential processes or one single process (see Abelson and Levi
1998:235). Different combinations of processes are possible — judgement without
choice and choice without judgement. The authors argue that judgement is neither
necessary nor sufficient for choice.
Analysing different models of assessment and choice, Cadwallader (1989) summarizes that most models of choice presume that people are capable of considering
simultaneously all characteristic indicators of a certain region. An alternative would
be a model according to which people leave unconsidered most of the regional indicators and consider only a few characteristics. In this situation, the problem is
divided into stages of decision-making, thus avoiding the need to consider all indicators simultaneously (see Figure 3.2). According to hierarchic choice, alternatives
are not assessed all together, but rather by groups of alternatives and then by certain
objects or sets within the chosen group. Proponents of hierarchical structure view
spatial information as too complex to be coded at only one level (Howard and Kerst
1981, McNamara 1986, McNamara et al 1989). Boundaries of those clusters may be
physical, administrative, or subjective.
Upper limit of individuals processing capacity
Choice set subdivided
All alternatives evaluated
Choice set further subdivided
Only alternatives in selected
cluster evaluated
Figure 3.2 Hierarchical information processing (Source: Stillwell 1991)
146
In addition, many empirical studies show that the search process consists of at least
two parts. Predominantly, the choice is being made mentally, without actually being
acquainted with the future place of residence (Hartshorn 1992, Johnson and Salt
1992, Fortheringham and Curtis 1999). Lloyd and Dicken (1972) call it a special
filter. In the subsequent assessment process, relatively few options are given a thorough consideration. The duality of the search process has been most tested in
in-town migrants. For example, Cahill (1994), studying Denver house owners, concluded that the behaviour of educated and wealthy searchers corresponded most to
the model of dual search. Vyvere et al (1989) analysed different models of assessment and found that parameters of housing utility functions, derived from different
models, were not statistically different across experiments, nor were they different
from those derived from non-hierarchical models.
Abelson and Levi (1998) believe that individuals command a repertoire including
most of the choice rules and might in different circumstances use different rules.
The choice style may depend on the number of alternatives, amount of information
and time pressure:
1. In the case of 6–12 alternatives, people tend to reduce quickly the number of
alternatives using a non-compensatory screening process.
2. The more alternatives there are in the initial set, the smaller is the proportion of
information searched and the greater the tendency that choices are based on
simple rather than complex decision rules.
3. People tend to reduce their information search as the number of attributes
increases. Increasing the number of attributes has even greater effect on the proportion of information searched than the increase in the number of alternatives
4. Under time pressure people tend to dichotomize dimensions into accept and
reject regions; the quality of the decision is dependent on the amount of data
used simultaneously (Abelson and Levi 1998:261).
To recapitulate, one could argue that the choice process and the type and number of
cognitive processes are closely related to the cost of thinking. “More elaborate and
costly decisions rules will to be used only when the decision maker considers it
“worth it” — time constraints are lenient, decision problem is significant, outcomes
are irreversible, decision maker is personally accountable for decision outcome”
(Abelson and Levi 1998:266). Decision cost may be measured in terms of time,
energy, and mental stress (see Hollis and Nell 1975). Adopting routine behaviour
could significantly reduce the cost. Although in the case of relocation it is usually
known, what strategies are available for information, and this could shorten the process and improve the results, different households still find themselves in very different situations. For example, a family with many children has different minimum
requirements, simply because of the number of household members, and, therefore,
one could assume that their process of choice making lasts longer and they have less
options. According to a classic migration rule, single people move more than larger
households. The prospects look very different also for those people who already have
migrated once and for those who need to learn everything from scratch or have less
resources for the process.
There can be different criteria for choice between alternatives after evaluation.
Non-compensatory rule does not allow trade-offs between alternatives and, thus, it
is suitable when commensurability is absent. It requires only an ordinal ranking of
147
attractiveness of values. The conjunctive rule assumes that the decision maker
defines minimum cut-offs for each dimension. If the alternative does not exceed all
the cut-offs, it is rejected.
When more than one alternative exceeds the cut-offs on all dimensions, the conjunctive
rule will yield more than one acceptable alternative. A decision maker may then proceed
either by making the cut-offs more stringent and applying the conjunctive rule again or by
using a different choice rule that will yield a single alternative. This situation will not arise,
however, if the decision maker stops examine alternatives as soon as an acceptable alternative is found (Abelson and Levi 1998:260).
According to the lexicographic rule, the decision-maker compares the alternatives
on one attribute at a time rather than examine each alternative separately as a whole.
For example, the decision maker considers the attribute of price as first in importance, safety as second, and economy as third. Using the compensatory rule,
attributes can be traded off against one another.
One of the milestones of migration research is the interpretation of migration as
an assessment of the utility of different locations that can be personally measured
with final place utility (Wolpert 1966, Brown and Moore 1970). The list of variables used to assess the utility of places can vary from person to person, as can the
relative significance that is attached to each variable or variables. In oder to get the
final place utility, it is important to multiply the place utility matrix by individual
weights of importance of regional characteristics, one for each variable, which will
yield a weighted place utility matrix. The sums of the values give the sum score for
each place. The individual will then use the score to judge the utility of different
places and will choose the place with highest value of utility.
3.4 Individual differences
The approach developed here is based on the assumption that human behaviour is
influenced by personal needs, resources and perception particularities. In that
respect, this approach is different from explanations based on household needs
(Brown and Moore 1970). Although a household is an essential in-group for a person
and the closest social capital resource, individual motivations within one household
might be different. Accompanied family migration could be a good example. Household members, who accompany others, could have a different set of needs, which
motivates them to follow the initiator of movement or agree with a democratic
decision of the whole household. Those needs could be a need for economic or social
safety, love or belonging or other (see Table 2.1). Not all households move together,
because individual needs might contradict with the needs of other household members. Accordingly, migration behaviour and migration decision of household can and
should be monitored on individual level. Individual differences influence the process
in all stages of a migration process (Figure 3.1), but also individual characteristics
may change during the process: they are influenced by new information flows, by
learning, reshaped values. Migration process can be time-consuming and it is probable that all individual characteristics change in time irrespective of migration process.
148
Why are some people more mobile? One could assume that they are either more
sensitive towards individual or environmental changes, there are more changes
taking place in their lives, or they might be more inclined towards solving their problems through migration. It is also possible that they see more profit in migration than
others, or that it is easier for them to relocate because of their resources.
Important for the whole process of migration are also the perceived environment
and the differences in perceptions. The process of perception can receive impulses
from several sources: direct experience (actual visits), indirect experience (verbal
account from other people, newspapers) or selective information processing according to the needs, attitudes and values. Hence, already the perception can be differentiated by cultural group, socio-economic status, age, experience, education, personality, prejudice or temperament. Perception can be correct but can also be quite
vague and inaccurate. It depends on individual characteristics and the capacity for
human cognition. The influence of an external stimulus on individual imagination is
related to the subjective intensity of that stimulus (Cadwallader 1989). Consequently,
the influence of the stimulus depends both, on its objective intensity and on its
subjective perception. Additionally, human social stratification leads to selective
information gathering: people select more complex information from in-group members and less differentiated information from out-group members (Howard 1994).
Authors emphasize the situatedness, flexibility and social embeddedness of cognition. For example, Ellaway et al (2001) found that neighbourhoods are the main
location, where people make social comparisons that could affect their sense of wellbeing. From this, a reversed result could be derived. Namely, people, who for some
reason do not compare themselves to other places and other people, should be more
satisfied with their present situation and less open to the idea of migration. The
importance of comparisons and alternatives for dissatisfaction has been confirmed in
various experiments, which have shown that the degree of discontentment is higher
when people know about other alternatives (Sabini 1992).
Tolerance threshold, tolerance of stress and adjustment capacities are different
for different individuals and could depend on many personal characteristics like
needs, personal resources, perception or optimism towards place habitat (Wolpert
1966, Moon 1995). According to Abelson and Levi (1998), the cumulative amplitude of challenging stimuli must reach or exceed the decision maker’s action threshold in order to motivate people to act consciously. However, action threshold is not
fixed, but shifts continuously, depending on information loaded and the number and
type of decision problems already in process. The more burdened an individual is
with ongoing decisions, the higher his or her action threshold will be for a new decision. Mansfield (1992:517) argues also that, presumably, the “value stretch” is narrower for the higher socioeconomic groups than for the lower.
Attitudes have been seen as an important link in the process of activity formation.
Attitudes towards acts are produced by persons’ perception of the consequences of
these acts, and evaluation of the consequences. Evaluation is formed according to
the relation of the consequences to personal drives and values, i.e., values are goals
that people have (Sabini 1992:693). Consequently, one should ask, what are individuals’ attitudes towards changing their place of residence. Those who answer positively are more likely to be migrants. Bagozzi (1992) argues that attitudes and
subjective norms are not sufficient determinants of intentions and that intentions
149
are not a sufficient impetus for action. The attitude-intention link is hypothesized to
depend on connate processes and on certain coping responses directed at the emotional significance of evaluative appraisal. Still, the author considers attitude research
necessary and even one of the key questions. But attitudes are not constant and
could be continuously modified. Some important factors in the formation of attitudes could be the following: subjective norms — the role of “important others” and
social rules that serve as behavioural guides (Hollis and Nell 1975, Abelson and Levi
1998)8 . Subjective norms often emerge through the influences from a reference
group. “Someone’s strength as an influence agent depends on the status, expertise,
and power of this agent. The influence of reference groups has been found to be a
noticeable factor also in the cases when people feel deprived, because other people
in their social group apparently do better (Sabini 1992:413). Comparing oneself to
others and a sense of community are important factors that influence peoples’
mobility. An interesting phenomenon, found by Ellaway et al (2001), indicated that
community cohesion was strongest for the people who believed that their status was
same than that of their neighbours. Community cohesion was least felt by those who
thought they were better off. It is likely that these people had a higher opinion of
their possibilities in other locations and they were more receptive to the idea of
relocation. Based on the influence of subjective norms, we can pose the hypothesis
that relocation into a certain area can be influenced by a group affiliation and the
impulse to move might come from the social reference group.
Finally, different individuals have different resources for adjustment alternatives,
information processing and consequently for decision-making. The most important
resources in migration process are skills, information, human capital, social capital,
and economic capital. Lack of any of those resources will influence individual freedom of behaviour.
In summary, it could be argued that if earlier behavioural analysis of migration has
paid quite a lot of attention to place utility and this conception is relatively welldeveloped, the future behavioural analysis of migration could focus more on the role
of individual differences in making the migration decision. While some people do
not give serious early consideration to their act of moving at all (Morrison 1972),
others could be highly effective and systematic “easy movers”.
3.5 Interviews with people
3.5.1 Method
The goal of the interviews was to study the strategies of decision-making and search
processes of the people who changed their place of residence. Both inductive and
deductive approaches were used during interviews in order to capture all variety of
Subjective norm approach corresponds with the social impact theory. According to Bibb Latane
(1981) theory, the social influence in any setting is a multiplicative function of the strength, immediacy, and number of people exerting the influence.
8
150
models. First, the respondents were asked to talk freely about their relocation. At
that stage, researchers used only some keywords as guidance if required. At the end
of the interview, respondents were more thoroughly questioned as regards to the
reasons of their relocation in order to capture all possible factors influencing it. Main
emphasis in the study was laid on the following stages: emergence of the idea of
relocation, search process for the new place of residence, and satisfaction with the
new dwelling. Retrospective verbalisation method was used in process of survey.
Respondents were asked about cognitive processes that occurred before and during
migration. Problems resulting from this method have been rather well acknowledged.
They can by classified under four groups:
1. Limitations in reporting. Nisbett and Wilson (1977) summarised many studies
and concluded that people cannot report accurately on the effects of particular
stimuli on their responses. People’s conscious awareness is limited to the products of mental processes; sometimes the problems themselves are unconscious
and not directly retrievable from the memory. However, in decisions about
migration a greater proportion of conscious decisions can be expected.
2. Limitations of memory — individuals are often unable to report the contents
previously held in short-term memory.
3. Misreporting. For example, Sabini (1992:677) emphasises that there are several
signs that people tend to answer more positively and their behaviour differs from
their oral statements, when questions about their attitudes are too general.
4. Interpretation of protocols by the researcher.
24 interviews were conducted with help of three specially trained interviewers in
the winter and spring of 1995. Brief information on the goal of the study was given to
the interviewees as well. The interviewees were contacted through the information
from real estate companies, house visits and personal contacts. The aim was to find
interviewees with different household structure who had moved in different directions and over different distances. Appendix 1 presents an overview of the age,
gender and social background of the interviewees. In all, 17 women, 5 men and
2 couples in age scope from 19–73 were interviewed. All different types of migration
— international, urban-rural, rural-urban and in-town movement — were represented. Sometimes informants also told the stories of their previous moves. In
coding system and in text the previous moves are marked with small letter ‘a’.
People were asked to talk about their motives of relocation, the emergence of
ideas, the search process and the choices. Later some control questions were used to
elaborate the motives leading to relocation. One of the interviewers used
selfreflection that is recognizable within the collected material. The interviews were
recorded and immediately transcribed. Since earlier relocations were also discussed
during the interviews, the actual recordings contain information on more than twentyfour cases of migration. Three cases were such in which the change of the place of
residence had not yet taken place, but people were considering it. After transcription of materials, some ambiguous points were concretised with the interviewees
and notes were added to earlier transcriptions. In the later process of thematic
coding of data, the following coding classes were used: triggering reason, reasons,
search, stress, selection, time, and initiator. In part, the data was also approached
biographically.
151
In order to understand better the general background situation during the interviews, it is important to give a brief description of the general situation on the real
estate market in 1995 in Estonia. The year 1995 could be viewed as a transition
period that incorporated many opportunities still preserved from the Soviet time,
but at the same time a real estate market based on market economy was already
operational. For example, residences were allocated through workplace and through
housing administration. Rising rents were a new phenomenon and many people
experienced problems because of another new phenomenon — unemployment. Characteristic features of this period were the last stages of privatization of housing for so
called “yellow cards” — an instrument whose value depended on one’s labour time
record. The freeing up of Soviet military flats that were generally in poor condition
enlivened the situation on the apartment market. There was a general lack of apartments. Building of new housing space had almost completely ceased. For a awerage
family, obtaining an apartment was very expensive. Consequently, having an own
apartment was in itself a valuable asset for many households. Many families used the
new opportunities, if they could afford it financially, and acquired flats in poor
condition and renovated them so that they would be habitable. Newly emerged
settlement loans enlivened the apartment market.
3.5.2 Methodological results
The interviews justified themselves. The benefits of a longer interview compared to
a fully structured questionnaire emerged in many cases. Namely, often people
initially gave standard responses concerning the motivations and reasons of their
migration and only after a longer interview, the actual backgrounds became clear. In
some cases, these new responses were even contradictory to the original statements.
Some stimuli that led people to relocation were revealed somewhat randomly and
sometimes only at the end of the interviews. This demonstrated a phenomenon of
learned responses that is also a danger in standard migration surveys.
In addition, the form of an open interview showed that people were often not
conscious of the choice process as a whole. The actual background of relocation was
sometimes revealed only after targeted questioning. This concerns, for example, the
exclusion of certain areas from the search process. Often in initial interviews, only
the positive task was formulated, although in reality there was an immediate negative exclusion of some options. The discrepancy between non-targeted and structured interviews can be the result of people’s inability to direct their attention to
certain points or to represent behavioural processes consciously. Consequently,
interviews are a helpful source for studying the migration process, but they should
have structured or half-structured setup.
3.5.3 Appraisal, needs, triggering reasons and other reasons
The interviews indicated that mostly the change of residence was caused by changing needs. Even though the need for change usually emerged before the opportunity of relocation, there were some reversed cases where the change of perceived
152
opportunities caused relocation, although the desire to relocate had not yet been
perceived. Still, it could be claimed that in all cases there existed a need for change.
Consequently, it could be argued that in several cases the triggering factor was the
change of perceived options and not the change of living environment or needs.
The chance came before the idea. Of course, the conditions were poorer than here and it
would have been nice to have something better, but as we had received the flat from the town,
we had no right to expect anything better from them any too soon. So we were as yet of no
mind to apply for one. Now, well, there had certainly been some talk of a future need for
something bigger and better, but she would hardly have started applying for one as soon as
that. (12, received a surprise offer from the workplace).
They just came and asked if we agreed to exchange with them. For this is what many people
are doing now. And then I thought about it for half a year or even a bit more. I didn’t want to
do it at all, having always wanted to keep it for my children. Yet later, when privatisation
came along, everything had to be done quickly (15).
It was possible that perceived needs emerged together with opportunities and
resources.
The necessity arose together with the possibility as soon as I attended to the business when
all (money) was there (20).
In many cases, the life circle reasons of other family members, from children to the
dog and their needs, also shaped the decisions of the people. Under the influence of
the needs of important in-group members, these needs were attributed to the whole
household and were accepted as individual needs.
The children were growing up and they began to realise that this wasn’t quite the right place
(1).
Well, I consider myself something like an unmarried father, or parent, you see, I’ve got a dog
I’ve grown up with and for him I’m responsible. But for him I could live a bohemian life in any
town, I wouldn’t need any money, or flat, for that matter (22).
Children? – To a certain extent, yes. Either wanted a room of their own (19).
I mean, I would like to move all right, but the older son would not. I should wait until the boy
grows up, so that he could stay in Tartu and I could simply exchange the flat for two. It’s
just... Having lived all the time in the same place in Tartu that’s where he’s got all his friends
(9).
Several classical sources of migration literature assume that migration is a result of
stress. Therefore, the interviews specifically focused on the fact of presence or lack
of stress. However, the meaning of stress was not specified and, therefore, the interpretation of this term in the interviews may be subjective. Some respondents claimed
that they did not feel stress, others classified their plight as stressful.
Well, I didn’t feel any stress in the old place. I actually liked it very much there and I still
would (10).
153
Perhaps not stress, only some inconvenience maybe (11).
No, there wasn’t any stress. Things had not been let go far enough to become stressful (13).
Yes, there I used to feel not at home all the time as there was no protection of home, or any
hometown people around, well, with strangers, you know, in that age, after high school it isn’t
so easy any more to find friends and make acquaintances. True, I got some workmates, but I
still felt something like an emigrant (14).
Oh yes, I did feel it. But they are good friends to me, very good friends. The stress wasn’t that
bad. It was still possible to live there. Yes, I felt it, I don’t know what they thought, but a room
of my own is important to me as well (3).
W: Certainly. M: to some extent. W: Say, hosting was a problem, for whoever had a visitor,
should have considered whether any of the roommates happened to be preparing for an exam
or what their condition was. We had everything just there, crowded in a couple of square
meters (18).
Somewhat surprising to researchers, but otherwise expected, were the reports about
experience of stress caused by the relocation itself …
Now (present move) we did not get any stress. There was stress when we got the previous flat,
as we had lived in a house without conveniences and we struggled real hard to get the previous
flat, but now, moving from the 2-room flat to the 3-room one – no (12).
… or stress due to adjusting to the new location.
At that moment – sure, there was. Inside, you mean? Yes, I was terribly sorry to leave my own
flat (2).
With the previous place, yes, as that was a much older house and, I mean, at first it was real
shocking, before renovation. This was the kind of feeling, when everyone kept asking where
the hell have you moved to (9, felt the stress when she already was a “happy” owner).
Triggering factors appeared both, as single reasons caused mainly by the change of
life stage, or in combination with other factors related to the living conditions and
economic reasons. The following cases are some typical examples of stories about
triggering forces:
As father died and I couldn’t keep the house up any more, this made it simpler for me (4).
First time I moved because the flat was too small, the second time I moved because I didn’t
like the house and neighbourhood. When the elder son grows up and wants to leave or needs
a flat we will change for a smaller one. There’s no sense to stay here just for the two of us (9).
(The place had been an emergency variant enabling the couple to move away from mother.)
There arose a necessity to make our own home. It’s just that the demands are growing all the
time. First I was happy to have that one-room apartment, then the feeling began to grow that
it was too small, that it was really impossible to live, sleep and study all in the same room,
there was just a need to have a bigger one (18).
I moved from Keila to Haapsalu. Once I had left Haapsalu to study, then remained to live in
154
Keila, but I’d always yearned to be back home. My husband didn’t like the idea, but later he
agreed and as soon as a possibility opened up, I returned here. …. Any other reasons? Well,
our marital relations had turned somewhat..., we had become so distant to each other that I
came back home and there was nothing to hold me back there. I had been away for his sake,
but now I came home (14).
There was just a need for a flat. At that moment I would ... I and my husband would have
agreed to anything to be able to live separately (5a).
Often relocation was triggered by coincidence of several factors. Although the need
was present before, it became a migration reason only after coincidence or accumulation of several factors. Consequently, even though the change of the place of residence had been considered before, the actual change was caused by some triggering
event that gave the last push that exceeded the tolerance threshold. In such cases,
the decision was often made under joint influence of actual needs and opportunities,
and the active search period was very short or was altogether absent.
We moved because the living conditions and conveniences are much better here. We are four,
after all. Back there we had a small flat. Somehow it happened that suddenly we had enough
money, really suddenly, as a female relative of mine came to help and thanks to her and some
connections (advice) we found a cheap flat and so we came (1).
There was something the matter all the time. It was so disturbing that in a big house you
cannot send the child out to play in one’s own yard or garden. I felt the pressure ever since the
baby was born. There was the pram to be carried down from the fourth floor and it had to be
carried up and down all the time as from the corridor it would have been stolen in a minute.
This is what started pushing me at once (21).
I moved where I had intended to. Well, perhaps ‘intended’ is not the right word as actually it
came quite suddenly. I came to Tartu from Tallinn, some acquaintances offered me a wonderful possibility of acquiring a small flat in a rather decrepit wooden house. So, here we are
now. It happened about a month ago. Some acquaintances provided the information. Well, we
were having a party, so I don’t remember all the details, but the idea was that the flat where
an acquaintance had lived – or were there several? – had become vacant, and so it happened.
Perhaps I wouldn’t have done it otherwise, I’d rather kept on vegetating at my mother’s side
for a very long time. Probably. For I didn’t have such an enormous need to leave, it was just
that the opportunity cropped up and I took it and I’m still happy and proud of having coped
with it (22).
Sometimes the reason for migration was prevention of future problems: people
behaved in forward-looking manner and a projection of life events took place. The
problem did not presently exist, but the decision of migration was made based on
household or life events forecasts and perceived needs in the future.
The first ideas, or discussions with my husband emerged a couple of months before the baby
was born. There the room was so tiny, and so was the corridor, that when we imagined how
the child would start crawling or stand up, it would one moment just fall and kill itself (11).
If I had remained living there, the house would have tumbled down long ago. Now it has a
fine owner and the house is in good order again. I’m happy that my parents’ home is living on.
Myself with my knowledge and my income wouldn’t have been able to do anything for it. The
house would simply have kept falling apart bit by bit as before (4).
155
It would be a pity to give the flat up. When my son once has a family of his own, where will he
live? There’s hardly any housing construction going on anywhere (8).
In some cases, the influence of a reference group or other important people could
be felt as the triggering reason of relocation. This is a rather sparsely noticed area in
migration research. However, the triggering ideas from “important others” were not
a direct reason but rather a push force.
A colleague of mine changed residence and I thought why not. With my own house being still
far from finished, who knows when it will be ready, if at all (26).
All our acquaintances lived in private homes, so everyone had something on their hands, like.
Sitting in my two-room flat and having nothing to do, I grew sort of restless (21).
It was my nephew who mentioned that I’d better file my application in, as there were many
there already and so it might happen that the queue grows too long and I’d be left emptyhanded (10).
Migration is only one possible strategy for fulfilling the emerging needs. During the
interviews, a case of strategy transfer was also recorded. People changed from one
relocation strategy to another. This shows that the whole process is very flexible and
open to all kinds of influences.
I changed for a bigger flat. Actually we’re building a home, but we found that we couldn’t
finish it now and so we decided to change flats for the time being (6).
Triggering reasons are factors, which give the final push to the process, but there can
be several reasons for migration before and after this triggering event. During the
interviews, the interviewees were asked to count all reasons of relocation. There
could be only one reason ...
Only for more space (19).
For financial reasons. Otherwise I would never have chosen to change, as we lived on our own
(23).
… or several reasons:
Well, we changed, first, because my husband worked in the police. He had been working there
for 4 years and wanted to quit. He had, after all, quite a different speciality – a builder’s,
and he wanted to work as one. He also wanted to continue his studies in Tallinn, extramurally. As it is really far from Valga and it is closer from here – this was a good thing.
Second, there was no builder’s job available in Valga. Here there is a company who hired him
for quite decent wages. We could live on it and this was why we came, too. And as I study in
Tallinn, it is also better for me. In addition, my sister lived here. So she took care of finding
the flat. She also had small children. So I can ask her to sit in with my baby, back there wasn’t
anybody to ask. In a word, there were enough reasons to move here. The main reason was still
the one connected with the husband, that is, he got a more satisfactory job in Haapsalu; he
liked it more and got paid better and it was the work he had been trained for. The rest of the
reasons were less important. Back in Valga we could sell our flat for quite a good price as we
had a privatised flat. And here we could buy cheaper than usual. So we won financially.
Actually we wanted to get away from the parents, too. Our parents live all in Valga county,
not in town, but in the country, and they are, in general, rather pushy. We got along, but they
156
tended to force their view of life upon us, while we wanted to live independently, not doing as
they thought better all the time, that’s why. (Later:) We came here to live in better conditions
altogether. The flat in Valga used to be much smaller, and much colder, there were walkthrough rooms, the kitchen was smaller – well, the conditions were worse (13).
It’s hard to say, the first reason could be poor environment – for myself as well as the children,
more even for the children. Second, the wife’s job, of course, this is certainly a very important
aspect. Third, well, all those living conditions – a stove with heat depending on myself, fourth
the money and all the rest to it (17).
From the point of view of quantitative structured interviews, it was important to
note that many actual reasons were specified only in the course of a longer conversation. It might point to possible shortcomings in the quality of very short questionnaires.
– In our case the previous owner got his house back and we had to move out and so we got a
flat here (The owner is referred to as the reason) .
– But wasn’t there a possibility that the owner could have kept you on as tenants.
– Oh, but there was. He was even cross at our leaving, why, he wondered, but we still wanted
to go, because we felt we were getting older and the water was far to fetch from and there were
stoves to heat and the firewood to obtain – it had all become too much for us.
– Well, so it was not the only reason that the house was privatised, but also the living conditions?
– Yes, just the living conditions, otherwise we could have gone on living there. We are elderly
people and with the water to be fetched from very far and the serious difficulties in obtaining
firewood – that was the reason (10).
I don’t know. Well, the hostel life, I’d say. ...........At work my position improved. Marriage
occurred (20).
Even though most of the arguments related to migration were rational, there were
some references to emotional decisions …
I took such a fancy to Tartu that I thought whatever happens, I would never leave Tartu.
Finally I was sent to Jõgeva to practice and then I took a liking to Jõgeva, commuting every
day by train (11).
... and influence of emotional attachment. Emotional attachment is quite unfamiliar in the quantitative research.
Now, thinking back the place looks really old and poor, but still I was sorry to leave, because
it was a stone house, after all, a warm flat, and somehow it wasn’t easy to leave (19).
I wanted to go back home, all the time something kept inviting me back. Just to have another
look. My acquaintances had all stayed there. Later I got accustomed, of course, but it wasn’t
easy. I missed my mother and sister. Maybe I would have felt the same if I had come here. But
this time I didn’t leave anybody and there’s no one to miss (1a).
I’d got used to the neighbourhood, everything had been mine there right from childhood on
(1a).
Moving is, well, how to put it, sort of troublesome, but it was still connected with the joy of
getting back home and so I didn’t notice the difficulties. I was home, back in my home town
(14).
157
Consequently, all reasons that led to relocation were related to the change of personal needs or personal resources. There were almost no reasons resulting from the
change of environment. Owing to the nature of qualitative research, it can, however,
not be claimed that this possibility is completely excluded. The changes of environment are more rare and generally slower than transformation of individual needs and
resources, which is the reason why it might have less influence on migration.
It appeared from the interviews that people had very different skills to assess and
analyse the environmental conditions. The number of factors considered in the process of relocation also varied significantly. We propose as hypothesis that amount of
aspects taken into account depends on individual’s ties to the environment (e.g.,
more factors need to be considered when there are children in the family), as well as
on education, age, and the number of different options.
3.5.4 Search for a new place
The following section presents the interview results with respect to the migration
decision process after the decision to change the residence had been made. Primarily, it is related to the process of search for a new residence.
It appeared during the interviews that people were prompted to think about change
or to abandon it by a preliminary assessment of their potential without actually
examining the particular situation. I.e., they did not waste energy on additional search
that, in their perception, did not pay off but, instead, they applied the mechanisms
of perceived behavioural control.
There’s no sense for me in moving to the country, for who would take on an old person like
myself. If I was still 30, with kolkhoses and sovkhoses still around there would be some jobs
available, but now even the local people are unemployed (15).
It was worse than this and of course it would have been nice to have something better, but as
we had got it from the town, we had no right to expect to be given anything better any too
soon. So we were of no mind to apply for one as yet (12).
This is the only variant, but it’s very expensive. Actually too expensive, but there wasn’t
anything else available (desisted from looking for other variants at once). The place is really
neat, ver y well renovated. And I knew I’d be asked a higher price than Estonians. I think I’ll
bring the spaghettis and ketchup now (3).
The whole process of relocation does not necessarily have to be intense, but can be
comprised of different stages covering long time periods: (1) emergence of the idea,
(2) search or passive waiting, (3) emergence of opportunities, a new activation and
evaluation of options. In the meantime, people might cope with the situation without migration. At other times, however, the change can take place very rapidly.
For three years we followed newspaper ads. Ever since the baby was born (active search
began). There was the pram to be carried down from the fourth floor and it had to be carried
up and down all the time as from the corridor it would have been stolen in a minute. This is
what started pushing me at once. This makes four years in all (21).
158
– Did it take you long to look for one?
– Oh, yes. Death seemed closer than finding anything. About three years ago. W: When we
first met. M: started making plans together (7).
This was about last turn of the year. At first we set about it quite actively, but then we failed
on the first try, being refused at the last moment. After that our enthusiasm subsided for a
while. There was a period when we practically weren’t looking. Yet later, when we weren’t
even planning to change any more, some variants were suggested to us, ‘if we were still interested’. Of course we were, as this house is newer than the previous place (11).
The need was first. I kept looking for quite a time. For a long time I had to search newspapers.
I don’t remember exactly, but I may have advertised myself as well (17).
It seemed that in many cases the process of search for a new dwelling was driven by
a couple of aspirations and restricted by some limiting factor (often economic
resources). These main non-compensatory criteria formed the basis of the search
process, and the possibilities within these criteria were then given a more thorough
consideration. Possibilities that did not fall within the range of determined criteria
were immediately disregarded. For example, a criterion could be the prospect of
getting a job.
I lost my job in Tartu. This was the reason, for with my speciality one can find work only
either in Tallinn or in Tartu. In Tallinn I can hope to find a job, but to get it I must first change
residence (Busy changing residence from Tartu to Tallinn) (5).
Town as such was important just because of my position I couldn’t find employment in the
country (11).
Often the maximum wishes of people would be greater, but from the beginning,
they set individually perceived realistic goals. Sometimes, when the possibilities
were not exactly known, a market research was carried out. Such market research
carried the same meaning as a search for additional information.
– Do you use newspaper advertising or some other way?
– At first, yes, I needed to get the overall picture of just what the offers might be, if any (5).
Almost in all cases, people considered their economic possibilities. Economic
resources were one of the major limitations on the freedom of choice.
– No money. There was the need for renovation, the costs would have been very high. Too
high. One has to look a little further and decide within one’s means.
– If you could have chosen your place of residence without any financial considerations,
what kind of a place would it have been? Any different from what you have now?
– Well, the same quarter would have been ideal, so home-like and quiet. But the same conditions as now? No (7).
This is just a cheap place, that’s why. If we could have afforded normal prices as they are in
town, we would have bought a different flat, but at the moment we agreed to this one because
of the (low) price (1).
One thing was evidently improvement of the living conditions as one can hardly live in a
hostel all one’s life. That’s for sure. Having no resources to build a home, what is there but
buying a flat. This is what I can afford (20).
159
(Would have preferred something better), but then I’d have had to pay a higher price (4).
Often, based on previous knowledge, certain preliminary search regions were
defined, i.e., geographic preferences were determined in order to optimise the search
process. According to the hierarchic choice model, the first decisions made to
decrease the amount of information were selection of price and geographical scope.
(The important considerations are:) where it is situated and in what state the place is and
the rent, too (laughing). This is what counts most. The centre, yes, I wouldn’t like living
anyplace else (3a).
Pärnu seemed likeable, neither did we have anything against Tartu as those towns seemed to
be relatively quiet. Tallinn, for example, was very tiring for my wife as she had to spend
fortnights there taking exams and by the end of the session she would be totally exhausted of
all that commotion. (At the same time) she considers Tartu and Pärnu as youthful towns
where life is still busy and developing, not stagnating (13).
Tallinn would tire one out in half an hour. First, there are so many Russians, and even more
Finns. Estonians are very few, there’s no one to talk to, and nobody knows anybody. One feels
so lonely, if not exactly walking hand in hand with someone. But in Tartu it’s cool. Tartu is a
sufficiently small place, quiet and green, too. As for the district of Annelinna it feels sort of
safer. Somehow one gets a feeling that there are people looking out of those windows all the
time and therefore it is safe to walk under the windows. In other neighborhoods one feels
apprehensive of getting lost somehow (11).
There were also other criteria for making the preliminary choice in addition to economic and geographical factors. The choice seemed to depend on different resources
and their assessment and availability of options. Sometimes the decision was made
based on minimum criteria and some desired non-compensatory criteria were
established.
It seemed that the range of options was much greater if the previous habitats of
the individual had been in a poor condition, i.e., the individual had more tolerance
towards potential dwellings. This seems to prove the hypothesis about upward housing careers as a general route.
I advertised in Radio Tartu. As well as at an agency and in papers, everywhere. I mean, it
was the radio announcement that brought me the flat.... I advertised and then I went to see all
kinds of flats, but this one suited me best. Of course, it was the cheapest of all offers, too. The
design somehow reminded me of my grandmother’s place, so it felt homey to me. – Did you
have many variants to choose from? – Yes, I had. As for the price, Elva would also have suited
me, but as I’ve always lived in Tartu, I chose Tartu. – In short, first you considered the price,
then picked two to choose from? Well, not exactly. I mean, the price was primary. I can’t jump
over my shadow, can I? I would have preferred a 4-room condominium, but I couldn’t just
afford one (4).
Yes, we, for example, are patriots of Annelinna of principle. We didn’t want to cross the river
(Annelinna is not regarded as a too prestigious neighborhood). So we selected from offers to
Annelinna only. We also excluded nine-stored houses. Up to the fifth floor was acceptable, but
not the first or second floors. So it was to be Annelinn, a five-stored house (11).
– Of course it would be good to have a grocery near, but this is not the first thing I pay
attention to. – But what is it that you do note when considering an offer? – The documents
160
should be in order. And it should better be not too dilapidated. – But the neighbourhood, is it
less important? – Less indeed, at least in the beginning, in the face of moving over. – And the
floor, first or second etc.? – This is quite irrelevant (5).
Several respondents claimed that they left many factors unconsidered. Although it
was not specified, why they did not consider them, it was an overall impression that
the search process was based primarily on a few decisive criteria (their number
varied), the presence of some criteria was considered useful, and some criteria were
not important at all. It is also possible that only some criteria were excluded and all
other remained in the search process.
I don’t think there were any preferences of principle. We didn’t want to get worse living
conditions, or a smaller space, neither was the (interviewee’s) husband ready to move into a
flat with stove heating (she was). Neither would they have agreed to move anywhere too far
from the bus station, or to any distant outskirts (13).
(Some quarters were eliminated) Those places we didn’t even consider. Just skipped them in
the ads. Picked only Mustamäe and places like that (7).
The nature of migration barriers can be psychological or of some other character.
Although a qualitative survey is not the best means to collect reports on individual
differences, it might be important to observe the general attitudes towards migration. Many people prefer to be settled and remain at the same location if possible.
For them the psychological or other costs of relocation might be higher. Barriers
could also be viewed as a lack of certain resources or a result of high emotional
attachment.
But there are such big problems with changing residence, it’s not just that you change and
that’s it. The moving and all, it is so difficult and then... (10).
I am truly so attached to my neighbourhood that I don’t even imagine moving any where else
within Estonia. There should be a real good reason for me to move (12).
Generally, the hypothesis of living conditions career found some support in that
with each move people tried to develop their living conditions career towards better
conditions, i.e., they did not move back to poorer conditions. However, this was not
always a one-way movement. Moving into poorer conditions could take place when
there were no more resources left to retain the current standard of living. Consequently, people acted flexibly combining their perceived resources, options and needs.
I never thought that I would have to look for something. It just happened that life turned
harder and I was forced to (15).
How long does the search last? It often lasted only until finding the first satisfactory
option that would meet certain minimum requirements. It can be explained in terms
of rationality, that people tried to optimise their resources, which were spent on the
whole process. Thus, often not the maximum, but only a satisfactory option was
accepted and looked for.
161
No, I advertised in a newspaper and got an answer right next week. And it’s with those ver y
people that I changed flats with (14).
There weren’t any other variants, like that one, I mean. Well, once there was another in
Tähtvere district, but nothing came of it (17).
I was interested in the centre and this was the only offer. I just asked if anybody knew of a
suitable place and that was the first possible variant. So I found it by the help of the same
friends. My landlady knew of the place. There aren’t many flats available here in Tartu. Well,
maybe Estonians have more information, but I don’t have the money to pay to a bureau to
have them find a flat for me (3).
Some people spent more time on search, trying to make the best possible choice,
although they sometimes returned to some lower level options (7).
Some used more active, others more passive search styles. Strategies can be
extremely diverse, including many different techniques.
We also discussed the possible whereabouts and talked to relatives and acquaintances. We
started looking for variants via the relatives, who then asked of their acquaintances if anybody knew of something. An ad was also submitted to the local exchange bureau, but the
main activity concerned relatives and acquaintances. The search didn’t last long as the donors were Russian ex-militaries who had to get rid of the flat quickly and so had advertised
in the street. I happened to be in Haapsalu, staying at my relatives to get a better view of the
situation, and I noticed the slip of paper stuck to the lamp-post (13).
We haven’t advertised in a newspaper, but I’ve been watching out for ads and talking with
people, especially this year... Early this year I looked for something and then a little period
came when I tried and tried (2).
– Did you try newspaper advertising or was it somewhere else? – This was in the beginning.
I just wanted to see what is available, if anything. – But you wouldn’t question your
acquaintances, would you? – No, I haven’t tried that way. I don’t think we would really.
Well, of course one should ring up everyone bringing up the problem. I just haven’t thought of
doing so. Next week my ad will be in a Tallinn paper, too (5).
– How did you find information on those availabilities? – That wasn’t difficult. It was even
exciting, I’d say, having a hope alive all the time. – Yes, but where did you find information
on vacant spaces? – The press, relatives, acquaintances. There’s been quite a lot of walking
around.
– Didn’t you ever feel like quitting, just because you couldn’t take any more? Didn’t you give
up your plans?– No, no, I never give up my plans (7 looked for place for several years).
We saw that there can be different search strategies. At the same time, it was
revealed that search strategies might depend on resources. Most often mentioned
in the interviews were the economic resources.
In Finland I’ve changed residence four times. It depends how exactly you do it: if you’re hard
up, you ask your friends if they know of something, or you advertise in a paper. If you have
money, you turn to a real estate agency and pay a month’s rent in advance and they will find
a flat for you. I’ve tried all of the ways (3a).
Whatever you do you need money, because without money no-one will run for you or care.
162
You’ve got to pay for everything. Well, no, I didn’t advertise in a newspaper. It’s like a oneday advertisement and this would cost something, wouldn’t it (8).
The resources could influence not only the choice strategy but also the whole process. We pose the hypothesis that large resources facilitate a more flexible choice. In
many cases, wealthier people were more relaxed and had more choices; they also
seemed to consider more factors during the search process. However, the qualitative
survey is not sufficient to prove this hypothesis.
I wouldn’t have the flat, because there’s no hot water, while old people want hot water and a
bath and so on. The director got even angry mentioning that we didn’t have any bath or hot
water in the old place either, but I said we were getting older and had our needs to consider.
What life would we live here without a bath? So I returned the key. The nephew came to see
us and I told him about it saying I didn’t dare return for the key, so he brought the key back.
This house has no plumbing for hot water at all, everybody has their own gas appliances and
stoves. My nephew got us an electric boiler, so now we do have hot water. There wasn’t any
gas range here either, so my nephew brought us an electric cooker, too. The doors and windows
had been removed and carried away and many things were missing here and there (the
Russians had left the house in such a condition), but now I’m so happy to be able to spend my
old days here (10).
We nearly gave up the exchange. First we sold our car to buy the flat. The flats here in
Haapsalu turned out to be more expensive than we had first thought. So we sold the car and
bought a worse one to get more money (13).
Even though usually the change of residence was carefully planned and the whole
process could take years, there also figured a type of behaviour similar to fortune
seeking.
– Did you have a clear vision of what you’d like to have?– No, not in detail. We just wanted
something in a house with conveniences, but no details. We didn’t even associate our actions
with any concrete consequences. Well, it was a blind stroke of luck. As I said, I got the ad in
the paper on the third, got an answer on Tuesday, and by January 28 the exchange had taken
place, money paid and papers in order. It was absolutely accidental. There was that newspaper ad. And the letter that followed was perfectly detailed – the number of rooms, which floor,
how to contact them – everything was perfect at once. The others were more vague, like ‘come
and see, let’s discuss it’. Some didn’t even give the address, it was all so hazy. Well, I’d say
there had been a subconscious need, though, we had even done a bit of renovation (18).
... Maybe just because I wanted to get away from home. It was a moment’s impulse with
several coincidences. First, my acquaintances offered me that house in Tartu, second, the
situation at home wasn’t too calm, being rather, say, tense. It came as a moment’s impulse,
my leaving Tallinn (22).
Analyses of the search process gave some hints that the behaviour of people could
depend on individual characteristics. It could depend on resources (time, money,
information, assessment of one’s own potentials) and needs. The number of factors
considered could also depend on the inevitability of the choice and the strength of
motivation to find a new residence.
163
3.5.5 Evaluation and choice
According to the needs theory, the person must simultaneously consider several of
his/her needs and seek at least a minimum level of satisfaction. The theories distinguished four different types of value evaluation strategies: the non-compensatory
rules, the conjunctive rule, the lexicographic rule, and the compensatory rule. As
different needs are not equally important to individuals, we can assume that during
the evaluation process every factor gets its weight. In summary, some factors are
more important for a person than others. Compensatory values could be exchanged
in the process of choosing a dwelling for other, more important values. The person
will be content if the subjective final benefit exceeds the subjective costs of all
changes. Consequently, often some circumstances at the place of destination may
cause as much discontent as at the origin, but the decisive role is played by the
triggering reasons, which probably are the most highly evaluated and, therefore,
still give a positive value to migration.
– Well, as the rent was lower there of course it would have been more affordable, but I don’t
care. Now that I’ve got a real good flat I don’t regret. – But losing your gardening plot, didn’t
that give you a pause?– Of course it did. It’s true that if you’ve always had a garden to
cultivate and now you’ve got to live without one – this is really bad. But one will overcome
anything (10).
I used to like the location there very much. Everything was close at hand, the centre and all,
telegraph, railway station, polyclinic. Here we are farther from town and you can’t simply
walk there. The park was also near, we used to walk there and almost lived in the park. But
in general one lives in a flat, after all, not in a neighbourhood. It’s inside you spend most of
the time and who wouldn’t like to live better and more nicely. Back there were, for example,
such general inconveniences as a common loo, with nothing pleasant about it (1).
I do regret, every night and every morning. When I was still working I used to oversleep every
morning (smiling slightly) as there wasn’t anybody to wake me up, the dog was a good sleeper
too. When I came home I’d got to prepare my own food, wash my own dishes and take the dog
out. While I was just back from work, tired, feeling like putting my feet up at once and
turning on some music (22).
The number of criteria used for making the choice about a new place may vary and
sometimes a coincidence of many favourable factors can prove to be decisive. It
seemed that different factors were summed up in decision-making process.
The other variants would have done, too. Only there weren’t any close relatives, only
acquaintances and friends (13).
Sometimes the final choices were strongly influenced by habitat indicators, economic calculations (4, 18) as well as by the values and needs of the individuals (18)
or by their emotional attachment (7, 9).
It was the cheapest of all offers. My grandmother had also had something with the same
design, so it felt homey to me (4).
164
I was born and used to live in a garden suburb, a very nice neighbourhood that has become
very expensive now. There are woods, there’s calm. I even remember my preparing for school
exams – all in the woods..., Well, it may sound funny, but the trees steadied my nerves and
everything became clear. Of course, there are trees in the new place, too, woods, peace and
quiet. It’s beautiful there (7).
No, but I simply wanted to Annelinna (a district with not a very good reputation), as I’ve got
used to living there. I just like it better on this side (of the river). Well, I would have been
ready to move to Veeriku, too, because I’ve once lived there, and the conditions were acceptable (9).
In the present case, if you go to university, you don’t have much time. You can concentrate
better on your studies if you needn’t keep worrying for all those buses from Aardla and back
(18).
Among other factors of finally choosing or rejecting a residence were, for example,
location in downtown too close to the street (26), at the first floor (22), in different
districts, whereas the choices of different people were quite divergent. A location
that was not suitable for one was the most agreeable for another. Other factors,
which were mentioned, were security (21), lack of relatives, friends (21), motherin-law (23), and cheaper living (22).
Although mostly people possessed good information on migration destination
they still made errors in a situation of limited time, or if they for some reason were
not careful enough in their choices. In such cases, people often moved again. However, generally, the change of the place of residence was a process during which
people acquaint themselves with the situation before making the final decision and,
thus, reducing the risk to a relatively low level.
When I moved, it seemed a little farther and not a good neighbourhood, but now it’s turned
out even a better location – with the bus stops next to the house. But when the opportunity
came to change, one just had to take it and such things were overruled (12).
(Tried several variants which all fell off.) Then I took the one that was left over. They kept on
ringing and insisting, too, that we change. I just went and looked in, didn’t even care to see
the bathroom. Back home I began looking for new possibilities at once. I also started renovating my own place. For it wouldn’t have been conceivable to change otherwise. Well, after that
I began active changing again. In November I advertised and got a new offer and I moved on
March 26 (9).
Finally, the respondents were asked to evaluate their satisfaction after the move. It
seems that when making a decision on the goodness of the new situation, people
compare the situations in relation to the main criterion of the change. Consequently,
satisfaction seemed to depend on the motives of the move and the role of the individual in the whole process, i.e., if he or she was a co-migrant or an initiator.
165
Conclusions: The behavioural theory
The theoretical migration model constructed in this chapter was based on the
assumption about interaction between individual needs, resources and perceived
environment. It was concluded that although the needs of the household can be
accepted by individuals, the analysis of migration should be based on individuals.
The important steps in order to analyse migration behaviour are perceived need for
changes, adjustment strategies, decision about the intention of migration, search,
evaluation and decision among possible options. All those stages are influenced by
individual needs and resources. Individual characteristics in turn are moulded by
adjustment and behaviour process.
Migration
Migration
Search
Search
Search
Intension to move
Intension to move
Intension to move
Perceived needs,
opportunities, dissatisfaction
Perceived needs,
opportunities, dissatisfaction
Perceived needs,
opportunities, dissatisfaction
All population
All population
All population
Migration
a
b
c
Figure 3.2 Share of people in different mobility stages: virtually mobile (c), low
mobility with migration barriers (a) and limited mobility (b) societies
Treating migration as a process would be a better way to explain the presence in the
society of people at different stages of migration (Figure 3.2). From the total population only a part is dissatisfied with the given environment, i.e., perceive the need
and opportunity for change. Relocation, in turn, is only one possible option for coping with dissatisfaction. Not everybody, who has in principle made the decision that
it would be good to relocate, will actually engage in an active search process and even
fewer will make the actual decision to migrate. The whole process of migration at its
different stages is limited by many individual and environment-related barriers to
the freedom to move. Consequently, there are many different imaginary societies
and in some of them people are extremely mobile (Figure 3.1 c) and in others they
are extremely settled (Figure 3.1 a). The question, whether people are settled or
mobile, is determined by the combination of the perceived needs, opportunities and
freedom to move.
The results of the interviews with people, who had changed the place of residence or were considering it gave an independent confirmation to several theoretically assumed connections about behavioural strategies. The behaviour of the people
showed several rational elements: limiting costs on time and resources, abandoning
the process of relocation when no potential result could be seen, selection from
among a limited group of options. Although mostly the arguments for the change of
the place of residence were rational, there were some references to emotion-based
166
decisions. However, it could be difficult to determine the type of some emotional
components: for example, emotional attachment, because emotions might produce
rational results for individuals. In sum, people acted flexibly combining their
resources and perceived options. The amount and quality of information seemed to
be very crucial at all the stages of the process.
Based on the interviews, two relocation strategies could be pointed out:
1. Pull or positive selection type of migration. There was a perceived need, and the
idea of migration emerged simultaneously with perceived options. Even though
the need for change usually emerged before the opportunity of relocation, there
were some reversed cases where the opportunity emerged before the active
perception of needs. However, the interviews showed that relocation did not
take place without reasons.
2. Push type migration. Relocation took place mainly because of poor conditions in
the current place of residence. The whole search process could often be quite
prolonged consisting of active and passive search periods. The choice of the new
place of residence was based on certain criteria that included possibilities, which
were assessed from the viewpoint of actual commensurability. The search process
was based on certain criteria of which some were non-commensurable. Other
factors might turn out to be commensurable. The amount of commensurable and
non-commensurable factors could vary.
Interviews revealed large diversity of options for behaviour. Triggering factors
appeared alone, as well as in combination with many other reasons. Similarly, there
could be one or many reasons of relocation. Sometimes a reason of relocation was
the wish to prevent future problems. It was a general tendency that people considered their future besides the present when making their life plans. Although large
amount of migration literature claims that stress is a sign of the beginning of migration behaviour, some respondents claimed that despite discomfort they did not feel
any stress. However, some informants stated that the presence of stress was the
reason of migration.
Changing the place of residence was only one possible option among several coping strategies. The interviewees also mentioned strategies in which the intention of
relocation was replaced by some other strategy. The search for a new dwelling was
often limited by a criterion that was given a more thorough consideration. In most
cases, this limiting criterion was price or region. Sometimes, however, a large range
of criteria was used in search and evaluation process. Strategies of searching for the
new place of residence seemed to depend on individual resources (time, money,
information, actual needs, assessment of one’s own potentials). In assessing different
possibilities both, non-compensatory rules and conjunctive rule were used. Different criteria were also used in the decision-making process: from maximum option to
the first suitable variant. Although the interviews confirmed several expected interactions drawn from the theoretical model, the large variety of combinations can
form many behavioural patterns. Different people apparently use different strategies.
Behavioural individual approach seems promising, because it could well meet all
requirements for an ideal migration model proposed by Zelinsky (1983). We argue
that it can be used for different flows of migration: in all societies, cultures and
167
historical periods, for different forms of spatial mobility, individual approach will
account for extreme, aberrant behaviour as well as for “normal” events. Consequently,
behavioural model of migration could be a theory that is able to meet the requirements set for an ideal theory of migration. However, the model still requires a more
thorough empirical verification. There still remain many intriguing ideas as well:
1. The presented model followed the idea of sequential ordering of certain events.
The process developed step-by-step. Would it be possible that there exist other
combinations of decision-making order, or that the whole process is intertwined
like the ‘primordial soup’? Several psychological theories would indicate such
possibility.
2. The current model gives a quite good explanation of the push forces and of the
migration resulting from the dissatisfaction of the person’s perceived needs. Would
all movements caused by pull forces fit under the reason of changing perception
of the environment?
One of the further challenges could be to test empirically the model as a whole. For
this task, the quantitative approaches should be applied.
168
Appendix 3.1 Description of respondents
1. Woman of 32. Ethnic Russian. Changed within town using an agency. Economist. University
education. Income 4000 EEK. Third change of residence in her life. Family of 4, with 2 little boys.
2. Woman of 29. Used a real estate agency. Believes to be considered a materialist. Human relations
are important. Secondary education, university unfinished (broke off in the fourth year). Chief
accountant. Income 5000–6000 EEK for the two of them.
3. Woman of 22. Finnish. After one year at another university is studying Law at Tartu University.
Friends are important. Income 2000 FIM, rent 2000 EEK. Communal rates and travels to and from
Finland subtracted will leave her with 600 EEK for living. Has experienced several changes of
residence abroad.
4. Woman of 25. Unmarried. After her father’s death moved away from a prestigious neighborhood of private homes into a block of flats in the same town. Shopgirl. 900 kroons. Earlier used to
live with her father only.
5. Woman of 43. Divorced, architect. Moving from Tartu to Tallinn for work. Unemployed at the
moment. University education. Has an earlier experience of changing residence together with her
former husband.
6. Woman of 51. Moving together with her husband whose children have grown up. Moving to a
concrete block of flats. 4 previous changes. Education is important. University education. Head of
department. Income 2000 EEK per head. In addition, the couple has a private house.
7. Man of 35, woman of 23. Both work as state officials. Planning to get married. Earlier either used
to live with their parents in Tallinn. Now wishing to move into a block of flats in a small settlement
50 km off Tallinn. Mutual understanding is important. Income 2500 EEK.
8. Elderly Russian woman in the quarter of blocks of flats (Annelinn). A 2-room flat shared with
her son, his girl and “the old man”. Her sister lives near lake Peipsi and provides them with
eatables. Financial difficulties. Sold their former flat 5 years ago. Has been trying to exchange the
flat, but failed. Considering to change the flat, because it is too expensive for them.
9. Woman of 29. Special secondary education. Family of 3–4. Earns 1480 EEK net. Husband is a
dependant, but he is not with the family all the time. Having often moved between hostels, moving
does not bother her at all. Two moves within the current year. Quiet and money are important.
10. Elderly woman with husband. Changed within town, because the flat they did not like anyway
was returned to the earlier owner.
11. Woman of 27. Married. On maternity leave with a baby. Moved recently within the same town,
having moved earlier, too. Lives in a block of flats. University education, prosecutor, salary 9000
EEK. Thinks highly of material security, friends and life without conflict.
12. Man of 50. Married, with 2 sons. University education. Schoolteacher, salary 5000 EEK. Changed
from a 2-room flat into a 4-room one. Got the flat via the working place. Has changed before. Born
in the country. All value education, work, home and family. Especially now that unemployment has
turned a reality in Estonia, work has become ever more important.
13. Woman of 21. Moved with husband and child from one end of Estonia (Valga) to another
(Haapsalu). Trained as a seamstress, but the husband’s profession also had an important role in
169
changing residence. They wanted to get away from the parents. Values family and normal financial
level. Money was a means to get on in life. Both want to get on in life, to make a career, to live well.
14. Woman of 43. Seamstress. Moved (with children) from Keila to Haapsalu as this was where she
came from. She had yearned back there all the time, but the husband had been keeping her back.
Freedom arrived with divorce. Moved to a stove-heated wooden house. Three rooms, very small
ones, divided between two or three families, near the sea. Several earlier changes of residence
around Tallinn: a hostel near Lake Harku, then a flat at Keila. The first period of hostel life was
connected with her leaving Haapsalu to study in Tallinn.
15. Woman of 48. Ethnic Russian, financial problems. Living in a 3-room flat in a block with two
grown-up sons and a grandmother. Married, but the husband, who was a military, is now missing
(lost in Russia). Works in hospital as an orderly, earlier a stamper and turner in a factory. The
greatest wish — no war. The flat had once been given to them by the Soviet army.
16. Woman of 73. Income 1015 EEK. 2-room flat in a block. Changed within the same district in
order to divide living space between the family members more rationally.
17. Man of 37. Married, with 3 children. Many occupations, works in car repairs, income 6000–
7000 EEK. A short try at university education. Two previous changes of residence. Comes from the
country. Home is important. Moving from a 4-room flat in a block into a wood-heated house in the
centre.
18. Cohabiting couple. Woman a registered pharmacist, man a university student. As a child, the
man has lived in Tallinn, the woman in Viljandi. Changed residence within the same town to
improve the living conditions. Important are work, studies and no everyday trouble with the flat.
19. Woman of 42. Married, with two children. Secretary and typist. Income 2000 EEK a month.
Values home. Exchanged their 2-room flat for a 4-room one within the same district.
20. Man of 27. Recently married, lives with wife. Assistant director. Income 5000-6000 EEK. Left
university without graduating. Studied in a private school. No previous changes of residence. Changed
within the town. Before buying the present flat he lived in a hostel. Money and friends are important in his life.
21. Woman. Used to live in a small town (Tõrva) and wished to move to the country. Now living 25
km off Tartu, with husband and baby.
22. Man of 19. Bohemian. On a moment’s impulse and in order to become independent moved
away from mother’s in Tallinn to Tartu Supilinn, which is a decaying quarter with a bohemian aura.
Dog is an important member of the household.
23. Woman of 26. Divorced, lives with her son at the mother’s of her new husband. Coach and
university student. Husband is a businessman. Her monthly wages are 1500 EEK, husband’s
income unknown. Four previous changes of residence. Appreciates privacy. Her ideal is a home
with a small garden, but although she has something like that she is not happy.
24. Married couple — husband 32, wife 24. Wife works as a bookkeeper, secondary education;
husband works at a commercial company, special secondary education. Pooled income 5000 EEK.
Wife says that material values are irrelevant, whereas husband argues that home and things are
necessary. Moved to live apart from her mother. For the wife it was important to have everything
her own.
170
171
Regional poverty and freedom to move
4
172
173
4 Regional poverty and freedom to move
Often poverty and social mobility are studied as indicators of equality or inequality
in a society. Territorial mobility can be seen as a part of local justice. We assume that
the inequality of outcome is more legitimate if people have more equal opportunities
to make choices and less legitimate in a situation of missing choices. This principle is
valid for both social and territorial mobility. This chapter investigates links between
regional and individual poverty and freedom to move in Estonia at the end of the
1990s. Freedom to move is measured with standardised outmigration rate by local
municipality units. Drawing parallels from political science, we argue that
outmigration is a sign of freedom or “voting with feet” (see Fidlay and Rogerson
1993). The concept of poverty is used as a basis to classify the lack of inequality and
resources. Our particular interest is in barriers to movement caused by regional
poverty.
Introduction — rationality for equality
Ideas of social justice have a geographical dimension, however, there are different
understandings as to how geographical justice or justice on space is defined and measured. It is particularly important because space itself seems to have an influence on
humans’ abilities and equality. Perhaps it is also the reason why the studies of
regional justice have gained popularity (see Merrifield and Swyngedouw 1996,
Holloway 1998). At the same time, spatial equity might be controversial concept for
spatial development and “it does not necessarily lead to more social equity or social
justice. Still, the dream of equity will remain in the (good) society, and spatial equity
and balanced spatial development … will remain the dream of spatial planners who
wish to contribute their part to achieving this vision” (Kunzmann 1998:104:118).
The role of migration can be seen as a liberator of individuals from space and different types of poverty. At the same time, regional and individual poverty can become
a barrier to migration. Different studies have given hints about the existence of individual poverty barriers in Estonia (see chapter 2, also Kutsar and Trumm 1993), but
there are no previous country-wide empirical studies on regional effects of poverty
on the freedom to move.
The aim of this chapter is to investigate the possible presence of a poverty trap
in migration on regional level. Therefore, we will map the poverty and study the
174
influence of poverty on migration in the 1990s in Estonia. Indicators both for individual and regional municipality level poverty data are presented for that purpose.1
Poverty lines are used here as tools to classify inequality differences. The situation
can be interpreted as more just in the case of absent migration barriers. Indicators of
outmigration from unfavourable conditions will show the individual’s ability to adjust to the conditions and improve well-being.
Discussion about spatial equity and equality in society is based on a mixture of
compromises between justice, rights and reasonable choices. Regional equity and
equality shape fulfilment of individual rights on space. To summarise the basic differences in the views about territorial equality we will follow the three different standpoints described by Harvey (1973): (1) free-market competition and mobility
encouragement approach, (2) regional equality approach, (3) compromise between
the two approaches. It is interesting that the same principal arguments about
economic well-being and moral principles are used to justify regional diversity and
equity. In the next sections, we will look at the three approaches through these
economic and moral arguments.
According to competition approach, economic justice is produced by the freemarket competition and competition differences in ability and merit result in more
productive people and localities with greater rewards. It is argued that social wellbeing will be higher in highly productive regions than in less productive regions. It
often entails concentration of resources. Many arguments in favour of this approach
are based on endogenous growth theory. The main elements of endogenous growth
theory are (1) accumulation of capital goods or investments, (2) increase in the
quality of labour force, (3) relocation of resources from low to high productivity
sectors, and (4) technological change (Durlauf et al 1996). The main economic
effects of this phenomenon lay in the so-called multiplier effect (Myrdal 1956, Mera
1981, Vining 1985, Durlauf et al 1996, Cervero 2001). It has been found that economic concentration of resources and agglomerations is the most effective way of
production. In contrast, equal distribution of settlement is believed to lead to a
situation where people are enticed to remain in low-productivity areas and in the
end, there may be neither people nor place prosperity. The reason for that is that the
marginal productivity of labour might be higher elsewhere (Courchene 1987:182).
It has given grounds to argue that the aim of governments should be, instead of
distributive regional policy, to mobilise people for a full-scale utilisation of their
resources and to induce movement of people to areas where their labour is likely to
be the most productive (Hansen 1972, Cameron 1972, Mabogunje 1981). Mobility
deprivation has been seen as an essential type of social deprivation (Pacione 1984). It
can be a reason why many European countries have used some form of subsidy assistance to persons migrating from rural to urban areas, despite the doubts concerning
its positive effects (Hansen 1981).
1
Many empirical studies report on the impact of government policies on migration (Simmons 1984,
Bolton and Chalkey 1989, Söderling 1988, Bengtsson and Johanson 1995, Vining 1985, Goldstein et al
1997), however there are contradicting results as well (Bontje 2001). In this chapter, we will not analyse
the impact of government policies to migration, but assume that they influence through the alleviation of
income poverty (see Ainsaar 2003).
175
Greater freedom to move also supports equality. Migration tends to reduce regional
differences in growth (Hietala 1973) and promote balanced regional growth in
accordance with the theories of balanced growth and international trade. According
to labour supply-demand curve, migration reduces income and demand for labour in
the region of arrival and thus decreases the production costs of enterprises in the
region. At the same time, migration reduces the supply of labour in the region of
departure. According to endogenous growth theory, migration is seen as a balancing
force between the demand and supply of labour forces. Empirical research within
this approach is loaded by the competition-based utility assumption (Rossi 1980,
Simmons 1984, Richardson 1973, Temple 1994, Pekkala and Kangasharju 1998,
Fischer 1998). It has been found to be beneficial also to migrants to move. There is
evidence about the influence of migration increasing the incomes of migrants
(Morrison 1972, Toomet 1991), opening up paths for occupational mobility (Morrison
1972), and unemployment equilibrium (Decressin and Fatas 1995). At the same
time other researchers have found that migration works as an equalising mechanism
only in the case of low barriers (Pekkala and Kangasharju 2000) and people tend to
remain still less mobile than goods (Morrill 1970). Several empirical results show
that migration fails to adjust to market forces, at least in the short run (Groenewold
1997, Martin and Sunley 1998).
Nevertheless, migration is not the smooth equaliser at all periods. Migration seems
to be less effective equilibrator of regional disparities in periods of recession (Cheshire 1981, Gordon 1985). Macro level effectiveness of migration is dependent on the
level of economic well-being. The explanation of relationship between migration and
economy is based on fluctuations of migration, economic barriers and on change of
prospects. For example, Pekkala (1998) analysed the connection between aggregate
economic fluctuations and regional productivity convergence in Finland during 1988–
1995. He found that the high intra-distribution mobility occurred during booms,
conversely regression years were characterised by much lower mobility and more
divergent regional pattern. The same principle can be true also in societies and areas
with different levels of other aspects of well-being.
Moral rationality of unequal development can be backed by the fact that concentrated economic growth can be beneficial in longer perspective also to other regions.
It will be in accordance with Rawls’ (1976) principles of justice: to the greatest
benefit of the least privileged. More conservative critics hold that the pursuit of
equality is incompatible with freedom. Strive for equality destroys the incentives on
which the market economy relies, and it is ultimately futile since new forms of
inequality will inevitably arise to replace those that have been suppressed (Letwin
1983, Miller 1993). Finally, Powell and Boyne (2001) concluded that much of the
existing work takes a centralist perspective assuming that all geographical inequalities are defect, ignoring at the same time the issue of local autonomy and politics.
Inversely, detected inequality may not be bad and greater spatial equity may not
necessarily be “good”. Other authors (Regan and Stewart 1982, Ashford 1990) have
also found that some degree of inequality may be a price worth paying for local
democracy and freedom. Nevertheless, there is always a good reason to ask how big
differences could be tolerated for the sake of democracy and when could we start to
see those big gaps as an essential threat to justice and well-being. In order to presume
176
just outcome in a situation of total local democracy at least two conditions should be
fulfilled: (1) fully equal participation of all members of the community in decisionmaking, and (2) full and free mobility of the population (Curtis 1989).
Regional equality supporters develop the second group of approaches. They are
the people who tend to believe in social justice, based on the broad principle of the
equality of outcome. According to this view, it is expected that some level of social
rights and equality should be guaranteed for all inhabitants, irrespective of their
place of living. The main economic argument against free development is the slow
adjustment towards a long-run equilibrium (Pissarides and McMaster 1989, Tervo
1991). There can be also other economic benefits accompanying the regional equality: saved costs of migration of individuals, reduced urban externality costs. Even
efforts to avoid migration might serve as a principle of regional equality (Tervo 1991).
It is a paradox of the migration process that, although movers may benefit individually, their actions collectively could lower the aggregate life quality of the population
(Morrison 1972). Chronic population decline confronts policy makers with (1) a
problem of people and (2) a problem of place. Outmigration areas are no longer
competitive: they have little prospect of attracting economic activities, although subsidies sometimes prolong their decline. This might be a reason why deprived areas
require positive discrimination (Curtis 1989).
The third group of people remain between the two previous attitudes. They are the
ones who would accept a certain amount of inequality of outcome as the price for
economic efficiency. Still, this approach is made more complicate by different
understandings about inequality, merits, outcomes and different combinations of
principles of justice and equity.
4.1 Regional equality and equity
Generally, people seem to support the idea that unjust inequalities should be
reduced and, if possible, eliminated. Often distinction is made between equality and
equity (see Champernowne and Cowell 1998). Equity is more dependent on standards and beliefs that are the best incentives for fairness and equality in a society.
Curtis (1989:4) states that “welfare could be seen as a situation in which all members of the society enjoy similar level of well-being in terms of the material and nonmaterial resources. Equity would prevail if the distribution of individual well-being
or real income in a society was considered appropriate and socially just”. Deciding
about equality and equity is complicated because of variety of different principles of
justified distribution of goods: utilitarian (maximising the overall benefit), pareto
optimum (distribution towards improving the benefits without worsening the position of the members) and Rawlsian (under the veil of ignorance to the greatest benefit of the least privileged ones). As in society generally, the spatial strategy of deciding about equity and equality itself is quite unclear because of the large variety of
criteria. We can differentiate between equality of expenditure, provision, opportunity, access, use, or outcome (Le Grand 1982, O’Higgins 1987, Mooney 1994,
177
Heywood 2000). Even deciding about equal results, we can take into account several
criteria: desert, worth, work, need, contribution to the common good, rights, entitlement, need (Runciman 1966, Harvey 1973, Titmuss 1974, Miller 1976, Plant and
Taylor-Gooby 1980).
In an empirical research, Smith (1998) suggests using an approach of equalisation
rather than equality. He states that achieving equality is virtually impossible. Equity
is a moral approach, encouraging to be good in moral sense and believing that more
equal world guarantees more equal division of more and less fortunate places (Smith
2000).
In social analyses, the most favoured indicators for measuring equality seem to be
income, lack of resources, social status and education. The broader concept of wellbeing is less often used (see for example Raunio 1983, Ringen 1995). Analysing a
range of papers on territorial justice Powell and Boyne (2001:183) come to the conclusion that “most of geographical studies examine equity rather than equality and
focus on the distribution of need, or “territorial justice””. This is partially because of
the fact that a lot of such research focuses on the analysis of the equality of services
instead of spatial equity or equality (Curtis 1989, Powell and Boyne 2001). Servicebased approach also leads to estimations of needs for services. From the point of
view of equity, services are only one part of the total well-being of inhabitants and,
therefore, the approach is clearly restricted. Often the concept of territorial justice
is used as a synonym of equity in regional studies. A great deal of literature measures
territorial justice within regional natural environment or through natural risks to
different population groups (see Morello-Frosch et al 2001, Stretesky and Hogan
1998).
Other researchers have doubts in the success of measuring regional inequality.
Powell and Boyne (2001) came to the conclusion that geographical inequality is problematic for three main reasons. Firstly, because of unclear geographical aims of the
welfare state with regard to the spatial strategy of equality. Secondly, the authors
criticise the lack of wider context for understanding the link between spatial equity
and local autonomy and, thirdly, because of the short-sighted interpretation of local
differences, which need not always be “bad”. Authors correctly point out the need to
differentiate between the context of welfare delivery and the reasons of inequality.
They justify the inequality with the argument that equity is opposed to equality (see
also Plant et al 1980, Wheale 1983) and with a variety of different measures for
equality (Powell and Boyne 2001:182). Powell and Boyne (2001) recommend analysing local disparities in the wider context of local autonomy and territorial justice.
They do not clarify the meaning of territorial justice in their paper, but the statement is interesting in the light of Sen’s (1999) argument for democracy as the main
value to be achieved, because there seems to be a contradiction between the general
strive towards equality and the right of local governments to self-determination.
Increasing local autonomy is usually associated with greater variations, and greater
centralisation with greater uniformity and equality (Page 1982, Brown and Jackson
1986, Ulst 1996).
178
4.2 Regional versus individual poverty
Poverty is often defined as the lowest stratum of inequality. Individuals and regions
are both sources of regional differences, but we are at this point more interested in
whether regional and individual poverty are interrelated. For example, Courchene
(1987:181) speaks about “people prosperity versus place poverty”. Powell et al
(2001) define people poverty as a situation when low-income people occupy certain
parts of space by virtue of low income, but their incomes are not low because of the
place where they live. Space can provide opportunities for people, as well as impose
limitations on them.
Environment has a considerable influence to human activity since it limits man’s choice.
Within these limits, however, man is free to choose. It is here that his personal nature
becomes decisive: the way he perceives and evaluates the environment and how he behaves
under different conditions Whynne-Hammond (1979:9).
Other authors see the place as having greater impact on individuals. Smith (1977)
believes that place poverty emerges when other benefits or penalties confound the
advantages or disadvantages of particular groups by virtue of where they live. The
idea that individuals are prisoners of their environment is presented in many studies.
For example, Townsend (1979) found that people with identical cash incomes differ
greatly in support they may obtain from free public services, depending on the area
they live in. Jargowsky (1997) believes that problems of distressed urban areas are
not only income problems, but a mix of different economic, social, environmental
circumstances. According to those findings, space and community can strengthen or
give a relief to individual poverty. This is a reason why it is often found (Glennerster
et al 1999, Berman and Philips 1999, Powell et al 2001) that the concept of social
exclusion has several advantages — including greater focus on local community rather
than on people — compared with the concept of poverty. Still, the concept of poverty is more often used among researchers because of its simpler measurement and
socially sensitive character.
Some researchers (Harvey 1973, Galbraith 1992, Smith 1999) have found that
the place and individual poverty are connected but they do not necessarily coincide. Curtis (1989:141) mentions that it is not clear, for example, “whether territorial needs are simply the sum of the individual needs of the population in the area. In
some cases, there are attributes of the locality itself, which influence needs for
welfare resources, and the concentration of vulnerable groups in some deprived
areas may tend to produce a multiplicative effect of deprivation”. Smith (1999)
measured place poverty (lack of necessities, spending below the aggregate level that
delivers necessities) and people’s poverty (weekly income per capita) with aggregated data on county level 1981–1995 in England. According to this study:
1. Pattern of people poverty was more stable over time, than place poverty.
2. Areas with low income tended to have higher local spending.
3. High need areas tended to have higher levels of expenditure.
4. Areas with high level of people poverty did not have high level of place
poverty.
179
One key element of intermediating regional and individual poverty is free movement of people. For example, Courchene (1987) writes that place prosperity is closer
to people prosperity if mobility of individuals is limited. Migration might be an
essential determinant in economic, urban and demographic processes. Because of
age- and skill-selective character of migration, positive net-migration is often seen as
a sign of prosperity and negative migration predicts regression (Morrison 1972). Push
factors, forcing people to move, are important to the well-being of those people.
According to the rational behavioural approach, people move to improve their wellbeing conditions, and in many cases to improve their individual prospects. However,
outmigration has a double meaning. On the one hand, it reveals shortcomings of the
regional well-being, and, on the other hand it is a sign of missing barriers. The ability
to leave is an important precondition for the freedom to fulfil personal needs. Lower
barriers in migration will contribute both, to economic growth and better possibilities of achieving social justice. We can expect that people with less economic barriers
for migration have fewer constraints than others and are therefore less vulnerable to
unfavourable social conditions.
Obstacles in migration are usually labelled as barriers. Summarising previous
research, we can conclude that the ability to migrate could depend on person’s social
ties (the less ties, the greater the freedom, people without families are more
mobile), age (younger people have more freedom), social capital (the more social
capital, the easier to move), economic capital (more capital gives a greater freedom
of choice) and environmental capital (information, infrastructure, mobility conditions). Barriers are not only the results of individual characteristics, but are dependent also on the general level of development of the society — availability of free
housing stock, transport, availability of information, labour situation and ultimately
even traditions. In this research, we concentrate our attention only on economic
well-being and migration barriers. Many researchers (Whynne-Hammond 1979,
Jagijelski 1980, Cadwallader 1989) believe that with economic progress people will
have more freedom to choose the living place and to move.
4.3 Data and method
According to classical migration laws, on regional level the interrelationship between
migration and poverty is quite simple2 . To be more transparent we use standardised outmigration data for analysis and assume that migration is influenced by
economic motivations. In a situation of free movement, outmigration and income
level will produce linear relationship, because people will try to move out from
regression regions and push forces diminish in economically more advanced places
(Figure 4.1). We are particularly interested in the impact of low income. Low
income can be a push factor for higher mobility, but can also produce individual
barriers to movement. In the case of barriers, the relationship between outmigration
and economic well-being would look like reversed U curve (Figure 4.1 line a).
2
This is not the case on individual level.
180
The explanations of expected outb
comes from the combination of
regional economic poverty and
Free movement
outmigration are presented in
Figure 4.2. In two combinations
out of four, we see high mobility
(2, 4), which is a clear sign of
missing barriers. The explanation
of the third combination — nona
Barriers
poverty and settled people is
probably satisfaction with present
conditions and missing incentives
Incomes
for outmigration. Our particular
Figure 4.1 Income and outmigration curve
interest lies in the first group
(a – mobility barriers reduces outmigration in
where poverty occurs together
low income groups, b – missing barriers and
with settled people. This combipush forces in low income areas)
nation is most open to the risk of
barriers. In our empirical data, we
will evaluate existence of possible poverty trap for migration in low-income local
units.
Poverty
Non-poverty
1. No outmigration
poverty trap, barriers
2. Outmigration
no barriers, push forces
3. No outmigration
satisfaction with a region
4. Outmigration
non-economic moves
Figure 4.2 Theoretical explanations of combinations of outmigration and regional wealth
Studies show that the selection amongst alternative indicators might influence the
result (Kangas and Ritakallio 1995, 1998; Quadrado et al 2001, Stewart 2003). In
addition, the unit of measurement also influences the results. We believe that local
municipality unit is the best level for analysis. Municipalities in Estonia are small
enough to reflect small territory disparities and, having also some degree of autonomy
they contribute to the production of local well-being. Functional commuting based
areas could be even better statistical units, but unfortunately it would be too labourintensive to use them in our analyses. Local municipality data are good for our
analyses as functionally justified and small enough research units to reflect the
differences. Table 4.1 gives an overview of the indicators used in our study. All indicators are municipality averages.
Relatively simple indicators were used in the analysis to measure equality. Our
analytical capability was limited by the short time period of the municipality-level
background data. Background data were derived mostly from the middle of the 1990s.
However, the data used here are the earliest data available. Despite some time
181
Table 4.1 Data
Migration
Individual
poverty
Indicator
Source
Net migration rate
Census
Internal outmigration rate 1989–2000 standardised by Census,
age structure of capital 1989 in age 0–80
calculations
Average social assistance benefit per capita 1997, 1998, Ministry of Social
1999, 2000 in municipality
Affairs
Poverty assistance poverty
Income poverty (60% from median). Average income of Calculations
those who got incomes 1999–2000.
Poverty of local Resource poverty (60% from median). All resources of Calculations
municipalities local municipalities per capita 1996–2000, average per
inhabitant
Background
Registered unemployment rate from 1995–1996
indicators
(divided by people in working age)
Average income of person receiving income 1999–2000 Tax Office
Distance from capital
Local municipality
data base
Distance from county centre
Local municipality
data base
Type of local municipality (capital, big towns and county Local municipality
centres, satellite town, other towns, hinterland of bigger data base
towns, hinterland of smaller towns, rural municipalities
with roads, rural municipalities with railway, periphery).
Demographic composition 1989, share of 15–35 year old Census,
in population
calculation
limitations for data, we believe that it can be justified to combine this data if the
municipalities have preserved their general ranking within the total sample throughout the years. Kivilaid et al (2002) have shown differentiation of regional incomes
during 1995–2000. Inhabitants of towns and their surroundings had comparatively
higher incomes than those of other municipalities at the end of the period. Comparison of background data showed that income ranks and total municipality fiscal
resources were quite stable, but municipality level unemployment fluctuated more
sharply (See also Chapter 1).3
The second justification for combining indicators from different periods is the
particularity of census-based migration data: not all migration acts are recorded, but
only final destination is registered. We have good reason to believe that the moves at
the end of the 1990s have greater impact on final migration indicators of census than
those made during previous years and, accordingly, all our results are most relevant
for the later period of the 1990s.
3
Income data from 1999–2000 are used here because of their higher reliability, but models were
controlled also with data from 1996. Data from 1996 gave the similar results.
182
Distanc e from Tallinn
300
To describe the mobility of people,
two indicators — net migration and
standardised internal outmigration
rate — are used. All migration indica200
tors are based on internal population
sample of census — individuals who
were alive and in Estonia both during
the last census in 1989 and in 2000.
100
This sample excludes the impact of
international migration and creates
closed country model. Outmigration
0
is used as the main indicator of free-30
-20
-10
0
10
dom to move. Because of the high
emigration rate - standardised emigration rate
dependency of migration on the age
structure, we standardised the Figure 4.3 Change of outmigration rate
outmigration rate. Age structure of after standardisation and distance
the capital as the municipality with from capital
biggest population was used as a
standard. The results of standardisation showed that the more distant the municipality was from the capital, the larger the hypothetical outmigration compared with
real outmigration was after standardisation (Figure 4.3). It might indicate that real
outmigration from more distant regions is more dependent on the lack of demographic migration potential.
All poverty measures used in this survey are based on the 60% from median
level. We use two principal types of regional level poverty measures — individual and
municipality resources. Two individual poverty indicators are poverty assistance per
inhabitant and average individual income per income receiver. Poverty assistance is a
means tested in Estonia, as well as in many other countries. During the initial stage
we considered including also an indicator of poverty assistance per receiver, but after
analysing the spread and essence of the indicator, it turned out to reflect more the
price level in the region than poverty. Therefore, it was excluded from subsequent
analyses. In addition, continuous indicators of income, assistance and resources are
used in the models.
Municipality resources are reflected by total fiscal resources per inhabitant
during 1996–2000. Proportion of the resources allocated from central funds to local
municipality is believed to be the fundamental issue in the relationship between
central and local governments (Curtis 1989). Appendix 1 displays equations and
description of the assistance delivery rules during the 1990s in allocation of money
to local municipalities from the central government budget. 56% of all the income
tax goes directly to the local municipality budget, but there was no correlation
between average income and municipality resource data. We assume that the
municipality fiscal resources reflect indirectly the local infrastructure level, which
influences well-being of the inhabitants.
Assuming that outmigration is not only an outcome of poverty, we added to the
models several background indicators: registered unemployment in 1995–1996,
distance from centres, and the share of 15–35 year old people in the population.
183
According to our reasoning, factors which show an influence to outmigration cannot
at the same time be barriers to movement.
Unemployment is clearly an economic indicator, but with a rather controversial
meaning (Kitching 1990, Cross 1990, Westerlund 1995). In some countries and
occasions, it diminishes and in other situations it increases outmigration. As in the
case of general economic development data, it seems to depend on economic cycles.
The distance from centres might be related to information flows and equality of
access to services. It is a very broadly used indicator in migration studies. However,
it does not have a very evident explanation. The share of 15–34 year old people is
used as an indicator of people most in risk of migration. Analyses revealed that the
share of 15–34 year old people had a strong positive correlation with the share of
children and all working people in municipality and strong negative correlation with
the share of retired people. Therefore, it can be used as substitute indicator for the
whole age structure.
In analysing the influence of poverty on migration, we run into methodological
problems with the collinearity of data. One of the most obvious disturbing relationships in the case of Estonia is the self-correlation of income level, population size,
settlement type and migration. Several previous studies (Ainsaar 1994, 2002a,b)
show that people in large towns tend to be more settled and with higher incomes.
Part of it seems to be an administrative statistical effect and another part the effect
of “environmental capital” — concentration of housing and services in rather tight
territory, which allows people to fulfil several needs on rather small territory without
leaving the statistical unit of migration. In order to solve this problem, we tried to
avoid using population size and settlement type in the same models. Still, the type of
settlement turned out to be so powerful and good explanatory indicator that we
included it at a later stage. In order to produce better understanding of the interrelationship between indicators, we use simple correlations simultaneously with
regression.
4.4 Regional poverty in Estonia
Different poverty measures gave different results about the spread of regional
poverty (Table 4.2). According to aggregated individual income and municipality
resource, Estonia is regionally more equally developed. Both approaches distinguish
only one municipality, which remains below poverty line. Segregation by poverty is
the highest, if using an indicator of poverty assistance per capita. According to 1997–
2000 data, in 39 municipalities could be classified as poor using this approach.
Income poverty map (Figure 4.4) shows that higher average incomes are concentrated into bigger towns and to their hinterland. Division of individual poverty by
average poverty assistance per inhabitant (Appendix 4.2) shows generally the same
south-north difference but is much more scattered.
Correlation analyses of data revealed that average individual income holds the
central role compared with other poverty indicators in the system of regional
economic and demographic resources (Figure 4.5). Individual income had strong
184
Table 4.2 Three different regional poverty (60%) types
Poverty assistance
per capita
1997–2000
Poverty
Median-poverty
Above median
Total
N
39
84
124
247
%
15.9
34.0
50.1
100
Individual income
1999–2000
N
1
122
124
247
%
0.4
49.4
50.2
100.0
All fiscal resources of
municipality
1996–2000
N
%
1
0.4
121
49.0
125
50.6
247
100.0
links with distance from capital, net migration rate and type of settlement. Incomes
were higher and there was less poverty the closer the municipality was to the capital
and the higher status the municipality had in the hierarchy of settlements. Net
migration rate has been more positive in municipalities with higher incomes. Net
migration was negatively correlated with outmigration.
From different poverty indicators, only the correlation between income poverty
and poverty assistance was revealed. Total local municipality resources did not have
any impact on poverty or migration indicators. It shows that regional resources and
average individual poverty were not related.
Tallinn
Narva
Tartu
Pärnu
Above median
Between 60% of median and median
Below 60% of median
Figure 4.4 Regional income poverty 1999–2000 (Average income in municipality)
185
Share of 15-35
year old
0.43
Distance from
capital
-0.57
Average
individual income
-0.51
Type of
settlement
0.39
0.33
Poverty assistance
per capita
0.36
Net migration
rate
Distance from
county centre
-0.63
0.59
0.19
Standardised
emigration rate
Unemployment
All local
resources
Figure 4.5 Main correlation path between poverty, migration and distance
(Pearson correlation, municipality averages)
cumulatice % of all municipalities in this type
Different poverty areas can overlap in territory. Cumulative individual and regional
poverty might create the highest barriers to mobility. Figure 4.6 reflects the cumulative distribution of income, poverty assistance and municipality resource poverty by
settlement types. It reveals the increase of income poverty in direction of periphery.
Consequently, the share of poor municipalities increased towards rural and peripheral municipalities. Municipality fiscal resources did not grow in linear way towards
settlement hierarchy. The majority of municipalities in the hinterland of bigger towns
experience the greatest shortage of fiscal resources per capita: at the same time they
were municipalities with greatest population growth during the 1990s (Chapter
1.4.2). Assistance poverty was quite equally spread in all municipality types. It seems
to be statistically highest in the capital, but this is a statistical result, because this
200
180
160
140
120
100
80
60
40
20
0
capital
county
centre
satellite
towns
other towns
resource
around
bigger towns
around
smaller
towns
assistance
rural with
roads
rural with
railway
rural
periphery
income
Figure 4.6 Cumulative share of municipalities below median by average
municipality individual income, poverty assistance and resources by municipality types
186
group included only one unit. Assistance poverty was the only poverty group, which
included the capital. The cumulative sum of poverty measures was the lowest in
satellite towns and tended to increase towards lower hierarchy settlement types.
Cumulative sum was mainly shaped by the income poverty.
4.5 Income and migration
Analyses of migration and poverty always run to the problem of interaction (Pekkala
and Kangasharju 1998). Migration is a determinant of regional development and at
the same time, it is itself affected by regional changes. Figure 4.7 reveals that municipalities with higher individual incomes had also higher positive net migration. If
this is quite an expected result, the margin of the curve with lower income municipalities is more interesting. It shows that also the poorest income municipalities
gained from migration and suggests that the overall macro level net migration can be
hardly constructed only in economic terms.
migration per 1000 inhabitants
60
50
40
standard outmigration
30
netmigration rate
20
10
0
-10
15000
25000
30000
35000
40000
45000 >50000
Average individual icome in municipalities
Figure 4.7 Net migration and standardised outmigration rate by average incomes in municipalities
Standardised outmigration curve drops, as expected, in richer areas and is higher in
lower income municipalities, but it does not increase any more in very poor areas.
Interesting is that the outmigration rate does not change from certain point and stays
steady at the same level. This fixed position might be a sign of slight migration barrier. In a situation of strong barriers the curve would drop.
Figure 4.8 illustrates division of municipalities according to income poverty and
standardised outmigration rate. Median line of incomes and outmigration was used
as a tool to organise all municipalities into four groups. It was assumed that the group
of municipalities with the combination of low income and low outmigration rate
might be the most open to the risk of migration barriers. 30% of rural periphery
municipalities were classified to this group, but the risk of barriers was quite impor-
187
% of municipalities from this group
100%
90%
80%
70%
60%
no poverty+high outmigration
50%
no poverty+low outmigration
40%
poor+high outmigration
30%
poor+low outmigration
20%
10%
periphery
rural with
railway
rural with
roads
hinterland of
smaller town
hinterland of
bigger town
other towns
satellite
county centre
capital
0%
Figure 4.8 Municipality types by income poverty and standardised outmigration
rate
tant also in other rural municipalities and smaller towns. Interesting was the position
of rural municipalities in the hinterland of bigger towns. They were more similar
here, as in many other occasions, to towns rather than to rural areas. The second
group of municipalities we analyse here were those with high outmigration but
with low incomes. These were poor municipalities with low migration barriers. The
share of those municipalities also increased towards periphery. The third group of
local units with high incomes and small outmigration were so-called satisfaction
municipalities. In all municipality types also some share of local territories were
characterised by high incomes, but also rather high outmigration rates. This
type was the largest in county centres and less represented in periphery and seemed
to increase towards the top of the settlement hierarchy. Regional map of those four
types (Figure 4.9) reveals strong inclination of the country towards North-West.
Bigger centres and their hinterland were clearly the regions with more favourable
economic conditions and smaller outmigration. The majority of country territory can
be described as outmigration areas and there were scattered local municipalities with
low outmigration despite of economic poverty.
In order to investigate the interrelationship between outmigration and economic,
demographic, geographic resources, we used linear regression model.4 Table 4.3 shows
differences of outmigration model between and within settlement types. Only these
results are presented in the table, which gave statistically significant models. Four
municipality groups out of nine had statistically significant outmigration models. All
other municipalities were later unified under the group “other”. We will interpret
all factors contributing to emigration as factors not setting the barriers to freedom of move.
At first, three different general models were constructed. The first model included only income and distance indicators. Unemployment, demographic potential
4
Linear model was used because of linear dependency.
188
Tallinn
Narva
Tartu
Pärnu
Poor and low outmigration
Poor and high outmigration
No poverty and low outmigration
No poverty and high outmigration
Figure 4.9 Municipalities according to income (1999–2000) and standardised
autmigration rate (1989–2000)
Table 4.3 Factors influencing standardised outmigration rate (Linear regression
betas, division of municipalities see Appendix 1.4, Local Municipality Database)
All
Average income
Distance from capital
Distance from county c
Unemployment 95–96
15–35 % 1989
All local resources
Type
Poverty assistance 5
N
R Square
Small
towns,
Model 1 Model 2 Model 3 except
satellite
Rural,
close to
big
towns
Rural
with big
roads
Periphery
Other
–0.45*
–0.13*
–0.14*
Not inc
Not inc
Not inc
Not inc
Not inc
–0.46*
–0.197*
–0.176*
0.04
–0.14*
0.07
Not inc
Not inc
–0.63*
–0.13
–0.10
0.01
–0.15*
0.05
0.40*
Not inc
–0.21
0.08
–0.18
0.05
–0.46*
0.41
Not inc
Not inc
–0.63*
–0.05
–0.04
0.03
–0.04
0.07
Not inc
Not inc
–0.54*
–0.62*
0.01
0.32*
–0.37*
-0.29*
Not inc
Not inc
–0.29*
–0.39*
–0.12
0.09
0.14
0.07
Not inc
0.29*
–0.45*
–0.07
–0.18
–0.17
–0.08
0.07
Not inc
Not inc
245
0.167
245
0.200
245
0.300
16
0.704
34
0.417
41
0.427
68
0.207
85
0.144
Not inc = not included
Used only in the model with peripheral municipalities. In all other cases did not fit to the model and
made it worse.
5
189
and local government resources were added to the second outmigration model. The
third model was similar to the second one, but additionally the settlement type was
added. Different models gave slightly different results. Standardised outmigration
rate was negatively related with income in all models. Average income was also the
most powerful determinant of outmigration. According to the first two models, municipalities with lower incomes and closer to bigger towns had higher outmigration
rates. In the third model, after adding the type of settlement, the direct impact of
distance disappeared. Probably the type itself covers part of distance’s explanative
power.
The second and the third model also revealed the dependency of outmigration on
demographic composition of settlement. Outmigration was higher in municipalities
with less people in the age 15 to 35. The influence of 15 to 35-year-old people might
seem to be rather contradictive. People in that age are usually most exposed to the
risk of leaving according to migration age curve. The explanation might lie in the
standardised emigration rates, which eliminate the age structure’s impact on migration. The other explanation is that incomes and demographic composition were interrelated. Selective economic and demographic development of municipalities had
produced the outcome that in places with less 15 to 30 year old people there were
also lower incomes (see Figure 4.5).
Unemployment, municipality fiscal resources per inhabitant and poverty assistance did not generally influence the outmigration. The missing influence of unemployment might seem surprising, but earlier works had also found controversial results. However, in different municipality types the outmigration was shaped by different factors. Local fiscal resources and unemployment had an influence in rural
municipalities, which did not have other advantages except big roads crossing their
territory.
Our particular interest was in peripheral municipalities with highest risk of poverty trap. Although poverty assistance did not have any connections with outmigration
in other models, it turned out to be important in peripheral municipalities. Municipalities that used more poverty assistance per inhabitant had also the highest individual poverty and the highest outmigration rates. It might be an evidence of freedom to move, despite of poverty.
190
Conclusions: Did it exist, freedom to move?
In this chapter, we studied a relationship between equality and migration in Estonia
at the end of the 1990s. We were particularly interested in poverty-caused migration
trap. Economic poverty, geographic and demographic indicators were used in order
to study the impact of resources on age-standardised outmigration.
The comparison of regional level poverty indicators revealed that Estonia was
regionally more equally developed according to aggregated individual income and
municipality fiscal resources per capita. Segregation by poverty was the highest according to poverty assistance per capita. Individual and regional fiscal poverty did
not have a correlation in Estonia. The more peripheral the municipality, the higher
was the probability to fall into some kind of poverty category. Estonian data confirmed the previous findings by Le Grand (1982) that the income might be the best
indicator of regional equity.
General outmigration models showed that the lower the incomes were and the
closer the municipality was to bigger towns the higher was standardised outmigration
rate during 1989–2000. No clear evidence on economic barriers for migration on
regional level was found. However, our analyses were limited only to municipality
level aggregated data.
191
Appendix 4.1 Regional aid and rules of distribution it in Estonia 1990–2000
1990
1996
2000
Equitation for resource allocation
52% individual income tax, incomes
compared with average incomes in
Estonia, number of population,
number of children at school,
distance of public roads, territory of
parks, distance of streets with lights,
territory of housing
56% income tax, number of
inhabitants, gap between average
country level and local income tax
Assistance
Small islands, South – East Estonia, transportation between
mainland and islands, additional assistance
1995-1999 project based government investment program,
priorities were given to poorer and municipalities which
had co financing.
Projects of Foundation of Regional Development
1996-2000 to compensate not received income tax
1996-2000 for public activities, additional
1997-2000 to small islands, additional
1997, 1998, 1999 for persons in care centres
1998-2000 to students to compensate transportation costs
1999 to EU integration
1999 art and music schools
1999 school allowances
1999 for roads
2000 - Ministries became responsible of making decisions
about regional investments. In decisions suggestions from
country governors were taken into account.
192
Appendix 4.2. Poverty assistance poverty
Tallinn
Narva
Tartu
Pärnu
Less than median
Between median and 40% above median
More than 40% above median
Appendix 4.3. All fiscal resources of local municipality per inhabitant
Tallinn
Tartu
Pärnu
More than median
Between 60% of median and median
Below 60% of median
193
Conclusions
194
195
Conclusions
Personal freedom to move can fulfil the role of agent in securing equal well-being and
opportunities in society in a situation where well-being is not evenlly distributed
across space. Therefore, the studies of migration can be important from the perspective of individual and overall well-being. This book studies human migration on macro,
individual and aggregated individual level, as well as human freedom to move. Migration studies reveal migration barriers, which might become well-being barriers. Based
on previous mobility studies, we could assume that individual opportunities for migration can be restricted because of the lack of information, economic barriers, unfavourable housing market, or the burden of some kind of capital. However, the dominant barriers could also be different for different societies. From the viewpoint of
social and regional policy it would be essential to direct attention to the factors that
are understood to constitute the greatest barriers to peoples’ ability to handle their
lives without assistance.
Chapter 2 in this book studies factors affecting individual migration more profoundly in Estonia between 1989–2000. Results revealed possession of economic
capital supported mobility – wealthier people were more mobile. The wealthiest
persons mostly moved because of housing-related reasons. However, the hypothesis
that economic desert was a large barrier to migration was not confirmed by this
research. Standardised out-migration models did not indicate clear evidence of economic barriers to migration on regional level. In the case of some other factors the
results were quite controversial. For example social capital had dual impact on
migration. Children complicated the change of residence, but having a partner made
it easier.
General trends of settlement development
Most of the history of population settlement development draws a picture of population concentration and growth of towns. A remarkable phenomenon in the 20th
century settlement history is the outflow of people from towns, which has led to
more even distribution of population in more advanced societies. Although in different countries the deconcentration processes emerged at different times, most developed countries have passed this stage during some period of their development.
The urban-rural development in Estonia in the 20th century can be divided into
two periods: urbanisation up to 1983 and deconcentration from 1983–2000. Deeper
insight into the internal migration processes in Estonia between 1989–2000 revealed
continuous outflow of people from towns into the rural areas. This direction of
migration has persisted in spite of economic, political and demographic changes since
1983. Population census data from 1989–2000 showed clear depopulation of bigger
196
towns and an inflow of people into the nearest rural hinterland of the towns. Satellite towns and closer hinterland of bigger centres were the most popular settlement
types among internal migrants. The whole process could be labelled as suburbanisation
which reached the distance of up to 60–70 km from the centres. The main age
groups supporting outflow from towns were people from the age 25 and up, who
mainly moved to the closer hinterland of bigger towns, and people from the age 30
and up, who moved also to more distant areas.
Reasons for deconcentration
Academic research analyses presented in this book, point to the dominance of lifestyle-related preferences and housing factors among the reasons used to explain the
migration turnaround. Economic reasons were less frequent, which places them only
to the second group of motivations. However, different authors use varying methods
and sometimes conclusions about reasons are based more on assessments than on
scientific evidences. The first and second chapters of this book present a more thorough analysis of migration before and after turnaround in one country (Estonia),
following the same methodology.
The country study suggested that the main forces behind the migration reverse
were non-economic reasons. Most remarkable change in the reasons of migration
during migration transition was the increase of housing-related migration. Economic
reasons preserved their third position after family and housing. In order to analyse
the mechanism of deconcentration, it could be fruitful to pay attention to the fact
that the migration turnaround usually begins from the bigger towns. This might indicate that those towns possess some characteristics that promote deconcentration.
Simultaneously, the hinterland of those towns experience often the highest growth
rate because of migration. It gives us reason to believe that migration turnaround
might be a result of at least two different factors: (1) increased freedom of people
from concentrated production functions (2) some extra pull or push forces which
makes life outside the urban centres more attractive compared with towns.
The analysis of the reasons of migration in Estonia in the 1990s gave the
following results: the biggest changes took place in migration flows with the capital.
Even though the overall employment-related migration decreased, the employmentrelated migration among in-migrants into the capital increased. The percentage of
those, who left because of housing-related reasons, was doubled among the outmigrants from the capital.
Theory of migration
Different chapters of this book employ a variety of approaches to migration analyses.
We can divide them into two broad categories: macro level and individual level
explanations.
197
On the macro level, three basic components can be distinguished as main forces
that determine the migration outcome: potential of different groups of migrants,
regional differences in well-being and alternatives for migration.
1. Population potential
Different population groups have different needs and therefore also different preferences. For example results from Estonia showed that different age groups emphasised different factors of well-being. The importance of the natural environment
increased with age. Although men were more active in moving for employmentrelated reasons and women in moving for family-related reasons in 1989–1999, there
have been certain changes taking place over time. General migration trends are shaped
by different groups of population and their demographic potential determines the
dominance of preferences among flow of migrants. Chapter 2 provided eleven types
of migration factors, which were identified on the basis of detailed pull, push and
individual age-specific data from the Estonian survey. Cumulative amount of individual well-being possibilities within a region makes some areas more and others less
attractive for certain groups. In the second chapter, it was argued that the traditional
aggregated individual level migration analyses could have more explanatory power
within the theories of well-being or needs. In addition, the concepts of well-being
and ill-being enable easy integration of the push and pull forces that are well known
in migration research. Different groups of people have different values that are
important for their well-being. In this book it is assumed that migrants take into
account several aspects of well-being. Empirical analyses of the Migrant Survey
suggested that the character of those aspects tends to vary depending on the triggering reason of migration.
2. Geography of well-being resources and place utility
Most theories of regional dislocation have emphasised the concept of production
space (economic space model, spatial growth model, endogenous growth theory) in
regional development. Geography of production is important as long, and to such
extent, as human settlement is influenced by the means of production. The role and
influence of different sources of well-being vary in time and, therefore, the attributes
of a successful region change as well. For example, ties between economic activities
and settlement differ in an era with developed communication technology compared to an agrarian. Human needs and preferences shape the place utilities, and the
characteristic features of a successful region can change over the course of time. The
most important factors from the viewpoint of migration magnitude are the place
utility disparities produced by well-being differences. A hypothesis was posed that
the primary reason of migration on macro level could be the factor, which is the most
deficient among regionally differentiated well-being values. For example, housing
reasons would prevail in a society, where there is a high degree of aspirations to
change the living conditions and at the same time major differences exist in the
regional housing markets, which allows fulfilling those needs.
198
3. Access to alternatives to migration
Such alternatives could be commuting, information technology or other processes
that offer alternatives to migration. General economic development is often associated with improvement of transportation and communication, which in turn increases
individual options. This factor could be called freedom from production space. In
order to benefit from those alternatives, people should have some access to them.
For example, commuting leads to greater freedom in choosing the place of residence
and the place of employment. Accordingly, people will have greater ability to fulfil
their individual well-being preferences. Still, alternatives to migration, including commuting, would require some extra resources, compared to a settled lifestyle.
Migration
Freedom to move
Place utility
Geography of wellbeing resources
Alternatives to
migration
Needs and potential
of population
Figure 4. Important factors of migration
On the macro level, migration flows are results of a joint interaction between demographic potential of population subgroups, regional differences in well-being, available alternatives to migration and, finally, individual freedom to move. Population
preferences, migration alternatives and perceived well-being differences determine
the place utility (Figure 4).
Behavioural approach
The third chapter constructs a theoretical model for migration behaviour. The theory
was based on the assumption that migration behaviour originates from individual
needs, and migration is only one adjustment option among several coping strategies
to improve wellbeing. Settled people live in a situation of equilibrium between
perceived environmental differences, personal needs and personal resources for
adjustment. Change in any of these components can lead to a change of equilibrium
and to the appraisal situation. The important steps in order to analyse migration
behaviour are perceived need for changes, adjustment strategies, decision about the
intention of migration, search, evaluation and decision among possible options. All
those stages are influenced by individual needs, resources and characteristics.
199
Several relocation theorists have pointed out the importance of the stress as a trigger
of the process. However, in the qualitative survey many respondents claimed that
despite discomfort they did not feel any stress. At the same time, there were cases
where the presence of stress was a reason of migration. Interviews with migrants
revealed large amount of possible combinations in behaviour. For example triggering
factors of migration appeared individually, as well as in combination with many other
reasons. Similarly, there could be one or many reasons of relocation. Some reasons
originated from past, but some were constructions of future life histories of people.
Even though the need for change usually emerged before the opportunity of relocation, there were some reversed cases where the stimulus (opportunity) emerged
before the appraisal of internal needs. This could be interpreted, as a change of
equilibrium caused by perceived opportunities elsewhere. Two principally different
relocation strategies were found:
1) Pull migration, or migration mainly towards higher well-being without a preceding feeling of serious ill-being.
2) Push or ill-being migration, where relocation took place mainly because of poor
conditions in the current place of residence.
There have been long discussions on the rationality of human beings. Interviews
indicated that people’s behaviour carried several rational elements that were used to
decrease costs in terms of time and resources. Although most of the arguments for
the change of the place of residence were rational, there were some references to
emotion-based decisions. The general rational approach was also somewhat supported
by the results of the Migrant Survey, which revealed that, irrespective of directions
or factors of migration, all respondents were more satisfied after the movement than
before. Although several theorethical assumptions were supported by information
from interviews, the future challenge would be to test the model with quantitative
data.
200
Postsciptum
This book was finished in the year 2004. Although the aim of the book was to highlight migration in the 1990s in Estonia, we would like to make a brief remark on the
subsequent years. Official statistics, which is still highly unreliable, shows an
increase in urban population from 2000. This could indicate that the period 1983–
2000, which is analyzed in this book, remains a unique period of deconcentration in
the history of settlement. In a situation of general depopulation, increased commuting and competition of municipalities for population, the battle for personal income
tax and central resources is fiercer than ever. In the year 2004, the official population number in municipalities is influenced by social and local benefits, and partly by
rational economic choices based on free library cards, bus tickets, or the presence of
water centres, which makes the migration analyses based on the official statistics
implausible.
The Spirit of Tartu tells to a cat: You can catch small herring and bream with
worms and bread, but our City Government will catch students with AURA
Source: “Tartu Postimees” (3.12.2003)
The Tartu City Government promised that every student, who will exchange the
monthly transport benefit of EEK 400 for a registered residency in Tartu, receives a
free entry to Aura water centre. It was City Government’s response to the growing
tendency among the students to register their official place of residence outside of
town in order to receive the transport benefit and subsistence benefit. Similar “stick
and carrot” measures and other methods influence the statistics of migration
between settlements.
201
References
Abelson, R.P. & Levi, A. (1998) Decision making and decision theory, Chapter 5, Lindzey,
G. & Aronson, E. (eds) The Handbook of Social Psychology, Random House, New
York, pp 231–309.
Ahnström, L. (1980) Turnround-trenden och de nordiska huvudregionernas utveckling
efter 1950. Nord-REFO, 3–4, pp 43–100.
Ainsaar, M. (1990) Maa-linn rändest Eestis. Diploma thesis, Tartu University Institute of
Geography.
Ainsaar, M. (1994) Noorte ränne Eestis ühe põlvkonna näitel. MSc thesis, Tartu University, Institute of Geography.
Ainsaar, M. (1995) Migration of Estonians and non-Estonians in Estonia. Dahl, O. (ed)
Selected Papers Presented at the Nordic Symposium in Stavanger 24–26. Nov. 1994,
Misjonshogskolens Forlag, Stavanger, pp 206–215.
Ainsaar, M. (1997a) Estonian population from Liber Census Daniae up to nowadays,
Tartu University Press.
Ainsaar, M. (1997b) Eestlaste lahkumist Ida-Virumaalt mõjutavad tegurid. Eesti
Geograafia Seltsi Aastaraamat 30, pp 145–162.
Ainsaar, M. (1999) Changes in internal migration: are there any changes and what counts?
Ainsaar, M. (ed) Regional policy and migration, Tartu University Press, Tartu, pp 114–
126.
Ainsaar, M. (2002a) Migration trends and reasons in Estonia in the 1990’s. Kutsar, D.
(ed) Living Conditions in Estonia Five Years Later. NORBALT II , Tartu University
Press, pp 148–163.
Ainsaar, M. (2002b) Eestlaste territoriaalne mobiilsuse võrdlus kahe
keskhariduspõlvkonna näitel. Titma, M. (ed) 30- ja 50-aastaste põlvkonnad uue
aastatuhande künnisel, Tartu Univerity Press, pp 97–120.
Ainsaar, M. (2003) Migration and poverty in local government units in the 1990s.
Servinski, M. (comp) Cities and Rural Municipalities in Figures, Estonian Statistical
Office, pp 22–27.
Ainsaar, M., Kõre, J., Hendrikson, M., Ernits, P. & Raagmaa, G. (1996) Habitat. Järve, P.,
Toomel, K. and Viik, L. (eds) Estonian Human Development Report 1996, pp 51–68.
Ajzen, I. (1991) The theory of planned behaviour. Organizational Behaviour and Human
Decision Processes 50, pp 179–211.
Ajzen, I. & Fishbein, M. (1980) Understanding Attitudes and Predicting Social Behaviour. Engelwood Cliffs, Prentice-Hall, New York.
Ajzen, I. & Madden, T.J. (1986) Predictions of goal-oriented behaviour: Attitudes,
intensions, and perceived behavioural control. Journal of Experimental Social Psychology 22, pp 453–474.
Amrhein, C.G. (1985) Interregional labour migration and information flows. Environment and Planning A 17, pp 1111–1126.
Amundsen, E.S. (1985) Moving costs and the microeconomics of intra-urban mobility.
Regional Science and Urban Economics 15, pp 573–583.
202
Arsdol, M.D. van (1986) Retrospective and subsequent metropolitan residential mobility. Demography 5, pp 245–257.
Aro, T. (1997) Muutoliike keskittyy korkeakoulukaupunkeihin. Siirtolaisuus 2, pp 5–11.
Auriat, N. (1991) Who forgets? An analyses of memory effects in a retrospective survey
on migration history. European Journal of Population 7, pp 311–342.
Ashford, D. (1990) Intergovernmental transfers in the welfare state. Environment and
Planning C 8, pp 217–232.
Baccaini, B. & Courgeau, D. (1996) The spatial mobility of two generations of youth
adults in Norway. International Journal of Population Geography 2, pp 333–359.
Baccaini, B. & Puman, D. (1996) Migration in France between 1975 and 1990: A limited
degree of Decentralization. Rees, P., Stillwell, J., Convey, A. & Kupiszewski, M. (eds)
Population Migration in the European Union, John Wiley & Sons, pp 191–206.
Bach, R.L. & Smith, J. (1977) Community satisfaction, expectations of moving and migration. Demography 14 (2), pp 147–167.
Bagozzi, R.P. (1992) The self-regulation of attitudes, intensions, and behaviour. Social
Psychology Quaterly 55 (2), pp 178–204.
Barr, J.B. (1997) The law of market areas and the size distribution of urban centres,
Papers in Regional Science: The Journal of the RSAI 76 (1), pp 43–68.
Basu, K. (1993) Welfare economics. Outhwaite, W. & Bottomore, T. (eds) The Blackwell
Dictionary of Twentieth — Century Social Thought. Blackwell. pp 707.
Beale, C.L. (1977) The recent shift of United States population to nonmetropolitan
areas, 1970–75. International Regional Science Review 2 (2), pp 113–122.
Beale, C.L. & Fuguitt, G.L. (1990) Decade of pessimistic nonmetro population trends
ends of optimistic note. Rural Development Perspective 6, pp 14–18.
Becker, G. (1993) Human Capital. Third edition, The University Chicago Press.
Beggs, J.J., Hurlbert, J.S. & Haines, V.A. (1996) Community attachment in a rural setting: A refinement and empirical test of the systemic model. Rural Sociology 61 (3),
pp 407–426.
Bell, M.M. (1992) The fruit of difference: The rural-urban continuum a system of
indentity, Rural Sociology 57, pp 65–82.
Bengtsson, G. (1981) The unpublished writings of Torsten Hägerstrand. Lund Studies in
Geography B 48, pp 237–245.
Bengtsson, T. & Johanson, M. (1995) The new migration transition — the case of postindustrial Sweden. Lundh, C. (ed) Demography, Economy and Welfare, Scandinavian
Population Studies 10, Lund Studies in Economic History, Lund University Press, pp
217–235.
Bentler, P. M. & Speckart, G. (1979) Models of attitude-behaviour relations. Psychological Review 86, pp 452–464.
Berman, Y. & Philips, D. (1999) Indicators of social quality and social exclusion at national and community level. Social Indicators Research 50, pp 329–350.
Berry, B.J.L ed. (1976) Urbanization and Counterurbanization. Sage Publications, Beverly
Hills.
Berry, B.J.L. (1988) Migration reversals in perspective: The long-wave evidence. International Regional Science Review 11, pp 245–251.
Berry, B.J.L. & Dahman, D.C. (1977) Population redistribution in the United states in
the 1970s. Population and Development review 3 (4), pp 443–471.
Black, M. (1983) Migration of young labour force entrants. Socioeconomic Planning 17,
pp 267–280.
203
Bogin, B. (1988) Rural-to-urban migration. Mascie-Taylor C.G & Lasker, G.W (eds) Biological Aspects of Human Migration, Cambridge University Press, pp 90–129.
Bogue, D.J. (1969) Principles of demography. Wiley, New York.
Bolton, N. & Chalkley, B. (1989) Counter-urbanization — disposing the myths. Town
and Country Planning, 58 (9), 249–250, In Robinson, V. (ed) Geography and Migration, An Elgar Reference Collection, pp 283–284.
Bolton, N. & Chalkey, B. (1990) The rural population turnaround: a case study of North
Devon. Journal of Rural Studies 6, pp 29–43.
Bonaguidi, A, & Abrami V.T. (1996) The pattern of Internal migration: the Italian case.
Rees, P., Stillwell, J., Convey, A., Kupiszewski, M. (eds) Population migration in the
European Union, John Wiley & Sons, pp 231–245.
Bonifazi, C., Chieppa, A. & Heins, F. (1997) Inter-provincial migration in Italy 1955–
1993. Paper presented in EAPS Conference, Cracow.
Bontje, M. (2001) Dealing with deconcentration: Population deconcentration and planning response in polynucleated urban regions in North-West. Urban Studies 38 (4),
pp 769–785.
Borgegård, L.-E. & Håkansson, J. (1995) Spridning och koncentration av befolkningen I
Sverige kommuner 1972–92. Layton, I. (ed) Geografiska uppsaster till minnet av
Ingvar Jonsson, Geografiska Institutionen, GERUM 2, Umeå.
Borgegård, L.-E., Håkansson, J. & Malmberg, G. (1995) Population redistribution in
Sweden — long term trends and contemporary tendencies. Geografiska Annaler 77B
(1), pp 31–45.
Boudon, R. (1994) The art of self-persuasions. Cambridge, Polity Press.
Böventer, E. von (1970) Towards a united theory of spatial economic structure, Chapter
20. Dean, R.D., Leahy, W.H. & McKee, D.L. (eds) Spatial Economic Theory, The
Free Press, New York, pp 325–355.
Boyle, P. (1995) Rural in-migration in England and Wales 1980–1981. Journal of Rural
Studies 11 (1), pp 65–78.
Boyle, P., Halfacree, K. & Robinson, V. (1998) Exploring Contemporary Migration,
Longman.
Boyne, G. & Powell, M. (1991) Territorial justice: a review of theory and evidence. Political Geography Quaterly 10, pp 163–181.
Brown, C.V. & Jackson, P.M. (1986) Public Sector Economics, Third edition. Blackwell.
Brown, L.A. & Moore, E.G. (1970) The intra-urban migration process: A perspective.
Geografiska Annaler B 52, pp 1–13.
Brown, L.A, Pavri, F. & Lawson, V. (1998) Gender, migration and the organisation of
work under economic evolution: Ecuador, 1982–90. International Journal of Population Geography 4, pp 259–274.
Bruijn, B.J (1992) The concept of rationality in social science. Postdoctorale
Onderzoekersopleiding Demografie Paper 9, NiDi.
Burt, R.S. (1992) Structural Holes. The social structure of competition, Harvard Cambridge University Press.
Cadwallader, M. (1989) A conceptual framework for analysing migration behaviour in
the developed world. Progress in Human Geography 4, pp 494–511.
Cahill, D.J. (1994) A two-stage model of the search process for single-family houses: A
Research Note. Environment and Behaviour 26 (1), pp 38–48.
Cairncross, F. (1997) The Death of Distance. Orion Business Books.
Cameron, G.C. (1972) The relevance to the United States of British regional population
strategies, with a note on the French Experience. Mazie, S.M. (ed) Population distri-
204
bution policy, The commission on population growth and the American future, Government Printing Office, pp 335–352.
Campell, A., Converse, P.E., & Rodgers, W.L. (1976) The Quality of American Life, Sage,
New York.
Campell, C. (1996) On the concept of motive in sociology. Sociology 30 (1), pp 101–
114.
Carlstein, T., Parker, D. & Thrift, N. (eds, 1978) Timing Space and Spacing Time 1–3,
Edward Arnold, London.
Cervero, R. (2001) Efficient urbanisation: economic performance and the shape of the
metropolis. Urban Studies 38 (10), pp 1651–1671.
Chamberlain, K. (1988) On the structure of subjective well-being. Social Indicators Research 20 (6), pp 581–604.
Champernowne, D.G. & Cowell, F.A. (1998) Economic Inequality and Income Distribution, Cambridge University Press.
Champion, A.G. (1981) Population trends in rural Britain. Population Trends 26, pp 20–
23.
Champion, A.G. (1987) Population deconsentration in Britain 1971–84. Seminar Papers
Department of Geography University of Newcastle upon Tyne No 49.
Champion, A.G. (1989a) Introduction: The counterurbanization experience. Champion,
A. G. (ed) Counterurbanization: The Changing Pace and Nature of Population
Deconcentration, Edward Arnold, London, pp 1–18.
Champion, A.G. (1989b) Conclusion: Temporary anomaly, long term trend or transitional phase. Champion, A. G. (ed) Counterurbanization. The Changing Pace and
Nature of Population Deconcentration, Edward Arnold, London, pp 231–244.
Champion, A.G. (1989c) Counterurbanization in Europe. Geographical Journal 155 (1),
pp 52–80.
Champion, A.G. (1992) Urban and regional demographic trends in the developed world.
Urban Studies 29 (3–4), pp 461–482.
Champion, A.G. (2002) Population change and migration in the British urban system.
Geyer, H. (ed) International handbook of urban systems, Cheltenham, Elgar, pp 87–
120.
Champion, A.G. & Illeris, S. (1990) Population redistribution in trends in Western Europe. Hebber, M. & Hansen, J.C. (eds) Unfamiliar Territory; the Reshaping of European Geography, Aldershot, Avebury, pp 236–253.
Champion, T. (2001) Urbanization, suburbanization, counterurbanization and
reurbanization. Paddison, R. (ed) Handbook of urban studies, London, Sage Publications, pp 143–161.
Chang, C-O., Chen, S-M. & Somerville, T. (2003) Economic and social status in household decision-making: Evidence to extended family mobility. Urban Studies 40 (4),
pp 733–746.
Chapman, K. (1979) People, Pattern and Process, An Introduction to Human Geography. Edward Arnold.
Charney, A.H. (1993) Migration and the public sector: A survey. Regional Studies 27 (4),
pp 313–326.
Cheshire, P.C. (1981) Labour market theory and spatial unemployment. Martin, R.L.
(ed) Regional wage inflation and unemployment, Pion, London, pp 189–207.
Christakopoulou, S., Dawson, J. & Gari, A. (2001) The community well-being
quastionnaire: Theorethical context and inititla assessment of its reliability and validity. Social Indicators Research 56, pp 321–351.
205
Christaller, W. (1933) Die zentralen Orte in Süddeutschland. Jena, Gustav Fischer Veralg.
Clark, E. (1984) Vacancy chains and life-cycle theory, Scandinavian Population Studies 6,
1, Springfeldt & Warnström (eds) The Sixth Scandinavian Demographic symposium
1982, Studies in Migration, pp 85–106.
Clark, W.A.V. (1976) Migration in Milwaukee. Economic Geography 52, pp 48–60.
Clark, W.A.V. (1982) Recent research on migration and mobility: A review and interpretation. Progress in Planning, Pergamon Press, Los Angeles.
Clark, W.A. & Cadwallader, M. (1973) Locational stress and residential mobility. Environment and Behaviour 5, pp 29–41.
Clark, W.A., Deurloo, M.C., & Dieleman, F.M. (2003) Housing careers in the United
States, 1968–93: Modelling the sequencing of housing states. Urban Studies 40 (1),
pp 143–160.
Clark, W.A.V. & Onaka, J.L. (1983) Life cycle and housing adjustment as explanations of
residential mobility, Urban Studies 20, pp 47–57.
Clark, W.A.V. & Smith, T.R. (1982) Housing market search behaviour and expected utility theory: 2. The process of search. Environment and Planning A 14, pp 717–737.
Coleman, J.S. (1988) Social capital in the creation of human capital. American Journal of
Sociology 94, pp 95–120.
Coll, A.G. & Stillwell, J. (1999) Inter-provincial migration in Spain: Temporal trends and
age-specific patterns. International Journal of Population Geography 5, pp 97–115.
Colman, A.M. (2001) Stress. Dictionary of Psychology. Oxford University Press.
Conger, R.D., Lorenz F.O., Elder, G.H., Simons, R.L. & Xiaojia G. (1993) Husbands and
Wife Differences in Response to Undesirable Life Events Journal of Health and Social
Behavior 34 (3), pp 71–88.
Corden, A. & Wright, K. (1993) Going into a home: Where can an elderly person choose.
Champion, T. (ed) Population matters, The local dimension, pp 117–135.
Courchene, T.J. (1987) Political structure and regional balance: some analytical explorations. European Journal of Political Economy 3 (1–2), pp 173–198.
Courgeau, D. (1985) Interaction between spatial mobility, family and career lifecycle: A
French survey. European Sociological Review 1 (2), pp 139–162.
Cross, D.F.W. (1990) Counterurbanization in England and Wales, Avebury.
Curtis, S. (1989) The Geography of Public Welfare Provision, Routledge, London.
Dahms, F.A. (1995) “Dying villages”, “counterurbanization” and the urban field – a Canadian perspective. Journal of Rural Studies 11 (1), pp 21–33.
Dahms, F.A. & McComb, J. (1999) ‘Counterurbanization’, interaction and functional
change in a rural amenity area – a Canadian example. Journal of Rural Studies 15 (2),
pp 129–146.
DaVanzo J. (1978) Does unemployment affect migration? Evidence from microdata.
Review of Economic Statistics 60, pp 504–514.
DaVanzo, J. (1981) Microeconomic approaches to studying migration decision. DeJoung,
G. F. & Gardner, R. W (eds) Migration decision making: multidisciplinary approaches
to microlevel studies in developed and developing countries, Pergamon Press, pp 90–
129.
Deakon, B. (2000) Eastern European welfare states: the impact of the politics of globalization. Journal of European Social Policy 10 (2), pp 146–161.
Dean, K.G. (1987) The dissaggregation of migration flows: The case of Brittany, 1975–
1982. Regional Studies 21 (4), pp 313–326.
Decressin, J. & Fatas, A. (1995) Regional Labour market dynamics in Europe. European
Economic Review 39, pp 1627–1665.
206
DeJong, G.F. (2000) Expectations, gender, and norms in migration decision-making.
Population Studies 54, pp 307–319.
DeJong, G.F.& Fawcett, J.T. (1981) Motivation for migration: An assessment and a valueexpectancy model. DeJoung, G. F. & Gardner, R.W. (eds) Migration Decision Making: Multidisciplinary Approaches to Micro Level Studies in Developed and Developing Countries, Pergamon Press, New York, pp 13–58.
Doyal, L. & Gough, I. (1991) Theory of Human Need. London, Macmillan.
Durlauf, S., Helliwell, J.F. & Raj, B. (1996) Long-run economic growth. Empirical Economics 21, pp 1–10.
Dziewoñski, K. & Korcelli, P. (1981) Migration and settlement: Poland. RR-81-20, International Institute for Applied Systems Analysis, Laxenburg, Austria.
Dövényi Z., Kok, Z., & Kovacs, Z. (1998) A szuburbanizâció, a lokâlis târsadalom és a
helyi önkormânyzati politika összefüggései a budapesti agglomerâcióban. Illés, S. &
Tóth, P. (eds) Migrâcio, Volume 1, KSH Népességtudomânyi Kutató Intézet, Budapest, pp 229–237.
Eamets, R. (1994) Labour market and employment issues of transition economies: the
case of Estonia. Communist Economies & Economist Transformation 6 (1), IEA, pp
55–73.
Ellaway, A., Macintyre, S. & Kearns, A. (2001) Perception of place and health in
socially contrasting neighbourhoods, Urban Studies 38 (12), pp 2299–2316.
Elliot, J.R. & Perry, M.J. (1996) Metropolitanizing nonmetro space: Population redistribution and emergent metropolitan areas, 1965–90. Rural Sociology 61 (3), pp 497–
512.
Elster, J. (1993) Some unresolved problems in the theory of rational behaviour. Acta
Sociologica 36, pp 179–190.
Fernandez, R.R. & Dillman D.A. (1979) The influence of community attachment on
geographic mobility. Rural Sociology 44, pp 345–360.
Fidlay, A. & Rogerson, R (1993) Migration, places and quality of life: voting with their
feet. Champion, T. (ed) Population matters, The local dimension, pp 34–49.
Fielding, A. (1982) Counterurbanization in Western Europe. Progress in Planning 17 (1),
pp 1–52.
Fielding, A. (1987) Counterurbanisation in the Western Europe. Findlay, A.M. & White,
P.E. (eds) West European Population Change, Croom Helm, pp 35–49.
Fielding, A. (1989) Interegional migration and social change of the study of South-East
England based upon data from the Longitudional Study. Transactions of the Institute
of the British Geographers 14, pp 24–36.
Fielding, A.J. (1993) Migration and metropolis: Recent research on the causes of migration to Southeast England. Progress in Human Geography 17 (2), pp 195–212.
Filkins, R., Allen, J.C. & Cordes, S. (2000) Predicting community satisfaction among
rural residents: an integrative model. Rural Sociology 65, pp 72–86.
Findlay, A. & Rogerson, R. (1993) Migration, places and quality of life: Voting with their
feet. Champion, T. (ed) Population Matters, The Local Dimension, pp 34–49.
Findlay, A. & White, P. (1986) The course of future change in European population.
Findlay, A. & White, P. (eds) West European Population Change, Croom Helm, pp
233–245.
Fischer, P.A. (1998) Exclusion, migration and Baltic regional development in an integrating Europe. Söderling, I. (ed) A Changing Pattern of Migration in Finland and its
Surroundings, Publications of the Population Research Institute, series D 32, the Population Research Institute, pp 51–72.
207
Fisher, T. (2003) Differentiation of growth processes in the peri-urban region: An Australian case study. Urban Studies 40 (3), pp 551–565.
Ford, T. (1999) Understanding population growth in the peri-urban region. International
Journal of Population Geography 5, pp 297–311.
Ford, T. & Champion, T. (2000) Who moves into, out of and within London? An analysis
based on the 1991 census 2% sample of anonymised records. Area 32 (3), pp 259–
270.
Forrest, R. & Murie, A. (1990) Moving strategies among home owners. Johnson, J.H. &
Salt, J. (eds) Labour migration: The internal geographical mobility in the developed
world, David Fulton Publishers, pp 191–209.
Fortheringham, A.S. & Curtis, A. (1999) Regularities in spatial information processing:
implications for modelling destination choice. Professional Geographer 51 (2), pp
227–239.
Foss, O. (1984) Life-spam perspective on migration, Scandinavian Population Studies 6,
1, Springfeldt & Warnström (eds) The Sixth Scandinavian Demographic Symposium
1982, Studies in Migration, pp 51–60.
Fredericks, A.L. & Dossett, D.L. (1983) Attitude-behaviour relations: A comparison of
the Fishbein-Ajzen and the Bentler-Speckart models. Journal of Personality and Social Psychology 45, pp 501–512.
Frey, W.H. (1987) Migration and depopulation of the metropolis: regional restructuring
or rural renaissance? American Sociological Review 52, pp 240–257.
Frey, W.H. (1988a) Migration and metropolitan decline in developed countries: A comparative study. Population and Development Review 14 (4), pp 595–628.
Frey, W.H. (1988b) The re-emergence of core region growth: a return to the metropolis?
International Regional Science Review 11, pp 261–268.
Frey, W.H. (1989) United States: counterurbanization and metropolis depopulation.
Champion, A.G. (ed) Counterurbanization: The Changing Pace and Nature of Population Deconcentration, Edward Arnold, pp 34–61.
Frey, W.H. (1995) The new geography of population shifts: Trends towards balkanization.
Farley, R. (ed) State of the Union: America in the 1990s, Russell Sage Foundation,
New York, pp 271–336.
Frey, W.H. & Speare, A. (1992) The revival of metropolitan population growth in the
United States: An assessment of findings from the 1990 census. Population and Development Review 18 (1), pp 129–146.
Frost, M.E. & Spence, N.A. (1981) Policy responses to urban and regional economic
change in Britain. The Geographical Journal 147 (3), pp 321–349.
Fuguitt, G. (1991) Commuting and the rural-urban hierarchy. Journal of Rural Studies 7,
pp 459–466.
Fuguitt, G.V. & Beale, C.L. (1996) Recent trends in nonmetropolitan migration: Towards
a new turnaround? Growth and Change 27, pp 156–174.
Galbraith, J.K. (1992) The Culture of Contentment. Sinclair-Stevenson, London.
Gallaway L.E. (1969). Age and labour mobility patterns. Southern Economic Journal 36,
pp 171–180.
Gang, I.N. & Stuart, R.C. (1999) Mobility where mobility is illegal: Internal migration
and city growth in the Soviet Union. Journal of Population Economics 12 (1), pp 117–
134.
Geyer, H.S. & Kontuly, T.M. (1993) A theoretical foundation for the concept of differential urbanisation. International Regional Science Review 15 (12), pp 157–177.
208
Gibbs, J. (1963) The evolution of population concentration. Economic Geography 39
(2), pp 119–129.
Glaser, K. & Grundy, E. (1998) Migration and household change in the population aged
65 and over, 1971–1991. International Journal of Population Geography 4, pp 323–
339.
Gleave, D. & Cordey-Hayes, M. (1977) Migration dynamics and labour market turnover.
Progress in Planning 8 (1), Pergamon Press, Oxford.
Glennerster, H., Lupon, R. Noden, P. & Power, A. (1999) Poverty, social exclusion and
neighbourhood. CASE Paper 22, London School of Economics and Political Science.
Gold, J.R. (1980) An Introduction to Behavioural Geography. Oxford University Press.
Goldstein, A., Zhigang, G. & Goldstein, S. (1997) The relationship of migration to changing household headship patters in China, 1982–1987. Population Studies 51, pp 75–
84.
Goldthorpe, J.H. (1998) Rational action theory for sociology. The British Journal of Sociology 49 (2), pp 167–192.
Golledge, R.G. & Stimson, R.J. (1990) Analytical Behavioural Geography, Routledge.
Golledge, R.G. & Stimson, R.J. (1997) Spatial behaviour: A geographical perspective,
behaviour in space versus spatial behaviour: A geographic perspective. The Guilford
Press, pp 4–7.
Gonzalez, J.R. & Puebla, J.M.A. (1996) Spain: Return to the South, metropolitan
deconsentration and new migration flows. Rees, P., Stillwell, J., Convey, A. &
Kupiszewski, M. (eds) Population Migration in the European Union, John Wiley &
Sons, pp 175–189.
Goodin, R.E., Headey, B., Muffels, R. & Dirven, H.-J. (1999) The Real World of Welfare
Capitalism. Cambridge University Press.
Gordijn, H. & Eichperger, L. (1996) The Netherlands: From interegional to international
migration. Rees, P., Stillwell, J., Convey, A. & Kupiszewski, M. (eds) Population Migration in the European Union, John Wiley & Sons, pp 247–259.
Gordon, I.R. (1985) The cyclical interaction between regional migration, employment
and unemployment: a time series analysis for Scotland. Scottish Journal of Political
Economy 32, pp 135–158.
Gordon, P. (1978) Deconcentration without a “Clean Break”. International Institute for
Applied Systems Analyses RM-78–39.
Gottlieb, P.D. & Lentnek, B. (2001) Spatial mismatch is not always a Central-city problem: An analysis of commuting behaviour in Cleveland, Ohio, and its suburbs. Urban
Studies 38 (7), pp 1161–1186.
Goudy, W.J. (1982) Further consideration of indicators of community attachment. Social Indicators Research 11, pp 181–192.
Green, A.E. (1995) The geography of dual career households: A research agenda and
selected evidence from secondary data sources for Britain. International Journal of
Population Geography 1, pp 29–50.
Green, A.E., Owen, D.W., Champion, A.D., Boddard, J.B. & Coombes, M.G. (1986)
What contribution can labour migration make to reducing unemployment. Hart, R.
(ed) Unemployment and Labour Market Policies, Gower, pp 52–74.
Green, M.B. & Meyer, S.P. (1997) An overview of commuting in Canada with special
emphasis on rural commuting and employment. Journal of Rural Studies 13 (2), pp
163–175.
Greenwood, M. (1975) Research on internal migration in the United States: A survey.
Journal of Economic Literature 13, pp 397–433.
209
Greenwood, M. (1985) Human migration: Theory, models, and empirical studies. Journal of Regional Science 25, pp 521–544.
Groenewold, N. (1997) Does migration equalise regional unemployment rates? Evidence
from Australia. Papers on Regional Science: The Journal of the RSAI 76 (1), pp 1–20.
Grundy, E.M.D. & Fox, A.J. (1985) Migration during early married life. European Journal of Population 1, pp 237–263.
Halfacree, K.H. (1995) Household migration and the structuration of patriarchy: evidence from the USA. Progress in Human Geography 19 (2), pp 159–182.
Hall, P. & Hay, D. (1980) Growth Centres in the European Urban Systems. Heineman
Educational Books, London.
Halliday, J. & Coombest, M. (1995) In search of counterurbanisation: Some evidence
from Devon on the relationship between patterns of migration and motivation. Journal of Rural Studies 11 (4), pp 433–446.
Hansen, N.M. (1972) The case of government assisted migration. Mazie, S.M. (ed) Population distribution policy, The commission on population growth and the American
future, Government Printing Office, pp 679–695.
Hansen, N.M. (1981) A review and evaluation of attempts to direct migrants to smaller
and intermediate- size cities. Population distribution policies in development planning, Papers of the Un/UNFPA workshop on population distribution policies in development planning Bangkok, 4–13 Sept 1979, pp 113–127.
Hart, J. (1973) Economic expectations and the decision to migrate: an analysis by socioeconomic group. Regional Studies 7, pp 271–285.
Hartshorn, T.A. (1992). Interpreting the City: An Urban Geography. 2nd edition, John
Wiley & Sons.
Harvey, D. (1973) Social Justice and the City. Edward Arnold, London.
Hechter, M. & Kanazawa, S. (1997) Sociological rational choice theory. Annual Reviews
of Sociology 23, pp 191–214.
Herm, A. (1996) Internal migration. 1994 Population of Estonia II. Estonian Statistical
Office, p 13.
Herm, A. (1999) Ränne, Rahvastik 1988. Demograafilised üldandmed, Tallinn: ESA, pp
31–34.
Herzog H.W. & Schlottmann A.M. (1984) Labor force mobility in the United States:
migration, unemployment and remigration. International Regional Science Review 9,
pp 43–58.
Herzog, H.W., Schlottmann, A.M. & Boem, T.P. (1993) Migration as Spatial Job Search:
A Survey of Empirical Findings. Regional Studies 27 (4), pp 327–340.
Heywood, A. (2000) Key Concepts in Politics. Maxmillan Press LTD.
Hietala, K. (1973) The economic causes and consequences of emigration. Migration Research in Scandinavia. Proceedings of the Nordic seminar on migration research, Helsinki.
Hogarth, T. & Reder (eds; 1986) Rational Choice, Chicago, University of Chicago Press.
Hollis, M. & Nell, E.J. (1975) Rational economic Man. A philosophical critique of neoclassical economics, Cambridge University Press, London.
Holloway, S.L. (1998) Geographies of justice: Preshool- childcare provision and the
conceptualisation of social justice. Environment and planning C: Government and
Policy 16, pp 85–104.
Howard, J.A. (1994) A social cognitive conception of social structure. Social Psychology
Quarterly 57 (3), pp 210–227.
210
Howard, J. & Kerst, S. (1981) Memory and perception of cartographic information for
familiar and unfamiliar environments. Human Factors 23, pp 495–504.
Hägerstrand, T. (1957) Migration and Area. Lund.
Illés, S. (2000) Changing levels of spatial mobility in Hungary, Kovacs, Z. (ed) Hungary
towards 21th century: the human geography of transition. Studies in Geography in
Hungary 31, Geographical Research Institute, HAS, Budapest, pp 137–149.
Illeris, S. (1980) Research on changes in the structure of urban network. A.K.F. Copenhagen.
Illeris, S. (1996) Changing pattern of net migration in Denmark: An explanatory analysis.
Rees, P., Stillwell, J., Convey, A. & Kupiszewski, M. (eds) Population Migration in the
European Union, John Wiley & Sons, pp 105–122.
Ioffe (1987) – Èîôôå, Ã.Â. (1987) Óïðàâëåíèå ðàññåëåíèåì: âîñìîæíîñòè è
îãðàíè÷åíèÿ. Ëàïïî, Ã. Ì. & Øóïåð, Â.À. (eds) Ìåòîäû èçó÷åíèÿ ðàññåëåíèÿ,
Moskva, pp 181–196.
Jagijelski (1980) – ßãèåëñêèé, A. (1980) Ãåîãðàôèÿ íàñåëåíèÿ. Ìîñêâà Ïðîãðåññ.
Jargowsky, P. (1997) Poverty and Place, Russell Sage Foundation, New York.
Johnson, J.H. & Salt, J. (1992) Population migration. Johnson, J.H & Douglas, J. (eds)
Aspects of Geography, Nelson.
Johnson, J.H., Salt, J. & Wood, R. (1974) Housing and the Migration of Labour in England and Wales. Saxon House, Fareborough.
Johnston, G. & Percy-Smith, J. (2003) In search of social capital. Policy & Politics 31 (3),
pp 321–224.
Johnson, K.J. & Fuguitt, G.V. (2000) Continuity and change in rural migration patterns.
Rural Sociology 65 (1), pp 27–49.
Johnston, K. & Beale, C.L. (1994) The recent revival of widespread population growth in
nonmetropolitan areas of the United States. Rural Sociology 59, pp 655–667.
Johnston, R. (2000) Counterurbanization. Johnston, R.J., Gregory, D., Pratt, G. & Watts,
M. (eds) The dictionary of Human Geography. Fourth Edition, Blackwell, Oxford.
Jolkkonen, Arja., Jolkkonen, Ari, Soininen, T. (2003) Pitkät työmatkat ja työn
kuormittavuus. University of Joensuu.
Jolkkonen, A. & Kilpeläinen, R. (2002) Occupational and regional mobility of labour as a
means to find employment. Koistinen, P. & Sengenberger, W. (eds) Labour flexibility:
A factor of the economic and social performance of Finland in the 1990s, pp 101–
121.
Jones, H.R. (1990) Population Geography. Second Edition, Paul Chapman Publishing
Ltd. London.
Jones, R.C. (1978) Myth maps and migration in Venezuela. Economic Journal of Social
Issues 22, pp 92–102.
Käärik, E., Tiit, E.-M, Vähi, M. & Valton, A. (2001) Lastega perede majanduslik toimetulek
1999–2000, http://www.riik.ee/rahvastik/ 20.06 2003, Office of Minister of Population Affaires in Estonia.
Kaldaru, H. & Päll, A. (2003) Regionaalarengu ja -poliitika teooria. Kaldaru, H. (ed) Eesti
regionaalarengu sotsiaalmajanduslik käsitlus, Tartu University Press, pp 23–46.
Kangas, O. & Ritakallio, V.M. (1998) Different methods — different results? : Approaches
to multidimensional poverty. Andreß, H.-J. (ed) Empirical poverty research in a comparative perspective, Ashgate, pp 167–203.
Kangas, O. & Ritakallio, V.M. (1995) Different methods — different results? : Approaches
to multidimensional poverty. Helsinki, National Research and Development Centre
for Welfare and Health.
211
Karjahärm, T. (1992) Eesti sajandivahetusest kuni Esimese maailmasõja alguseni. Õispuu,
S. (comp) Eesti ajalugu ärkamisajast tänapäevani, Koolibri, Tallinn, pp 41–67.
Karjalainen, E. (1989) Migration and regional development in the rural communes of
Kainuu, Finland in 1980–85. Publicationes Instituti Geographici Universitatis
Ouluensis, Oulu.
Katus, K., Puur, A., Põldma, A., & Sakkeus, L. (1999) Ränne ja paiksus. Rahvastiku
vananemine Eestis. EKDK, Tallinn, pp 117–128.
Kauppinen, J., Rissanen, E., Naukkarinen, A., Muilu, T. & Colpaert, A. (1997) Migration
as a function of population. Nordia Geographical Publications 26 (1), pp 17–27.
Kawabe, H. & Liaw, K.-L. (1992) Marriage and migration in Japan: An explanation by
personal factors and ecological variables. NUPRI Research Paper Series No 60, p 51.
Keddie, P.D. & Joseph, A.E. (1991) The turnaround of the turnaround? Rural population
change in Canada 1976 to 1986. The Canadian Geographer 35, pp 367–379.
Keeble, D.E. (1989) The dynamics of European industrial conterurbanisation in the 1980s.
Geographical Journal 155, pp 70–74.
Kelley, A. & Williamson, J.G. (1984) Population Growth, Industrial Revolutions, and the
Urban Transition. Population and Development Review 10 (3), pp 419–441.
Kephart, G. (2003) Economic restructuring, population redistribution, and migration in
the United States. Gottdiender, M. & Pickvaner, C. G. (eds) Urban life in transition.
Urban Affaires Annual Reviews 39, pp 12–34.
Kikas, E. (2002) Mõtlemine. Chapter 8, Allik, J. & Rauk, M. (eds) Psühholoogia
gümnaasiumile, Tartu University Press, pp 129–147.
King, R., Shuttleworth, I. & Walsh, J. (1996) Ireland: The human resource warehouse
Europe. Rees, P., Stillwell, J., Convey, A. & Kupiszewski, M. (eds) Population Migration in the European Union, John Wiley & Sons, pp 207–229.
King, R., Warnes, A. & Williams, A.M. (1998) International retirement migration in Europe. International Journal of Population Geography 4, pp 91–111.
Kitching, R. (1990) Migration behaviour among the unemployed and low – skilled.
Johnson, J.H. & Salt, J. (eds) Labour migration: The internal geographical mobility in
the developed world, David Fulton Publishers, pp 172–190.
Kivilaid, M., Lehto, K., Maurer, K., Nurmsaar, R., Pikner, T., Ristkok, P., Servinski, M.,
Sõmer, H.-M. & Tischler, G. (2002) Eesti Regionaalareng 1995–2000. Estonian Statistical Office, Ministry of Internal Affaires.
Knox, P. & Pinch, S. (2000) Urban Social Geography, Fourth edition, Prentice Hall.
Koivukangas, O. (1986) Vousina 1981–85 Suomessa tapahtuneet muuttoliikeilmiöt ja
tutkimustilanne. Koivukangas, O., Narjus, R. & Söderling, I. (eds) Muutoliikesymposium 1985, Siirtolaistutkimuksia A 14, Siirtolaisuusinstituutti, pp 53–70.
Kok, H. (1999) Migration from the city to the countryside in Hungary and Poland.
GeoJournal 49 (1), pp 53–62.
Kontuly, T. & Vogelsand, R. (1989) Federal Republic of Germany; the intensification of
the migration turnround. Champion, A. (ed) Counterurbanization, Edward Arnold.
Korcelli, P. (1984) The turnaround of urbanization in developed countries. International
Conference of Population 1984, Population Distribution, Migration and Development.
Prodaction of the Expert Group; UN New York 1984, pp 349–378.
Kõre, J. & Hendrikson, M. (2000) Elamuehitus. Purju, A. (ed) Elamistingimused, Estonian Statistical Office, pp 38– 41.
Kuddo, A. (1988) Rahvastikupoliitika põhisuunad maal. Raig, I. (comp) Maaelu. Sotsiaalse
arengu probleeme Eesti külas, Tallinn: Valgus, pp 60–68.
212
Kümmel, T. (1986) Proportsioonid Eesti linnade arengus. Hallemaa, H. (ed) Geograafia
rakendusprobleeme Eesti NSV-s, pp 123–128.
Kümmel (1986) – Kþììåëú, Ò. (1987) Ñòàäèàëíàÿ êîíöåïöèÿ óðáàíèçàöèè:
ìåòîäîëîãèÿ è ìåòîäû àíàëèçà. Ëàïïî, Ã. Ì. & Øóïåð, Â. À. (eds) Ìåòîäû èçó÷åíèÿ
ðàññåëåíèÿ, Moskva, pp 82–100.
Kunzmann, K. (1998) Planning for spatial equity in Europe. International Planning Studies 3 (1), pp 101–120.
Kupiszewski, M., Durham, H. & Rees, P. (1998) Internal migration and urban change in
Poland. European Journal of Population 14 (3), pp 265–290.
Kurs, O. & Toots, A. (comp) (2000) Ann Marksoo 70: scientific and didactic activities.
Publicationes Instituti Geographici Universitatis Tartuensis 88, Tartu University Press.
Kutsar, D. & Trumm, A. (2003) Provision of social assistance. The case of the two municipalities in Estonia. Project Social Policy and Social Exclusion in the Baltic countries. Fafo Paper 1.
Laakso, S. (1998) Alueiden välinen muutto Suomessa. City of Helsinki Urban Facts,
Research Series 4.
Laas, K. (1978) Rahvastikuprotsessidest Eestis. Tallinn.
Laas, K. (1987) Rahvastiku ränded Eestis. Eesti Geograafia Seltis Aastaraamat, 22, Valgus, Tallinn, pp 81–97.
Ladinsky, J. (1967) The geographical mobility of professional and technical manpower.
Journal of Human Recourses 2, pp 475–494.
Lanaspa, L., Pueyo, F. & Sanz, F. (2003) The evolution of Spanish Urban structures during the twentieth century. Urban Studies 40 (3), pp 567–580.
Law, C.M. & Warnes, A.M. (1976) The changing geography of the elderly in England and
Wales. Transactions, New Series 1 (4).
Law, C. M. & Warnes, A.M. (1982) The destination decision in retirement migration.
Warnes, A.M. (ed) Geographical perspectives on the elderly, Chichester, John Wiley
& sons Ltd, Chp 3, 53–81, In Robinson, V. (ed) Geography of Migration, pp 238–266.
Lawton, M.P. (1983) The varieties of well-being. Experimental Aging Research 9, pp 65–
72.
Lawton, M.P, Kleban, M.H. & diCarlo, E. (1983) Psychological well-being in the aged.
Experimental Aging Research 6, pp 67–97.
Lazarus, R.S. (1991) Emotions and Adaption. Oxford University Press, New York.
Lazarus, R..S. & Folkman, S. (1984) Stress, Appraisal, and Coping. Springer, New York.
LeClere, F.B. & McLaughlin, D.K. (1997) Family migration and changes in women’s
earnings: A decomposition analysis. Population Research and Policy Review 16, pp
315–335.
Lee, E.S. (1969) A theory of migration. Jackson, J.A. (ed) Migration, Cambridge University Press, pp 282–297.
LeGrand, J. (1982) The Strategy of Equality. George Allen and Unwin, London.
Letwin, W. (ed.) (1983) Against Equality. London, Macmillan.
Levin, K. (1951) Intensions, will, and need. Rapaport, D. (ed) Organization and Pathology of Human Thought, Columbia University Press, New York, pp 95–153.
Leyens, J.-P. & Dardenne, B. (1996) Basic concepts and approaches in social cognition.
Hewstone, M., Stroebe, W. & Stephenson, G.M. (eds) Introduction to Social Psychology, Blackwell Publishers, pp 109–134.
Liao, P-S. (2001) Contextual Analysis of rural migration intention: A comparison of Taiwanese and Pennsylvania data. International Journal of Comparative Sociology 42
(5), pp 435–460.
213
Lin, G. (1999) Assessing changes in interstate migration patterns of the United States
elderly population, 1965–1990. International Journal of Population Geography 5 (6),
pp 411–424.
Lin, N. (2001) Social Capital: A Theory of Social Structure and Action. Cambridge University Press.
Linnemann P. & Graves, P.E. (1983) Migration and job change: a multinominal logit approach. Journal of Urban Economy 14, pp 263–279.
Liska, A.E. (1984) A critical examination of the causal structure of the Fishbein-Ajzen
model. Social Psychology Quarterly 47, pp 61–74.
Lloyd, P.E. & Dicken, P. (1972) Location in space: Theoretical approach to economic
geography. Harper & Row, London & New York.
Long, L. (1988a) Migration and residental mobility in the United States. Journal of the
American Statistical Association 68, pp 37–43.
Long, L. (1988b) Migration and Residential Mobility in the United States. Russell, Sage.
Long, L. & Nucci, A. (1997) The Hoover Index of Population Concentration: A correlation and update. Professional Geographer 49 (4), pp 431–440.
Lõo, A. (1987) Majandite tööjõu olukorra ja seda mõjutavate tegurite uurimisest
ekspertiisküsitluse vahendusel. Manuscript. TAMI, Tallinn.
Lopez, R. & Hynes, H.P. (2003) Sprawl in the 1990s, measurement, distribution and
trends. Urban Affaires Review 38 (3), pp 325–355.
Lösch, A. (1954) The economics of location, Translated from the 2. rev. ed. by William
H. Woglom, New Haven, Yale University Press, 1959.
Louviere, J.J. (1984) Hierarchical information integration: A new method for the design
and analysis of complex multiattribute judgement problems. Advances in Consumer
Research 11, pp 148–155.
Loyd, P.E. & Dicken, P. (1972) Location in space: Theoretical approach to economic
geography, Harper & Row, New York.
Luuk, A. & Tulviste, P. (2002) Motiivid ja vajadused. Chapter 10, Allik, J. & Rauk, M.
(eds) Psühholoogia gümnaasiumile, Tartu University Press, pp 162–176.
Lyons, M. & Simister, J. (2000) From rags to riches? Migration and interregional change
in London’s housing market, 1971–91. Area 32 (3), pp 271–285.
Mabogunje, A.L. (1970) System approach to a theory of rural-urban migration. Geographical Analyses 2, pp 1–18.
Mabogunje, A.L. (1981) Objectives and rationales for regional population redistribution
in developing countries. Population distribution policies in development planning,
Papers of the Un/UNFPA workshop on population distribution policies in development planning Bangkok, 4–13 Sept 1979, pp 19–29.
MacCrimmon, K.R. & Taylor, R.N. (1976) Decision making and problem solving,
Dunnette, M.D. (ed) Handbook of Industrial and Organizational Psychology, Rand
McNally, Chicago.
Maier, G. (1990) The economics of information in the context of migration. Johnson,
J.H & Salt, J. (eds) Labour migration: The internal geographical mobility in the developed world, David Fulton Publishers, pp 115–136.
Mangalam, J.J. & Morgan, C. (1968) Human Migration. A Guide to Migration Literature
in English 1955–1962. University of Kentucky Press.
Mann, M. (1973) Workers on the move. Cambridge University Press.
Mansfield, Y. (1992) Tourism: towards a behavioural approach. Progress in Planning 38,
pp 1–92.
214
Margolis, J. (1977) Internal migration: measurement and models. Brown, A. & Neuberger,
E. (eds) Internal migration, a comparative perspective, Academic Press, New York,
pp 135–144.
Marksoo (1986) – Màðêñîî, À. (1986) Âçàèìîçâÿçàííîñò òðóäîâîé ìàÿòíèêîâîé è
áåçâîçâðàòíîé ìèãðàöèè íàñåëåíèÿ. TRÜ Toimetised 726, pp 64–84.
Marksoo, A. (1987) Naiste rände seaduspärasusi Eestis. Eesti Geograafia Seltsi
Aastaraamat 22, Valgus, Tallinn, pp 98–112.
Marksoo (1974) – Màðêñîî, À. (1974) Ïðè÷èííàÿ ñòðóêòóðà ìèãðàöèîííûõ ïîòîêîâ â
ñèñòåìå ðàññåëåíèÿ Ýñòîíñêîé ÑÑÐ. TRÜ toimetised. Geograafia-alaseid töid XII,
431–C, pp 32–45.
Marksoo, A. (1988) Rände suunamisest. Raig, I. (comp) Maaelu. Sotsiaalse arengu
probleeme Eesti külas. Valgus, Tallinn, pp 68–76.
Marksoo, Ü. (1989) Eesti asustussüsteemi seaduspärasustest rahvastiku dünaamika näitel.
Manuscript.
Martin, R. (2001) Spatial mismatch and costly suburban communes: Can commuting
subsidies help? Urban Studies, 38, 8, pp 1305–1318.
Martin, R. & Sunley, P. (1998) Slow convergence? The new endogenous growth theory
and regional development. Economic Geography 74, pp 201–227.
Maslow, A.H. (1954) Motivation and Personality. New York. Harper & Row.
McAuley, W. & Nutty, C.L. (1985) Residential satisfaction, community integration, and
risk across the family life cycle. Journal of Marriage and Family, pp 125–130.
McCarthy, K. (1976) The household lifecycle and housing choices. The Rand Corporation, pp 55–65.
McNamara, T. (1986) Mental representation in spatial relations. Cognitive Psychology
18, pp 87–121.
McNamara, T., Hardy, J. & Hirtle, S. (1989) Subjective hierarchies in spatial memory.
Journal of Experimental Psychology: Learning, Memory and Cognition 15, pp 271–
287.
Mellers, B.A., Schwartz, A. & Cooke, A.D.J. (1998) Judgement and decision. Annual
Reviews in Psychology 49, pp 447–477.
Mera, K. (1981) Population distribution policies: the need for caution. Population distribution policies in development planning, Papers of the Un/UNFPA workshop on population distribution policies in development planning Bankok, 4–13 Sept 1979, pp 33–
41.
Merrifield, A. & Swyngedouw, E. (1996) The urbanization of injustice. Lawrence &
Wishart, p 245.
Miller, D. (1976) Social Justice. Clarendon Press, London.
Miller, D. (1993) Equality and Inequality. Outhwaite, W. & Bottomore, T. (eds) The
Blackwell Dictionary of Twentieth — Century Social Thought. Blackwell, pp 200–
2002.
Miller, D.W. & Starr, M.K. (1967) The Structure of Human Decisions. Engelwood Cliffs,
Prentice Hall, New York.
Millington, J. (2000) Migration and age: The effect of age on sensitivity to migration
stimuli. Regional Studies 34 (6), pp 521–533.
Mincer, J. (1978) Family migration decisions. Journal of Political Economy 86, 749–73.
Mintzberg, H, Raisinghani, D. & Thoret, A. (1976) The structure of unstructured decisions. Administrative Science Quarterly 21, pp 246–275.
Mitchelson, R.L. & Fisher, J.S. (1987) Long-distance commuting and population change
in New York State. Urban Geography 8 (3), pp 133–211.
215
Molho, I. (1986) Theories of migration: a review. Scottish Journal of Political Economy
33, pp 396–419.
Moissejenko & Mukomel (1982) — Moèñååíêî, Â.Ì. & Ìóêîìåëú. Â.È. (1982)
Èíòåíñèâíîñò ìèãðàöèîííûõ ïðîöåññîâ. Óðáàíèèçàöèÿ è äåìîãðàôè÷åñêèå
ïðîöåññû, Moskva, pp 129–144.
Moon, B. (1995) Paradigms in migration research: exploring “moorings” as a schema.
Progress in Human Geography 19 (4), pp 504–524.
Mooney, G. (1994) Key Issues in Health Economics. Hemel Hempstead, Harvester
Wheatsheaf.
Morello-Frosch, R., Pastor, M. & Sadd, J. (2001) Environmental justice and Southern
California’s “riskscape”: the distribution of air toxic exposures and health risks among
diverse communities. Urban Affairs Review 36 (4), pp 551–578.
Morgan J.N. & Robb E.H. (1981) The impact of age upon interregional migration. Annales
Regional Science 15, pp 31–45.
Morrill, R.L. (1970) The Spatial Organization of Society. Wadsworth Publishing Company.
Morrison, P. (1971) Chronic movers and the future redistribution of population: a longitudinal analysis. Demography 8 (2), pp171–184.
Morrison, P. (1972) Movements and the shape of urban growth: Implication for public
policy. Mazie, S.M. (ed) Population distribution policy, The commission on population growth and the American future, Government Printing Office, pp 280–322.
Mueser, P.R, White, M.J. & Tierney, J.P. (1988) Pattern of net migration by age for US
counties 1950–1980: The impact of increasing spatial differentiation by life cycle.
Canadian Journal of Regional Science 11, pp 57–76.
Mulder, C.H. (1991) Age- period-cohort models of short and long distance migration in
The Netherlands, PDOD-paper no 6, Postdoctorale Onderzoekersopleiding
Demografie, University of Amsterdam.
Mulder, C.H. (1993) Migration dynamics: A life course approach, Thesis Publishers,
Amsterdam.
Mulder, C.H. & Hooimeijer, P. (1999) Residential relocations in the life cource. Van
Wissen, L.J.G. & Dykstra, P.A. (eds) Population Issues: An Interdisciplinary Focus,
Kluwer Academic Press.
Mulder, C.H. & Manting, D. (1993) Strategies of nest-leavers: “Settling down” versus
flexibility. PDOD Paper No 15.
Mulder, C.H. & Wagner, M. (1993) Migration and marriage in the life course: a method
for studying synchronized events. European Journal of Population 9, pp 55–76.
Musil, I. (1993) Changing urban systems in post-communist societies in Central Europe:
Analysis and predictions. Urban Studies 30 (6), pp 899–905.
Myers, D., Choi, S. & Woo L.S. (1997) Constraints of housing age and migration on
residential mobility. The Professional Geographer 49 (1), pp 14–27.
Myhren, G. (1998) Planleggingen av det gode stedet. Et eksempel fra Landås. Plan 98
(4), pp 39–43.
Myrdal, G. (1957) Economic theory and underdeveloped regions, Methuen, London.
Natosteen, R.A. & Zimmer, M.A. (1982) The effects on earnings in interregional and
interindustrial migration. Journal of Regional Science 22, pp 325–341.
Naukkarinen, A. (1969) Population Development in Northern Finland 1950–1965. Oulu.
Newbold, B.K. (1998) Outmigration from California: The role of migrant selectivity.
Geographical Analysis 30 (2), pp 138–152.
216
Nieminen, M. (1983) Internal Migration in Finland in 1977–78. A Survey Investigating
the Reasons for Migration. Yearbook of Population Research in Finland 1983 XXI.
The Population Research Institute. Helsinki.
Nijkamp, P., Wissen, L.V. & Rima, A. (1993) A household life cycle model for residential
relocation behaviour. Socio-Economic Planning 27, pp 35–53.
Ni Laoire, C. (2000) Conceptualising Irish rural youth migration: A biographical approach.
International Journal of Population Geography 6, pp 229–243.
Nilsson, K. (2001) Migration among young men and women in Sweden. Doctoral thesis,
Department of Social and Economic Geography, Umeå University, GERUM.
Nisbett, R.E. & Wilson, T.D. (1977) Telling more than we can know: verbal reports on
mental processes. Psychology Review 84, pp 231–259.
Nordvik, V. (2001) Moving costs and the dynamics of housing demand. Urban Studies 38
(3), pp 519–523.
Nucci, A. & Long, L. (1995) Spatial and demographic dynamics of metropolitan territory
in the United States. International Journal of Population Geography 1, pp 165–81.
Ogden, P.E. & Hall, R. (2000) Households, reurbanisation and the rise of living alone in
the principal French cities, 1975–1990. Urban Studies 37 (2), pp 367–390.
Oh, J.-H. (2003) Social bonds and the migration intentions of elderly urban residents:
The mediating effect of residential satisfaction. Population Research and Policy Review 22, pp 127–146.
O’Higgins, M. (1987) Egalitarians, equalities and welfare evaluation. Journal of Social
Policy 16 (1), pp 1–18.
Oja, U. & Tiit, E.-M. (2002) Recalculating expansion factors of the Living Condition
Survey samples based on the data of the 2000 census. Kutsar, D. (ed) Living conditions in Estonia five years later, Tartu University Press, pp 164–176.
Olsson, G. (1965) Distance and human interaction a review and bibliography. Bibliography series number two. Regional Science Research Institute.
Pacione, M. (1984) Rural geography. Quality of Life, Harper & Row, Publishers, pp
199–217.
Page, E. (1982) The value of local autonomy. Local Government Studies 8 (4), pp 21–42.
Pagozzi, R.P. & Yi, Y. (1989) The degree of intention formation as a moderator of attitude-behaviour relationship. Social Psychology Quarterly 52 (4), pp 266–279.
Pandit, K. (1997) Demographic cycle effects on migration timing and the delayed mobility phenomenon. Geographical Analysis 29 (3), pp 187–199.
Parr, J. (1987) Interaction in an urban systems: Aspects of trade and commuting. Economic Geography 63, pp 223–240.
Parysek, J.J. & Kotus, J. (1995) Development patterns of Polish towns in the years 1950–
1990. Geographia Polonica 66, pp 87–110.
Pavlik, Z. & Kucera, M. (2002) Population development in the Czech Republic 1990–
2002. Department of Demography and Geodemography, Charles University in Prague.
Peltonen, T. (1982) Terveellinen asumisympäristö. Koivukangas, O., Lindström, K., &
Narjus, R. (eds) Muutoliikesymposium 1980, Siirtolaistutkimuksia A 8,
Siirtolaisuusinstituutti, pp 311–318.
Pekkala, S. (1998) Aggregate economic fluctuations and regional convergence: The Finnish case 1988–95. Working Paper no 185 University of Jyväskylä School of Business
and Economics.
Pekkala, S. & Kangasharju, A. (1998) Migration and regional convergence among the
Finnish subregions 1975–95. Pellervo Economic Research Institute Working Papers
no 12.
217
Pekkala, S. & Kangasharju, A. (2000) Regional Labour Market Adjustment in Finland,
Working Paper 218, University of Jyväskylä School of Business and Economics.
Pettai, Ü. & Sõstra, K. (1997 comp) Estonian Labour Force 1989–1997. Statistical Office of Estonia.
Pissarides, C.A & McMaster, I. (1989) Unemployment and the inter-regional mobility of
labour. Economic Journal 99, pp 739–755.
Plane, D.A. (1993) Demographic Influences on Migration. Regional Studies 27 (4), pp
375–383.
Plane D.A. & Rogerson P.A. (1991) Tracking the baby boom, the baby bust, and the echo
generations: how age composition regulates US migration. Professional Geographer
43, pp 416–430.
Plant, R.L. & Taylor-Gooby, P. (1980) Political Philosophy and Social Welfare. Routledge
and Kegan Paul, London.
Poulain, M. (1996) Migration in Belgium: Temporal trends and counterurbanization. Rees,
P., Stillwell, J., Convey, A. & Kupiszewski, M. (eds) Population Migration in the European Union, John Wiley & Sons, pp 91–104.
Powell, M. & Boyne, G. (2001) The spatial strategy of equality and the spatial division of
welfare. Social Policy & Administration 35 (2), pp 181–194.
Powell, M., Boyne, G. & Ashworth, R. (2001) Towards a geography of people poverty
and place poverty. Policy & Politics 29 (3), pp 243–258.
Pragi, U. (1988) Population distribution and redistribution in rural settlements of the
Estonian SSR. Punning, J-M. (ed) Geographical Researchers, Tallinn, pp 101–109.
Pred, A. (1967) Behaviour and location: Foundations for a geographic and dynamic location theory. Part 1. Lund Studies in Geography. Glweerup.
Pullat, R. (1978) Linnad kodanlikus Eestis. Eesti Raamat, Tallinn.
Putnam, R. (1993) The prosperous community: social capital and public life. American
Prospect 13, pp 35–42.
Quadrado, L., Heijman, W. & Folmer, H. (2001) Multidimensional analysis of regional
inequality; The case of Hungary. Social Indicators Research 56, pp 21–42.
Rannikko, P. (1986) “Tietoyhteiskunnan” muutoliike: omien arvojen toteuttamista vai
taloudellisiin pakkoihin sopetumista? Työryhmäalustus. Koivukangas, O., Narjus, R. &
Söderling, I. (eds) Muutoliikesymposium 1985, Siirtolaistutkimuksia A 14,
Siirtolaisuusinstituutti, pp 119–130.
Raunio, K. (1983) Hyvinvointi ja taloudelliset muutokset: Tutkimus suomalaisen
yhteiskunnan hyvinvoinnin sisällöstä ja taloudellisten muutosten hyvinvointivaikutuksista vuosina 1950–1977. Sosiaalipoliittinen Yhdistys, Turku.
Ravenstein, E.G. (1885) The laws of migration. Journal of Statistical Society 48, pp 167–
227.
Rawls, J. (1976) A Theory of Justice. Oxford University Press.
Rayer, S. & Brown, D.L. (2001) Geographic diversity of inter-county migration in the
United States, 1980–1995. Population Research and Policy Review 20, pp 229–252.
Rees, P. & Kupiszewski, M. (1999) Internal Migration and Regional population dynamics
in Europe: a synthesis. Population Studies 32, EU Directorate of Social and Economic
Affairs, Council of Europe Publishing.
Regan, D. & Stewart, J. (1982) An essay in the government of health: the case for local
authority control. Social Policy and Administration 16 (1), pp 19–43.
Reggiani, A. (1991) Spatial interaction models: New directions, PhD Dissertation. Free
University, Amsterdam.
218
Reijo, M. & Valkonen, T. (1993) Population development and the changes in the economic structure of Finnish urban areas. Yearbook of Population Research in Finland
31, pp 25–46.
Reiman, H. (1927) Rahvaliikumine üle piiri ja sisemaal 1925–1926. Eesti Statistika 68
(7), pp 384–389.
Richardson, H. (1973) Regional Growth Theory, Macmillian, Lodon, Basingstoke.
Richardson, H. (1980) Polarization reversal in developing countries. Papers of the Regional Science Association 45, pp 67–85.
Ringen, S. (1995) Well-being, measurement, and preferences. Acta Sociologica 38 (2),
pp 3–15.
Rogers, A. (1992) Elderly migration and population redistribution in the United States.
Rodgers, A. (ed) Elderly migration and population redistribution, Belhaven Press, London, pp 226–248.
Rogers, A. & Rajbhandary, S. (1997) Period and cohort age patterns of US migration,
1948–1993: Are American males migrating less? Population Research and Policy Review 16, pp 513–539.
Rodgers, A. & Sweeney, S. (1998) Measuring the spatial focus of migration patterns.
Professional Geographer 50 (2), pp 232–242.
Rogerson, P.A (1987) Changes in US National mobility levels. The Professional Geographer 39, pp 344–351.
Rook, K.S. (1984) The negative side of social interaction: Impact on psychological wellbeing. Journal of Personality and Social Psychology 46, pp 1097–1108.
Roseman, C. (1971) Migration as a spatial and temporal process. Annales of the Association of American Geographers 61 (3), pp 589–598.
Rossi, P.H. (1980) Why Families Move. Sage, London, Beverly Hills.
Runciman, W. (1966) Relative Deprivation and Social Justice. Routledge and Kegan Paul,
London.
Rybakovsky (1987) — Pûáàêîâñêèé, Ë.Ë. (1987) Ìèãðàöèÿ íàñåëåíèÿ: ïðîãíîçû,
ôàêòîðû, ïîëèòèêà, Moskva.
Saar, E. (2003) Sotsiaalne ebavõrdus Eestis: tegelikkus ja inimeste hinnang. Vetik, R. (ed)
Eesti inimarengu aruanne 2003, Tallinna Pedagoogikaülikool, Rahvusvaheliste ja
Sotsiaaluuringute Instituut, Tallinn, pp 52–57.
Sabini, J. (1992) Social Psychology. W.W. Norton and Company, New York, London.
Sandefur, G. & Scott, W. (1981) A dynamic analysis of migration: an assessment of the
effects of age, family and career variables. Demography 18 (3), pp 355–368.
Sanderfur, R.L. & Laumann, E. (1998) A paradigm for social capital. Rationality and
Society 10 (4), pp 481–501.
Schoemaker, P.J. (1982) The expected utility model: Its variants, purposes, evidence and
limitations. Journal of Economic Literature 30, pp 529–563.
Schofield, B. (1990) Individual and organisational dimensions in the migration of school
teachers. Johnson, J.H & Salt, J. (eds) Labour migration: The internal geographical
mobility in the developed world, David Fulton Publishers, pp 85–98.
Scheider, J. & Kasper, B. (2003) Lifestyles, shoice of housing location and daily mobility:
the lifestyle approach in te context of spatial mobility and planning. International
Social Science Journal 176, pp 319–332.
Schulze, R. Artis, J., & Beegle, J.A. (1963) The measurement of community satisfaction
and the decision to migrate. Rural Sociology 28, pp 279–283.
219
Sell, R.R. ( 1990) Migration and job transfers in the United States. Chapter 2, Johnson,
J. & Salt, J. (eds) Labour migration the internal geographical mobility of labour in the
developed world, David Fulton Publishers, pp17–31.
Sen, A. (1999) Development as Freedom. Alfred A. Knopf, New York.
Shumway, J.M. & Davis, J.A. (1996) Nonmetropolitan population change in the mountain West: 1970–1995. Rural Sociology 61 (3), pp 513–529.
Shuval, J.T. (1982) Migration and stress. Goldberger, L. & Bresznitz, S. (eds) Handbook
of stress: Theoretical and clinical aspects, Free Press, New York, pp 677–691.
Simmons, A.B. (1984) Migration and rural development: conceptual approach research
findings and policy issues, Population distribution, migration and Development. Proceedings of the Expert Group on Population Distribution, New York, pp 156–191.
Simon, H.A. (1957) Models of Social and Rational Man. John Wiley, New York.
Skinner, J.L., Rampa, H. & Springs, J. (1983) Push and pull factors influencing migration
to Queensland: 1961–1981. Migration in Australia: Proceeding of Symposium, 1–2
Dec. Royal Geographical Society of Australia and Australian Population Association,
Brisbane, pp 101–144.
Skrede, K. (1977) Yrkesmobilitet — strategi eller tilfeldighet? Tidsskrift for
samfunnsforskning No 5–6, Oslo.
Smailes, P.J. (1996) Demographic reasons to rural restructuring and counterurbanisation
in South Australia, 1981–1991. International Journal of Population Geography 2, pp
261–287.
Smith, D.M. (1977) Human Geography: A Welfare Approach. Edward Arnold, London.
Smith, D.M. (1998) How far should we care? On the spatial scope of beneficence.
Progress in Human Geography 22 (1), pp 15–38.
Smith, D.M. (2000) Moral progress in human geography: transcending the place of good
fortune. Progress in Human Geography 24 (1), pp 1–18.
Smith, G. (1999) Area-based initiatives: The rationale for and options for area tergeting.
CASE Paper 25, London School of Economics and Political Science
Smits, J. (2001) Career migration, self-selection and the earnings of married men and
women in the Netherlands 1981–93. Urban Studies 38 (3), pp 541–562.
Snaith, J.G. (1990) Migration and dual career households. Johnson, J.H. & Salt, J. (eds)
Labour migration: The internal geographical mobility in the developed world, David
Fulton Publishers, pp 155–171.
Söderling, I. (1983) Maassamuutto ja muuttovirrat. Vuosina 1977–78 kunnasta toiseen
muuttaneiden elinolosuhdetutkimus. Siirtolaisuustutkimuksia A 11, Migration Institute, Turku.
Söderling, I. (1984) Internal migration-migration flows. Scandinavian Population Studies 6, 1. Springfeldt & Warnström (eds) The Sixth Scandinavian Demographic Symposium 1982, Studies in Migration, pp 35–47.
Söderling, I. (1988) Massamuuton ulottuvuudet, Annales Universitatis Turkuensis C 65,
Turun Yliopisto.
Spatial-growth model (1997) A Dictionary of Geography. Susan Mayhew. Oxford University Press, Oxford Reference Online, Oxford University Press. http://
www.oxfordreference.com/views/ (17.06.2003).
Spatial-interaction theory (1997) A Dictionary of Geography. Susan Mayhew. Oxford
University Press, 1997. Oxford Reference Online, Oxford University Press. http://
www.oxfordreference.com/views/ (17.06.2003).
Speare, A. (1971) A cost-benefit model of rural to urban migration in Taiwan. Population
Studies 15, pp 117–130.
220
Stankuniene & Sipavitsiene (1989) – Ñòàíêóíèåíå, Â. & Ñèïàâè÷åíå, À. (1989)
Ïðîáëåìû óïðàâëåíèÿ ñåëúñêî-ãîðîäñêîé ìèãðàöåé íàñåëåíèÿ. Êàòóñ, Ê. (ed)
Ïîäõîäû ê óïðàâëåíèþ ìèãðàöèîííûì ðàçâèòèåì. Tallinn, pp 106–123.
Stark, O. (1988) On Marriage and Migration. European Journal of Population 23, pp 23–
37.
Stewart, K. (2003) Monitoring social inclusion in Europe’s regions. Journal of European
Social Policy 13 (4), pp 335–356.
Stillwell, J. (1991) Spatial interaction models and the propensity to migrate over distance, Chapter 3. Stillwell, J. & Congdon, P. (ed.) Migration models (Macro and Micro Approaches), Belhaven Press. London & New York.
Stillwell, J., Rees, P. & Duke-Williams, O. (1996) Migration between NUTS level 2 Regions in the United Kingdom. Rees, P., Stillwell, J., Convey, A. & Kupiszewski, M.
(eds) Population Migration in the European Union, John Wiley & Sons, pp 275–307.
Stinner, W.F. & van Loon, M. (1992) Community size preference status, community
satisfaction and migration intensions. Population and Environment 14, pp 177–195.
Stockdale, A., Findlay, A. & Short, D. (2000) The repopulation of rural Scotland and
threat. Journal of Rural Studies 16, pp 243–257.
Stouffer, S.A. (1940) Intervening opportunities: A theory relating mobility and distance.
American Sociological Review 5, pp 845–867.
Stretesky, P. & Hogan, M.J. (1998) Environmental justice: An analysis of superfund cities
in Florida. Social Problems 45 (2), pp 268–287.
Suburbanization (1998) A Dictionary of Sociology. Ed. Gordon Marshall. Oxford University Press, 1998. Oxford Reference Online. Oxford University Press. http://
www.oxfordreference.com/views/ (17.06.2003)
Swanson, L.E., Luloff, A.E. & Warland, R.H. (1979) Factors influencing willingness to
move: An examination of nonmetropolitan residents. Rural Sociology 44 (4), pp 719–
735.
Syrjänen, O. (1982) Aluerakenteen kehittäminen ja palveluken saavutettavuus.
Koivukangas, O., Lindström, K., & Narjus, R. (eds) Muutoliikesymposium 1980,
Siirtolaistutkimuksia A 8, Siirtolaisuusinstituutti, pp 215–219.
Tallman, I., Leik, R.K., Gray, L.N. & Stafford, M.C. (1993) A Theory of problem-solving
behaviour, Social Psychology Quarterly 56 (3), pp 157–177.
Tamm, E. (2001) 2000. aasta rahva ja eluruumide loenduse kaetuse hindamine
järelküsitluse põhjal. Estonian Statistics 10, pp 10–15.
Tamm, E. (2003) Quality estimates of the 2000 population and housing census. Estonian
Statistics 12, pp 144.
Tammaru, T. (2001a) Linnastumine ja linnade kasv Eestis nõukogude aastatel.
Dissertationes Geographicae Universitatis Tartuensis 13. Tartu University Press.
Tammaru, T. (2001b) Suburban growth and suburbanisation under central planning: The
case of Soviet Estonia. Urban Studies 38 (8), pp 1341–1357.
Tammaru, T. (2001c) Eesti linnade mõjualad ja pendelränne (www.riik.ee/rahvastik/)
6.12.2003, Office of Minister of Population Affaires in Estonia.
Tammaru, T., Kulu, H. & Kask, I. (2003) Siserände üldsuunad. Tammaru, T. & Kulu, H.
(eds) Ränne üleminekuaja Eestis, Statistical Office of Estonia, pp 5–27.
Tammaru, T. & Sjöberg, Ö. (1999) On the move: explaining migration patterns in Estonia
during the transition period. International Journal of Population Geography 5, pp
241–260.
Temple, M. (1994) Regional Economics, St. Martin Press, New York.
221
Tervamäki, E. (1987) Migration in Finland, a multi-level system of regions. Fennia 165
(1), pp 3–87.
Tervo, H. (1991) Studies on the economic case for and effects of regional policy. Jyväskylä
Studies in Computer Science, Economics and Statistics 17.
Tervo, H. (2001) Työvoiman liikkuvuus aluetaouksien näkökulmasta Suomessa. Heikkilä,
E. (ed) Muutoliikkeet vuosituhannen vaihtuessa – halutaanko niitä ohjata?,
Muutoliikesymposium, Siirtolaisuustutkimuksia A 24, pp 101–107.
The Size of Municipalities, Efficiency and Citizen Participation (1995) Local and Regional Authorities, Europe No 56.
Theodori, G.L. (2001) Examing the effects of community satisfaction and attachment
on individual well-being. Rural Sociology 66 (4), pp 818–824.
Thynen, J.H. von (1895) Der Isolierte Staat in Beziehung auf Landwirtschaft und
Nationalökonomi, Berlin.
Titmuss, R.M. (1974) Social Policy. George Allen and Unwin, London.
Tomberg, A. (1930) Rahvaliikumine üle piiri. Eesti Statistika 135 (2), pp 110–113.
Toomet, O. (2001) Eesti-sisese migratsiooniga kaasnev vahetu palgatous. Master Thesis,
Tartu University, Faculty of Economics.
Townsend, P. (1979) Poverty in the UK. Penguin, Harmondsworth.
Treivish, A., Brade, I. & Nefedova, T. (1999) Russian cities at a crossroad. GeoJournal 49
(1), pp 117–129.
Triandis, H.C. (1971) Attitude and attitude change, New York, Wiley.
Ulst, E. (1996) Riigirahandus I. Tartu University Press.
Valkonen, T. & Martelin, T. (1986) Influence of household and area characteristics on the
residental mobility of households. Yearbook of Population Research in Finland 14,
Population Research Institute, pp 14–28.
Vining, D.R. & Kontuly, T. (1978) Population dispersal from major metropolitan regions:
An international comparison. International Regional Science Review 3 (1), pp 49–73.
Vining, D.R. & Pallone, R. (1982) Migration between core and peripheral regions: A
description and tentative explanation of the patterns in 22 countries. Geoforum 14
(4), pp 339–410.
Vining, D.R.Jr (1985) The growth of the core urban regions in developing countries.
Population and Development Review 11 (3), pp 495–514.
Vyvere, Y. Oppewal, H. & Timmermans, H. (1989) The validity of hierarchical information integration choice experiments to model residential preference and choice. Geographical Analysis 30 (3), pp 254–272.
Wagner, M. (1989) Räumliche Mobolität im Lebensverlauf, Eine empirische
Untersuchung sozialer Bedingungen der Migration, Enke, Stuttgart.
Walker, R., Ellis, M. & Barff, R. (1992) Linked migration systems: Immigration and internal labor flows in the United States. Economic Geography 68 (3), pp 234–248.
Walmsley, D.J. (1982) Personality and regional preference structures: a study of introversion-extroversion. Professional Geographer 34, pp 279–288.
Walmsley, D.J., Epps, W.R. & Duncan C.J. (1998) Migration to the New South Wales
North coast 1986–1991: Lifestyle motivated counterurbanisation. Geoforum, 29 (1),
pp 105–118.
Wardwell, J.M. (1977) Equilibrium and change in nonmetropolitan growth. Rural Sociology 42, pp 156–179.
Warnes, A.M. (1986) The residential mobility histories of parents and children, and relationship to present proximity and social integration. Environment and Planning A 18,
pp 1581–1594.
222
Warnes, A.M. (1992) Age-related variations & temporal change. Chapter 3. Rogers,
E.A. (ed) Elderly Migration and Population Redistribution, Belhaven Press, London.
Weber, A. (1929) Alfred Weber’s Theory of the location of industries. Translated with an
introduction and notes by Carl J. Friedrich, Chicago University of Press.
Weibull, J.W. (1978) A search model for microeconomic analysis with spatial applications. Karlqvist, A., Lundqvist, L., Snickars, F. & Weibull, J. W. (eds) Spatial Interaction Theory and Planning Models, North-Holland, Amsterdam, pp 47–53.
Westerlund, O. (1995) Economic Influences on migration in Sweden. Umeå Economic
Studies No 379, University of Umeå.
Wetherly, P. (1996) Basic needs and social policies. Critical Social Policy 46 (16), pp 46–
65.
White, P. (1989) Internal migration in the nineteenth and twentieth centuries. Ogden,
P.E. & White, P.E. (eds) Migrants in Modern France, Unwin Hyman, pp 13–33.
White, P. (1990) Labour migration and counterurbanisation in France. Johnson, J.H. &
Salt, J. (eds) Labour migration: The internal geographical mobility in the developed
world, David Fulton Publishers, pp 99–111.
White, P.E. & Woods R.I. (1980) The foundations of migration study. White, P.E. & Woods
R.I. (eds) The Geographical impact of migration, pp 1–20.
Whynne-Hammond, C. (1979) Elements of Human Geography. Hyman.
Wiessner, R. (1999) Urban development in East Germany — specific features of urban
transformation processes. GeoJournal 49 (1), pp 43–51.
Willekens, F. (1987) Migration and development: a micro-perspective. NIDI working
paper No 62. Voorburg.
Wilson, F.D. (1988) Aspects of migration in an advanced industrial society. American
Sociological Review 53, pp 113–126.
Wiseman, R.F. (1980) Why older people move: theoretical issues. Research on Aging 2,
pp 141–154.
Wolpert, J. (1965) Behavioural aspects of the decision to migrate. Papers and Proceedings, Regional Science Association 15, pp 159–169.
Wolpert, J. (1966) Migration as an adjustment to environment stress. Journal of Social
Issues 22, pp 92–102.
Wood, B. & Carter, J. (2000) Towns, urban change and local government. Halsey, A.H. &
Webb, J. (eds) Twentieth-Century British Social Trends, pp 412–436.
Woods, R. (1982) Theoretical Population Geography, Longman.
World Urbanization Prospects (2001) The 1999 Revision, United Nations, New York.
Yoshitaka, I. (1999) Contribution of the demographic factor to the migration turnarounds
in Japan, Sweden and Canada. International Journal of Population Geography 5, pp
1–17.
Zaiontskovskaia (1985) – Çàéîí÷êîâñêàÿ, Æ.À. (1985) Íåêîòîðûå íàïðàâëåíèå
ýâîëóþöèè ðàññåëåíèÿ. Äîñòèæåíèÿ è ïåðñïåêòèâûû. Ðåãèîíàëúíîå ðàçâèòèå è
óïðàâëåíèå 52 (5), pp 43–51.
Zelinsky, W. (1971) The hypothesis of the mobility transition. The Geographical Review
61 (2), pp 219–249.
Zelinsky, W. (1983) The impasse in migration theory: A sketch map for potential escapees. Morrison, P.A. (ed) Population movements: Their form and functions in urbanization and development, Ordina, Bruxelles, pp 21–49.
223
224