Labour market disadvantage and the experience of recurrent poverty

Labour market disadvantage and
the experience of recurrent
poverty: Comparing Britain and
Germany
Mark Tomlinson
University of Sheffield
Robert Walker
University of Oxford
Overview



Dualisation of labour market and economic circumstances
Long tradition of dualisation theories
Analysis of longitudinal panel data from Great Britain and
Germany to identify outsiders in the labour market and explores
how this affects their future well-being in terms of poverty
experience.
 The panels are largely comparable with each other enabling a
direct comparison of two welfare regimes: one liberal and one
conservative.
 Other factors are also taken into account such as previous
poverty, family circumstances, age, gender and educational
endowments.
Page 2
Main hypothesis



It is postulated that labour market divisions (by which we refer
to the existence of upper and lower strata in employment
conditions and opportunities) have an impact on social
divisions.
Those in unstable labour market positions may have an
increased tendency to experience recurrent poverty episodes.
The existence of these dualist structures within different policy
regimes can then be explored by comparing equivalent
longitudinal panel data from different countries
Page 3
Segmented labour market theories and
dualisation



Standard economic theory generally postulates that there is
one labour market and all buyers and sellers compete on the
basis of perfect information in this market,
Another set of institutional theories argues that there is not a
single labour market at all, but generally two.
“Neoclassical economists have been unable to explain satisfactorily the structure
and functioning of labour markets in North America and Europe. A rich nonneoclassical labour market analysis has evolved to fill this void. It goes under the
rubric of "labour market segmentation". Segmented labour market theories argue
that distinct labour market segments characterise the advanced industrial
economies. This is in contrast with traditional competitive labour market analysis
which pictures the labour market as relatively unified.” Rosenberg 1989.
Page 4
Dual Labour Markets

Rosenberg shows how segmentation theory developed out of
the initial American dual labour market theory (DLM).
 DLM theory was initially concerned to address the problem of
ethnic minority underemployment in the USA.
 The labour market consisted of ‘primary’ and ‘secondary’
sectors.
 The primary sector - good working conditions, high wages, job
security and promotion prospects
 The secondary sector consisted of poorly paid, unstable work
with generally poor working conditions.
 This was the view put forward by Piore (1970) and Edwards
(1975) among others.
Page 5
Segmented Labour Markets and CorePeriphery Economy



We can think of the two main developments of this early theory
of DLM leading to two different and more sophisticated though
obviously related strands.
Piore, with Doeringer, developed a more advanced
segmentation theory (see Doeringer and Piore, 1971; Piore,
1975, referred to as segmented labour market theory)
Another theory of a ‘core’ and ‘peripheral’ economy was also
developed (see Bluestone, 1970; Harrison, 1972).
Page 6
Contemporary views





Building on this early literature, later commentators have noted
that, in the era of post-industrialization, labour markets have
become even more segmented.
Unemployment has become much more prevalent in postindustrial societies and the rise of the temporary contract has
become a central feature of many European countries
(Standing, 1993).
Part-time employment is also a more prominent feature of
European labour markets than before; especially for women
(Esping-Andersen, 1999).
Thus segmentation is now a more entrenched feature of
modern economic structures than was previously the case.
Atypical employment and unemployment determine more
contemporary positions on the determinants of outsiderness
(Rueda 2005, 2006; Emmenegger 2009).
Page 7
Analogous social policy debates
The “Low Pay No Pay” cycle



There is evidence that recurrent poverty is often linked to the
so-called ‘low-pay-no-pay cycle’, labour market conditions and
to specific household level changes (Smith and Middleton,
2007).
Occupying the lower strata of the labour market or being
excluded altogether from labour market activity has a tendency
to exacerbate the occurrence of poverty in the following years.
This can be compounded by family structure. For instance,
being part of a single parent household or having a family with
large numbers of dependents may combine to make matters
worse.
Page 8
UK policy deficiency?

Smith and Middleton have also drawn attention to the neglect
of the ‘low-pay-no-pay cycle’ in the UK and the associated
prevalence of ‘recurrent poverty’.
 The policy emphasis has been more on job entries than the
quality of the jobs themselves
 Even though one of the guiding principles underpinning the
previous Government’s welfare reform proposals was ‘retention
and progression, not just job entry’ (DWP, 2007),
Page 9
Relating labour market dualisation to
poverty

If the existence and identification of different segments within
the labour market can be operationalized using panel data then
the relationships between these labour market strata and the
repeated occurrence of poverty are then open to detailed
statistical analysis
 Thus labour market outsiderness can be analyzed in relation to
dimensions of economic outsiderness.
 The comparative dimension that an investigation of analogous
panel data between countries can bring, will allow us to
ascertain the impact of different policy regimes on poverty
alleviation after other factors have been taken into
consideration.
 Here we will be comparing a Liberal regime (Great Britain) with
the Corporatist regime of Germany (Esping-Andersen, 1990,
1999).
Page 10
Hypotheses – differences between
regimes



In the British regime where benefits are thought to be meanstested and focused on poverty alleviation we might expect that
people trapped in peripheral employment or unemployment
eventually move out of poverty as they change jobs or move
back into better labour market positions over time.
In the German regime we might expect to see more
segmentation due to the corporatist nature of business where
trade unions assist in perpetuating and protecting insiders
(Palier and Thelen)
The restrictions in social rights generated by this type of
system might perpetuate inequalities rather than diminish them.
Thus the impact of segmentation might be greater in Germany
than in Great Britain.
Page 11
Or convergence?



On the other hand Fleckenstein et al. (2011) argue that several
changes in the German system have allowed a degree of
convergence with the British system from the 1990s onwards.
Although the German system is still different there are several
changes taking place in both the composition of the labour
force and the welfare system.
Deindustrialization has also potentially altered the assumptions
implicit in the derivation of traditional typologies of welfare
states (Estevez-Abe et al., 2001; Iversen and Stephens, 2008).
Page 12
Deindustrialization

There has been a polarization of skills in the service sector
which has generated a group of highly skilled professionals
along with an army of lower skilled white collar workers

predominantly women and often part-time rather than full-time.

This, coupled with the trend towards increased means testing
and the institutionalization of an employment-oriented family
policy in Germany, has had the potential to alter the
assumptions underlying the Conservative welfare regime (cf.
Fleckenstein et al., forthcoming; Bleses and Seeleib-Kaiser,
2004).
 If there has been a degree of convergence between the British
and German systems then we might not expect to see very
significant differences in the impact of labour market divisions
on economic hardship between the two countries and levels of
segmentation might in fact be similar in the two regimes.
Page 13
Data from Britain:BHPS





The analysis utilizes data from the British Household Panel
Study (BHPS)
The BHPS commenced in 1991 with an initial sample of around
10,000 individuals resident in some 5,000 households.
These individuals have subsequently been re-interviewed each
year and the sample has also been extended to include more
households from Scotland and Wales and to embrace Northern
Ireland.
While the data can be weighted to provide an accurate picture
of life in Great Britain or the United Kingdom at different points
in time, this analysis is restricted to Great Britain.
The analysis covers the period 1999 to 2005 where it is
comparable to the German data. Individuals under 18 years of
age are excluded from the sample analyzed and also those
over 60.
Page 14
Data from Germany: SOEP

The German SOEP data commenced in 1984 and comprised
around 4500 households in West Germany.
 Over time eight additional panels have been merged with the
original (including an East German panel and a panel of high
earners).
 The panel currently has over 20,000 individuals included for
analysis.
 Several of the questions in the SOEP are identical or very
similar to those in the BHPS thus allowing some comparative
panel data analysis to be undertaken.
Page 15
Economic dualism: Defining poverty

Poverty is defined as falling into the bottom quartile of the
distribution of equivalized household income. Equivalization
was undertaken by dividing total household income by the
square root of the number of household member (OECD
Method)
Page 16
Real numbers




This gives us a relative and comparable poverty measure
between the two regimes although it must be borne in mind that
the UK had higher rates of both overall poverty and in-work
poverty in the early 2000s (the period we are analyzing here).
For example, in 2001, Germany had a poverty rate of 11%
compared to the UK's 15% and in work poverty rates were 4%
and 6% respectively (Eurostat, 2005).
The UK was in fact very close to the EU-15 average at this
time.
Poverty here was measured by Eurostat as being below 60% of
equivalized median income.
Page 17
Defining dualisation as labour market
segmentation


The employed respondents in the BHPS and SOEP are
categorized into various labour market strata based on their
current job or the lack of one.
An insider in the BHPS is defined as having a permanent job
and either or both of the following job characteristics:

a pay structure that includes an annual incremental pay rise
 self-reported real prospects for promotion
Page 18
SOEP




In the BHPS these questions are asked every year.
In the SOEP the questions on contractual status are collected
annually, but the other two questions are asked in broadly
alternate years and differ slightly in wording.
The question on pay asks for the likelihood of a pay rise in the
next two years (as a percentage from 10%, 20%, 30% and so
on) and the question on promotion also asks for percentage
likelihood rather than a yes/no answer.
We have treated these questions as equivalent to the British
data if the respondent stated a percentage greater than zero.
As the data are not collected annually in Germany we are
therefore restricted in our analysis to alternate years (1999,
2001, 2003, 2005).
Page 19
Outsiders

The insider status defined here therefore represents the better
strata of the labour market in terms of prospects for
advancement and job stability. An outsider on the other hand is
defined as someone who has either:

no permanent employment contract
 a permanent job, but with none of the benefits mentioned above
(what might be termed a 'dead-end job')
 someone who is unemployed (and looking for work)
 someone who is out of work and not looking for work (referred to
hereafter as non-employed)

The impact of these segmentation variables on poverty
experience will reveal the extent to which social divisions are
perpetuated by and reinforced by labour market divisions.
Page 20
Overall levels of segmentation 2003-5
Insider
Britain
Germany
2005
2005
Outsider
Unemplo
Non-
Outsider
Unemploy
Non-
employed
yed
employed
employed
ed
employed
3141
900
78
192
4311
3664
1471
164
218
5520
73%
21%
2%
4%
100%
67%
27%
3%
4%
100%
Outsider
850
2318
88
208
3464
Outsider
1293
3015
302
422
5032
employed
25%
67%
3%
6%
100%
employed
25%
60%
6%
8%
100%
Unemploy
77
91
99
113
380
Unemploy
145
207
418
224
994
21%
24%
26%
30%
100%
14%
21%
42%
23%
100%
Non-
165
227
80
1500
1972
Non-
268
466
152
1742
2628
employed
8%
12%
4%
76%
100%
employed
11%
18%
6%
66%
100%
Total
4233
3536
345
2013
10127
Total
5373
5159
1036
2606
14174
42%
35%
3%
20%
100%
38%
36%
7%
18%
100%
2003
Insider
Total
Insider
2003
Insider
Page 21
Total
Summary: Persistent divisions




The transition matrix tells us that labour market divisions are
persistent in both countries.
The diagonal percentages for the different employed segments are
all higher than 50% implying that it is relatively difficult to leave the
outsider categories and relatively stable for the insiders of both
countries.
The other conspicuous findings are the transition probabilities of
the unemployed and non-employed in 2003.
In GB a quarter of unemployed respondents moved into the
outsider employed group while even more moved into
economically inactive positions (30%). A quarter remained
unemployed after 2 years. Only 21% had acquired an insider job in
2005.
Page 22
Germany



In Germany the situation with respect to the unemployed was similar
although a much greater proportion remain in unemployment (42%), but
again only 14% make it into the insider group and 23% drop into nonemployment.
The fact that the Germans can remain unemployed for longer (rather than
stopping looking for work and becoming non-employed) suggests that
there is more leeway in the German system for an unemployed person to
wait until a suitable job can be found.
The higher percentage remaining in unemployment in Germany does
reinforce the idea that the benefit system still protects and allows a
degree of flexibility in job choice. Whereas in Britain people are more
likely to stop looking for work and become economically inactive.
Page 23
By occupation
Great Britain
Germany
Insider level 3
25.8
26.7
Insider level 2
27.2
22.1
Insider level 1
2.4
2.4
Outsider level 3
18.3
21.0
Outsider level 2
23.3
23.1
Outsider level 1
3.1
4.1
Level 1 refers to unskilled and semi-skilled labour (ISCO group 9)
while level 3 refers to associate professionals, professionals and
managerial occupations (ISCO groups 1, 2 and 3). All other
occupations are referred to as level 2 (intermediate ISCO groups 4,
5, 6, 7 and 8).
Page 24
Occupational segmentation

Table 2 reveals once more that the levels of segmentation
appear to be very similar between the two countries even when
taking occupation into consideration.
 Furthermore the relationship between occupational level and
segmentation is not straightforward.
 There is a significant number of employees in high level
occupations in the outsider segment of the labour market (18%
of British and 21% of German workers).
 So there is a high degree of uncertainty and risk involved in the
labour markets of both regimes even among many
professionals and associated technical and managerial
occupations.
 Thus having elevated skills and relatively prestigious jobs does
not necessarily guarantee stable and favorable working
conditions and benefits.
Page 25
Probit analysis

As we stated above we are interested in predicting poverty as a
function of labour market divisions, human endowments and
family circumstances.
 We essentially take the poverty position of someone in a given
year and use this to predict their poverty experience during the
next period of observation (which in this case is every two
years).
 We include labour market divisions in two ways. First of all as a
simple insider/outsider dichotomy and second by
disaggregating the outsiders into employed, unemployed and
non-employed to ascertain whether there are differences
between these outsider sub-groups.
 It might be the case that even being an outsider with
employment is somewhat less risky than being out of work
altogether. This is tested in the models that follow.
Page 26
Independent variables
Previously poor (lagged dependent variable)
Female
Germany (country dummy in pooled models)
Age 18-24
Age 25-34
Age 35-44
Reference: Age 45+
Higher level education (ISCED 5)
College level (ISCED 4)
Ordinary level (ISCED 3)
Elementary or no education (less than ISCED 3)
Reference: Postgraduate (ISCED 6)
Couple no children
Couple with dependent children
Mixed adult households
Single parent
Reference: Single adult
Insider versus outsider
labour market status:
Outsider employed
Unemployed
Non-employed
Reference: Insider
Page 27
Models predicting poverty 1999-2005 (all
adults aged 18-60)
Independent variable
Great Britain
Germany
Pooled
Great Britain
Germany
Pooled
Previously poor
Female
Germany
1.65***
0.05 ns
–
1.92***
0.00 ns
–
1.86***
0.02 ns
–0.06**
1.63***
–0.00 ns
–
1.87***
–0.02 ns
–
1.82***
–0.01 ns
–0.10***
Age 18-24
Age 25-34
Age 35-44
Reference: Age 45+
Higher level education
College level
Ordinary level
Low or no education
Reference: Postgraduate
0.38***
0.16***
–0.07 ns
0.18***
0.37***
0.20 ***
0.21***
0.32***
0.13***
0.33***
0.16***
–0.05 ns
0.19***
0.36***
0.23 ***
0.20***
0.32***
0.16***
0.23 ns
0.52***
0.58***
0.88***
0.28***
0.36***
0.51***
0.68***
0.19***
0.40***
0.50***
0.69***
0.22 ns
0.52***
0.58***
0.85***
0.26***
0.30***
0.45***
0.63***
0.16***
0.36***
0.45***
0.64***
Couple no children
Couple with dependent
children
Single parent
Mixed adult households
Reference: Single Adult
Outsider vs insider
–0.78***
–0.42***
–0.76***
–0.79***
–0.77***
–0.71***
–0.77***
–0.40***
–0.77***
–0.81***
–0.77***
–0.72***
0.29***
–1.08***
0.02 ns
–0.79***
0.07 ns
–1.00***
0.28***
–1.08***
0.00 ns
–0.83***
0.06 ns
–1.03***
0.60***
0.50***
0.52***
–
–
–
0.30***
0.98***
0.77 ***
0.17***
0.96***
0.64***
0.20***
0.98***
0.67***
19489
45919
65408
Peripheral
Unemployed
Non-employed
Reference:
Insider
N
19489
45919
65408
Page 28
Models predicting poverty 1999-2005 (employed
and economically active aged 18-60)
Independent variable
Great Britain
employed
Germany
employed
Pooled
employed
Great Britain
Economically
active
Germany
Economically
active
Pooled
Economically
active
Previously poor
Female
Germany
1.70***
–0.00 ns
–
1.92***
–0.04 ns
–
1.87***
–0.03 ns
–0.07**
1.70***
–0.03 ns
–
1.96***
–0.07**
–
1.89***
–0.06**
–0.01 ns
Age 18-24
Age 25-34
Age 35-44
Reference: Age 45+
Higher level education
0.23***
0.13*
–0.06 ns
0.45***
0.49***
0.31***
0.36***
0.39***
0.21***
0.28***
0.15**
–0.07 ns
0.25***
0.35***
0.22***
0.24***
0.31***
0.15***
0.24 ns
0.26***
0.18***
0.27 ns
0.26***
0.19***
College level
Ordinary level
Low or no education
Reference: Postgraduate
0.41**
0.55***
0.80***
0.28***
0.49***
0.64***
0.32***
0.48***
0.67***
0.46***
0.60***
0.88***
0.33***
0.51***
0.70***
0.35***
0.51***
0.71***
Couple no children
Couple with dependent children
–0.73***
–0.28***
–0.80***
–0.69***
–0.79***
–0.59***
–0.72***
–0.32***
–0.79***
–0.74***
–0.77***
–0.64***
Single parent
Mixed adult households
0.46***
–0.95***
0.15**
–0.69***
0.21***
–0.94***
0.41***
–0.99***
0.09 ns
–0.72***
0.16***
–0.96***
Outsider vs insider
0.41***
0.23***
0.29***
0.48***
0.40***
0.42***
N
15607
32390
47997
16171
35738
51909
Reference: Single Adult
Page 29
Conclusions

We have shown that it is possible to partition employment into
strata along the lines of labour market segmentation theory
using data from the BHPS and SOEP.
 This reveals a reasonable degree of similarity between the two
countries. Unlike what might be expected from a standard
typology along the lines of Esping-Andersen (1990, 1999) there
is a degree of correspondence between the two regimes in
both the proportions of insiders and outsiders and also the
relative probability of transfer between insider and outsider
status within the two regimes.
 This may be because the time period available for analysis is
post 1999 when many of the policies adopted by Germany
could have converged to a greater degree with a liberal
economy like Britain. Unfortunately we cannot assess the pre1999 state of affairs so this hypothesis remains unexplored.
Page 30
Conclusions

However, it still remains the case that there is a small, but
significant difference between German and British systems in
terms of the likelihood of experiencing poverty.
 The pooled statistical models show that the German welfare
system is statistically superior with respect to the prevention of
poverty spells (as it is more generous) after controlling for the
other factors we have considered.
 Thus despite any convergence that may have taken place the
German welfare structure is still more effective in this sense.
 In both nations, previous poverty experience remains the
strongest determinant of future poverty experience.
Page 31
Outsiders are in a vulnerable position



We have demonstrated that when labour market divisions are
used to predict poverty experience there are significant effects
that imply that superior labour market conditions such as those
characterized by being an insider rather than an outsider
decrease the likelihood of becoming poor even after accounting
for a person’s human capital and family situation
However, those in employed, but outsider labour market
positions generally do better on average than those who are
not working at all at any point in time.
In both countries the transition matrices reveal that there
appears to be a reinforcing and entrenchment of outsiderness.
Page 32
Labour market policy



What this analysis confirms is that structural factors,
opportunities presented by the labour market, are as important,
and often more important, than personal attributes and
circumstances in determining the risk of poverty.
Policies that simply encourage people to find work, without
paying attention to the kind of jobs that are available, cannot
secure a marked reduction in recurrent poverty or a sustained
decline in the poverty rate.
The analysis underlines the importance of seeking to ensure
the availability of high quality core jobs offering security and
prospects as well as policies that foster job search and
improved skills whether in a Liberal or Conservative regime.
Page 33
Human capital



Education and human capital also undoubtedly play a role as
predicted by human capital theory. But the educational effect
appears to be stronger in the British context.
This is perhaps because the return on investment in a Liberal
market economy would be higher than in a Conservative one
where skills are more attached to a corporation and labour
mobility is supposedly lower (see Estevez-Abe et al., 2001, for
example).
But the transition matrices also call into question the notion that
Germany has a more rigid labour market than Britain. There
was actually slightly more fluidity in the German matrix than the
British one. Nevertheless education is a significant offsetting
component of poverty alleviation in both regimes.
Page 34
Convergence?

The structural reinforcement of dualisation seen in the two
countries, be it founded on an economic or a labour market
basis, provides evidence for the possibility of a shared political
or social identity that cuts across traditional class lines and is
founded on poor labour conditions (insecurity) and low income
(relative poverty) rather than occupation.
 There are many more similarities than differences. There is
some diversity, but it is not as extensive as might be expected.
 Therefore the ideas proposed by Fleckenstein et al.
(forthcoming) that there has been a degree of convergence
between the welfare systems and employment structures of the
two countries and that the impact of deindustrialization in the
1990s and 2000s has had profound effects on the assumptions
implicit in traditional typologies of welfare regimes seems to be
plausible. A longer time frame for analysis would have to be
employed to verify these issues.
Page 35