The Cultural Transmission of Environmental
Preferences: Evidence from International Migration
Anastasia Litina, Simone Moriconiy, Skerdilajda Zanajz
January 27, 2014
Abstract
This paper theoretically and empirically advances the hypothesis that di¤erences
in environmental preferences can be traced to cultural di¤erences. In particular, we
argue that environmental attitudes such as willingness to pay for the environment
and beliefs about environmental attitudes are not solely formed by the interaction of
individuals with local environmental conditions but can also be accounted for by cultural
di¤erences across countries. To explore our hypothesis we exploit the natural experiment
of international migration ‡ows and establish that the environmental culture of migrants,
as has been formed in their country of origin and transmitted across generations, is still
prevalent in the host country despite the fact that all migrants in a country are exposed
to the same local environment. In the presence of multiple environmental problems that
require collective action, comprehending the driving forces behind the formation of an
environmental culture, that can a¤ect policies, is critical.
Keywords: Cultural Transmission, Migration, Environmental Preferences
JEL Classi…cation Numbers: Q50; Q58; R23
CREA, Université du Luxembourg, 162A, avenue de la Faïencerie, L-1511, Luxembourg. Email:
[email protected].
y
Università Cattolica di Milano, Via Necchi, 5, 20123, Milano, Italy and CREA, University of Luxembourg.
Email: [email protected];
z
CREA, Université du Luxembourg, 162A avenue de la Faïencerie, L-1511, Luxembourg. Email:
[email protected]
1
Introduction
The profound role of culture on economic outcomes and the formation of public policy has
been at the center of a recent debate that explores the intergenerational transmission of
cultural traits. Several aspects of the economy and the society such as the fertility rate,
female labor force participation and preference for redistribution have been argued to manifest
a cultural component that frames individual economic behavior and ultimately economic
policies. As part of culture, these traits are transmitted across generations through social
or formal learning. In the light of major ecological problems that requite immediate collective
action, the current research explores to what extend environmental preferences manifest and
cultural component and whether this cultural trait is transmitted across generations.
This paper theoretically and empirically advances the hypothesis that di¤erences in
environmental preferences can be traced to cultural di¤erences. In particular, we argue that
environmental attitudes such as willingness to pay for the environment and beliefs about
environmental attitudes are not solely formed by the interaction of individuals with local
environmental conditions but can also be accounted for by cultural di¤erences across countries.
To explore our hypothesis we exploit the natural experiment of international migration ‡ows
and establish that the environmental culture of migrants, as has been formed in their country
of origin and transmitted across generations, is still prevalent in the host country despite the
fact that all migrants in a country are exposed to the same local environment.
To de…ne culture is a challenging task. As R. Borofsky (2008) puts it: " we talk
about culture as something real,..., but culture is in fact an intellectual construct used for
describing a complex cluster of human behaviors, ideas, emotions and artifacts. ... e¤orts
to de…ne culture are akin to trying to encage wind".1 For the purpose of this paper, we
de…ne environmental preferences as a set of values that embody beliefs, social norms and
individual attitudes towards the natural environment, which rule individual behavior towards
the environment. The environmental economics literature has so far neglected the analysis of
the role of environmental culture. Motivated by the upheaval in public debate on the fragility
of the environment and the urgent need for the formation of public policy we believe that
addressing this issue will help to better understand what crystallizes environmental awareness.
Consequently, governing bodies will be in a better position to design and widely implement
environmental friendly policies, as for instance, an environmental education system that will
adequately prepare the future generations to protect the future they will inherit.
As far as policy is concerned, comprehending whether environmental preferences are
part of culture will further improve our knowledge about the status quo of international
economic agreements. It could be plausibly argued that the di¢ culties in …nding a consensus
on greenhouse gas emission targets for the period 2013-2020 could partly re‡ect the fact that
1
Another related issue is how culure a¤ects human behaviour. For this see Bowles and Gintis (2008).
1
country members of the United Nations Framework Convention on Climate Change manifest
highly heterogeneous attitudes towards environmental protection. Our research suggests that
the views expressed by each government are not only directed by purely economic incentives
but also re‡ect the social preferences prevalent in the country.
Our analysis addresses all these issues both theoretically and empirically. To clearly
capture the transmission of cultural traits we exploit the natural experiment of international
migration ‡ows. The advantage of such an approach is that since migrants are faced with a
di¤erent natural environment in the host country than that of their country of origin, we can
isolate the cultural elements of their environmental preference.
In the theoretical part of the paper we present a model of transmission of environmental
preferences following Bisin and Verdier (2001). We de…ne the environmental cultural trait as
the disutility from pollution, which ultimately determines the marginal willingness to pay for
the reduction of pollution. Agents live in two homogeneous social groups whose populations
mix if migration takes place. Agents consume goods whose production causes pollution. When
the social groups are still homogeneous, agents select the amount of consumption, and thus
the level of pollution, by maximizing their utility function. When migration takes place due
to income di¤erences, …rst generation migrants carry their cultural traits and preferences
with them in the destination group. Crucially though they are now facing the possibility that
their o¤spring may acquire the cultural trait of the destination group. Therefore, if individuals
want their culture to prevail in the new environment they should invest in family socialization.
We …nd that the environmental trait is successfully transmitted to the next generations if the
marginal disutility from pollution of migrants is inferior to the marginal disutility of the native
population.
The theoretical analysis generates one main testable hypothesis, i.e. that environmental
preferences may be culturally transmitted to future generations. Given that the natural
environment is di¤erent in the host country and the country of origin, the persistence of
environmental attitudes similar to those in the country of origin suggests the presence of a
cultural component in the formation of environmental preference.
Our empirical analysis aims to identify the cultural component of environmental preferences exploiting the natural experiment of international migration ‡ows. The focus on
immigrants allows to distinguish the e¤ect of “environmental culture”, as formed by individuals’ experience in their region of origin (…rst generation migrants) or by their parents’
experience (second generation migrants), from the local environmental incentives, which arise
in the country of destination. To establish our testable hypothesis we use survey data
from the European Values Study (EVS). Our sample comprises 47 European countries and
studies the attitudes, beliefs, cultural aspects and preferences of Europeans towards a range
of issues such as the environment, religion, politics, the economy etc. Our study focuses
on the environmental culture of individuals. In particular we focus on two main aspects
2
of environmental attitudes. The …rst aspect is the willingness of individuals to pay for the
environment and the second is their beliefs with respect to their environment.
As to their willingness to pay for environmental causes, respondents are given the
statement "I would give part of my income if I were certain that the money would be used to
prevent environmental pollution" and they are asked if they "strongly agree-agree-disagreestrongly disagree". To capture the individual beliefs with respect to the environment, the
respondents are asked whether they agree or disagree with the following statement "Humans
were meant to rule over the rest of nature".
The purpose of the empirical section is to explore whether attitudes towards the
environment are driven by a cultural component, by inertia or whether they are simply a¤ected
by the natural environment in which an individual resides. To address this issue we exploit
the natural experiment of migration that allows us to trace attitudes of migrants that have
left their country of origin and currently reside in a host country. The EVS dataset allows
us to trace …rst and second generation migrants, i.e. people who have been born in any of
the 47 countries of our sample and currently reside in a di¤erent host country, or people who
have been born in the host country but whose parents have migrated to the host country
from another country. After excluding all people who do not provide information about their
country of origin, or their parents’country of origin and keeping all individuals above the age
of 18 we have a sample of 4295 migrants coming from the 47 countries or our sample having
migrated to a di¤erent country of our sample. Of those migrants, 2554 are …rst generation
migrants whereas 1741 are second generation migrants.
Our …ndings suggest that the average environmental preference in an immigrant’s
country of birth has a large and signi…cant e¤ect on her own environmental preference.
Immigrants born in a country with a low preference for the environment tend to have
lower preferences for the environment than the natives of the countries in which they reside.
These results are robust to introducing in our analysis a large set of individual controls, host
country …xed e¤ects (that take care of any unobserved heterogeneity associated with the
country of residence) and a number of controls associated with the home country. Moreover
we run a horse-race between average beliefs in the origin country and average level of pollution
in the origin country and reassuringly we …nd that our results survive suggesting that the role
of beliefs is critical for the formation of individual environmental preferences.
Our …ndings have a number of important implications with respect to the formation
of public policy both at the local and at the international level. First, they highlight the fact
that whereas pollution and environmental conditions have a direct e¤ect on the formation
of an environmental culture, nevertheless we show that environmental preferences have a
cultural component that should be taken into account. Practically, that would imply that
government should not only aim at abatement and pollution reducing policies but also at
social learning activities that increase environmental awareness. This issue becomes even
3
more critical in an era where international migration ‡ows are rather vast and therefore
the median voter is not necessarily a native person but a weighted average" of di¤erent
cultures. Addressing the concerns and satisfying the needs of such a heterogenous group
will not only improve environmental quality but also facilitate the assimilation of migrants.
Finally, understanding the role of culture is particularly important at the international level
since international environmental agreements are reached by leader who represent not only
the economic incentives of their countries but also the average preferences of their country.
Therefore understanding the driving forces behind the formation of an environmental culture
is critical for reaching consensus at all levels.
The paper is organized as follows. Section 2 explores the related theoretical and
empirical literature and highlights the contribution of our paper. Section 3 presents a formal
model that explores the transmission mechanism. Section 4 presents the data, the empirical
strategy and the testable implication of the model. Section 5 subjects our analysis to a number
of robustness tests. Finally Section 6 concludes.
1.1
Related literature
Our study adds to a growing economics literature that looks at the joint role of parental and
social channels of cultural transmission. Initiated by Bisin and Verdier (2001), transmission is
modelled as the result of a decision of parents within the family (direct vertical socialization)
as well as the result of an indirect processes that takes place in society by the means of social
imitation and learning (oblique socialization). A heterogeneous society in which individuals
develop and transmit di¤erent cultural traits is a stable outcome of the process if the oblique
socialization is a substitute of the direct vertical socialization. In this case, direct transmission
acts as a cultural substitute to oblique transmission since parents have less incentives in
children’s education the larger the share of population sharing the same trait. Then, the
higher are the chances that the child acquires the same trait by means of oblique socialization.
There exist few theoretical articles in environmental economics that study social norms
and their impact on the environment. Sethi and Somanathan (1996) study the endogenous
evolution of social norms in a local common-property resource setting using evolutionary
game theory. They …nd that with a su¢ ciently large number of individuals that are enforcers,
the society can reach and remain in a norm-guided society rather then individualistic one.
Schumacher (2013) investigates cultural dynamics of environmental preferences, as in Bisin and
Verdier (2001), including a feedback from pollution to the cultural dynamics. Pollution a¤ects
the proportion of the two cultural traits that exist - environmentalists (greens) and browns.
The dynamic transmission is such that green preferences are less likely to be transmitted
intergenerationally for low levels of pollution, whereas they are likely to be transmitted for
high levels of pollution. Behadj and Tarola (2013) study social norms and their e¤ect on
4
environmental awareness. Authors consider consumption choice between a green and a brown
product made under social norms in‡uence. Individuals su¤er if they buy a brown product
when their similar select a green version. This mechanism allows to build a market demand
that embodies social norms that concern the environment. Then, the market equilibrium that
arises depends on the extent of such social norms. Authors show that brown …rms could exist
the market due to such consumption externalities.
Our theoretical analysis provides a simple and intuitive mechanism via which environmental culture can be transmitted across generations. Building upon the baseline cultural
transmission model and by plausibly assuming that individuals may migrate we capture the
conditions under which the environmental culture is transmitted and we generate a clear
hypothesis to be tested in the empirical section of the paper.
Empirically, our paper locates in the quantitative cultural economic literature. While
the idea that culture a¤ects economic outcomes is quite old and much debated in other …elds as
anthropology, the quantitative measures of how much culture matters for economic phenomena
starts only in early nineties. Borjas (1992) is a seminal paper that relates the labor outcomes of
immigrant workers with their ethnic capital in the USA. Nevertheless, it cannot be considered
as the beginning of the quantitative analysis to culture because ethnic capital encompasses
many features that cannot be all attributed to culture. The oversight of the quantitative
economists may be due to a series of technical di¢ culties that arise in a quantitative analysis
of culture. First of all, data restrictions. The seminal paper that attempts to identify a
cultural component in the propensity to save (Caroll et al., 1994), fails to …nd a systematic
e¤ect of culture probably due to the data restrictions, as admitted by the authors. A large
body of literature uses presently the internal surveys as World Value Survey, or the European
Surveys which o¤er a rich set of individual data. A second issue is causality. The …rst attempts
to link culture and growth (for instance, Tabellini, 2005) had to use historical literacy rates as
instruments for the link between individual e¤ort and economic success. To assess causality,
Alesina and Giuliano (2010), look at the behavior of second generation immigrants rather
then …rst generation migrants. They show a signi…cant in‡uence of the strength of family ties
on economic outcomes. Using data on second generation migrants and also multilateral data
among all countries in the sample, is also a way to deal with the selection problem (Luttmer
and Singhal, 2011).
De…ning culture and then isolating the cultural component in the empirical analysis
is another major issue. The strategy used in the literature exploits the natural experiment
of international migration ‡ows. In the same vein, in the attempt of quantifying the cultural
compound on redistribution preferences Guiso, Sapienza, and Zingales (2006) proxy culture
by a …xed e¤ect of the father’s origin country. A di¤erent approach is used in Fernández and
Fogli, 2009, who show the e¤ect of culture on fertility and female labor participation; and
also in Luttmer and Singhal, 2011, who analyze the taste for redistribution. In these last two
5
papers, authors use a country of birth dummy arguing that this cannot related individual
behavior to a speci…c determining factor in the origin country. Hence, in both papers culture
is proxied by the mean behavior in the country of origin.
Interestingly, whereas the literature on the transmission of several cultural traits is
expanding, our paper is the …rst to empirically explore and establish the presence and the
intergenerational transmission of environmental traits. In doing so, we implement a number
of alternative assumptions and conduct various robustness tests that address the challenges of
this literature. Our results are robust to all these tests and con…rm the presence of a cultural
component in the formation of the environmental culture.
2
The Model
Inhere, we brie‡y describe a model à la Bisin and Verdier (2001) to explain how environmental
preferences are transmitted intergenerationally. Assume two population of individuals that
in period one constitute two within-homogeneous groups.2 The …rst group lives in a polluted
environment, while the second one enjoys a better environmental quality. Denote these two
social groups B and G; correspondingly. Each population is assumed to have a speci…c
environmental trait di¤erent from the population of the other group. This di¤erence in
trait may have been developed through time and the di¤erence between the two has been
augmented due to di¤erent levels of income, di¤erent levels of technology, etc. Denote these
two traits as Brown and Green. The environmental trait can be seen as the disutility from
pollution, which determines the marginal willingness to pay for the environment, as de…ned
below. A representative agent’s utility function writes as:
u(c)
hi (p) + (Pii Vii + Pij Vij ) ; i; j = G; B; i 6= j
where c is consumption and p is pollution, u( ) is the subutility from consumption and h( )
captures the disutility from pollution. Both functions are assumed monotonic increasing. The
production of the consumption good(s) generates emissions as a by-product. We assume that
each unit of consumption good produced emits fi (c); i = B; G units of pollution, thus pollution
technology writes as p = fi (c); f 0 (c) > 0; i = B; G: Each individual has an endowment of
income Mi ; i = B; G: Neglecting for the time being the last part of the utility, the optimal
choice of consumption yields
1
h0i (p)
= 0 ; i = G; B
(1)
0
u (c)
fi (p)
2
A group is a generic set of individuals (a country, a village, an ethnic group), who share a common cultural
trait.
6
h0 (p)
The expression ui0 (c) captures the marginal willingness to pay of the representative agent
for a unit reduction of pollution. The two social groups show di¤erent willingness to pay,
governed by the shape of the disutility from pollution hi (p): At the optimal choice, the
marginal willingness to pay is equal to the inverse of the marginal productivity of the polluting
technology. Last, by assumption, u(c) hi (fi (c)) is a concave function c: Hence, the marginal
increase of consumption always dominates the corresponding disutility from the marginal
increase in pollution up to a certain level of pollution but once that level of pollution is
exceeded disutility from pollution exceeds utility from consumption.
The last part of the utility function, Pii Vii + Pij Vij ; concerns the cultural transmission
of the environmental trait. It captures the empathy of the representative agent for her prole
where Vii is the subutility function of a parent of type i having a child of type i; and Vij is
the subutility function of a parent of type i having a child of type j: Pii is the probability
that a child from a family with trait i is socialized to trait i; and Pij is the probability that a
child from a family with trait i is socialized to trait j: Before any migration takes place and
assuming that mutations within an homogeneous social group are absent, the event that a
new born within a family of type i has a di¤erent trait, namely j; never occurs, i.e., Pij = 0.
This implies that families need not to invest in within-family socialisation, which is a costly
activity. Let the cost of this investment, when it takes place, be Ii (e); I 0 > 0; I 00 > 0, where e
denotes resources devoted to within-family education.
In period 1, countries are in autarchy, so no migration takes place. Then, without loss
of generality, it is assumed that the trait Green is developed in group G and the trait Brown is
developed in group B. Hence, group Green is made of individuals of type G and group Brown
is made of individuals of type B. Our de…nition of types is determined by the cultural trait
(the type of function h). The level of pollution and consumption in each group is determined
by the interplay of the cultural trait G or B, income Mi and the polluting technology fi ;
which are all assumed to be group-speci…c variables. Clearly, were income endowments and
polluting technologies the same between the groups, the level of pollution in each group will
di¤er only due to cultural di¤erence, hi 6= hj :
Since before migration, the two populations were homogeneous, then family transmission and the obliquel transmission are complementary. Parents can refrain from socializing
with their descendants since the process is costly because they will acquire the trait from their
friends, hence I = 0. This is also known as the Social conformity. The representative agent
will maximize
7
max u(c)
c
s:t: c
hi (p) + (Pii Uii + Pij Uij )
(2)
Mi
p = fi (c)
Then, substituting these expressions in (2), the …rst order condition obtains again as
1
h0i (p)
= 0 ; i = B; G
0
u (c)
fi (c)
(3)
which determines the implicit optimal solutions of consumption and pollution, ci (Mi ) and
p (Mi ), respectively. Notice that since Pij = 0 in the period that precedes migration, then, the
…rst order condition is the same as in a classical model with no intergenerational transmission
of traits (1). At the optimal choice of agents, the marginal willingness to pay to reduce
pollution shall equalize the marginal productivity of the polluting technology. As expected,
ci (Mi ) and p (Mi ) depend on the shape of hi and the productivity of the polluting technology.
Assume now that in the second period, migration takes place from population i to j;
i = B; G; i 6= j. This migration can take place, as explained in the seminal paper of Roy
(1951), due to income di¤erences: Mi < Mj . Now, a fraction q i ; i = B; G of the population
in group j shows a trait i which is di¤erent from the native’s population trait j: A child born
in the migrant family receives the same trait as the parent through the socialization within
the family with probability di (q i ):
If the socialization within the migrant family is not successful, with probability 1
di (qi ), then with probability q i the trait i is acquired by the socialization in the society, and
trait j with probability qj = 1 qi : Then, a child of a migrant family shows the trait of his
family with probability di (qi ); when the trait is acquired at home, plus (1 di (qi )) qi ; when
the trait is acquired within the migrant minority. Hence,
Pii = di (qi ) + (1
di (qi )) qi :
Then, a second generation migrant will not show the same trait as his family with probability
Pij = 1 Pii :
8
After migration, the migrant families, with trait i; maximize the following utility
function
max u(c)
e
s:t: c + e
hi (p) + (Pii Uii + Pij Uij )
Ii (e)
(4)
Mj
p = fj (c)
The …rst order condition now obtains as
(1 + Pii ) h0i + Pij h0j
u0
2
Ii0 1
=
0
u0 fj
fj0
(5)
When facing an heterogeneous population, at the optimal choice, the marginal willingness
to pay for the environment of a migrant of type i changes with respect to the marginal
willingness to pay if he had stayed in his own group. This change occurs even though his
cultural trait hi (p) remains the same. After migration, a migrant’s marginal willingness to
pay for the environment encompasses the e¤ect of pollution on the utility of the child whether
he is of type i or j. Secondly, part of the resources is now devoted to within-family education.
Furthermore, the migrant family now receives a di¤erent income and …nally, she recognizes
that the polluting technology is di¤erent. This …rst order condition gives the implicit solutions
ci (Mj ; qi ) and p (Mj ; qi ): Totally di¤erentiating the above …rst order condition, we obtain the
following results:
Proposition 1 The optimal level of investment in direct socialization is decreasing in the
cultural composition of the population if and only if h0i (p) h0j (p) < 0: This implies that when
h0i (p) h0j (p) < 0; environmental preferences are successfully transmitted from one generation
to the other.
Proof. Totally di¤erentiating the …rst order condition (5) with respect to e and q;
yields:
de
=
dq
fj0 (c) [d0 (qi ) (1
2u00 (c)
qi ) + (1
(1 + Pii ) h00i (p)fj0 (c) + h0i (p)fj00 (c)
d(qi ))] h0i (p)
h0j (p)
0
Pij h00j (p)fj (c) + h0j (p)fj00 (c)
I 00
Since the denominator is negative for the concavity condition of the utility function with
respect to e, then, the sign of de
is given by the sign of h0i (p) h0j (p): If h0i (p) h0j (p) < (>)0
dq
then de
< (>)0 !
dq
The above proposition shows that the level of investment in family socialization depends
crucially on the di¤erence in marginal disutility of pollution hi versus hj ; i; j = B; G; i 6= j: If
9
the destination group is characterized by individuals with a very high marginal disutility from
pollution, then the higher the ‡ow of migrants, the lower the investment of families in withinfamily socialization because children will obtain the trait from oblique transmission. This is
the condition of cultural substitutability settled by Bisin and Verdier (2001) which guarantees
that the environmental trait is a cultural trait that will be successfully transmitted from
generation to generation and the population in the destination country will be heterogeneous.
Two remarks are in order. First, the condition h0i (p) ? h0j (p) does not imply that the
level of disutility from pollution of individuals in group i is higher or lower from the disutility
from pollution of those in group j: It can well be that individuals in j su¤er more pollution
with hi (p) < hj (p); and vice versa. The condition in Proposition 1 determines a relationship
between the marginal disutility from pollution at the optimal level of pollution in the group
j:
dc
< 0 i¤
Second, totally di¤erentiating the …rst order condition for c; we …nd that dq
0
0
hj (p) > hi (p) holds. Hence, the higher the ‡ow of migrants qi , who will transmit their trait
successfully (as established in Proposition 1), the lower the level of the pollution pj produced
in the destination country by migrants as compared to the level of pollution they would have
produced if they had stayed at the origin country. Furthermore, it is worth noticing that the
optimization of the native representative agent of group j is similar to (4) : It follows that
migration can determine a decrease of the level of pollution in the destination social group
depending on the elasticity of the demand for consumption. More precisely, this happen if the
dc
< 0) dominates the increase of consumption
total decline of the per capita consumption ( dq
demand due to migrants arrival.
3
3.1
Empirical analysis
Data and variables
We use data from the European Values Study (EVS). The EVS is a large scale cross national
survey spanning over the period 1981-2008 and currently has 4 waves (1981, 1990, 1999 and
2008). It comprises a large number of European countries and regions with the maximum
number being 47 countries for the fourth wave. The countries in the sample are Albania, Armenia, Austria, Azerbaijan, Belarus, Belgium, Bosnia-Herzegovina, Bulgaria, Croatia, Cyprus,
Czech Republic, Czechoslovakia, Denmark, Estonia, Finland, France, Georgia, Germany,
Germany West, Great Britain, Greece, Hungary, Iceland, Ireland, Italy, Kosovo, Latvia,
Lithuania, Macedonia, Malta, Moldova, Netherlands, Norway, Poland, Portugal, Romania,
Russia, Serbia, Slovakia, Slovenia, Spain, Sweden, Switzerland, Turkey, USSR, Ukraine and
Yugoslavia (Socialist Federal Rep.).
10
In line with our theoretical analysis, we focus on two main aspects of environmental
attitudes i.e. the preferences for the environment and the willingness of individuals to pay
for the environment. We proxy the individual preferences for the environment by means of
an ordered variable, which measures the extent of disagreement of the respondent to the
statement “Humans were meant to rule over the rest of nature". The variable takes the value
of 1 for “agree strongly", 2 for “agree", 3 for “disagree" and 4 for “disagree strongly". This
variable proxies the group speci…c disutility from pollution hi (p); (where i = B; G) in the
theoretical model. An increasing disagreement captures the individual belief that the nature
is valuable ‘per se’as opposed to the view that nature is valuable only insofar as it can be
exploited by humans to material purposes. Conversely, the stronger is the agreement with the
statement ‘humans are meant to rule over nature’, the smaller is the disutility from pollution
because the agent appreciates nature only to the extent that if he can exploit it for economics
reasons.
We measure the willingness to pay for environmental causes, by means of an ordered
variable, which measures the extent of agreement to the statement “I would give part of my
income if I were certain that the money would be used to prevent environmental pollution".
The variable takes the value of 1 for “strongly disagree", 2 for “disagree", 3 for "agree" and 4
for "strongly agree". This variable corresponds to the derivative of the utility function with
respect to pollution in the theoretical model. 3
Additional individual controls are available by the EVS such as age (18+), gender
(male-female), employment status (employed-unemployed), highest educational level (8 classi…cations: inadequately completed elementary education, completed elementary education,
incomplete secondary school (technical), complete secondary school (technical), incomplete
secondary school (university preparatory), complete secondary school (university preparatory), some university without degree, university with degree) and monthly household income.
In order to address the main purpose of the paper, we need to trace attitudes of
migrants that have left their country of origin and currently reside in a host country. We thus
focus only on the last wave of the EVS i.e. the 2008 wave, which is the only one to trace back
immigrants. We are able to distinguish …rst and second generation migrants, i.e. people who
were born in any of the 47 countries of our sample and currently reside in a di¤erent host
country, from people who were in the host country but whose parents have migrated to the
host country from another country. After excluding all people who do not provide information
about their country of origin, or their parents’country of origin and keeping all individuals
above the age of 18, we have a sample of 4295 migrants coming from the 47 countries. Of those
migrants, 2554 are …rst generation migrants whereas 1741 are second generation migrants. We
3
Given the level of pollution p, the higher the materialist attitude captured in the variable ‘humans are
meant to rule over nature’, namely the smaller h(p), the lower the marginal willigness to pay for the enviroment.
11
Table 1: Sample Summary Statistics
Va ria b le
(1 )
(2 )
(3 )
(4 )
(5 )
N um b er of
M ean
S ta n d a rd
M in
M ax
Obs
D e v ia tio n
W illin g n e ss to P ay fo r th e E nv iro n m e nt
4295
2 .2 2 8
0 .8 8 6
1
4
M e a n W illin g n e ss to P ay fo r th e E nv iro n m e nt - H o st C o u ntry
4295
2 .7 4 4
0 .2 7 5
2 .0 9 0 6 2 9
3 .2 2 2 1 3 9
B e lie f th a t H u m a n s R u le ove r N a tu re
4295
2 .8 7 2
0 .8 6 3
1
4
M e a n B e lie f th a t H u m a n s R u le ove r N a tu re - H o st C o u ntry
4295
2 .7 4 4
0 .2 7 5
2 .0 9 0 6 2 9
3 .2 2 2 1 3 9
A ge
4295
4 6 .8 8 5
1 7 .3 4 1
18
97
H ig h e st E d u c a tio n a l L e ve l
4295
5 .1 6 1
2 .0 1 4
1
8
M o nth ly In c o m e H o u se h o ld
4295
1 .9 4 3
0 .7 9 1
1
3
E m p loy m e nt
4295
0 .5 4 4
0 .4 9 8
0
1
Fe m a le
4295
0 .5 6 0
0 .4 9 6
0
1
also drop observations for which environmental preferences are missing or any other individual
control so as to ensure that our results are not being driven by sample selection.
We proxy environmental culture by the average attitudes of natives in the country
of origin. We use the same wave of the EVS, and compute the weighted average value of
environmental preferences and marginal willingness to pay of all natives.
Table 1 reports the demographic characteristics of the sample of migrants, as well as
the mean values for the environmental preferences of migrants. For instance, the mean age in
our sample is a migrant of 46 years, who has at least completed secondary education. More
than half or our sample of migrants are women and employed.
Tables 2 and 3 below report the migration ‡ows within the EVS sample of countries.
Columns (1)-(4) for both tables describe the migration ‡ows from each country. For instance in
our sample we have 54 individuals who have an Albanian origin. These migrants currently live
in 5 destination countries with the most prevalent destination country being Greece hosting 45
migrants out of the 54 in total. Columns (5)-(8) describe the migration patterns as viewed from
the host country perspective. For example in Switzerland currently reside 232 migrants from
24 countries from our sample, with the prevalent birth country of migrants being Germany
(51 migrants of German origin from our sample currently live in Switzerland).
12
13
(2 )
(3 )
12
3
7
16
16
10
E sto n ia
F in la n d
Fra n c e
G re a t B rita in
G e o rg ia
7
C z e ch o slova k ia
S p a in
4
S w itz e rla n d
6
8
B e la ru s
20
12
B o sn ia -H e rz e g ov in a
D e n m a rk
14
B u lg a ria
G e rm a ny
7
B e lg iu m
3
7
A z e rb a ija n
12
12
A u stria
C z e ch R e p u b lic
8
C y p ru s
5
46
91
126
53
10
57
29
226
44
8
14
6
122
283
40
102
62
31
67
54
C o u ntry
C o u ntry
A rm e n ia
D e stin a tio n
fro m B irth
C o u ntrie s
M o st
A rm e n ia
G re a t B rita in
L u x e m b o u rg
Sweden
Sweden
Fra n c e
N o rw ay
L u x e m b o u rg
S lova k ia
G re e c e
H u n g a ry
L u x e m b o u rg
L a tv ia
C ro a tia
Tu rke y
L u x e m b o u rg
A rm e n ia
S w itz e rla n d
A z e rb a ija n
G re e c e
P re va le nt
N um b er of
Im m ig ra nts
D istin c t
(4 )
29
21
76
38
5
21
9
61
18
6
5
3
60
116
20
86
49
13
59
45
D e stin . C o u ntry
P re va le nt
M ig ra nts to
N um b er of
Im m ig ra tio n F low s by B irth C o u ntry
D e stin a tio n
A lb a n ia
C o u ntry
(1 )
3
18
13
4
10
10
16
24
10
7
24
8
3
7
22
5
16
9
4
C o u ntrie s
B irth
(6 )
(7 )
(8 )
10
41
105
5
328
28
43
102
67
25
232
190
33
24
148
70
115
148
6
C o u ntry
in D e stin .
Im m ig ra nts
N um b er of
R u ssia
Ire la n d
Ita ly
R u ssia
R u ssia
R o m a n ia
G e rm a ny
Tu rke y
S lova k ia
R o m a n ia
G e rm a ny
R u ssia
S e rb ia
G re e c e
Ita ly
A rm e n ia
G e rm a ny
A z e rb a ija n
K o sovo
C o u ntry
B irth
P re va le nt
M o st
7
12
26
2
236
14
17
18
50
7
51
123
22
8
48
59
23
49
3
P re va le nt C o u ntry
fro m M o st
Im m ig ra nts
N um b er
Im m ig ra tio n F low s by C o u ntry o f D e stin a tio n
D istin c t
(5 )
Table 2: Migration Flows
14
4
16
8
Ic e la n d
Ita ly
K o sovo
5
8
19
23
Sweden
Tu rke y
U k ra in e
5
S lova k ia
S love n ia
27
20
21
R o m a n ia
S e rb ia
9
P o rtu g a l
R u ssia
7
N e th e rla n d s
27
9
M a lta
P o la n d
1
M o ld ova
N o rw ay
7
14
M a c e d o n ia
8
6
Ire la n d
10
14
H u n g a ry
L a tv ia
14
L ithu a n ia
14
232
383
21
18
64
191
786
77
223
123
13
48
2
45
18
21
36
35
240
5
24
59
109
51
C o u ntry
C o u ntry
C ro a tia
D e stin a tio n
fro m B irth
C o u ntrie s
M o st
B e la ru s
C y p ru s
N o rw ay
C ro a tia
C z e ch R e p u b lic
M o nte n e g ro
E sto n ia
S p a in
L u x e m b o u rg
G e rm a ny
Sweden
B e lg iu m
G re a t B rita in
R u ssia
U k ra in e
E sto n ia
L a tv ia
M a c e d o n ia
L u x e m b o u rg
N o rw ay
G re a t B rita in
S lova k ia
R u ssia
M a c e d o n ia
P re va le nt
N um b er of
Im m ig ra nts
D istin c t
D e stin a tio n
G re e c e
C o u ntry
N um b er of
P re va le nt
M ig ra nts to
52
172
10
7
50
54
236
14
171
17
5
18
2
14
9
11
19
9
77
2
12
21
41
11
D e stin . C o u ntry
Im m ig ra tio n F low s by B irth C o u ntry
11
6
27
14
9
8
6
5
4
5
19
14
5
9
7
11
5
6
12
4
9
8
14
C o u ntrie s
B irth
D istin c t
180
43
120
99
54
123
43
8
5
19
62
39
19
43
76
281
64
6
24
21
32
148
146
C o u ntry
in D e stin .
Im m ig ra nts
N um b er of
R u ssia
B u lg a ria
F in la n d
B o sn ia
H u n g a ry
B o sn ia
U k ra in e
M o ld ova
R o m a n ia
R u ssia
Sweden
G e rm a ny
G re a t B rita in
S e rb ia
U k ra in e
R u ssia
R u ssia
G e rm a ny
G e rm a ny
G re a t B rita in
R o m a n ia
B o sn ia
Tu rke y
C o u ntry
B irth
P re va le nt
M o st
144
20
38
35
21
57
26
3
2
7
10
15
15
17
39
144
28
1
5
14
12
116
69
P re va le nt C o u ntry
fro m M o st
Im m ig ra nts
N um b er
Im m ig ra tio n F low s by C o u ntry o f D e stin a tio n
Table 3: Migration Flows (ctd)
3.2
Empirical strategy
We follow the recent literature on the economic e¤ects of culture (Luttmer and Singhal (11),
Fernandez and Fogli (9)) and estimate speci…cations of the following form for immigrant
preferences:
ni = Hb + Xi + r + i ;
(6)
where ni denotes immigrant i’s attitudes towards the environment, Hb is the cultural
attitude towards the environment among natives in immigrant i’s birth country b, Xi is a
vector of individual characteristics, r is a …xed e¤ect for residence country r of immigrant i,
and i denotes the error term. Our coe¢ cient of interest is . If the attitudes towards the
environment of the immigrant were only a¤ected by the relevant economic and institutional
factors in the country of residence, we should expect = 0. However a signi…cantly di¤erent
from 0 signals an e¤ect of culture on the environmental preferences of the immigrant. Our
identifying assumption is that there are no omitted factors correlated with environmental
preferences in the birth region other than culture that a¤ect immigrant’s preferences in the
country of destination. Based on this assumption, estimates of by ordinary least squares
(OLS) can be given a causal interpretation (Luttmer and Singhal (11), Fernandez and Fogli
(9)).
Measurement of Hb and endogeneity issues. Luttmer and Singhal (11) and Fernandez
and Fogli (9) discuss extensively the issues of endogeneity that arise in a speci…cation such as
equation (6). There is a primary issue of sample selection since workers with strong preferences
for the environment can migrate out of countries characterized by little environmental protection. If the environmental di¤erences were the only driving force for migration, all individuals
would move to countries exactly aligned with their own environmental preferences. Then, we
would not be able to estimate any e¤ect of culture even if cultural e¤ects were in fact strong.
However, we believe that selective migration is not an issue in our estimates for at least three
reasons. This type of selective migration is already limited for preferences for redistribution
in ESS and EVS type of survey data (see Luttmer and Singhal (11)). Moreover, it is well
known that the main driving force of migration is the income di¤erence between the origin and
the destination country. Hence, selection is economically motivated selective migration. This
argument is enough to rule out selection for environmental preferences provided that lower
income countries are not the more polluted ones.4 Finally, following Fernandez and Fogli (9),
focussing on 2nd generation migrants provides a tool to minimize selective migration since
the 2nd generation migration status is only determined by parents’migration decision, thus
4
It is well-known that income and pollution across countries follows a non-monotonic relationship known
as the Kuznets curve. Hence, it is likely that migrants move for income reasons and often they live behind
cleaner countries to reside in more polluting ones.
15
exogenous with respect to migrants’attitudes towards environmental conditions in the country
of origin.
There is a second issue of omitted variable bias in so far as preferences of 2nd generation
migrants can be a¤ected by other factors than culture (e.g. social segregation in highly
polluted areas of the destination country), which may in‡uence their environmental preferences
(Algan et al. (2)). Following Luttmer and Singhal (11) and Fernandez and Fogli (9), we
assume that the set of personal, demographic and family characteristics available in the data
fully captures the e¤ect of such unobserved factors. We also carry out an extensive set of
robustness checks to control for any omitted factor which may confound our main estimates.
3.3
Results
Table 4 and Table 5 report estimates for the impact of culture on environmental preferences,
and the marginal willingness to pay for the environment, respectively. In column [1] of each
table, we only include the host country dummies. In column [2] we add controls for income in
the country of birth (measured by the log of purchasing power parity adjusted GDP in 2000)
as well as relevant demographic, socioeconomic and household characteristics. Finally, in
column [3] we add an indicator of environmental quality in the country of birth (nitrous oxide
emissions, in logs). Results in Table 4 show that an one unit increase in the mean preference
for the environment in the individual’country of birth is associated with a 0:18 0:25 unit
increase in the individual’own preference for the environment. Similarly, an one unit increase
in the mean marginal willingness to pay in the individual’country of birth is associated with
a 0:14 0:19 unit increase in the individual’own marginal willingness to pay.
16
17
All Immigrants
First Generation Second Generation
[1]
[2]
[3]
[4]
[5]
[6]
[7]
Birth Country: Humans Rule Over Nature
0.25*** 0.16*** 0.18*** 0.19**
0.21**
0.13*
0.13*
(0.06)
(0.05)
(0.05)
(0.08)
(0.08)
(0.07)
(0.07)
Birth Country: GDP per Capita (log)
0.06*** 0.05**
0.05**
0.04
0.04*
0.04
(0.02)
(0.02)
(0.02)
(0.02)
(0.03)
(0.03)
Birth Country: Nitrous Oxide Emissions (log)
0.02**
0.04**
0.00
(0.01)
(0.01)
(0.01)
Age
0.01*
0.01*
0.01
0.01
0.01
0.01
(0.00)
(0.00)
(0.01)
(0.01)
(0.01)
(0.01)
Age Squared
–0.00** –0.00** –0.00** –0.00** –0.00
–0.00
(0.00)
(0.00)
(0.00)
(0.00)
(0.00)
(0.00)
Female
0.08*** 0.08*** 0.11*** 0.10***
0.03
0.03
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
Secondary education
0.09**
0.09**
0.09*
0.08*
0.09
0.09
(0.04)
(0.04)
(0.05)
(0.05)
(0.07)
(0.07)
Tertiary Education
0.12**
0.11*
0.09
0.08
0.21**
0.21**
(0.06)
(0.06)
(0.05)
(0.05)
(0.09)
(0.09)
Employed
–0.03
–0.03
–0.02
–0.02
–0.04
–0.04
(0.04)
(0.04)
(0.04)
(0.04)
(0.06)
(0.06)
Household Income (log)
0.05*** 0.05*** 0.07*** 0.07***
0.03
0.03
(0.02)
(0.02)
(0.03)
(0.03)
(0.03)
(0.03)
R sq.
0.09
0.10
0.10
0.10
0.10
0.14
0.14
N
4295
4295
4295
2554
2554
1741
1741
Host Country FE
yes
yes
yes
yes
yes
yes
yes
Summary:This table establishes that the mean belief at country of origin on whether humans
should rule over nature is positively correlated with the corresponding individual belief of the
migrant residing at the host country.
The analysis controls for individual characteristics,
income per capita at home country and mean level of pollution at the home country..
Notes:(i) Robust standard error estimates are reported in parentheses; (ii) *** denotes statistical signi…cance at
the 1 percent level, ** at the 5 percent level, and * at the 10 percent level, all for two-sided hypothesis tests.
Table 4: Predictors of Individual Preferences Towards the Environment Predictors of the Marginal
While columns [1]-[3] report estimates for all immigrant types, we also carried out
separate estimates for …rst generation immigrants (reported in columns [4] and [5] of each
table) and second generation immigrants (reported in columns [6] and [7] of each table).
The coe¢ cient of mean environmental attitudes in the country of birth remain positive and
signi…cant for both immigrant categories. In particular, the positive and signi…cant coe¢ cient
for the second generation migrants is reassuring that a cultural transmission mechanism is at
work, and that results in columns [1]-[3] of Table 4 and Table 5 are not driven by selective
migration.
Estimates in Table 4, show that nitrous oxide emissions in the country of birth are
associated with a higher preference for the environment of immigrants (Table 4, column [3]).
However, this positive impact is driven by …rst generation immigrants (Table 1, column [5]),
while emissions in the birth country do not have any impact on the preferences of second
generation immigrants (Table 4, column [7]). This …nding indicates that some inertia of
preferences exists such that pollution in the country of origin increases individual preferences
towards the environment, even after individuals migrate to a di¤erent country. However this
e¤ect is not transmitted to the next generation, which is only exposed to the local environmental conditions in the host country. This inertia of preferences e¤ect is not transmitted
from preferences to marginal willingness to pay. Results in Table 5 (column [3], [5], [7]) show
that pollution in the birth country does not hold any signi…cant association with the marginal
willingness to pay.
18
19
All Immigrants
First Generation Second Generation
[1]
[2]
[3]
[4]
[5]
[6]
[7]
Birth Country: Marginal Willingness to Pay
0.14*** 0.19*** 0.19*** 0.20*** 0.20*** 0.17**
0.16*
(0.05)
(0.05)
(0.04)
(0.06)
(0.06)
(0.08)
(0.08)
Birth Country: GDP per Capita (log)
–0.01
–0.01
–0.01
–0.01
0.00
–0.01
(0.02)
(0.02)
(0.02)
(0.02)
(0.02)
(0.02)
Birth Country: Nitrous Oxide Emissions (log)
–
0.00
–
0.00
–
0.02
(0.01)
(0.01)
(0.01)
Age
–0.00
–0.00
–0.00
–0.00
0.00
0.00
(0.01)
(0.01)
(0.01)
(0.01)
(0.01)
(0.01)
Age Squared
–0.00
–0.00
0.00
0.00
–0.00
–0.00
(0.00)
(0.00)
(0.00)
(0.00)
(0.00)
(0.00)
Female
0.04
0.04
0.06*
0.06*
0.02
0.02
(0.03)
(0.03)
(0.04)
(0.04)
(0.04)
(0.04)
Secondary education
0.10*** 0.10*** 0.10**
0.10**
0.09*
0.09*
(0.03)
(0.03)
(0.05)
(0.05)
(0.05)
(0.05)
Tertiary Education
0.25*** 0.25*** 0.26*** 0.26*** 0.23*** 0.23***
(0.04)
(0.04)
(0.06)
(0.06)
(0.07)
(0.07)
Employed
0.00
0.00
0.03
0.03
–0.03
–0.03
(0.03)
(0.03)
(0.05)
(0.05)
(0.04)
(0.04)
Household Income (log)
0.08*** 0.08*** 0.06**
0.06** 0.11*** 0.11***
(0.02)
(0.02)
(0.02)
(0.02)
(0.03)
(0.03)
R sq.
0.09
0.10
0.10
0.10
0.10
0.14
0.14
N
4295
4295
4295
2554
2554
1741
1741
Host Country FE
yes
yes
yes
yes
yes
yes
yes
Summary:This table establishes that the mean willingness to pay for the environment at
the country of origin is positively correlated with the corresponding individual belief of the
migrant residing at the host country.
The analysis controls for individual characteristics,
income per capita at home country and mean level of pollution at the home country..
Notes:(i) Robust standard error estimates are reported in parentheses; (ii) *** denotes statistical signi…cance at
the 1 percent level, ** at the 5 percent level, and * at the 10 percent level, all for two-sided hypothesis tests.
Table 5: Unwillingness to Pay for the Environment
Turning to controls, households’income increases the preferences towards the environment and the will to spend more to protect it. Females display a higher preference for the
environment, and are relatively more willing to protect it relatively to males. On a similar vein
and individuals with secondary or tertiary education have stronger environmental attitudes
as compared with individuals that only completed primary education. Age and employment
status do not display any signi…cant association with environmental attitudes. Finally the
log of purchasing power parity adjusted GDP in the birth country is included as a control to
capture economic di¤erences across countries, which may result in di¤erent preferences for the
environment. Size and signi…cance of the coe¢ cients is very similar to those of nitrous oxide
emissions. This suggests that the two variables have a very similar impact on environmental
attitudes. The countries with higher GDP levels are also the more polluted ones, thus a¤ect
individual preferences in the way described above.
3.4
Robustness
This section subjects our baseline analysis to a number of robustness tests. In particular we
address the issue of selection by limiting our analysis to second generation migrants.5 We
introduce additional controls to the baseline analysis such as parental and spouse education
and employment status. Finally, we cluster the standard errors both at the dimension of the
host and the origin country level. Reassuringly our results are robust to all these additional
tests, thereby con…rming the presence of an environmental culture.
3.4.1
Parental and Spouse Controls
This section explores the robustness of the analysis to a number of parental and spousal
controls. As the number of observations drops signi…cantly these controls are not employed
in the baseline analysis, however Table 6 establishes the robustness of our results. Columns
(1)-(3) explore the e¤ect of mean preferences on the willingness to pay at the home country
on the individual willingness to pay of the migrants. In particular Column (1) introduces in
the baseline analysis three more parental controls, e.g. educational level of father/mother,
whether the respondents’ parents were employed at the age of 16 and their occupational
status classi…ed under 7 categories. Similarly, Column (2) introduces in the baseline analysis
three spousal controls, e.g. educational level of spouse, the occupational status of the spouse
classi…ed under 7 categories and whether the spouse has experience unemployment for more
than three months. Finally Column (3) combines both sets of controls.
Columns (4)-(6) repeat the same steps for the beliefs of individuals on whether humans
rule over nature. Reassuringly the results in both cases remain intact to the addition of all
these controls despite the signi…cant reduction of the sample.
5
Results about the …rst generation migrants are reported as well for the sake of comparative analysis.
20
Table 6: Robustness: Parental and Spouse Controls
(1 )
(2 )
(3 )
H o st C o u ntry : M a rg in a l W illin g n e ss to P ay
B irth C o u ntry : M a rg in a l W illin g n e ss to P ay
E d u c a tio n a l L e ve l o f P a re nts
P a re nt E m p loye d w h e n R e sp o n d e nt w a s 1 6
O c c u p a tio n a l S ta tu s o f P a re nts
(5 )
(6 )
B irth C o u ntry : H u m a n s R u le O ve r N a tu re
0 .2 1 9 * * *
0 .1 8 2 * * *
0 .2 8 0 * * *
0 .1 7 6 * *
0 .1 0 0 * *
0 .1 4 6 * * *
(0 .0 5 3 )
(0 .0 6 7 )
(0 .0 7 5 )
(0 .0 6 7 )
(0 .0 4 0 )
(0 .0 5 3 )
0 .0 1 9
-0 .0 0 6
0 .0 1 4
0 .0 2 7
(0 .0 4 4 )
(0 .0 3 9 )
(0 .0 2 7 )
(0 .0 4 7 )
-0 .0 5 5
-0 .0 8 7
-0 .0 6 5 *
-0 .0 8 7
(0 .0 4 7 )
(0 .0 5 8 )
(0 .0 3 8 )
(0 .0 6 4 )
-0 .0 1 0 *
-0 .0 1 6 * *
-0 .0 0 0
-0 .0 0 1
(0 .0 0 5 )
(0 .0 0 6 )
(0 .0 0 6 )
(0 .0 0 8 )
E d u c a tio n a l L e ve l o f S p o u se
H a s S p o u se b e e n u n e m p loye d m o re th a n 6 m o nth s
O c c u p a tio n a l S ta tu s o f S p o u se
O b se rva tio n s
(4 )
0 .0 9 3 *
0 .0 9 0 *
0 .0 2 1
0 .0 1 4
(0 .0 4 9 )
(0 .0 4 5 )
(0 .0 4 4 )
(0 .0 4 3 )
0 .0 2 1
0 .0 1 5
0 .0 3 2
0 .0 3 6
(0 .0 4 9 )
(0 .0 5 1 )
(0 .0 4 3 )
(0 .0 4 2 )
-0 .0 0 1
0 .0 0 6
-0 .0 1 4 * *
-0 .0 1 4 * *
(0 .0 0 8 )
(0 .0 0 8 )
(0 .0 0 5 )
(0 .0 0 6 )
Ye s
Ye s
Ye s
Ye s
Ye s
Ye s
R -sq u a re d
3413
2301
1856
3413
2301
1856
H o st C o u ntry F E
0 .1 0 6
0 .1 1 8
0 .1 2 6
0 .0 9 9
0 .0 8 5
0 .0 9 0
Summary:This table establishes that the robustness of the baseline analysis while controlling for parental
and
spouse
characteristics.
Notes:(i) Robust standard error estimates are reported in parentheses; (ii) *** denotes statistical
signi…cance at the 1 percent level, ** at the 5 percent level, and * at the 10 percent level, all for two-sided
hypothesis tests.
21
3.4.2
Double Clustering
To account for the fact that correlation may arise both at the host and the home country
dimension we follow the approach of Cameron et al. (2011) and rely on two-way clustering
of the standard errors on the dimension of the host country and the country of origin. Their
estimator allows for an arbitrary correlation between errors belonging to the same group
thereby being applicable for cases where spatial correlation is a potential concern. The results
of this approach are also robust to two-way clustering as well as suggested by Table 7.
Columns (1)-(3) explore the e¤ect of mean preferences on the willingness to pay at
the home country on the individual willingness to pay of the migrants, whereas Columns (4)(6) repeat the same steps for the beliefs of individuals on whether humans rule over nature.
Reassuringly the results remain robust. In particular Column (1) explore the e¤ect of mean
preference at the home country on individual preferences of migrants while controlling only
for host country …xed e¤ects. Column (2) augments the analysis by introducing the full set
of individual controls (e.g. age, education, gender and employment status) whereas Column
(3) further introduces controls associated with the country of origin (e.g. GDP per capita
and local pollution levels). The results remain largely una¤ected with the exception of those
in Column (1) where the signi…cance of the coe¢ cient drops to an increase in the standard
errors, yet it is restore in Column (2).
3.4.3
Robustness to Sample Selection
To maximize the number of observations the baseline analysis has made the assumption that
the country of origin is determined by the father’s country of origin. However it could be
argues that if the respondent’s father and mother come from a di¤erent country of origin
or one of them is native then the transmission of cultural traits from the country of origin
could be a¤ected by mixing the two cultures (if parents come from di¤erent countries other
than the host country), or our results could be attenuated if one of the parents is native.
Columns (1)-(3) of Table 8 explore the e¤ect of mean preferences on the willingness to pay
at the home country on the individual willingness to pay of the migrants, whereas Columns
(4)-(6) repeat the same steps for the beliefs of individuals on whether humans rule over
nature. In particular Column (1) explores the e¤ect of mean preference at the home country
on individual preferences of migrants while controlling only for host country …xed e¤ects.
Column (2) augments the analysis by introducing the full set of individual controls (e.g. age,
education, gender and employment status) whereas Column (3) further introduces controls
associated with the country of origin (e.g. GDP per capita and local pollution levels).
Our results remain robust to this assumption, particularly after adding individual
controls despite the sharp reduction in our sample.
22
Table 7: Robsustness: Double Clustering
(1 )
(2 )
(3 )
(4 )
H o st C o u ntry : M a rg in a l W illin g n e ss to P ay
B irth C o u ntry : M a rg in a l W illin g n e ss to P ay
0 .1 3 7
0 .1 6 8 * *
0 .1 8 5 * *
(0 .0 8 9 )
(0 .0 8 2 )
(0 .0 7 8 )
B irth C o u ntry : H u m a n s R u le O ve r N a tu re
B irth C o u ntry : G D P p e r C a p ita (lo g )
B irth C o u ntry : N itro u s O x id e E m issio n s (lo g )
A ge
A g e S q u a re d
Fe m a le
S e c o n d a ry e d u c a tio n
Te rtia ry E d u c a tio n
E m p loye d
H o u se h o ld In c o m e (lo g )
O b se rva tio n s
(5 )
(6 )
B irth C o u ntry : H u m a n s R u le O ve r N a tu re
0 .2 5 2 * * *
0 .2 2 5 * * *
(0 .0 6 4 )
(0 .0 6 3 )
0 .1 8 0 * * *
(0 .0 4 4 )
0 .0 1 2
0 .0 4 9 * *
(0 .0 2 3 )
(0 .0 2 4 )
-0 .0 0 2
0 .0 2 3
(0 .0 1 5 )
(0 .0 1 6 )
-0 .0 0 0
-0 .0 0 0
0 .0 0 6
0 .0 0 6
(0 .0 0 9 )
(0 .0 0 9 )
(0 .0 0 5 )
(0 .0 0 6 )
-0 .0 0 0
-0 .0 0 0
-0 .0 0 0
-0 .0 0 0
(0 .0 0 0 )
(0 .0 0 0 )
(0 .0 0 0 )
(0 .0 0 0 )
0 .0 4 1
0 .0 4 1
0 .0 7 6 * *
0 .0 7 7 * *
(0 .0 3 5 )
(0 .0 3 7 )
(0 .0 3 1 )
(0 .0 3 0 )
0 .0 9 8 *
0 .0 9 8 *
0 .0 9 0 *
0 .0 8 6 *
(0 .0 5 1 )
(0 .0 5 2 )
(0 .0 4 7 )
(0 .0 4 5 )
0 .2 4 9 * * *
0 .2 5 0 * * *
0 .1 1 8 * *
0 .1 1 1 * *
(0 .0 5 7 )
(0 .0 5 7 )
(0 .0 5 7 )
(0 .0 5 6 )
0 .0 0 2
0 .0 0 2
-0 .0 3 2
-0 .0 2 9
(0 .0 4 6 )
(0 .0 4 7 )
(0 .0 5 8 )
(0 .0 5 7 )
0 .0 8 3 * * *
0 .0 8 2 * * *
0 .0 5 7 * *
0 .0 5 3 * *
(0 .0 2 6 )
(0 .0 2 9 )
(0 .0 2 5 )
(0 .0 2 5 )
Ye s
Ye s
Ye s
Ye s
Ye s
Ye s
R -sq u a re d
4295
4295
4295
4295
4295
4295
H o st C o u ntry F E
0 .0 8 7
0 .1 0 3
0 .1 0 3
0 .0 8 1
0 .0 9 2
0 .0 9 5
Summary:This table establishes the robustness of the baseline analysis to two way clustering of the standard errors at the dimension of the origin and of the host country.
Notes:(i) Robust standard error estimates are reported in parentheses; (ii) *** denotes statistical
signi…cance at the 1 percent level, ** at the 5 percent level, and * at the 10 percent level, all for twosided hypothesis tests.
23
Table 8: Robsustness: Both Parents from the Same Country
(1 )
(2 )
(3 )
(4 )
H o st C o u ntry : M a rg in a l W illin g n e ss to P ay
B irth C o u ntry : M a rg in a l W illin g n e ss to P ay
0 .1 0 1
0 .1 4 8 * *
0 .1 4 7 * *
(0 .0 6 4 )
(0 .0 5 6 )
(0 .0 6 2 )
B irth C o u ntry : H u m a n s R u le O ve r N a tu re
B irth C o u ntry : G D P p e r C a p ita (lo g )
(5 )
B irth C o u ntry : H u m a n s R u le O ve r N a tu re
0 .2 4 0 * * *
0 .2 0 2 * * *
(0 .0 7 3 )
(0 .0 7 0 )
0 .0 0 3
B irth C o u ntry : N itro u s O x id e E m issio n s (lo g )
A g e S q u a re d
Fe m a le
0 .1 6 9 * *
(0 .0 7 3 )
0 .0 5 0 * *
(0 .0 2 1 )
(0 .0 2 4 )
-0 .0 0 6
0 .0 4 4 * * *
(0 .0 1 3 )
A ge
(6 )
(0 .0 1 4 )
-0 .0 0 7
-0 .0 0 7
0 .0 1 0 * *
0 .0 1 0 * *
(0 .0 0 7 )
(0 .0 0 7 )
(0 .0 0 5 )
(0 .0 0 5 )
0 .0 0 0
0 .0 0 0
-0 .0 0 0 * * *
-0 .0 0 0 * * *
(0 .0 0 0 )
(0 .0 0 0 )
(0 .0 0 0 )
(0 .0 0 0 )
0 .0 3 6
0 .0 3 7
0 .0 7 3 * *
0 .0 7 1 * *
(0 .0 3 4 )
(0 .0 3 5 )
(0 .0 3 6 )
(0 .0 3 5 )
S e c o n d a ry e d u c a tio n
0 .1 1 8 * * *
0 .1 1 8 * * *
0 .0 9 9 * *
0 .0 9 1 * *
(0 .0 3 3 )
(0 .0 3 2 )
(0 .0 4 5 )
(0 .0 4 3 )
Te rtia ry E d u c a tio n
0 .2 7 9 * * *
0 .2 8 1 * * *
0 .1 4 6 * * *
0 .1 2 6 * * *
(0 .0 4 4 )
(0 .0 4 4 )
(0 .0 4 5 )
(0 .0 4 5 )
-0 .0 0 3
-0 .0 0 3
-0 .0 5 0
-0 .0 4 9
(0 .0 3 9 )
(0 .0 4 0 )
(0 .0 3 8 )
(0 .0 3 6 )
0 .0 8 3 * * *
0 .0 8 3 * * *
0 .0 4 8 * *
0 .0 4 1 * *
(0 .0 2 1 )
(0 .0 2 1 )
(0 .0 1 9 )
(0 .0 1 9 )
E m p loye d
H o u se h o ld In c o m e (lo g )
O b se rva tio n s
Ye s
Ye s
Ye s
Ye s
Ye s
Ye s
R -sq u a re d
2813
2813
2813
4295
4295
4295
H o st C o u ntry F E
0 .0 8 3
0 .1 0 1
0 .1 0 1
0 .0 8 1
0 .0 9 2
0 .0 9 5
Summary:This table establishes the robustness of the baseline analysis to limiting
the sample to migrants whose both parents come from the same country of origin.
Notes:(i) Robust standard error estimates are reported in parentheses; (ii) *** denotes statistical
signi…cance at the 1 percent level, ** at the 5 percent level, and * at the 10 percent level, all for two-sided
hypothesis tests.
24
Table 9: Robsustness: Both Parents are Migrants
(1 )
(2 )
(3 )
(4 )
H o st C o u ntry : M a rg in a l W illin g n e ss to P ay
B irth C o u ntry : M a rg in a l W illin g n e ss to P ay
0 .1 1 6 * *
0 .1 4 8 * *
0 .1 4 7 * *
(0 .0 4 8 )
(0 .0 5 6 )
(0 .0 6 2 )
B irth C o u ntry : H u m a n s R u le O ve r N a tu re
B irth C o u ntry : G D P p e r C a p ita (lo g )
(5 )
B irth C o u ntry : H u m a n s R u le O ve r N a tu re
0 .2 3 2 * * *
0 .2 0 2 * * *
(0 .0 5 8 )
(0 .0 7 0 )
0 .0 0 3
B irth C o u ntry : N itro u s O x id e E m issio n s (lo g )
A g e S q u a re d
Fe m a le
0 .1 6 9 * *
(0 .0 7 3 )
0 .0 5 0 * *
(0 .0 2 1 )
(0 .0 2 4 )
-0 .0 0 6
0 .0 4 4 * * *
(0 .0 1 3 )
A ge
(6 )
(0 .0 1 4 )
-0 .0 0 7
-0 .0 0 7
0 .0 1 0 * *
0 .0 1 0 * *
(0 .0 0 7 )
(0 .0 0 7 )
(0 .0 0 5 )
(0 .0 0 5 )
0 .0 0 0
0 .0 0 0
-0 .0 0 0 * * *
-0 .0 0 0 * * *
(0 .0 0 0 )
(0 .0 0 0 )
(0 .0 0 0 )
(0 .0 0 0 )
0 .0 3 6
0 .0 3 7
0 .0 7 3 * *
0 .0 7 1 * *
(0 .0 3 4 )
(0 .0 3 5 )
(0 .0 3 6 )
(0 .0 3 5 )
S e c o n d a ry e d u c a tio n
0 .1 1 8 * * *
0 .1 1 8 * * *
0 .0 9 9 * *
0 .0 9 1 * *
(0 .0 3 3 )
(0 .0 3 2 )
(0 .0 4 5 )
(0 .0 4 3 )
Te rtia ry E d u c a tio n
0 .2 7 9 * * *
0 .2 8 1 * * *
0 .1 4 6 * * *
0 .1 2 6 * * *
(0 .0 4 4 )
(0 .0 4 4 )
(0 .0 4 5 )
(0 .0 4 5 )
-0 .0 0 3
-0 .0 0 3
-0 .0 5 0
-0 .0 4 9
(0 .0 3 9 )
(0 .0 4 0 )
(0 .0 3 8 )
(0 .0 3 6 )
0 .0 8 3 * * *
0 .0 8 3 * * *
0 .0 4 8 * *
0 .0 4 1 * *
(0 .0 2 1 )
(0 .0 2 1 )
(0 .0 1 9 )
(0 .0 1 9 )
E m p loye d
H o u se h o ld In c o m e (lo g )
O b se rva tio n s
Ye s
Ye s
Ye s
Ye s
Ye s
Ye s
R -sq u a re d
3179
2813
2813
3179
2813
2813
H o st C o u ntry F E
0 .0 8 4
0 .1 0 1
0 .1 0 1
0 .0 8 0
0 .0 9 1
0 .0 9 7
Summary:This
table
establishes
the
robustness
of
the
baseline
analysis
to
limiting
the
sample
to
migrants
whose
both
parents
are
migrants..
Notes:(i) Robust standard error estimates are reported in parentheses; (ii) *** denotes statistical
signi…cance at the 1 percent level, ** at the 5 percent level, and * at the 10 percent level, all for two-sided
hypothesis tests.
Similarly Table 9 explores the assumption that both parents are migrants but may not
come from the same country. The table has a similar structure with the previous one and
further con…rms the …ndings of the baseline analysis.
3.4.4
Assimilation of Migrants
As already analyzed in the main text the baseline analysis builds upon the …ll sample of
migrants in order to bene…t from the increased number of observation. As the analysis of the
sample of second generation migrants suggests the results are robust to potential selection
concerns and con…rm the transferability of preferences. To further reinforce this result the
analysis controls for a number of individual characteristics that capture the assimilation of
the …rst generation of migrants. Table 10 con…rms our main …nding for the willingness to pay
for the environment after controlling for a number of characteristics that capture the degree
25
Table 10: Robustness: Assimilation of First Generation Migrants
(1 )
(2 )
(3 )
(4 )
H o st C o u ntry : M a rg in a l W illin g n e ss to P ay
B irth C o u ntry : M a rg in a l W illin g n e ss to P ay
Ye a r S p e nt in th e C o u ntry
0 .2 3 7 * * *
0 .1 7 7 * * *
0 .1 8 4 * * *
0 .2 3 4 * * *
(0 .0 6 8 )
(0 .0 4 5 )
(0 .0 4 3 )
(0 .0 6 9 )
0 .0 0 0
0 .0 0 1
(0 .0 0 2 )
(0 .0 0 2 )
C itiz e n sh ip
0 .0 9 2 * *
0 .1 2 8 * *
(0 .0 4 0 )
Im p o rta nt to S p e a k th e L a n g u a g e
O b se rva tio n s
(0 .0 4 8 )
-0 .0 3 0
-0 .0 3 3
(0 .0 2 1 )
(0 .0 3 7 )
Ye s
Ye s
Ye s
Ye s
R -sq u a re d
2453
2453
2453
2453
H o st C o u ntry F E
0 .0 9 9
0 .1 0 4
0 .1 0 4
0 .1 0 2
Summary:This table establishes the robustness of the baseline
when accounting for the assimilation of …rst generation migrants
Notes:(i) Robust standard error estimates are reported in parentheses;
(ii) *** denotes statistical signi…cance at the 1 percent level, ** at the 5
percent level, and * at the 10 percent level, all for two-sided hypothesis
tests.
of assimilation. In particular, Column (1) controls for the number of years a migrant has
spent in the country, Column (2) for whether the migrant has been granted citizenship or not,
Column (3) how important he consider to be able to speak the native language and Column
(4) for all the above. Whereas obtaining citizenship is crucial, however our results are largely
una¤ected.
Table 11 replicates the same analysis for the variable on the belief that humans rule
over nature and …nds similar results.
3.4.5
The Role of Preferences
This section explores whether a range of other preferences of the individual are a¤ecting her
individual preferences. In particular Table 12 replicates the baseline analysis for willingness
to pay of the migrants and consecutively introduces in Columns (1), (2), (3) controls for
individual preferences on trust to environmental organizations, religious denomination and
their belief as to whether people should stick to their own a¤airs. Column (4) combines all
the above mentioned controls. Whereas the analysis suggests that some of these controls are
signi…cant, nevertheless our results remain intact.
Table 13 replicates the same analysis for beliefs on whether humans should rule over
nature and the …ndings remain intact.
26
Table 11: Robustness: Assimilation of First Generation Migrants
(1 )
(2 )
(3 )
(4 )
H o st C o u ntry : M a rg in a l W illin g n e ss to P ay
B irth C o u ntry : H u m a n s R u le ove r N a tu re
Ye a r S p e nt in th e C o u ntry
0 .1 7 9 * *
0 .1 8 5 * * *
0 .1 8 1 * * *
0 .1 9 1 * *
(0 .0 8 2 )
(0 .0 5 0 )
(0 .0 5 2 )
(0 .0 8 7 )
0 .0 0 0
0 .0 0 0
(0 .0 0 2 )
(0 .0 0 2 )
C itiz e n sh ip
0 .1 0 1 * * *
0 .0 3 2
(0 .0 3 2 )
(0 .0 3 8 )
Im p o rta nt to S p e a k th e L a n g u a g e
O b se rva tio n s
0 .0 4 2 *
0 .0 4 7
(0 .0 2 2 )
(0 .0 3 5 )
Ye s
Ye s
Ye s
Ye s
R -sq u a re d
2453
2453
2453
2453
H o st C o u ntry F E
0 .0 9 1
0 .0 9 7
0 .0 9 7
0 .0 9 3
Summary:This table establishes the robustness of the baseline
when accounting for the assimilation of …rst generation migrants
Notes:(i) Robust standard error estimates are reported in parentheses;
(ii) *** denotes statistical signi…cance at the 1 percent level, ** at the 5
percent level, and * at the 10 percent level, all for two-sided hypothesis
tests.
Table 12: Robsustness: Willingness to Pay and Individual Preferences
(1 )
(2 )
(3 )
(4 )
H o st C o u ntry : M a rg in a l W illin g n e ss to P ay
B irth C o u ntry : M a rg in a l W illin g n e ss to P ay
Tru st in E nv iro n m e nta l O rg a n iz a tio n s
0 .1 6 5 * * *
0 .1 8 8 * *
0 .1 7 0 * * *
(0 .0 4 9 )
(0 .0 7 1 )
(0 .0 4 5 )
-0 .1 6 5 * * *
(0 .0 2 3 )
-0 .0 0 0
-0 .0 0 1
(0 .0 0 4 )
(0 .0 0 3 )
P e o p le S h o u ld S tick to T h e ir O w n A ¤a irs
O b se rva tio n s
(0 .0 5 8 )
-0 .1 4 4 * * *
(0 .0 2 3 )
B e lo n g in g to a R e lig io u s D e n o m in a tio n
0 .1 2 8 * *
0 .0 2 6
0 .0 0 5
(0 .0 1 7 )
(0 .0 2 1 )
Ye s
Ye s
Ye s
Ye s
R -sq u a re d
4052
3057
4229
2838
H o st C o u ntry F E
0 .1 2 3
0 .1 0 9
0 .1 0 5
0 .1 2 2
Summary:This table establishes the robustness of the baseline analysis
to controlling for a number of alternative preferences of the migratns
Notes:(i) Robust standard error estimates are reported in parentheses; (ii)
*** denotes statistical signi…cance at the 1 percent level, ** at the 5 percent
level, and * at the 10 percent level, all for two-sided hypothesis tests.
27
Table 13: Robsustness: Humans Rule over Nature
(1 )
(2 )
(3 )
(4 )
H o st C o u ntry : M a rg in a l W illin g n e ss to P ay
B irth C o u ntry : H u m a n R u le ove r N a tu re
Tru st in E nv iro n m e nta l O rg a n iz a tio n s
0 .1 9 2 * * *
0 .1 9 8 * * *
0 .2 0 3 * * *
0 .2 2 6 * * *
(0 .0 4 9 )
(0 .0 6 5 )
(0 .0 5 2 )
(0 .0 7 3 )
0 .0 3 4 *
0 .0 3 9 *
(0 .0 2 0 )
(0 .0 2 1 )
B e lo n g in g to a R e lig io u s D e n o m in a tio n
-0 .0 0 1
-0 .0 0 0
(0 .0 0 3 )
(0 .0 0 3 )
P e o p le S h o u ld S tick to T h e ir O w n A ¤a irs
O b se rva tio n s
0 .0 4 5 * * *
0 .0 3 9 * * *
(0 .0 1 1 )
(0 .0 1 4 )
Ye s
Ye s
Ye s
Ye s
R -sq u a re d
4052
3057
4229
2838
H o st C o u ntry F E
0 .1 0 0
0 .1 0 3
0 .0 9 7
0 .1 1 4
Summary:This table establishes the robustness of the baseline analysis
to controlling for a number of alternative preferences of the migratns
Notes:(i) Robust standard error estimates are reported in parentheses; (ii)
*** denotes statistical signi…cance at the 1 percent level, ** at the 5 percent
level, and * at the 10 percent level, all for two-sided hypothesis tests.
3.5
Conclusions
Fertility rate, female labor participation, or preference for redistribution are cultural attributes
that frame individual economic behavior and ultimately economic policies. As part of culture,
these traits are transmitted across generations from parents to children. Are environmental
preferences among these cultural traits? This is the main question we tried to answer in this
study.
We …rst presented a model of transmission of environmental preferences following
Bisin and Verdier (2001). We de…ned the environmental cultural trait as the disutility from
pollution, which ultimately determines the marginal willingness to pay for the reduction of
pollution. Agents live in two homogeneous social groups whose populations mix if migration
takes place. We found that the environmental trait is successfully transmitted to the next
generations if the marginal disutility from pollution of migrants is inferior to the marginal
disutility of the native population.
Then, we empirically tested our theoretical result using survey data on environmental
preferences for 47 European countries. We found that the average environmental preference
in an immigrant’s country of birth has a large and signi…cant e¤ect on her own environmental
preference. Immigrants born in a country with a low preference for the environment tend to
have lower preferences for the environment than the natives of the countries in which they
reside. This relationship is veri…ed in the regression analyses in which we include country of
residence dummies and rich controls for individual characteristics.
28
Knowing whether environmental preferences are part of culture improves our knowledge
about the status quo of international economic agreements. As a matter of fact, the di¢ culties
in …nding a consensus on greenhouse gas emission targets for the period 2013-2020 could
partly re‡ect the fact that country members of the United Nations Framework Convention on
Climate Change manifest highly heterogeneous attitudes towards environmental protection.
Our research suggests that the views expressed by each government also re‡ect the social
preferences in the country (i.e. preferences of the median voter).
References
[1] Alesina A. and Giuliano P., 2010. "The power of the family," Journal of Economic Growth,
Springer, vol. 15(2), pages 93-125, June.
[2] Algan, Y., Dustmann, C., Glitz, A., and A. Manning (2010). “The economic situation of
…rst and second-generation immigrants in France, Germany and the United Kingdom",
The Economic Journal, 120, February F4-F30.
[3] Belhadj N., and Tarola O., 2013. "Relative quality-related (dis)utility in vertically
di¤erentiated oligopoly with an environmental externality," mimeo.
[4] Bisin A. and Verdier Th., 2001. "The Economics of Cultural Transmission and the
Dynamics of Preferences," Journal of Economic Theory, Elsevier, vol. 97(2), pages 298319, April.
[5] Borjas G., 1993. "The Intergenerational Mobility of Immigrants," Journal of Labor
Economics, University of Chicago Press, vol. 11(1), pages 113-35, January.
[6] Borofsky R., 2008. "Studying "Culture" scienti…cally in an oxymoron: The interesting
question is why people don’t accept this", pg 281, in "Explaining culture scienti…cally",
edited by Melissa Brown, The university of Washinton Press.
[7] Bowles S. and Gintis H., 2008, "Culture matters: Inferences from comparative behavioral
experiments and evolutionaly models", chapter 10, in "Explaining culture scienti…cally",
edited by Melissa Brown, The university of Washinton Press.
[8] Carroll C.D., Rhee B.K. and C. Rhee, (1994), “Are There Cultural E¤ects on Saving?
Some Cross-Sectional Evidence ”, The Quarterly Journal of Economics, Vol. 109, No. 3
(Aug., 1994), pp. 685-699
[9] Fernandez, R. and A. Fogli (2009) “Culture, an Empirical Investigation of Beliefs, Work,
and Fertility"American Economic Journal: Macroeconomics, 1:1, 146-177.
29
[10] Guiso, L., Sapienza, P. and L. Zingales (2006). “Does Culture A¤ect Economic
Outcomes?" Journal of Economic Perspectives, 20(2): 238.
[11] Luttmer E.F.P. and M. Singhal (2011) “Culture, Context and the Taste for
Redistribution"American Economic Journal: Economic Policy, 3 (February): 157-179.
[12] Schumacher I., 2013. "The endogenous formation of an environmental culture," Working
Papers 13, Department of Research, Ipag Business School.
[13] Roy A. D., 1951, "Some Thoughts on the Distribution of Earnings", Oxford Economic
Papers , New Series, Vol. 3, No. 2 (Jun., 1951) , pp. 135-146.
[14] Sethi R. and Somanathan E., 1996. "The Evolution of Social Norms in Common Property
Resource Use" American Economic Review, American Economic Association, vol. 86(4),
pages 766-88, September.
[15] Tabellini G., 2010. "Culture and Institutions: Economic Development in the Regions
of Europe," Journal of the European Economic Association, European Economic
Association, vol. 8(4), pages 677-716, 06.
30
Appendices
A
Summary Statistics
31
Table A.1: Willingness to Pay for the Environment
(1)
(2)
Frequency
Percent
agree strongly
agree
disagree
disagree strongly
865
2,024
964
442
20.14
47.12
22.44
10.29
Total
4,295
100
environment:
giving part
of income
Table A.2: Humans are meant to Rule over Nature
environment:
humans were meant
to rule over nature
(1)
(2)
Frequency
Percent
279
1,073
1,858
1,085
6.5
24.98
43.26
25.26
agree strongly
agree
disagree
disagree strongly
Total
4,295
32
100
Table A.3: Classi…cation of Migrants
Country of Origin
Albania
Armenia
Austria
Azerbaijan
Belarus
Belgium
Bosnia-Herzegovina
Bulgaria
Croatia
Cyprus
Czech Republic
Czechoslovakia
Denmark
Estonia
Finland
France
Georgia
Germany
Germany West
Great Britain
Greece
Hungary
Iceland
Ireland
(1)
All Migrants
(2)
First Generation Migrants
(3)
Second Generation Migrants
54
67
31
62
122
102
283
40
109
8
44
14
29
10
53
126
46
225
1
91
51
59
5
24
50
44
17
42
75
73
188
25
75
7
24
5
16
4
32
81
27
139
4
23
14
20
47
29
95
15
34
1
20
9
13
6
21
45
19
86
1
29
29
39
1
9
62
22
20
4
15
33
Table A.4: Classi…cation of Migrants (ctd)
Country of Origin
Italy
Kosovo
Latvia
Lithuania
Macedonia
Malta
Moldova
Netherlands
Norway
Poland
Portugal
Romania
Russia
Serbia
Slovakia
Slovenia
Spain
Sweden
Switzerland
Turkey
USSR
Ukraine
Yugoslavia, Socialist Federal Rep.
(1)
All Migrants
(2)
First Generation Migrants
(3)
Second Generation Migrants
240
35
21
36
45
2
18
48
13
123
223
77
775
164
64
18
57
21
6
383
11
232
27
97
19
13
19
17
9
31
9
65
166
54
431
109
38
8
35
13
3
216
6
131
18
2,554
143
16
8
17
28
2
9
17
4
58
57
23
344
55
26
10
22
8
3
167
5
101
9
34