Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Explaining well-being: essays on the socio-economic factors accompanying life satisfaction
(USC Thesis Other)
Explaining well-being: essays on the socio-economic factors accompanying life satisfaction
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
EXPLAINING WELL-BEING: ESSAYS ON THE SOCIO-ECONOMIC FACTORS
ACCOMPANYING LIFE SATISFACTION
by
Malgorzata A. Switek
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirement for the Degree
DOCTOR OF PHILOSOPHY
(ECONOMICS)
May 2014
Copyright 2014 Malgorzata A. Switek
ii
DEDICATION
For my Mother, in memory of my Father.
iii
ACKNOWLEDGEMENTS
This dissertation would not have been written without the inspiration coming from my
parents. For this, and continuous support, I am grateful to them and to all those who I
consider to be my family, including Manuel Castro. I also want to thank my advisor,
Richard Easterlin, who has taught me the meaning of hard work, dedication, and true
scholarship. He has given me more than just advice and guidance, but a true example to
follow, and that I will never forget. There have been many other wonderful scholars
from whom I have learned in the past years. Jeffrey Nugent who gave me the opportunity
to work with him on joint projects, an experience that I found both valuable and
enjoyable. My happiness team-mates, Onnicha Sawangfa, Laura Angelescu, Jacqueline
Smith-Zweig, Robson Morgan, and Kelsey O’Connor, have all given me valuable
feedback as well as their friendships. Fellow Graduate students, including (though not
exclusively) Jaime Meza-Cordero, Fei Wang, Brijesh Pinto, Saurabh Singhal, Aleks
Giga, and Emily Page, have been a source of useful comments and immense support in
the past years. Last, though not least, I would also like to thank the whole administrative
staff of the Economics Department. Morgan Ponder, Young Miller, Shannon Durbin, and
Christopher Frias have been especially attentive and patient at solving any and all my
problems, and for that I am highly grateful.
iv
TABLE OF CONTENTS
DEDICATION .................................................................................................................... ii
ACKNOWLEDGEMENTS ............................................................................................... iii
LIST OF TABLES ............................................................................................................. vi
LIST OF FIGURES ......................................................................................................... viii
ABSTRACT ....................................................................................................................... ix
CHAPTER 1. Introduction.................................................................................................. 1
CHAPTER 2. Life Satisfaction in Latin America: A Size-of-Place Analysis .................... 8
2.1 Introduction ............................................................................................................... 8
2.2 Data description....................................................................................................... 12
2.3 Descriptive analysis of the patterns of life satisfaction and their explanations ...... 15
2.4 Econometric analysis............................................................................................... 25
2.5 Conclusions ............................................................................................................. 34
CHAPTER 3. Internal Migration and Life Satisfaction: Well-Being Paths of Young Adult
Migrants ............................................................................................................................ 36
3.1.Introduction ............................................................................................................. 36
3.2. Literature review .................................................................................................... 38
3.3. Data description...................................................................................................... 42
3.4. Methods .................................................................................................................. 48
3.5. Results .................................................................................................................... 53
3.5.1 Migration and life-satisfaction .......................................................................... 53
3.5.2 Life domains behind the migration and life satisfaction association................ 57
3.6. Conclusions ............................................................................................................ 67
CHAPTER 4. Explaining Well-Being Over the Life Cycle: A Look at Life Transitions
During Young Adulthood ................................................................................................. 69
4.1. Introduction ............................................................................................................ 69
4.2. Literature review .................................................................................................... 71
4.3. Data description...................................................................................................... 77
4.4. Methods .................................................................................................................. 82
v
4.5. Results .................................................................................................................... 88
4.5.1 Life satisfaction path during ages 22 to 40 ....................................................... 88
4.5.2 Common transition pattern and its association with life satisfaction ............... 91
4.5.3 A glimpse into the life domain changes between ages 22 and 40 .................... 99
4.6. Conclusions .......................................................................................................... 103
CHAPTER 5. Summary and Conclusions ...................................................................... 107
REFERENCES ............................................................................................................... 111
APPENDICES ................................................................................................................ 120
Appendix A: Additional analysis for Chapter 2 .......................................................... 120
A.1: Methods ........................................................................................................... 120
A.2: Description of countries in the analysis ........................................................... 125
A.3: Description of variables in the analysis ........................................................... 127
Appendix B: Additional analysis for Chapter 3 .......................................................... 128
B.1: Attrition in the Young Adult Panel Study ....................................................... 128
B2: Description of variables used in the study ........................................................ 139
B.3: Additional regression results ........................................................................... 141
Appendix C: Additional analysis for Chapter 4 .......................................................... 142
C.1: Description of variables in the analysis ........................................................... 142
C.2: Additional statistical analysis and regressions ................................................. 145
C.3: Further details on age interval construction ..................................................... 148
vi
LIST OF TABLES
Table 2-1: OLS regressions: life satisfaction as dependent variable, places 40k-100k as
reference ............................................................................................................................ 16
Table 2-2: OLS regressions: life satisfaction as dependent variable, rural-happy and
urban-happy grouping ....................................................................................................... 17
Table 2-3: Symbols for variables in Tables 2-4 to 2-7 ..................................................... 19
Table 2-4: Mean life satisfaction, financial status and percent of population employed in
different occupations, 40k-100k as reference, by size-of-place ....................................... 20
Table 2-5: OLS regressions: life satisfaction as dependent variable, test of different
explanations separately for life satisfaction patterns ........................................................ 28
Table 2-6: OLS regressions: life satisfaction as dependent variable, test of importance of
different types of public spending for life satisfaction patterns ........................................ 30
Table 2-7: OLS regressions: life satisfaction as dependent variable, test of different
explanations combined for life satisfaction patterns ......................................................... 32
Table 3-1: Descriptive statistics of migrants and non-migrants before and after the move,
by reason to move ............................................................................................................. 46
Table 3-2: OLS regressions: change in life satisfaction as dependent variable, migration
(all and by reason to move) as main explanatory variable ................................................ 54
Table 3-3: OLS regressions: job domain variables as dependent variables, migration (all
and by reason to move) as main explanatory variables .................................................... 58
Table 3-4: OLS regressions: financial domain variables as dependent variables, migration
(all and by reason to move) as main explanatory variables .............................................. 60
Table 3-5: OLS regressions: housing domain as dependent variable, migration (all and by
reason to move) as main explanatory variables ................................................................ 63
Table 3-6: Change in life satisfaction by migrant type and occupational trajectory ........ 64
Table 4-1: Descriptive statistics, by age, beginning and end of each age interval ........... 82
Table 4-2: OLS regressions: change in life satisfaction as dependent variable, age
intervals as explanatory variables ..................................................................................... 89
Table 4-3: Life transitions undergone, by age group ........................................................ 92
Table 4-4: OLS regressions: change in life satisfaction as dependent variable, main life
transitions and age intervals as explanatory variables ...................................................... 95
Table 4-5: OLS regressions: change in life domains as dependent variables, main life
transitions as explanatory variables - all age intervals pooled ........................................ 101
vii
Table A-1: Percent of urban population (living in cities of 5,000/10,000 or more
inhabitants) as surveyed by the Latinobarometro (weighted values used) ..................... 121
Table A-2: Percent of people classified as having indigenous origins using the
Latinobarometro data (weighted values) ........................................................................ 124
Table A-3: Number of observations used for analysis, per year and country ................. 125
Table A-4: Mean life satisfaction by size-of-place and country ..................................... 125
Table A-5: Questions and response categories for each of the explanations analyzed .. 127
Table B-1: Comparison of the characteristics at baseline (1999) of respondents who
consequently attrit (not interviewed in 2009) and do not attrit (interviewed in 2009) ... 129
Table B-2: Mobility by cohort: general population vs. YAPS non-attritors................... 133
Table B-3: OLS regressions: variables of interest (in levels) on future attrition ............ 134
Table B-4: OLS regressions: variables of interest (in 99-03 changes) on attrition ........ 136
Table B-5: Mean disposable income (hundreds of SEK), whole population (1968, 1972
and 1976 cohorts) and YAPS (non-attritors), by migration status, by year .................... 137
Table B-6: Number of people surveyed answering each question in both 99 and 09, by
migration status and reason to move ............................................................................... 139
Table B-7: Description of original survey questions used in the analysis ...................... 139
Table B-8: OLS regressions: results dividing respondents by occupational status
trajectory and reason for moving .................................................................................... 141
Table C-1: Description of all variables used in the analysis ........................................... 142
Table C-2: Life satisfaction changes associated with marriage and cohabitation ......... 145
Table C-3: OLS regressions: variables of interest (in changes) on future attrition ........ 145
Table C-4: OLS regressions: specification including time trends and cohort dummies 146
Table C-5: OLS regressions: specification dividing those going through school-to-work
transition by occupation type after education completion .............................................. 147
Table C-6: OLS regressions: specification dividing those going through school-to-work
transition by level of final education (postsecondary, or secondary and less) ................ 148
viii
LIST OF FIGURES
Figure 4-1: Life Satisfaction by age, by cohort and pooling cohorts by age interval ....... 90
Figure 4-2: Main life transitions, by age interval, age 22 to 40 ........................................ 93
Figure 4-3: Life Satisfaction by age, actual and predicted, adj. values (LS age 22=0) .... 99
Figure C-1: Method of pooling of cohorts for age interval creation ............................... 150
ix
ABSTRACT
The present work provides new knowledge on the relation between various life
circumstances and quality of life, as measured by life satisfaction. Chapter 2 analyzes
rural-urban life satisfaction differences in 14 Latin American countries using
Latinobarometer survey data for four years. The countries divide into two groups: those
where rural people are happier than urban people, and those where they are less happy. In
a multivariate analysis higher public social spending is found to increase rural relative to
urban life satisfaction through its positive influence on access to services such as health.
Chapter 3 examines life satisfaction changes following internal migration for young
adults in Sweden, using panel data for years 1999 and 2009. Life satisfaction increases
following an internal move, but the persistence of this increase depends on the reason for
moving. Migrants who move for work- related reasons experience a long term increase in
life satisfaction. Those who move for reasons other than work (such as housing),
experience a medium term, but no long term, well-being improvement. Occupational
status improvements are the main reason behind the long lasting increase in life
satisfaction of work migrants. Chapter 4 explores the life satisfaction path during young
adulthood, again using the panel data for Sweden. Life satisfaction increases slightly until
the age of 30 and then turns downward. This pattern is largely explained by major
transitions in life circumstances. Before 30 partnership formation, the school-to-work
transition, and having babies all tend to increase life satisfaction. After 30 life satisfaction
declines as children grow older and the breakup of unions becomes more prevalent.
1
CHAPTER 1. Introduction
This dissertation provides an empirical analysis of the social and economic conditions
contributing to people’s quality of life. Its main objective is to identify the factors
conducive to improving well-being, when well-being is measured using self-reported life
satisfaction obtained from individual level surveys. To achieve this, three original studies
are included analyzing how living in a rural or urban area, migration, and life transitions
each affect individual level life satisfaction.
The approach used throughout this dissertation combines the rigor of economic
analysis with new ways of measuring well-being that make use of subjective measures
such as life satisfaction. In recent decades concerns have risen in economics about the
focus on measures that capture exclusively the material aspects of life, such as income
growth and inflation. While such measures (also referred to as “objective” measures)
may capture important factors affecting material living standards of a population, they
fail to provide information on the non-material aspects of life, such as health and family
life. To provide a more comprehensive picture of well-being, some economists have
suggested a shift towards a different set of “subjective” measures of social progress such
as self-reported levels of happiness or life satisfaction. The relevance of these measures is
perhaps best illustrated by the support they have gained from, among others, Nobel Prize
winners (Stiglitz, Sen and Fitoussi 2009) and members of the UN General Assembly who
in July 2011 passed a resolution that invited member countries to collect subjective
measures and to use them to guide their policy decisions (Helliwell et al. 2013).
2
The research presented here is a response to the growing need to expand
knowledge on the conditions that lead to improving subjective well-being. The findings
contribute new insight into the nature and relative importance of the social and economic
components of well-being. The three main chapters of this dissertation provide an
analysis of the association between life satisfaction and various social conditions, such as
regional well-being differences, migration, and financial concerns during the transition
into adulthood. As will be shown, the results obtained could be used to design policies
leading to the improvement of overall satisfaction with life, as opposed to the pure
maximization of economic growth. What follows provides a brief description of these
three studies (chapters two, three, and four of the dissertation) outlining the questions
asked and methods used.
Chapter two asks two main questions: how does life satisfaction vary by size-of-
place, and what can account for these variations. Previous subjective well-being literature
has limited itself to the description of life satisfaction by size-of-place. Their results have
been mixed, and depend on the country under analysis. Little or no information about the
reasons for the life satisfaction differences between rural and urban places has been
offered. Focusing on fourteen countries in Latin America, this chapter contributes to
previous knowledge by analyzing the differences in rural-urban life satisfaction patterns
in this region, and providing an explanation for why these differences may exist. Data for
years 2003 through 2006 of the Latinobarometro survey are used. This data set is
important because it focuses on Latin America, a region with a fairly homogenous culture
3
and language, thereby reducing the bias that could arise when a more diverse sample of
countries is analyzed.
The methods used include both descriptive analysis and pooled-cross sectional
multivariate regressions in which interaction terms are used to identify the difference in
life satisfaction between rural and urban places. Four explanations for the relative life
satisfaction patterns between rural and urban places are considered: economic
development, public social spending, social values, and indigenous origins. The first two
of these are country level variables with a large dispersion in Latin America. The third
and fourth focus on individual level traits with rural-urban distributions that also differ by
country. Though all these have been mentioned in previous literature, the present study is
the first to considered them in a joint analysis, which proves to have an important impact
on the results obtained.
The main objective of the second analysis is to examine the relationship between
internal migration and the well-being of young adult migrants. Three questions are asked.
First, is internal migration accompanied by an increase in life satisfaction? Second, does
this increase depend on the reason for moving? And third, if life satisfaction changes
following an internal move, what are the aspects of life that underlie the change in well-
being of the migrants? The focus on life satisfaction as the main variable of interest is
especially relevant in the context of migration because, while various studies have
discussed the income benefits of moving, considerably less information exists on changes
in the non-material conditions of the migrants. Since migration is often accompanied by
important social and psychological changes, the full well-being effect of moving cannot
4
be established unless these non-material conditions are considered. Self-reported life
satisfaction provides a summary measure of all aspects affecting individual well-being,
providing a more comprehensive picture of the change in the overall conditions of the
migrants.
The few existing studies of life satisfaction changes following an internal move
arrive at mixed results. The present analysis suggests that these mixed results could be
clarified if a distinction is made between different types of migrants. Specifically, age
and reason for moving have been shown to affect the changes in income and housing
quality after migration, and as such they could also impact the change in life satisfaction.
To account for this, the focus here is restricted to a specific age group, young adults.
Additionally, migrants are divided into those who move for work related reasons, and
those who move for reasons other than work, such as housing or the desire to be closer to
one’s family.
To evaluate the change in life satisfaction following migration, surveys for the
years 1999 and 2009 of the Young Adult Panel Study (YAPS) are used. The YAPS
includes individual-level longitudinal information on place of residence and life
satisfaction for three young adult cohorts of the Swedish population. Because these data
were collected by statistical offices with access to register information (providing records
on all vital events such as a change in residence collected by the Swedish Tax Agency for
the entire Swedish population), they minimize follow-up issues that may result in attrition
selective on migration. As such they represent a highly reliable source of information for
the analysis of internal migration for a nationally representative sample of young adults.
5
A first-difference regression model is used to compare the change in life
satisfaction of those who moved municipalities of residence between 1999 and 2009
(migrants), to the change in life satisfaction of those who did not move (non-migrants).
To assure comparability between migrants and non-migrants, the main differences
between the two groups are accounted for. These differences include fixed individual-
level characteristics (such as personality traits), selective migration by region of origin,
and the main life experiences undergone during the period under analysis (such as
changes in marital status or the birth of a child). The possibility that migrants may
represent a select group of people with a high motivational profile is partially controlled
for by accounting for final level of education and by providing additional comparison of
migrants and non-migrants on similar occupational trajectories. The findings of this study
contribute important information on the difference in well-being improvements between
work and non-work migrants, as well as on the life aspects that underlie the change in life
satisfaction for the two groups of movers.
The last study of this dissertation centers around three main questions. First, what
is the path followed by life satisfaction during the young adult years? Second, what is the
typical pattern of life transitions undergone in those years? And finally, to what extent
can these transitions account in themselves for the overall life satisfaction changes? To
answer these questions, the association between the life satisfaction path and life
transitions undergone by young adults is examined, focusing again on Sweden. Four life
transitions are considered: partnership (marriage and cohabitation) formation, the school-
to-work transition, parenting, and partnership dissolution. Previous literature had focused
6
on analyzing each of these separately without considering their interaction. Given the
typically close timing of partnership formation, the school-to-work transition, and birth of
the first child, examining a single transition alone may create a bias due to the effects of
the other two on well-being. To avoid such bias, the present analysis considers all three
transitions jointly, aiming to disentangle the association between them. In addition to this,
the present analysis contributes to previous knowledge by providing the first attempt to
link life satisfaction changes over the life cycle to relevant life transitions. As will be
seen, its findings illustrate the degree to which the overall satisfaction path during young
adulthood may be explained by the common transition patterns experienced at those ages.
The data used are again obtained from the Young Adults Panel Study and include
surveys for 1999, 2003, and 2009. Using the 2003 survey reduces the amount of
observations available, but allows one to capture the life satisfaction changes during four
age intervals covering ages 22 to 40. The YAPS data set, by providing information on
self-reported life satisfaction combined with register data on timing of education
completion, changes in marital status, and birth of the first child, makes it possible to
conduct a detailed analysis of the timing and well-being effects of the main young adult
transitions. As mentioned, changes in satisfaction over four age intervals between ages 22
and 40 are analyzed to identify the average life satisfaction path during young adulthood.
Age intervals are constructed pooling observations for respondents from different cohorts
interviewed at the same age. Cohort and individual level effects are controlled for by
using first-difference regressions where change in life satisfaction is the main dependent
variable, and age and life transitions are the explanatory variables of interest. The results
7
of this analysis have important implications regarding the type of policies that could be
useful to assist young adults as they transition through different stages of life.
The remainder of this dissertation is organized as follows. Chapter two presents
how life satisfaction varies by size-of-place in fourteen Latin American countries and
explores the reasons behind these life satisfaction variations. Chapter three analyzes the
association between life satisfaction and internal migration for young adults in Sweden,
dividing migrants into those who move for work related and non-work related (housing
or other) reasons. Focusing again on Sweden, chapter four illustrates the life satisfaction
path followed during young adulthood and discusses how this life satisfaction path is
shaped by the concurrent young adult transitions. A summary of the main findings and
possible policy implications are outlined in Chapter five.
8
CHAPTER 2. Life Satisfaction in Latin America: A Size-of-Place Analysis
1
2.1 Introduction
How does life satisfaction vary with size-of-place? Little has been done to answer this
question. The few previous studies have not yielded uniform results. Some authors find
that rural places are as happy or happier than their urban counterparts (Berry and
Okulicz-Kozaryn 2009); others claim that rural places are less happy, although the gap
between the rural/urban life satisfaction becomes narrower in richer countries
(Veenhoven 1994). Focusing exclusively on rural/urban differentials, these studies omit
other size-of-place categories which could be related to happiness patterns. The present
study draws on recent Latinobarometro surveys to fill in this gap and re-evaluate the
earlier conclusions undertaking a more comprehensive analysis of life satisfaction trends
across a number of size-of-place categories in fourteen Latin American countries.
The study centers on two main questions. First, it investigates whether a general
life satisfaction pattern by size-of-place may be identified in the countries analyzed. The
relative cultural and economic homogeneity of the region is of particular value in
addressing this question, as previous studies have found that the size-of-place (mainly
rural/urban) happiness differentials are influenced by cultural characteristics of nations
(Berry and Okulicz Kozaryn 2009; Spellerberg, Huschka and Habich 2007). It is found
here that the four size-of-place categories studied fall naturally into a non-
1
This chapter has been published in the Journal of Development Studies, Volume 48, Issue 7,
2012, available online at http://www.tandfonline.com/10.1080/00220388.2012.658374
9
urbanized/urbanized division where the dividing line can be set at 40,000 inhabitants
2
: in
eight of the fourteen countries people in non-urbanized places are less happy relative to
those living in urbanized places while in the other six inhabitants of non-urbanized places
are happier than their urbanized counterparts. To simplify the language in what follows
the non-urbanized places (villages and small towns) are referred to as rural and the
medium and large cities as urban. Using this rural/urban division, countries where rural
inhabitants are less happy are referred to as urban-happy, and those where the rural
inhabitants are happier as rural-happy.
Given the existence of a dichotomous pattern – urban-happy and rural-happy – the
second main question is what are its causes? In what follows, four possible explanations
are examined: differences between the two groups of countries in level of economic
development; social values; public social spending; and the proportion of indigenous
population in the rural areas.
Some studies have found that the level of economic development of a country is
an important factor in the determination of the rural/urban happiness differentials, with
more developed countries having circumstances that increasingly favor the rural relative
to urban life satisfaction levels (Easterlin, Angelescu and Zweig 2011). It is therefore
possible that the patterns of life satisfaction observed in Latin America are due to
differences in the levels of development of the countries in this region. The GDP per
capita in 2006 (in constant 2000 dollars) of the fourteen countries considered presented a
2
This division is close to that suggested by the US Census Bureau which defines an Urban Area as
a census block that meets, among others, the criterion of having at least 50,000 inhabitants (Department of
Commerce: Bureau of the Census 2002). The present study adjusts the number of inhabitants to 40,000 so
as to include a larger sample size in the medium-sized city category.
10
great variance ranging from $1145 in the poorest (Bolivia) to $8700 in the richest
(Argentina) (World Bank 2010). Richer countries are more likely to have experienced
greater technological improvements in transportation, communication and production.
According to Frey (1992) these improvements lead to a convergence between the
availability of economic opportunities and income in different size-of-place areas. Since a
higher individual income at a point in time corresponds to higher life satisfaction
(DiTella, MacCulloch and Oswald 2003; Easterlin 2001a), the convergence of economic
opportunities suggests that if rural-happy countries are in general more developed than
urban-happy, this could account for their relatively higher rural life satisfaction.
The second possible explanation for the two patterns of happiness is related to the
higher social values that are said to characterize rural places (Veenhovenv 1994). In the
literature social values encompass such things as stronger moral cohesion (Berry and
Okulicz-Kozaryn 2009), a slowing down of the pace of life (Spellerberg, Huschka and
Habich 2007), more secure social relations, and a better quality of family life (Pichler et
al. 2006). However, difference in various circumstances might cause differences in levels
of social values among rural areas. In regard to crime, for example, it is possible that in
countries with a weak judicial system, the rule of law may not apply in rural places where
there are not enough resources available for enforcement. This lack of enforcement could
create higher levels of crime in the rural areas in these countries, influencing negatively
their security and moral cohesion, and decreasing life satisfaction. Therefore, differences
in social and judicial environments between countries in Latin America could potentially
account for different relative happiness levels in the rural areas. To assess the possible
11
relevance of the social values explanation, variables related to family, criminal activity,
and faith will be compared in different size-of-place categories in rural-happy and urban-
happy countries.
The third explanation refers to the availability of public services such as health
and education, which, in turn, depends largely on the level of public social spending.
Public social spending is defined here specifically as the per capita amount of money
spent by the State on education, health and nutrition, social security, work, social
assistance, housing, and water and sewage programs (Economic Commission for Latin
America and the Caribbean 2010a). In Latin America there exists a sizeable dispersion in
the amount of public spending by country, varying in 2003-2006 from a high of $1156
per capita (in constant 2000 dollars) for Brazil down to $94 in Ecuador (Economic
Commission for Latin America and the Caribbean 2010a). The rates of poverty in rural
areas in Latin America are very high, with over 50 percent of households in rural
communities being below the poverty line (Arriagada 2000). Hence, the people in rural
communities may not be able to afford private facilities, making their access to health,
education, and other basic services highly dependent on public social spending. Because
access to these services could be expected to increase a person’s a life satisfaction, the
different levels of public social spending by country might explain differences in the
relative happiness of rural places.
The last possible explanation considered is related to the presence of indigenous
population in rural areas. In Latin America, indigenous communities, which in certain
countries represent over one fourth of the total population and live mainly in rural areas,
12
have some of the worst socioeconomic conditions (Machinea and Hopenhayn 2005). As
long as the poor social and economic conditions in which indigenous people live have an
effect on their life satisfaction, countries with higher proportions of indigenous
population living mainly in rural areas would tend to have lower rural relative to urban
levels of happiness. Therefore, a fourth possible explanation of the size-of-place life
satisfaction differentials is the difference in the proportion of indigenous residents in rural
places in urban-happy compared to rural-happy countries.
Two main methods are used. The first is a detailed descriptive analysis of the
microeconomic data by size-of-place relating to life satisfaction and the four different
explanations under consideration. This descriptive analysis makes clear the variables
being studied and provides a preliminary impression of the relevant explanations of the
differing patterns. An econometric approach is then taken which combines individual and
country level data in a set of microeconomic regressions including size-of-place and
pattern type dummies. The econometric analysis confirms many of the preliminary
impressions from the descriptive analysis and tests the relative importance of the different
explanatory variables. In the end, public social spending proves to be the most important
factor behind the relatively higher life satisfaction in rural areas in rural-happy countries.
2.2 Data description
The main source of the microeconomic data is the Latinobarometro, which is a public
opinion survey performed in seventeen Latin American countries since 1995. The
13
Latinobarometro is meant to measure the social, political, and economic conditions of
each country, and therefore includes a broad variety of questions ranging from socio-
demographic characteristics to perceptions of the economic and political situations.
Ideally, one would like to include as many years and countries as possible; however,
certain data issues required a reduction in the amount of data analyzed
3
. The present
study is based on the 2003-2006 surveys, and covers fourteen Latin American countries –
Argentina, Bolivia, Brazil, Colombia, Costa Rica, Chile, Ecuador, El Salvador,
Honduras, Mexico, Panama, Peru, Uruguay and Venezuela.
A pooled cross section approach is used
4
because no single year is simultaneously
available for all countries. The dependent microeconomic variable is life satisfaction,
obtained from the following question: “In general, would you say that you are satisfied
with your life? Would you say that you are very satisfied, fairly satisfied, not very
satisfied or not satisfied at all?” The responses to this question are assigned values from 1
(not satisfied at all) to 4 (very satisfied). The main explanatory variable, size of town
where the person lives as measured by number of inhabitants, is reported by the
interviewer according to the following classification: less than or equal to five thousand
inhabitants; five to ten thousand; ten to twenty thousand; twenty to forty thousand; forty
to fifty thousand; fifty to one hundred thousand; more than one hundred thousand; capital
city. To assure a sufficient number of observations by size-of-place, these categories were
3
For a detailed description of these issues and the methodology used to select the data employed
in the analysis see Appendix A.1.
4
Country level descriptive statistics can be found in Appendix A.2. The least amount of
observations by country considered in the analysis is 1,000 (able A-3).
14
combined as follows: less than or equal to five thousand; five to forty thousand; forty to
one hundred thousand; and above one hundred thousand (which includes capital cities).
The following variables were identified as being related to each of the four alternative
explanations of the life satisfaction patterns:
1. For the development explanation: number of goods owned by the individual,
rating of economic situation, income/needs relationship, and the
occupation/employment category of the individual
5
;
2. for the social values explanation: marital status, crime victimization, participation
in corruption, and level of religiosity;
3. for the public spending explanation: satisfaction with access to education,
satisfaction with access to health services, and number of basic services available
in the individual’s household;
4. for the indigenous origins explanation: mother tongue of the person interviewed
6
.
In addition to the microeconomic data described above, two main macroeconomic
variables are used: mean log GDP per capita and public social spending per capita – total,
and subdivided into spending on education, health, housing, and social security. Both
GDP and public spending are reported in constant 2000 US dollars for the time period
2003-2006. GDP per capita is used to assess a country’s economic development. The
source for this variable is the World Bank’s online database (World Bank 2010). Public
social spending is employed to proxy for a country’s availability of public services. The
5
Ideally, one would want to include the income variable in this domain, but the Latinobarometro
survey does not report an individual’s income.
6
For a detailed description of the variables related to the four explanations, see Appendix A.3.
15
source for this data is the Economic Commission for Latin America and the Caribbean’s
statistical information database CEPALSTAT (Economic Commission for Latin America
and the Caribbean 2010a).
2.3 Descriptive analysis of the patterns of life satisfaction and their explanations
As previously mentioned, the main classification by size-of-place used is: less than five
thousand inhabitants, five to forty thousand, forty to one hundred thousand, and over one
hundred thousand (including capital cities). In what follows these groups will be referred
to as village, small town, medium sized city and large city respectively. As noted earlier,
villages and small towns, when spoken about jointly, will be referred to as rural areas.
To establish the significance level of the differentials in life satisfaction by size-
of-place, OLS regressions of life satisfaction on the different size-of-place categories are
analyzed, with the middle sized cities as the reference category. Because life satisfaction
is an ordinal variable, one might be concerned with the use of a linear regression model.
Two main problems with using OLS for ordinal variables exist. First, OLS coefficients
imply continuous estimates (not confined to the original categories of the dependent
variable). The present study, however, is not interested in prediction of the dependent
variable, which reduces the importance of this problem. Second, OLS regressions with
ordinal dependent variables will result in heteroskedastic error terms. To correct for this
heteroskedasticity, robust standard errors are used throughout the analysis.
16
The differences in average life satisfaction between rural areas and middle sized
cities are significant in most cases and, as noted, the countries fall into two groups –
rural-happy and urban-happy (Table 2-1). In contrast, the pattern between the middle
sized and the large cities is fairly consistent between the two sets of countries – for the
large cities the coefficients are mostly either negative or not significantly different from
the middle sized places
7
. Therefore the dichotomous pattern in life satisfaction by size-of-
place arises mainly from differentials between the villages/small towns and middle sized
cities. The rural-happy countries are Brazil, Colombia, Costa Rica, Chile, Uruguay, and
Venezuela; the urban-happy, Argentina, Bolivia, Ecuador, El Salvador, Honduras,
Mexico, Panama, and Peru. Both with and without country controls, in the pooled rural-
happy countries the difference between the rural areas and the middle sized cities is
significantly positive; for the pooled urban-happy countries this differential is
significantly negative (Table 2-2).
Table 2-1: OLS regressions: life satisfaction as dependent variable, places 40k-100k
as reference
7
The only exception is Colombia, for which the large cities are significantly happier than the
middle sized places.
Argentina Bolivia Brazil Colombia
Costa
Rica
Chile Ecuador
<5k -0.157 -0.244 0.019 0.032 0.042 0.04 -0.079
(2.90)** (4.29)** (0.4) (0.38) (0.6) (0.32) (0.93)
5k-40k -0.117 -0.111 0.089 0.211 0.019 0.07 -0.166
(2.16)* (1.82)+ (3.62)** (4.00)** (0.32) (0.43) (1.96)+
40k-100k reference reference reference reference reference reference Reference
>100k & -0.162 -0.154 -0.035 0.165 0.043 0.138 0.044
Capital (3.46)** (2.82)** (1.48) (3.61)** (0.57) (1.2) (0.54)
Constant 3.083 2.757 2.719 3.032 3.337 2.802 2.776
(69.58)** (53.96)** (145.02)** (73.44)** (56.23)** (25.58)** (35.60)**
17
Table 2-1 (Continued)
Table 2-2: OLS regressions: life satisfaction as dependent variable, rural-happy
and urban-happy grouping
Urban-happy countries Rural-happy countries
With country
controls
No country
controls
With country
controls
No country
controls
<5k -0.143 -0.238 <5k 0.082 0.065
(5.74)** (9.58)** (3.56)** (2.75)**
5k-40k -0.088 -0.084 5k-40k 0.099 0.159
(3.69)** (3.42)** (5.51)** (8.62)**
40k-100k Reference reference 40k-100k reference Reference
>=100k and -0.065 -0.137 >=100k and 0.043 -0.001
Capital (2.88)** (5.98)** Capital (2.41)* (0.08)
Argentina Reference reference Brazil reference Reference
Bolivia -0.337 Colombia 0.447
(18.53)** (24.61)**
Ecuador -0.199 Costa Rica 0.596
(9.77)** (34.00)**
El Salvador 0.084 Chile 0.169
(4.32)** (5.44)**
Honduras 0.208 Uruguay 0.097
(10.63)** (6.10)**
Obs 4779 4770 4786 4784 2996 1196 3593
R-sq 0 0.01 0.01 0.01 0 0.01 0.01
El
Salvador Honduras Mexico Panama Peru Uruguay Venezuela
<5k -0.04 -0.053 -0.182 -0.269 -0.044 0.134 0.114
(0.73) (0.51) (1.55) (5.27)** (0.47) (2.96)** (2.31)*
5k-40k -0.027 -0.017 -0.109 -0.171 0.005 0.184 -0.011
(0.44) (0.37) (0.98) (3.48)** (0.05) (4.15)** (0.31)
40k-100k reference reference reference reference reference reference Reference
>100k & 0.052 -0.077 0.003 -0.138 -0.071 0.046 -0.008
Capital (0.93) (1.68)+ (0.03) (2.74)** (0.8) (1.1) (0.21)
Constant 3.009 3.193 3.087 3.341 2.529 2.743 3.405
(60.03)** (80.30)** (29.59)** (76.75)** (29.92)** (73.73)** (122.16)**
Obs 4026 3998 3584 2995 2383 4771 3586
R-sq 0 0 0.01 0.01 0 0.01 0
Robust t statistics in parentheses
+ significant at 10%; * significant at 5%; ** significant at 1%
18
Table 2-2 (Continued)
Mexico 0.091 Venezuela 0.678
(3.98)** (40.14)**
Panama 0.24
(12.95)**
Peru -0.441
(18.50)**
Constant 3.016 3.033 Constant 2.678 2.998
(127.17)** (140.87)** (171.52)** (192.13)**
Observations 30128 30128 Observations 22119 22119
No. of
countries 8 8
No. of
countries 6 6
R-squared 0.07 0.01 R-squared 0.11 0.01
Robust t statistics in parentheses
+ significant at 10%; * significant at 5%; ** significant at 1%
Why do some countries have a higher level of happiness in rural areas relative to
middle sized cities, while in others rural areas are relatively less happy? The first possible
explanation is that the rural-happy countries are more economically developed than
urban-happy countries. This higher development allows them to have more technological
and infrastructural advances that in turn induce a greater spread of economic
opportunities and earnings from the cities to the rural areas. If this were the case, then the
people in villages and small towns in the rural-happy countries would be expected to be
on average richer relative to middle sized cities’ inhabitants than those in urban-happy
countries.
In fact, in rural-happy countries there is no significant difference in the average
economic situation and income/needs relationship between the villages and middle sized
cities. Moreover, the difference between the small towns and the middle sized cities
actually favors the small towns. In contrast, in urban-happy countries villages as well as
small towns do significantly worse than middle sized cities in both, economic situation
19
and income/needs relationship. Although in rural-happy countries the people living in
villages and small towns are poorer than those in the middle sized cities, the difference
between the amount of goods owned in rural areas and middle sized cities in this group of
countries is considerably smaller than in the urban-happy group (Table 2-4
8
). The
patterns of ownership of goods, economic situation and income/needs relationship
therefore suggest that people living in rural areas in rural-happy countries have, on
average, better economic conditions than those in urban-happy countries, supporting the
development explanation.
Table 2-3: Symbols for variables in Tables 2-4 to 2-7
Variable Explanation
Lifesat In general, would you say that you are satisfied with your life? Would you say
that you are very satisfied, fairly satisfied, not very satisfied or not satisfied at
all? (scale 1-4)
Goods Ordinal variable of the sum of the ownership of following goods: color tv,
refrigerator, home, computer, washer, phone, car, and holiday home (scale 0-8)
econ sit In general, how would you describe your present economic situation and that of
your family? Would you say that it is very good, good, about average, bad or
very bad? (scale 1-5)
income/
needs
Does you salary and the total of your family's salary allow you to satisfactorily
cover your needs? Which of the following situations do you find yourself in?
Covers them well, covers alright, does not cover, does not cover (scale 1-4)
other private 1 if the person reported working in the "other private" (other than professional
or owner) occupation, 0 otherwise
farmer/fisher 1 if the person reported working in the "farmer/fisher" occupation, 0 otherwise
Married 1 if the person reported being married, 0 otherwise
Single 1 if the person reported being single, 0 otherwise
Crime Have you, or someone in your family, been assaulted, attacked, or been the
victim of a crime in the last 12 months? (yes/no)
corruption Have you or someone in your family been aware of an act of corruption in the
last 12 months? (yes/no)
Religiosity How would you describe yourself? Very devout, devout, not very devout, or not
devout at all? (scale 1-4)
8
For reference on the variables and abbreviations presented in Tables 2-4 through 2-7, see Table
2-3.
20
Table 2-3 (Continued)
sat access
educ
Would you say that you are very satisfied, rather satisfied, not very satisfied or
not at all satisfied with the education to which you have access? (scale 1-4)
sat access
health
Would you say that you are very satisfied, rather satisfied, not very satisfied or
not at all satisfied with the health care to which you have access? (scale 1-4)
basic
services
Sum of the availability of following services in the household: water, hot water,
sewage (scale 0-3)
indigenous 1 an Indigenous language was the respondent’s mother tongue, 0 otherwise
Age Age of the person interviewed
age_sq Square of the age of the person interviewed
Male 1 if the person is male and 0 if female
y_edu Years of education of the person interviewed
Village 1 if the person lives in a town with a population of less than 5k and 0 otherwise
small town 1 if the person lives in a town with a population of 5k-40k and 0 otherwise
rural-happy 1 if the person lives in a "rural-happy" country and 0 otherwise
village*rural
-happy
1 if the person lives in a town with a population of less than 5k AND belongs to
a "rural-happy" country, 0 otherwise
small*rural-
happy
1 if the person lives in a town with a population of 5k to 40k AND belongs to a
"rural-happy" country, 0 otherwise
logGDP The logarithm of the GDP per capita of the country where the person lives
(source: WB)
Pub Spend Public social spending of the country where the person lives (ECLAC)
PS educ Public social spending on education of the country where the person lives
(ECLAC)
PS health Public social spending on health of the respondent’s country (ECLAC)
PS house Public social spending on housing of the respondent’s country (ECLAC)
PS Soc Sec Public social spending on social security of the respondent’s country (ECLAC)
Table 2-4: Mean life satisfaction, financial status and percent of population
employed in different occupations, 40k-100k as reference, by size-of-place
Urban-happy countries
Pop
( ‘ 000)
Life Financial status
Employment
distribution
Social values
lifesat Goods econ sit
income/
needs
other
private
farmer/
fisher
married Single
>=5 -0.143** -1.081** -0.148** -0.241** -0.063** 0.133** 0.035* -0.037**
5-40 -0.088** -0.69** -0.092** -0.126** -0.051** 0.076** 0.009 -0.024+
40-
100
ref ref ref ref ref ref ref Ref
>100 -0.065** 0.146** -0.052** 0 -0.013 -0.007+ -0.015 0
21
Table 2-4 (Continued)
Pop
( ‘ 000 )
Social values Availability of public services
Indige-
nous
crime
corrup-
tion
religio-
sity
sat
health
sat
educ
basic
services
indige-
nous
>=5 -0.098** -0.063** 0.011 -0.048+ -0.076** -0.58** 0.064**
5-40 -0.075** -0.046** -0.002 -0.006 -0.05* -0.286** 0.035**
40-
100
ref ref ref ref ref ref ref
>100 0.069** 0.023* -0.003 -0.065** -0.064** 0.074** 0.008+
Rural-happy countries
Pop
( ‘ 000 )
Life Financial status
Employment
distribution
Social values
lifesat Goods econ sit
income/
needs
other
private
farmer/
fisher
married Single
>=5 0.082** -0.55** 0.032 -0.008 0.01 0.082** 0.052** -0.032*
5-40 0.099** -0.143** 0.063** 0.06** -0.007 0.025** 0.012 -0.015
40-
100
ref ref ref ref ref ref ref Ref
>100 0.043* 0.587** -0.011 0.083** 0.019* -0.035** -0.027* 0.02*
Pop
( ‘ 000 )
Social values Availability of public services
Indige-
nous
pop
crime
corrup-
tion
religio-
sity
sat
health
sat
educ
basic
services
indige-
nous
>=5 -0.091** -0.061** 0.017 0.132** 0.071** -0.586** 0.004*
5-40 -0.051** -0.038** -0.025 0.102** 0.058** -0.084** 0.003
40-
100
ref ref ref ref ref ref ref
>100 0.104** 0.053** -0.089** -0.035+ -0.045* 0.311** 0.005**
Note: All regressions include country controls using Argentina and Brazil as reference for urban-
happy and rural-happy countries respectively.
One might also expect that if development were behind the happiness patterns,
rural-happy countries might have less people employed in the agricultural sector in rural
relative to urban areas than the urban-happy countries. In fact, the rural areas in rural-
happy countries do have more private workers and less farmers and fishers in relation to
the middle sized cities than the urban-happy countries, suggesting a redistribution of the
22
workers in rural areas from the primary to other sectors of the economy
9
. Thus, the
employment distribution also supports the idea that people living in rural areas in the
rural-happy countries are at a higher level of economic development than those in the
urban-happy countries.
A second possible explanation of the difference in happiness patterns between
rural-happy and urban-happy countries – the social values explanation – is that rural
family ties and security levels are higher in the rural-happy than in the urban-happy
countries. To assess the validity of this idea four main variables are considered: the
distribution of individuals by marital status (to account for family ties), the amount of
violent crime, the amount of corruption, and the religiosity of each area. It turns out that
these variables do not seem to be strongly related to the life satisfaction patterns (Table 2-
4). In both groups of countries there are significantly more married and less single
people in the villages and there are no significant differences in marital status between
the small towns and the middle sized cities. Crime and corruption are clearly lower in the
rural areas, but this is so for both rural-happy and urban-happy countries. Finally,
religiosity levels are not significantly different between the town sizes in either group of
countries. In general, therefore, the microeconomic data do not support the social values
hypothesis.
A third possible explanation is that higher public social spending levels increase
the availability of public services in the villages and small towns of rural-happy relative
to urban-happy countries. If this were the reason behind the life satisfaction patterns in
9
The other employment categories considered did not show any relevant patterns by size-of-place
in the two groups of countries analyzed and are therefore not included in the tables.
23
Latin America, then people in villages and small towns in the rural-happy countries
should be relatively more satisfied with their access to education and health than those
living in urban-happy countries. And, indeed, there does exist a clear relationship
between the patterns of satisfaction with access to education and satisfaction with access
to health and the country group. In rural-happy countries the access variables are rated
higher in rural areas relative to middle sized cities; in urban-happy countries both access
to education and health are rated higher in the middle sized cities. The pattern of
satisfaction with access to services corresponds perfectly to that of life satisfaction, with
life satisfaction being higher in rural areas for rural-happy countries and in middle sized
cities for urban-happy countries (Table 2-4). These matching satisfaction patterns support
the public social spending explanation.
Because satisfaction with access to education and to health, as well as life
satisfaction, all represent self-reported variables, one might argue that the similarity in
their patterns occurs because some people tend to report higher satisfaction levels,
regardless of the question asked, than others. But if this were the case, then the same type
of pattern should hold for any satisfaction variable available, not just those examined to
this point. To test this, an analysis was performed with variables rating personal
satisfaction with democracy and with the market system. These variables do not, in fact,
follow the same pattern as life satisfaction, implying that a reporting tendency is not
driving the association observed between life satisfaction and satisfaction with access to
services. The next section will further demonstrate the robustness of this relationship by
24
employing public spending information, external to the survey, to proxy for access to
services.
A fourth possible explanation of the life satisfaction patterns is that rural-happy
countries have smaller indigenous communities living in rural areas than urban-happy
countries. This explanation also seems supported by data. The percentage of indigenous
population (as measured by the amount of people reporting an indigenous language as
their mother tongue) seems to be quite strongly related to the life satisfaction patterns by
size-of-place. In rural-happy countries the proportion of indigenous population living in
small towns is not significantly larger than that living in middle sized cities. Though the
proportion of indigenous population is significantly higher in the villages than in middle
sized cities, the difference between the two size-of-place categories is relatively small in
rural-happy countries. In contrast, urban-happy countries have a significantly higher
proportion of indigenous population living in both villages and small towns than in
middle sized cities, and the differences between the rural and urban categories are
relatively high in comparison to those observed in rural-happy countries (Table 2-4). As
long as the indigenous population is less satisfied with life than the non-indigenous
population, the relatively smaller presence of indigenous communities in rural areas in
rural-happy countries could account for the observed difference in life satisfaction
between the rural-happy and urban-happy groups. This result is not driven by the higher
overall levels of total indigenous population in some urban-happy countries such as
Bolivia or Peru because it is based on relative differences between the various size-of-
place categories and not on the overall size of the indigenous community in a country.
25
2.4 Econometric analysis
The descriptive analysis suggests that the level of economic development, public social
spending, and indigenous population explanations might each account for the life
satisfaction by size-of-place differences between rural-happy and urban-happy countries,
and that only the social values hypothesis is unlikely to be supported by more rigorous
analysis. The descriptive analysis, of course, does not provide information about the
statistical significance of the individual explanations or their relative importance. As the
next step, therefore, an econometric assessment of the reasons behind the life satisfaction
patterns by size-of-place is carried out.
The analysis is performed using Ordinary Least Square regressions with life
satisfaction as the dependent variable. The regressions include dummies for the size-of-
place category (village or small town), with middle sized city as reference. The size-of-
place dummies capture the differences in life satisfaction between villages and small
towns respectively and middle sized cities. A dummy for the type of pattern of the
country in which the person lives (rural-happy or urban-happy) is also included, with
urban-happy countries as reference. The pattern type dummy captures the difference in
life satisfaction between rural-happy and urban-happy countries. Finally, an interaction of
the size-of-place with pattern type dummies is also included. This interaction captures the
difference in life satisfaction differentials by size-of-place between rural-happy and
urban-happy countries. For example, if villages are happier relative to middle sized cities
26
in rural-happy countries than in urban-happy countries, then this positive difference will
be captured by the interaction of the village and pattern-type dummy variables. The basic
regression used is the following:
(2.1) ls
ic
= α + β’x
ic
+ λ
1
village + λ
2
small + γ(rural-happy) + φ
1
(village*r_h) +
φ
2
(small*r_h) + ε
where:
- ls is the life satisfaction of individual i in country c,
- X is a vector of individual specific characteristics,
- village and small represent dummies for living in these two types of places,
- rural-happy is a dummy variable indicating a person lives in a rural-happy
country.
In regression (2.1) each λ
i
captures the mean differential in life satisfaction
between the respective size-of-place and the middle sized city for urban-happy countries
(the reference group), and each φ
i
indicates the difference in those differentials for rural-
happy and urban-happy countries. Based on the descriptive analysis φ
1
and φ
2
are
expected to be positive. Notice that regression (2.1) only includes observations for people
living three size-of-place categories – villages, small towns, and middle sized cities.
Observations for the large cities are excluded here, but are brought in later as a robustness
check.
27
In most of the analysis an expanded version of regression (2.1) is used, which
takes the form:
(2.2) ls
ic
= α + β’x
ic
+ λ
1
village + λ
2
small + γ(rural-happy) + φ
1
(village*r_h) +
φ
2
(small*r_h) + ω’z
ic
+ μ’s
c
+ ε
where z
ic
and s
c
are the individual and country specific characteristics. As more
explanatory variables are added to regression (2.1), φ
1
and φ
2
are expected to decrease in
magnitude. Consider the case where people in villages are more satisfied with life relative
to those living in medium sized cities in rural-happy countries strictly because they have
a higher ownership of goods than those in urban-happy countries. Then adding the
‘goods’ variable into the z
ic
vector should cause φ
1
to become insignificant. In what
follows both micro and macro level variables are added into z
ic
and s
c
respectively, to
identify their effects on φ
1
and φ
2
, and to assess their role in explaining the different
patterns observed in the two groups of Latin American countries.
In addition to the individual characteristics used in the descriptive analysis,
country level variables are also used in the econometric analysis in examining the
economic development and public social spending explanations. For development, the
country level variables included are the GDP per capita and its interactions with village
and small town. For public social spending, the country variables are public spending per
capita, total, and divided into spending on education, health, housing and social security.
28
The respective interactions of each spending type with village and small town are also
included.
The regressions testing each of the four explanations separately mostly confirm
the impressions from the descriptive analysis. The development, public social spending
and indigenous population explanations all reduce φ
1
and φ
2
, though this decline is
relatively small in the case of the indigenous population (Table 2-5, columns b and c-g).
In contrast, the social values variables do not reduce the difference in the size-of-place
differential coefficients (Table 2-5, column c).
Table 2-5: OLS regressions: life satisfaction as dependent variable, test of different
explanations separately for life satisfaction patterns
(a) (b) (c) (d) (e) (f) (g)
age
-0.013 -0.007 -0.015 -0.011 -0.011 -0.013 -0.012
(7.08)** (3.82)** (7.36)** (6.26)** (6.04)** (7.23)** (6.90)**
age_sq
0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001
(5.56)** (3.09)** (5.77)** (5.00)** (4.56)** (5.78)** (5.43)**
male
0.01 0.0004 0.025 0.013 0.011 0.009 0.011
(0.93) (0.04) (2.07)* (1.18) (1.03) (0.84) (0.96)
years_
educ
-0.0005
(0.31)
goods
0.041
(11.79)**
econ
sit
0.244
(33.56)**
mar.
0.027
(2.09)*
crime
-0.004
(0.3)
corrup
-tion
-0.108
(6.68)**
religio
-sity
0.071
(9.92)**
sat
educ
0.178
(28.06)**
sat
health
0.185
(29.50)**
29
Table 2-5 (Continued)
indige
-nous
-0.368
(14.14)**
village
-0.236 -1.17 -0.245 -0.212 -0.209 -0.351 -0.2
(9.48)** (4.65)** (9.15)** (8.39)** (8.71)** (11.69)** (8.03)**
small
town
-0.079 -0.06 -0.086 -0.063 -0.074 -0.115 -0.067
(3.26)** (0.26) (3.30)** (2.55)* (3.17)** (3.88)** (2.78)**
vill*
r-hap
0.307 0.144 0.338 0.262 0.238 0.209 0.272
(8.95)** (3.52)** (9.24)** (7.57)** (7.09)** (5.42)** (7.91)**
small*
r-hap
0.241 0.168 0.26 0.197 0.201 0.238 0.229
(7.87)** (4.77)** (8.04)** (6.40)** (6.74)** (7.14)** (7.50)**
lGDP*
vill
0.13
(4.01)**
lGDP*
small
0.004
(0.15)
PS*
vill
0.0003
(6.69)**
PS*
small
0.00006
(1.58)
rural-
happy
-0.032 -0.043 -0.014 -0.037 -0.027 0.015 -0.041
(1.2) (1.36) (0.49) (1.37) (1.03) (0.53) (1.54)
log
GDP
-0.023
(0.87)
pub.
spend.
-0.0002
(5.15)**
const. 3.325 2.485 3.195 2.814 2.798 3.401 3.327
(77.77)** (11.90)** (64.43)** (60.12)** (61.40)** (75.58)** (77.96)**
Obs. 28631 28439 25022 27653 28274 28631 28631
R-sq 0.03 0.1 0.04 0.06 0.07 0.03 0.04
Robust t statistics in parentheses
+ significant at 10%; * significant at 5%; ** significant at 1%
The regressions that include the specific public social spending variables provide
new insight into the effects of different services on the size-of-place differentials.
Spending on health, housing, and social security, are each behind the effects of public
spending on rural happiness. This is indicated by their influence on reducing φ
1
and φ
2
,
and by the positive and significant coefficients of their respective interactions with the
village dummy variable (Table 2-6, columns b-d). Spending on education, however, has
30
little influence on φ
1
, and its interactions with both village and small town lack
significance, suggesting that this type of public spending has a questionable relation to
the size-of-place patterns (Table 2-6, column a). Regressions including satisfaction with
access to health and education tend to support these findings: satisfaction with
access to health has a stronger effect on reducing the difference in the size-of-place
coefficients than satisfaction with access to education (Table 2-5, columns d and e)
10
.
Table 2-6: OLS regressions: life satisfaction as dependent variable, test of
importance of different types of public spending for life satisfaction patterns
(a) (b) (c) (d)
age -0.013 -0.013 -0.013 -0.013
(7.31)** (7.20)** (7.14)** (7.43)**
age_sq 0.0001 0.0001 0.0001 0.0001
(5.66)** (5.66)** (5.58)** (6.21)**
male 0.01 0.01 0.01 0.008
(0.92) (0.89) (0.93) (0.68)
village -0.227 -0.389 -0.361 -0.304
(6.17)** (12.53)** (11.39)** (11.17)**
small town -0.068 -0.164 -0.099 -0.105
(1.84)+ (5.35)** (3.26)** (3.96)**
village*rural-happy 0.258 0.231 0.26 0.219
(6.86)** (6.42)** (7.16)** (5.59)**
small*rural-happy 0.198 0.216 0.226 0.243
(6.06)** (6.89)** (7.10)** (7.14)**
PSeduc*village 0.0001
(0.44)
PSeduc*small 0.00004
(0.2)
PShealth*village 0.0015
(8.21)**
PShealth*small 0.0007
(4.65)**
PShouse*village 0.003
(4.89)**
10
The satisfaction with access to services variables were excluded from the later regressions run in
the analysis to avoid endogeneity concerns.
31
Table 2-6 (Continued)
PShouse*small 0.001
(1.09)
PSSocSec*village 0.0004
(6.46)**
PSSocSec*small 0.00004
(0.78)
rural-happy -0.072 -0.019 -0.031 0.083
(2.54)* (0.69) (1.12) (2.82)**
PS educ 0.001
(6.09)**
PS health -0.0006
(4.42)**
PS house -0.0001
(0.18)
PS Soc Sec -0.0005
(9.77)**
Constant 3.183 3.4 3.332 3.406
(66.28)** (74.75)** (73.54)** (78.09)**
Observations 28631 28631 28631 28631
R-squared 0.04 0.03 0.03 0.04
Robust t statistics in parentheses
+ significant at 10%; * significant at 5%; ** significant at 1%
Joint regressions on happiness including simultaneously the three explanations
that have, so far, been supported by the data, suggest that public social spending is the
most relevant factor behind the life satisfaction by size-of-place differences between
rural-happy and urban-happy countries. When the total public spending and the logarithm
of GDP variables, together with their interactions with the size-of-place categories, are
included in the regressions, the interactions of total public spending remain positive and
significant. In contrast, the interactions of the logarithm of GDP with villages and small
towns not only lose their positive significance, but become, in fact, significantly negative.
This suggests that an important channel through which the economic development of a
country influences the size-of-place life satisfaction differentials in Latin America is
32
through public social spending. Once this channel is controlled for the effects of the
development on the villages and small towns are no longer positive (Table 2-7, column
b).
Table 2-7: OLS regressions: life satisfaction as dependent variable, test of different
explanations combined for life satisfaction patterns
(a) (b) (c) (d) (e) (f)
age -0.013 -0.007 -0.006 -0.007 -0.007 -0.007
(7.08)** (4.09)** (3.58)** (3.79)** (3.67)** (4.23)**
age_sq 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001
(5.56)** (3.69)** (3.07)** (3.17)** (3.00)** (4.10)**
male 0.01 -0.002 -0.001 -0.00002 0.001 -0.003
(0.93) (0.14) (0.08) (0.88) (0.06) (0.3)
y_educ -0.001 0.002 -0.001 -0.001 0.001
(0.67) (1.08) (0.88) (0.81) (0.54)
goods 0.037 0.032 0.04 0.04 0.03
(10.80)** (9.20)** (11.29)** (11.33)** (8.70)**
econsit 0.253 0.242 0.249 0.248 0.254
(35.17)** (33.38)** (34.42)** (34.14)** (35.43)**
indigenous -0.251 -0.315 -0.295 -0.297 -0.248
(9.37)** (11.54)** (11.02)** (10.99)** (9.30)**
village -0.236 0.792 -2.71 0.909 1.114 0.437
(9.48)** (1.85)+ (6.48)** (2.29)* (2.45)* (1.18)
small town -0.079 1.166 -0.156 1.962 1.377 0.735
(3.26)** (3.00)** (0.41) (5.65)** (3.00)** (2.22)*
village* 0.307 0.219 0.086 0.261 0.248 0.197
rural-happy (8.95)** (5.26)** (2.05)* (5.88)** (6.11)** (4.88)**
small* 0.241 0.221 0.116 0.284 0.254 0.174
rural-happy (7.87)** (6.08)** (3.18)** (7.33)** (7.11)** (4.92)**
lGDP*village -0.138 0.371 -0.159 -0.183 -0.083
(2.36)* (6.28)** (2.92)** (2.91)** (1.67)+
lGDP*small -0.165 0.019 -0.283 -0.196 -0.102
(3.11)** (0.35) (5.95)** (3.10)** (2.31)*
PS*village 0.0003
(4.64)**
PS*small 0.0002
(3.29)**
PSeduc*village -0.002
(5.92)**
PSeduc*small -0.00002
(0.07)
PShealth*village 0.002
(6.35)**
33
Table 2-7 (Continued)
PShealth*small 0.002
(8.23)**
PShouse*village 0.004
(4.20)**
PShouse*small 0.003
(2.98)**
PSSocSec*village 0.0004
(4.09)**
PSSocSec*small 0.0002
(2.16)*
rural-happy -0.032 -0.072 0.017 -0.166 -0.144 0.019
(1.2) (2.21)* -0.53 (4.70)** (4.55)** (0.61)
logGDP 0.384 -0.308 0.293 0.235 0.299
(7.88)** (6.39)** (6.75)** (4.18)** (7.26)**
Pub. Spend. -0.0007
(12.09)**
PS educ 0.002
(7.21)**
PS health -0.002
(11.04)**
PS house -0.004
(4.71)**
PS Soc Sec -0.001
(13.76)**
Constant 3.325 -0.413 4.405 0.278 0.651 0.123
(77.77)** (1.17) (12.97)** (0.88) (1.61) (0.4)
Observations 28631 28439 28439 28439 28439 28439
R-squared 0.03 0.12 0.11 0.11 0.11 0.13
Robust t statistics in parentheses
+ significant at 10%; * significant at 5%; ** significant at 1%
When the individual components of public spending are included in a joint
regression with the logarithm of GDP per capita and its respective size-of-place
interactions, previous findings from the econometric analysis are confirmed. Spending on
health, housing and social security have a relatively higher positive effect on happiness in
rural as compared to middle sized places while public spending on education does not. In
the joint regressions, the positive interactions of public spending on health, housing and
34
social security with the villages and small towns remain significant, while the interactions
of the logarithm of GDP with the rural places become significantly negative (Table 2-7,
columns d-f). However, in regressions including both public spending on education and
the logarithm of GDP, the interactions of the spending variable with the villages and
small towns are found to be insignificant and the logarithm of GDP maintains its positive
effects on relative happiness in the rural areas (Table 2-7, column c). A possible reasons
why the public spending on education does not appear to raise relative life satisfaction in
rural areas is that higher levels of education may be accompanied by increasing
expectations regarding living standards. As expectations increase, life satisfaction is more
likely to remain constant in spite of improvements in objective conditions (Easterlin
2005). Another possibility is that the association of spending on education with happiness
of rural relative to middle sized city inhabitants, is reduced because public spending on
secondary and tertiary education in Latin America favors mostly the richer population
(Economic Commission for Latin America and the Caribbean 2006).
2.5 Conclusions
In Latin America two main patterns in happiness by size-of-place co-exist: in some
countries people in rural areas are happier than those living in middle sized cities; in
others rural inhabitants are less happy. Considered individually, higher levels of
economic development, higher public social spending, and a smaller proportion of
indigenous population in the rural areas all partially account for the higher rural life
35
satisfaction in the first group of countries, with the first two explanations having the
strongest association.
In joint regressions on life satisfaction that include the three explanatory factors,
public social spending is found to have the biggest influence on the size-of-place patterns,
with higher levels of spending being associated with higher relative life satisfaction in
rural areas. When different types of public social spending are considered separately, the
stronger effect of public spending than of GDP per capita on the happiness patterns holds
for spending on health, housing, and social security, but not for spending on education.
The results suggest that the relative improvement of happiness in rural areas is not an
automatic by-product of higher GDP. Rather, the relative improvement of happiness in
rural areas depends on public social spending policies.
The conclusions of this study are specific to Latin America, a region with
exceptionally high income inequality. Public social spending could possibly have a
stronger influence on the rural-urban differences in well-being in regions like Latin
America where the original inequality levels are high. An interesting issue for future
research is whether the present findings apply to countries with less initial income
inequality. Future studies of other regions could add insight on the interaction of
economic development and public policies in raising the relative well-being of rural
areas.
36
CHAPTER 3. Internal Migration and Life Satisfaction: Well-Being Paths of Young
Adult Migrants
3.1.Introduction
The life of a young adult is filled with changes and transitions. Choosing a community in
which to live, finishing education, getting married – these are important events
experienced early in life and that potentially influence future happiness. This chapter
discusses the association between one such life event – internal migration – and life
satisfaction. The particular relevance of studying migration is best illustrated by its
prevalence: in the United States, almost one-third of total population lived in a state
different from where they were born in 2009 (Molloy, Smith and Wozniak 2011). Is
internal migration of young adults accompanied by an increase in life satisfaction? Does
the change in migrants’ life satisfaction depend on the reason why the decision to move
was made? What are the aspects of life – such as housing or financial situation –
underlying overall life satisfaction that are altered following internal migration? These
are the questions addressed here.
A longitudinal survey of young adults in Sweden along with collated information
from Statistics Sweden are used. To evaluate the association between internal migration
and life satisfaction, life satisfaction levels of migrants and non-migrants are compared
before and after the move. To assure comparability between these two groups, the
analysis controls for the main differences between them such as fixed personality traits,
37
shocks to the community of origin, and various life transitions that often accompany
migration. Migrants who move for work related reasons and those who move for other
(non-work related) reasons are analyzed separately to allow for different outcomes
depending on the reason for moving. Other transitions that are characteristic of young
adults, such as education completion, changes in marital status, and the birth of a child,
are controlled for to avoid confounding effects. After investigating the change in life
satisfaction following migration, an analysis of changes in specific life aspects for the
movers is carried out to identify the possible factors underlying the association between
well-being and migration.
The decision to move is usually treated in economics as the result of a cost-benefit
analysis, and as such it is expected to increase an individual’s utility. Results of previous
studies suggest that internal migration is generally accompanied by improvements in
objective circumstances that partially depend on the reason for moving: income for
migrants who move due to work, and housing conditions for those who move for
residential reasons. The association between migration and subjective well-being has
only become a focus of study in recent years and considerably less research has been
conducted in this area. Findings to date suggest that internal migration is positively
associated with housing satisfaction, but its association with overall life satisfaction
remains unclear.
The present study contributes to previous literature by analyzing the association
between life satisfaction and internal migration for a specific group of migrants: young
adults ages 22 to 30 before the move. The focus on young adults is motivated by previous
38
findings that the destination and reason for moving may vary by age, and by the high
prevalence of internal moves among young cohorts. Additionally, the analysis
distinguishes between those who move for work, and those who move for other reasons.
Doing this further reduces the diversity in the migrant sample and provides new
information on the differences in well-being changes for work and non-work migrants.
Life satisfaction depends on many aspects of one’s life, such as the financial,
health, housing, and job situation. These different parts of everyday life are commonly
referred to in the subjective well-being literature as “life domains” (Rojas 2006). The
final step of the analysis acknowledges the importance of these domains in explaining life
satisfaction trends by examining the life aspects altered following a move. Specifically,
changes in the financial, housing, and job domains accompanying migration are studied
separately for work and for non-work migrants to clarify the possible factors underlying
the association between migration and life satisfaction.
3.2. Literature review
In both economics and demography migration is typically viewed as the result of a cost-
benefit analysis in which people evaluate various monetary and non-monetary aspects of
moving and make the decision to migrate if they believe this will maximize their utility
(De Jong et al. 1983; Harris and Todaro 1970; Sjaastad 1962; Speare 1974). The
monetary factors considered in this decision usually include income and labor market
39
opportunities (Bartel 1979; Ghatak, Levine and Price 1996); the non-monetary aspect
mostly considered has been residential satisfaction (Diaz-Serrano and Stoyanova 2009).
A vast empirical literature has been developed to evaluate whether internal
migration leads to an increase in income. Early studies based on cross-sectional analyses
in which income levels of migrants are compared to those of non-migrants (either from
the place of origin or the place of destination), provide mixed results possibly due to a
selection bias (Lansing and Morgan 1967; Weiss and Williamson 1972). More recent
studies have used panel data which allows to control for fixed differences between
migrants and non-migrants thus accounting for a large part of the selection bias. The
results of these panel studies suggest that the association between migration and income
gains is complex and depends on age, reason to move, gender, and marital status.
Generally, young males who move due to work-related reasons experience the highest
income gains from migration (Bartel 1979; Böheim and Taylor 2007; Finnie 1999).
However, the positive association between migration and income does not always hold,
as in the case of married women whose incomes may decline following a move
(Blackburn 2009; Cooke and Bailey 1996; Morrison and Clark, 2011).
The issue of residential mobility and its outcomes has also been vastly explored.
Residential dissatisfaction due to poor dwelling conditions is often found to be an
important predictor of migration (Diaz and Stoyanova 2009; Speare 1974). Whether the
decision to migrate improves a person’s housing conditions may depend on various
circumstances including life-cycle events and the reasons that trigger the move (Barcus
2004; and Rabe and Taylor 2010).
40
Analyses of the association between migration and well-being using life
satisfaction measures are less common. Even if income and housing conditions are found
to improve following migration, this may not translate into higher life satisfaction if
aspirations increase together with the objective circumstances (Easterlin 2001a; Easterlin
and Angelescu 2009). In empirical analysis, an ongoing debate about the long-term
importance of income in affecting life satisfaction changes remains unsettled (Frijters,
Haisken-DeNew and Shields 2004; Frijters et. al. 2006; Oswald 1997). Additionally, the
financial and housing domains are not the only aspects of life likely to change after
migration, implying the need for a more comprehensive well-being measure (such as life
satisfaction) to assess the final well-being changes of migrants.
Studies using cross-sectional analyses to compare life satisfaction of migrants
after the move to that of non-migrants typically find a negative association between
migration and life satisfaction (Bartram 2011; Knight and Gunatilaka 2011). Still, cross-
sectional comparisons may suffer from a selection bias due to differences between
migrants and non-migrants. To avoid this bias, De Jong, Chamratrithirong and Tran
(2002) used questions about the migrants’ own perception of how the move had affected
their satisfaction with employment conditions, living environment, and community
facilities. Their findings suggest that a non-trivial proportion of migrants report decreased
satisfaction levels after the move. These results, however, could be affected by the
existence of a recall bias in past satisfaction levels (Easterlin 2001a).
The results of panel studies assessing the relationship between internal migration
and life satisfaction are mixed. In a paper focusing on residential migrants, Nakazato,
41
Schimmack and Oishi (2011) find that while housing satisfaction does increase following
migration, overall life satisfaction does not. A different study considering migrants from
East to West Germany, however, does find a positive long-term association between
migration and life satisfaction (Melzer 2011). Finally, in two recent papers Nowok et al.
(2013) and Findlay and Nowok (2012) find that life satisfaction of migrants deteriorates
prior to the move and recovers at the time of the migration. Though these studies do not
find any long lasting effects of migration on life satisfaction (Nowok et al. 2013), they do
observe significant long-lasting improvements in housing satisfaction for the migrants
(Findlay and Nowok 2012).
Differences in the composition of the migrant sample may provide an explanation
for the mixed results of previous longitudinal studies. The main focus in the analysis by
Nakazato, Schimmack and Oishi (2011), which finds a positive association between
migration and housing (but not life) satisfaction, is on residential migrants. In contrast,
the studies by Nowok et al. (2013), and Findlay and Nowok (2012) do not impose any
restrictions on the migrant sample. Their findings of no change in life satisfaction may
therefore be due to confounding changes for migrants with different characteristics.
Finally, while Melzer (2011) does not restrict the migrant sample, considering the
circumstances of the German re-unification it is likely that the migrants in her analysis
were mostly young people moving for work reasons. The present study contributes to
previous literature by focusing on more homogenous groups of migrants: young adults
who move for work and those who move for non-work reasons.
42
3.3. Data description
The main data source is the Young Adult Panel Study (YAPS) collated with Swedish
register information (www.suda.su.se/yaps). The YAPS consists of a longitudinal survey
designed by Eva Bernhardt from Stockholm University, carried out in Sweden in the
years 1999, 2003, and 2009. Of these three years, two are used in the analysis,
corresponding to the surveys conducted in 1999 and 2009 respectively. Main socio-
economic characteristics (such as civil status or income) are obtained from the Swedish
register records which were linked with the survey information for respondents
interviewed in 2009. The analysis was restricted to years 1999 and 2009 because register
data was missing for a large portion of the population interviewed in 2003.
Although YAPS contains information for over 3000 individuals, only a portion of
the respondents participated in all consecutive surveys. The sample under study is
restricted to those interviewed in both 1999 and 2009, and for whom information on the
main variables of interest is available. From the 2820 people initially interviewed in
1999, 56% were re-interviewed ten years later reducing the sample of observations to
1575 individuals, a small portion of whom did not answer some of the relevant questions
and had to be dropped from the regression analysis
11
. The high attrition rate may create
worries about the possible existence of a selection bias. The methodology used
throughout the analysis, which controls for individual fixed effects, community-specific
shocks, and a number of time-varying observable characteristics, should account for an
11
For complete information on the number of observations available for each of the main
variables included in the study, see Table B-6, Appendix B.2.
43
important part of the differences between attritors and non-attritors. An additional
analysis of the remaining differences between attritors and non-attritors provides
reassurance that the residual selection bias is small in magnitude, and is therefore
unlikely to influence the results of the study (Appendix B.1).
The two main variables employed in the analysis are life satisfaction and
migration. Life satisfaction is measured in all waves of the YAPS using the answer to the
question: “How satisfied are you with your life in general?”. Response categories are
given on a scale from 1 to 5, with 1 corresponding to “very dissatisfied” and 5 to “ very
satisfied”. Migration status is established using information on the place of residence in
1999, 2003 and 2009. A person is classified as a migrant if he/she changed his/her
municipality of residence in the years under analysis (including those who reported a
different municipality in 2003 and later moved back), and as a non-migrant if no such
change took place. Sweden is organized into 290 municipalities grouped within 21
counties. Around half of those classified as migrants changed their county as well as
municipality of residence, of which a big proportion (69%) involved moves between
counties separated by 100 miles or more
12
. For those migrants who changed
municipalities within a county the distance traveled was smaller, averaging about 25
miles. Because of this difference, robustness tests separating the between- and within-
county movers are conducted.
The question used to divide migrants into work and non-work migrants was
included in 2009 only and asks: “What was the most important reason for you to move?”
12
The distance traveled by those who changed counties of residence was roughly approximated
using the distance between the centers of county of origin and county of destination.
44
The possible response categories for this question include “my work/studies” as well as
other seven options (Table B-7, Appendix B.2). Using the answer to this question,
migrants were classified as either work migrants if they chose “my work/studies” as their
main reason to move, or non-work migrants if they chose any of the other response
categories. Given the long time span between the two surveys, migrants are also divided
based on the self-reported year in which they moved into more recent migrants (if they
had moved less than six years before the 2009 survey), and less recent migrants (if they
had moved six years or more ago).
13
Because of limited domain information in the YAPS survey, the life domain
analysis is restricted to aspects related to the financial, job, and housing domains.
Variables used to assess changes in these domains include: work income, relative income,
and economic satisfaction for the financial domain; occupation status, and satisfaction
with what the person is currently doing for the job domain; and satisfaction with housing
for the housing domain. Work income is given on individual level and is adjusted for
inflation using the Consumer Price Index obtained from Statistics Sweden. Income from
the years previous to the survey (1998 and 2008 respectively) is used because interviews
were conducted at the beginning of a year and so satisfaction levels of the respondents
are likely to reflect their past years’ income. Relative income is constructed as the
difference between an individual’s income and the average income of his/her
13
The threshold of six years is chosen because it allows to split the movers into roughly two equal
sized groups, assuring an appropriate number of observations in both the less and more recent migrant
categories.
45
municipality of residence with robustness tests changing the reference category to
municipality of origin.
To create the occupation status, respondents are first classified into one of nine
occupational categories constructed combining survey questions on main occupation and
on main activity. Subsequently the nine occupation categories are divided into four
groups depending on the respective status associated with each occupation. The criteria
for this division are based on the Standard International Occupational Prestige Scale
(SIOPS) as updated by Ganzeboom and Treiman (1996). The final categories represent
an ordinal variable on a scale of 1 to 4, with 1 corresponding to the lowest and 4 to the
highest occupation status.
Three additional satisfaction variables are used in the domain analysis:
satisfaction with economic situation, with housing, and with what the person is currently
doing.
14
Though satisfaction with what the person is currently doing is used to capture
occupational satisfaction, it represents an imperfect measure of the job domain
15
.
Therefore occupational status represents the preferred job domain measure, and
satisfaction with what the person is currently doing (from now on referred to as
satisfaction with occupation) is used to complement the analysis.
The control variables used include birth cohort, change in marital status,
education completion, birth of a child, and the final education level. Given the young age
14
Satisfaction with relationship with partner, though available in the survey, is not used due to
high non-response rates in both years (Table B-6, Appendix B.2).
15
The response to this question measures satisfaction with any activity that the person was
currently doing, which should most often, but not always, be interpreted as occupation. Additionally, the
question prior to this changed in between 1999 and 2009 from one related to work (importance of being
successful at work) to one related to religion (importance of religion).
46
of the subjects surveyed, the widowed and divorced groups are very small and are
therefore combined for the purpose of the analysis. Education completion and birth of a
child are measured by bivariate variables with value 1 if the given event took place
between 1999 and 2009, and 0 otherwise. Lastly, final education level is a categorical
variable that represent the highest level of education achieved by 2009. For additional
description of these variables, see Appendix B.2.
Migrants in this study are mostly young, unmarried, and have higher final
education levels than non-migrants (Table 3-1). While in 1999 migrants are more likely
to be unmarried than non-migrants, in 2009 the marriage rates of the two groups are very
close. Similar patterns are true for education completion and parenting: in 1999 migrants
are less likely to have experienced either of these events, though by 2009 the likelihoods
of education completion and having a child for migrants and non-migrants are nearly the
same. Regarding the satisfaction levels, migrants are generally less satisfied with their
life, financial situation, and housing before (but not after) the move. Finally, important
differences between work and non-work migrants can be observed reflecting the need to
study the two groups separately.
Table 3-1: Descriptive statistics of migrants and non-migrants before and after the
move, by reason to move
Statistics before the move (1999)
Averages
Work
migrants
Non-work
migrants
All
migrants
Non-
migrants
Total
Life satisfaction 3.79 3.87 3.85 3.97 3.92
Work income (in ‘000 SEK) 96.57 109.53 109.84 125.14 118.69
Relative income (in ‘000 SEK) -27.92 -14.12 -16.23 2.95 -5.04
Income muni of residence 124.50 123.64 126.08 122.19 123.80
Occupation status 1.59 1.73 1.70 1.84 1.78
Mean economic satisfaction 3.07 3.06 3.08 3.13 3.11
Satisfaction with occupation 3.99 3.84 3.88 3.76 3.81
47
Table 3-1 (Continued)
Housing satisfaction 3.41 3.57 3.53 3.80 3.69
Percent male 50.0% 42.7% 45.5% 43.9% 44.6%
Percent unmarried 93.8% 89.3% 91.4% 81.8% 85.8%
Percent married 5.8% 9.2% 7.4% 17.4% 13.2%
Percent divorced/widowed 0.4% 1.4% 1.2% 0.9% 1.0%
Percent with educ complete 38.5% 49.3% 46.5% 63.4% 56.4%
Percent with post-sec educ 61.5% 50.4% 52.1% 36.4% 43.0%
Percent with child in household 6.2% 13.5% 11.0% 35.8% 25.6%
Percent from 1976 cohort 53.5% 45.8% 47.8% 29.9% 37.4%
Percent from 1972 cohort 27.9% 33.7% 32.3% 36.1% 34.5%
Percent from 1968 cohort 18.6% 20.5% 20.0% 33.9% 28.1%
Percent of total 14.37% 22.06% 41.39% 58.61% 100.00%
Statistics after the move (2009)
Work
migrants
Non-work
migrants
All
migrants
Non-
migrants
Total
Life satisfaction 3.99 3.99 4.01 3.92 3.96
Work income (in ‘000 SEK) 270.08 218.34 241.51 229.99 234.64
Relative income (in ‘000 SEK) 101.22 57.53 75.51 74.12 74.56
Income muni of residence 168.86 160.81 166.00 155.87 160.08
Occupation status 2.41 2.24 2.29 2.13 2.20
Mean economic satisfaction 3.68 3.47 3.58 3.49 3.53
Satisfaction with occupation 4.04 3.92 3.97 3.92 3.94
Housing satisfaction 3.80 4.03 3.97 3.99 3.98
Percent male 50.0% 42.7% 45.5% 43.9% 44.5%
Percent unmarried 54.9% 47.0% 50.2% 49.2% 49.6%
Percent married 41.6% 49.0% 45.9% 44.9% 45.4%
Percent divorced/widowed 3.5% 4.0% 3.8% 5.9% 5.0%
Percent with educ complete 91.2% 92.2% 92.0% 91.8% 91.9%
Percent with post-sec educ 77.9% 64.3% 67.1% 48.4% 56.2%
Percent with child in household 49.1% 70.3% 63.5% 69.5% 67.0%
Percent from 1976 cohort 53.5% 45.8% 47.8% 29.9% 37.4%
Percent from 1972 cohort 27.9% 33.7% 32.3% 36.1% 34.5%
Percent from 1968 cohort 18.6% 20.5% 20.0% 33.9% 28.1%
Percent of total 14.37% 22.06% 41.39% 58.61% 100.00%
Because of missing values for the reason-to-move variable, not all migrants could be classified as
work or non-work migrants.
The averages all calculated for all respondents available for 1999 and 2009, and answering the
question.
48
3.4. Methods
The main problem faced in the analysis of the association between migration and life
satisfaction is the lack of a perfect control group. Though optimally one would like to
compare the migrants’ life satisfaction to what it would have been had they not moved, in
practice this counterfactual is not observed. Therefore, one is left with the second-best
option: comparing the life satisfaction of migrants to that of non-migrants controlling for
the differences between the two groups. These differences may be either observable (such
as marital status) or unobservable (such as personality), and may explain some of the
relative well-being improvements following migration (Pekkala and Tervo 2002).
Observable differences may be accounted for using appropriate control variables.
Controlling for unobservable differences, however, may be more challenging.
The following analysis controls for all unobservable differences between migrants
and non-migrants that are either fixed at the individual level (such as personality traits),
or that represent one-time community-level shocks associated with migration. An
example of the latter is an economic crisis that induces massive layoffs. Massive layoffs
could permanently lower life satisfaction among the residents of the affected community
at the same time as making them more likely to migrate to a region not hit by the crisis.
For a shock of this type to affect both the change in life satisfaction and the likelihood of
migration it needs to take place between times 0 and 1 (implying its effect would be
present in time 1 but not 0). Such unobservable shocks to the migrant’s community-of-
49
origin, while representing a potentially important source of bias, have rarely been
controlled for in internal migration literature.
The following econometric model represents the life satisfaction of individual i, in
community c, at time t, taking into account the effects previously described:
(3.1) Y
ict
= μ
t
+ η
i
+ θ
co
*t + γM
i
*t + β’x
it
+ε
ict
Where Y
ict
is the outcome variable of interest (in this case life satisfaction); μ
t
is a time
effect; η
i
is the individual fixed effect; θ
co
captures the effect of the shock to the
community of origin; t is a time dummy; M
i
is a migration dummy equal to 1 for migrants
and 0 for non-migrants; x
it
is a vector of observable characteristics; and ε
ict
is an error
term that is allowed to be correlated for the same individual over time, and for different
individuals within a community. The effect of the shock to the community of origin on
life satisfaction captured by θ
co
is only present at time 1 (after the shock), which is why it
appears interacted with a time dummy in the model
16
.
With the two period approach used in the analysis (where 1999 and 2009
represent times 0 and 1 respectively), the fixed effects model is analytically equivalent to
a first-difference model. Therefore the above specification (3.1) may be implemented
using the following first-difference regression:
16
This statement holds under the assumption that the shock is related to the decision to migrate
and therefore migrants will have been present at community c during its occurrence and will only make the
decision to move after this event. If no shock occurs at a community between periods 0 and 1 or if a shock
takes place that is unrelated to the migration decision, then it would not be a source of endogeneity and so it
would not bias the results. In that case θ
c(t-1)
= 0.
50
(3.2) ΔY
ic
= λ
0,1
+ θ
co
+ γM
i
+ β’Δx
i
+ Δε
ic
Where λ
0,1
captures a time trend between periods 0 and 1; the individual fixed effects
have been eliminated; the community of origin shock is controlled for by including a
vector of community-of-origin dummies represented by θ
co
; Δx
i
controls for changes in
observable characteristics; and γM
i
captures the association between migration and the
change in life satisfaction. The community dummies denote the county, not municipality,
of residence because of the large number of municipalities (over 250) which complicates
the use of municipality dummies.
The estimation procedure employs first difference OLS regressions
17
. Standard
errors are clustered according to the community of residence at both times 0 and 1
18
and
robustness tests clustering standard errors at the level of municipality and county of
origin are conducted. To distinguish the different trends in life satisfaction depending on
the reason for moving all regressions are also run using separate dummies for work and
non-work migrants. Finally, given the long time span in between the two surveys (ten
years), separate regressions are run for more recent movers (moved within the past six
17
Though life satifsaction and some of the other dependent variables are ordinal, the first
difference OLS model is preferred due to the complications arising from assuming fixed-effects with
ordered models (Wooldridge 2002). Additionally, it has been shown that assuming either ordinality or
cardinality of satisfaction answers provides virtually the same empirical results, and that the benefits of
including fixed-effects exceed the losses of using a non-linear model in these estimations (Ferrer-i-
Carbonell and Frijters 2004).
18
This implies that with two communities, for example, four separate clusters would be used: for
those living in community c
a
at times 0 and 1, c
b
at times 0 and 1, c
a
at time 0 and c
b
at time 1, and c
b
at
time 0 and c
a
at time 1.
51
years) and less recent (moved six years or more before 2009) movers. Non-migrants are
used as the reference category throughout the analysis.
The observable characteristics included in (3.2) represent common life events that
are likely to influence both life satisfaction and the likelihood of migration. Specifically,
changes in marital status, completion of formal education, and the birth of a child are
controlled for. All these are potentially more likely to take place for migrants than non-
migrants, and to have significant effects on life satisfaction (Chen and Rosenthal 2008;
Myrskyla and Margolis 2012; Rabe and Taylor 2010; Zimmerman and Easterlin 2006).
Additional control variables include birth cohort of the respondent, and final level of
education. These allow for differences in life satisfaction trends depending on the
person’s age and educational achievement, both of which have been suggested to exist by
previous literature (Blanchflower and Oswald 2008; Easterlin 2001b). Though changes in
occupation may affect life satisfaction and migration, they are not included as control
variables since changes in the job domain (including improvements in occupation status)
are considered as a possible factor underlying the association of migration with life
satisfaction.
The main assumption behind (3.2), is that the individual and community effects
described are the only unobservable sources of endogeneity. In reality other sources, like
individual time-varying differences between migrants and non-migrants, may exist. For
example, migrants may represent a select sample of highly motivated respondents whose
life satisfaction would increase regardless of whether they had moved or not. The
analysis partially accounts for the higher motivation of the migrants in two ways. First,
52
controlling for final level of education should capture some of the effects of a person’s
motivational profile. Second, as an additional test, the well-being change of migrants
whose occupational status increased during the period under analysis is compared to that
of non-migrants with a similarly high occupational trajectory. Still, it should be
recognized that unobserved time-varying differences may remain a problem. To account
for these an instrumental variable could be used. However, suitable instruments for
migration are difficult to obtain and have rarely been found (for an example, see Munshi
2003). The use of life satisfaction as a dependent variable creates further complications as
few factors affecting a person’s decisions (such as a natural disaster, or the place where
they live) are likely to satisfy the second stage assumption of the instrumental analysis.
Since inappropriate instruments may lead to substantial biases (Wooldridge 2002), the
model used is considered to represent a suitable approach given the limitations.
Out of the 643 migrants in the analysis, 77 did not answer the reason to move
question. This implies a great loss of power when migrants are divided into work and
non-work movers. Two methods are used to deal with the missing data: likewise deletion
and multiple imputation (MI)
19
. The MI method used is imputation by chained commands
(ICE), in which imputed values for the missing variable are generated from a series of
multivariate models based on a group of personal features
20
, and only the imputed values
19
Out of the traditional techniques employed to treat missing data, likewise deletion has been
suggested to be as good as any of the other approaches. However, when large proportions of data are
missing more advanced methods, such as multiple imputation, have been found to work best (Scheffer
2002).
20
The exact model for the multiple imputation of reason to migrate (a binary variable for migrants
defined as work or other) included the following variables: birth cohort, dummies for completion of
education and birth of a child, changes in civil status, life satisfaction in 99 and 09, work income in 99 and
09, occupation status in 99 and 09, satisfaction with housing in 99 and 09, economic satisfaction in 99 and
53
for the main variable of interest are kept to avoid introducing further noise into the
estimation (method suggested in von Hippel 2007). ICE was preferred over multivariate
normal imputation because of its superiority imputing ordinal variables.
The factors underlying the relationship between migration and life satisfaction are
examined by considering changes in the different life aspects (or life domains) that
compose overall life satisfaction. The analysis of life domains is not new to the
subjective well-being literature (Rojas 2006; Easterlin and Sawangfa 2009). For each of
the three life domains considered – financial, housing, and job – its relationship with
migration is assessed using regressions with dependent variables related to this domain.
The main assumption is that if an increase in life satisfaction for migrants is accompanied
by improvements in a given life domain, then this domain represents a likely factor
underlying the migration/life satisfaction relationship.
3.5. Results
3.5.1 Migration and life-satisfaction
The change in life satisfaction following migration is generally positive, but whether or
not it remains significant six to ten years after the move depends on the reason for
moving. Pooling all migrants, a significant increase in life satisfaction (relative to non-
09, satisfaction with occupation in 99 and 09. For more information on the ICE method and how its results
compare to other imputation techniques see Ambler and Omar 2007.
54
migrants) is observed for both those who moved less than six years, and those who
moved six years or more prior to 2009 (Table 3-2, Panel A). These findings are robust to
the specification: the coefficients on migration are highest in a reduced form regression
where migration is the only explanatory variable, and fall slightly (but remain significant)
when variables allowing for differential time trends by final level of education and
cohort, as well as additional socio-demographic variables, are included.
Table 3-2: OLS regressions: change in life satisfaction as dependent variable,
migration (all and by reason to move) as main explanatory variable
Panel A: OLS regressions for all migrants pooled
Whole sample
More recent migrants
(<6y)
Less recent migrants
(6y+)
1 2 3 4 5 6
all migs 0.209 0.165 0.231 0.187 0.185 0.139
(3.65)*** (2.56)** (3.07)*** (2.16)** (3.06)*** (2.20)**
married_fd 0.003 -0.051 0.014
(0.05) (0.63) (0.15)
div/wid_fd -0.048 0.003 -0.051
(0.41) (0.03) (0.35)
educ comp 0.081 0.08 0.083
(1.61) (1.28) (1.42)
child birth -0.025 -0.04 -0.014
(0.38) (0.47) (0.18)
final educ 0.011 0.019 -0.007
(0.45) (0.69) (0.31)
cohort
dummies no yes no yes no yes
county of
origin no yes no yes no yes
Obs 1541 1532 1216 1208 1211 1203
R-squared 0.01 0.02 0.01 0.03 0.01 0.02
Panel B: OLS regressions for work and non-work migrants separately
Whole sample More recent migs (<6y) Less recent migs (6y+)
1 2 3 4 5 6
work migs 0.266 0.214 0.305 0.239 0.237 0.189
(3.13)*** (2.3)** (2.38)** (1.73)* (2.64)*** (2.11)*
non-work 0.175 0.139 0.199 0.166 0.144 0.101
migs (2.86)*** (2.11)** (2.52)** (1.89)* (1.92)* (1.34)
married_fd 0.003 -0.050 0.011
(0.04) (0.6) (0.12)
55
Table 3-2 (Continued)
div/wid_fd -0.051 0.007 -0.059
(0.44) (0.05) (0.41)
educ comp 0.081 0.081 0.082
(1.61) (1.31) (1.39)
child birth -0.018 -0.036 -0.009
(0.28) (0.43) (0.11)
final educ 0.008 0.018 -0.009
(0.35) (0.64) (0.38)
cohort
dummies No yes no yes no yes
county of
origin No yes no yes no yes
Obs 1541 1532 1216 1208 1211 1203
t-statistics in parentheses, standard errors clustered at change in county level
* significant at 10%; ** significant at 5%; *** significant at 1%
Dividing migrants into those who move for work related, and those who move for
non-work related reasons, interesting differences between the two groups are found.
While work migrants experience a significant long-lasting increase in life satisfaction, the
life satisfaction of non-work migrants increases significantly only for those who moved
within the last six years, but not for those who moved six years or more prior to 2009
(Table 3-2, Panel B). Focusing on work movers first, the positive association between
migration and life satisfaction is strongest using the reduced form specification,
regardless of time gone by since the move. Controlling for socio-demographic variables
such as final level of education and birth cohort, the positive coefficients on both more
and less recent work migration are slightly weakened but remain significant. (Table 3-2,
Panel B). Using the preferred specification (with the full set of control variables) the
differential increase in life satisfaction for the pooled sample of more and less recent
work migrants (relative to non-migrants) is approximately 0.21 which seems sizeable
56
considering that the change in life satisfaction for the average young adult over the same
time period was only 0.04 (Table 3-1).
For non-work migrants the story is slightly different. With the preferred
specification, only those who moved within the last six years display a significant
increase in life satisfaction above that of non-migrants. For those who moved six years or
more prior to 2009, the association between non-work migration and life satisfaction
remains positive, but loses its significance (Table 3-2, Panel B, Columns 7 and 11). For
those non-work migrants who moved within the last six years and for whom a significant
association is found, the magnitude of the increase in life satisfaction above that of non-
migrants is approximately 0.17. This magnitude is lower than that experienced by recent
work migrants (0.24), but remains significant and sizeable compared to the average
change in life satisfaction (0.04). For the less recent non-work migrants, however, the
positive association with life satisfaction loses its significance in all specifications except
for the reduced form regressions where it remains only marginally significant at 10%.
The findings regarding the association between internal migration and life
satisfaction suggest that a weaker long term increase in life satisfaction accompanies non-
work migration than work migration. The difference in the change in life satisfaction
between work than non-work migrants could be due to different factors underlying the
well-being improvement for the two groups. This possibility is further discussed in what
follows. Because of the overall robustness of results to the specification, the remainder of
the analysis uses the preferred specification that includes the full set of socio-
demographic and county of origin variables.
57
3.5.2 Life domains behind the migration and life satisfaction association
In the case of migrants who move for work reasons, changes in life domains following
the move are complex. In the short term, work migrants experience an improvement
relative to non-migrants in the job domain (specifically in occupational status) (Table 3-
3), but not in the economic or housing domains (Tables 3-4 and 3-5). In the long term, the
relative improvement in occupational status for work migrants remains, and is joined by
significantly higher levels of absolute and relative income and housing satisfaction as
compared to non-migrants.
Taking a detailed look at the job domain, relative improvements in occupational
status can be observed for both, more and less recent work movers suggesting that work-
related migration is followed by a lasting improvement in this domain (Table 3-3,
Columns 2, 4, and 6). The magnitude of this increase in occupational status for work
migrants is substantial. The status change associated with work migration represents
between one third and one half of the status improvement due to education completion,
and is stronger than the relation between status and final education level. Interestingly,
however, satisfaction with occupation does not improve for either the more, or the less
recent work migrants (Table 3-3, Columns 8, 10, and 12). The lack of a significant
change in satisfaction with occupation could be due to the long work hours associated
58
with jobs that provide a high relative status
21
as increased hours of work have been found
to decrease satisfaction (Clark and Oswald 1996; Rätzel 2012). However, it is also
possible that the absence of a significant change in occupational satisfaction is due to this
question’s limitations
22
. Though either reason could be true, the latter seems more likely
given the lack of an association between other factors that one could expect to affect job
satisfaction (such as education completion or final level of education) and satisfaction
with occupation as measured here.
Table 3-3: OLS regressions: job domain variables as dependent variables,
migration (all and by reason to move) as main explanatory variables
Changes in Occupational Status
Whole sample
(10 years)
More recent migs
(less than 6 years)
Less recent migs
(6 years or more)
all migrants 0.159 0.216 0.137
(3.37)** (2.34)* (2.34)*
work migrant 0.394 0.378 0.417
(4.76)** (3.31)** (3.58)**
non-work 0.031 0.15 -0.063
migrant (0.55) (1.38) (1.06)
cohort 1972 -0.32 -0.309 -0.381 -0.377 -0.261 -0.254
(4.63)** (4.68)** (5.26)** (5.31)** (3.84)** (3.91)**
cohort 1968 -0.309 -0.307 -0.307 -0.307 -0.227 -0.225
(6.26)** (6.24)** (5.60)** (5.61)** (4.18)** (4.07)**
married_fd 0.213 0.207 0.144 0.147 0.208 0.192
(4.37)** (4.29)** (2.47)* (2.55)* (3.40)** (3.13)**
div/wid_fd 0.165 0.151 0.169 0.18 -0.042 -0.085
(1.06) (0.98) (0.88) (0.94) (0.23) (0.46)
educ completion 1.181 1.168 1.164 1.160 1.175 1.155
(19.53)** (19.19)** (16.67)** (16.74)** (17.18)** (17.17)**
child birth -0.148 -0.121 -0.134 -0.124 -0.125 -0.099
(3.80)** (3.22)** (2.63)** (2.45)* (2.17)* (1.74)+
Constant 0.485 0.464 0.547 0.54 0.389 0.375
(6.58)** (7.26)** (6.71)** (6.86)** (8.45)** (9.73)**
21
While hours worked increased for those whose occupational status improved between 1999 and
2009 by 5.73, for those for whom status remained the same hours worked decreased by -2.62.
22
For a full discussion of this variable, refer to the data description section.
59
Table 3-3 (Continued)
Observations 1467 1467 1152 1152 1150 1150
R-squared 0.35 0.35 0.34
Changes in Satisfactin with Occupation
Whole sample
(10 years)
More recent migs
(less than 6 years)
Less recent migs
(6 years or more)
all migrants -0.061 -0.016 -0.079
(0.69) (0.14) (0.89)
work migrant -0.151 -0.047 -0.196
(1.36) (0.35) (1.44)
non-work -0.012 -0.002 0.006
migrant (0.12) (0.02) (0.06)
cohort 1972 -0.032 -0.035 -0.1 -0.101 -0.036 -0.038
(0.42) (0.47) (1.13) (1.13) (0.44) (0.47)
cohort 1968 0.043 0.043 0.025 0.0248 0.002 0.002
(0.43) (0.42) (0.19) (0.19) (0.02) (0.01)
married_fd -0.047 -0.044 -0.086 -0.086 -0.07 -0.063
(0.68) (0.64) (1.02) (1.03) (0.91) (0.8)
div/wid_fd 0.089 0.094 0.072 0.069 0.04 0.060
(0.66) (0.69) (0.32) (0.31) (0.25) (0.35)
educ completion 0.1 0.107 0.153 0.154 0.044 0.0519
(1.53) (1.59) (1.98)+ (1.98)* (0.55) (0.64)
child birth 0.095 0.084 0.097 0.095 0.092 0.081
(1.3) (1.16) (1.24) (1.24) (1.13) (1)
Constant 0.089 0.096 0.104 0.105 0.137 0.142
(1.11) (1.2) (1.34) (1.36) (1.92)+ (2.01)*
Observations 1519 1519 1195 1195 1190 1190
R-squared 0.02 0.03 0.03
t-statistics in parentheses, standard errors clustered at change in county level
t-statistics in parentheses, standard errors clustered at change in county level
+ significant at 10%; * significant at 5%; ** significant at 1%
Note: additional control variables for all regressions include county of origin, and change in
county of residence in between 99 and 09
Work migrants who moved six years or more prior to 2009 also experience a
significant increase in both absolute and relative income
23
as compared to non-migrants
23
The similarity in the differential (with respect to non-migrants) absolute and relative income
changes is due to the move patterns: for the average migrant the incomes of the municipalities of origin and
of destination are almost the same (160 vs 166 thousand kronas in 2009). This implies that the reference
incomes for migrants and non-migrants are very close in magnitude. If the reference incomes were exactly
the same at both time 0 and time 1, then the difference between migrants and non-migrants in absolute and
60
(Table 3-4, Columns 6 and 12), though this does not hold for the more recent work
movers (Table 3-4, Columns 4 and 10). For less recent movers, the increase in absolute
income associated with work-related migration is stronger than the association between
income and education completion. The strong relation between work migration and
income present six to ten years after the move suggests that the occupational status
improvement experienced by work migrants may positively influence their future career
and earnings paths. Despite these significant income changes, however, the economic
satisfaction of less recent work migrants does not increase above that of non-migrants
(Table 3-4, Column 18). The differential increase in absolute income may not be
accompanied by a relative change in economic satisfaction because of adaptation. The
more unexpected lack of similarities between the relative income and economic
satisfaction changes is likely the result of limitations of the reference group. Due to data
restrictions, the reference group here is composed of all those living in the respondent’s
municipality of residence with no consideration for age, gender, or other characteristics,
limiting the accuracy of the findings regarding relative income.
Table 3-4: OLS regressions: financial domain variables as dependent variables,
migration (all and by reason to move) as main explanatory variables
Changes in Work Income
Whole sample
(10 years)
More recent migs
(less than 6 years)
Less recent migs
(6 years or more)
all migrants 4.512 4.105 6.514
(0.7) (0.47) (0.75)
relative income changes would be the same. Numerically, where RY=relative income, and AY=absolute
income:
ΔRY
M
– ΔRY
NM
=[(AY
1
M
– c1)–(AY
0
M
– c0)]–[(AY
1
NM
– c1)–(AY
0
NM
– c0)]= ΔAY
M
– ΔAY
NM
61
Table 3-4 (Continued)
work migrant 33.198 32.902 38.018
(3.39)** (2)+ (3.37)**
non-work migrant -10.683 -7.745 -15.422
(1.39) (0.84) (1.19)
cohort 1972 -20.601 -19.319 -25.663 -25.044 -15.059 -14.256
(2.46)* (2.28)* (2.99)** (2.91)** (1.5) (1.4)
cohort 1968 -5.776 -5.325 -7.655 -7.693 -1.821 -1.466
(0.87) (0.81) (0.89) (0.9) (0.28) (0.22)
married_fd 12.421 11.72 4.287 4.941 12.404 10.576
(1.39) (1.35) (0.48) (0.56) (1.34) (1.2)
div/wid_fd 27.685 25.767 19.194 21.028 24.457 19.386
(1.33) (1.28) (0.76) (0.84) (1.2) (0.96)
educ completion 66.066 64.711 69.038 68.598 63.55 61.685
(7.07)** (6.92)** (7.56)** (7.45)** (5.95)** (5.83)**
child birth -19.799 -16.667 -11.681 -10.02 -20.713 -17.944
(3.50)** (2.88)** (1.5) (1.26) (3.47)** (2.97)**
Constant 117.331 114.987 118.284 117.084 113.158 111.566
(12.55)** (13.41)** (11.97)** (12.23)** (16.26)** (17.66)**
Observations 1570 1570 1237 1237 1234 1234
R-squared 0.12 0.11 0.1
Changes in Relative Income
Whole sample
(10 years)
More recent migs
(less than 6 years)
Less recent migs
(6 years or more)
all migrants 2.07 1.336 5.258
(0.31) (0.15) (0.61)
work migrant 31.705 31.885 36.528
(3.2)** (1.92)+ (3.02)**
non-work migrant -13.628 -11.233 -16.521
(1.65)+ (1.17) (1.3)
cohort 1972 -21.756 -20.432 -26.328 -25.67 -15.602 -14.804
(2.64)** (2.45)* (2.97)** (2.89)** (1.59) (1.5)
cohort 1968 -8.551 -8.083 -8.635 -8.675 -3.676 -3.322
(1.19) (1.14) (0.99) (1.01) (0.54) (0.47)
married_fd 10.474 9.749 2.377 3.071 11.5 9.686
(1.26) (1.22) (0.3) (0.39) (1.25) (1.1)
div/wid_fd 25.75 23.763 18.518 20.462 22.665 17.631
(1.26) (1.2) (0.74) (0.82) (1.12) (0.87)
educ completion 65.546 64.148 68.245 67.778 63.495 61.646
(6.65)** (6.49)** (7.37)** (7.25)** (5.91)** (5.76)**
child birth -20.578 -17.341 -11.599 -9.837 -21.852 -19.102
(3.57)** (2.94)** (1.47) (1.21) (3.79)** (3.27)**
Constant 78.638 76.212 78.939 77.667 74.137 72.555
(8.45)** (8.96)** (8.20)** (8.34)** (10.91)** (11.73)**
Observations 1570 1570 1237 1237 1234 1234
R-squared 0.1 0.09 0.09
62
Table 3-4 (Continued)
Changes in Economic Satisfaction
Whole sample
(10 years)
More recent migs
(less than 6 years)
Less recent migs
(6 years or more)
all migrants 0.021 -0.066 0.112
(0.41) (0.97) (1.52)
work migrant 0.039 -0.08 0.143
(0.45) (0.48) (1.37)
non-work migrant 0.012 -0.061 0.091
(0.18) (0.72) (0.92)
cohort 1972 0.037 0.038 0.053 0.053 -0.001 -0.001
(0.47) (0.48) (0.56) (0.55) (0.02) (0.01)
cohort 1968 -0.075 -0.075 -0.034 -0.034 -0.046 -0.046
(0.98) (0.98) (0.46) (0.46) (0.61) (0.61)
married_fd 0.061 0.061 0.074 0.074 0.015 0.014
(1.04) (1.04) (1.1) (1.1) (0.2) (0.18)
div/wid_fd -0.147 -0.149 -0.112 -0.113 -0.311 -0.316
(1.24) (1.26) (0.85) (0.87) (2.26)* (2.33)
educ completion 0.497 0.496 0.541 0.542 0.514 0.512
(5.72)** (5.66)** (6.75)** (6.73)** (4.78)** (4.66)**
child birth -0.039 -0.037 -0.025 -0.026 -0.073 -0.07
(0.76) (0.71) (0.41) (0.42) (1.1) (1.07)
Constant 0.222 0.221 0.218 0.218 0.26 0.259
(3.19)** (3.19)** (3.52)** (3.5)** (3.21)** (3.28)**
Observations 1552 1552 1223 1223 1220 1220
R-squared 0.08 0.07 0.08
t-stats in parentheses, standard errors clustered at change in county level
+ significant at 10%; * significant at 5%; ** significant at 1%
Note: additional control variables for all regressions include county of origin,
and change in county of residence in between 99 and 09
Finally, while no significant change is observed in the housing domain for the
more recent movers, work migrants who moved six to ten years prior to 2009 are
significantly more satisfied with their housing than non-migrants (Table 3-5, Columns 4
and 6). This finding could stem from the long term spill-over effects from the
improvement in occupational status into the financial domain. In the long term, work
migrants who are set on high achieving career and earning paths may be more likely than
63
non-migrants to make improvements in their residential conditions thereby materializing
their higher status.
Table 3-5: OLS regressions: housing domain as dependent variable, migration (all
and by reason to move) as main explanatory variables
Changes in Satisfaction with Housing
Whole sample
(10 years)
More recent migrants
(less than 6 years)
Less recent migrants
(6 years or more)
OLS MI ICE OLS MI ICE OLS MI ICE
1 2 3 4 5 6
all migrants 0.206 0.095 0.283
(3.30)*** (1.14) (3.38)***
work migrant 0.203 -0.083 0.356
(2.14)** (0.53) (2.64)***
non-work migrant 0.207 0.165 0.227
(2.99)*** (1.86)* (2.51)**
married_fd -0.089 -0.089 -0.008 -0.014 -0.186 -0.190
(1.01) (1.02) (0.08) (0.14) (1.87)* (1.89)*
div/wid_fd -0.230 -0.230 -0.111 -0.124 -0.349 -0.361
(1.49) (1.49) (0.57) (0.64) -1.65 (1.66)*
educ completion -0.041 -0.041 -0.055 -0.058 -0.020 -0.021
(0.62) (0.62) (0.6) (0.64) -0.24 (0.27)
child birth 0.138 0.138 0.093 0.081 0.194 0.203
(1.39) (1.38) (0.79) (0.67) (1.79)* (1.91)*
final educ level 0.038 0.038 0.061 0.067 0.038 0.035
(1.36) (1.38) (1.66) (1.79)* -1.48 (1.43)
Constant 0.121 0.120 -0.102 -0.119 0.214 0.221
(1.03) (1.02) (0.63) (0.74) (1.72)* (1.8)*
Observations 1535 1535 1209 1209 1205 1205
R-squared 0.03 0.03 0.04
t-statistics in parentheses, standard errors clustered at change in county level
* significant at 10%; ** significant at 5%; *** significant at 1%
Note: additional control variables for all regressions include cohort of birth and county of origin
The results for work migrants suggest that changes in occupational status are an
important factor underlying their increase in well-being. Given the low life satisfaction
level of work movers prior to migration (Table 3-1), however, concerns could be raised
regarding the need of a status improvement for the increase in life satisfaction to occur.
An alternative approach may suggest that work migrants are “catching up” to the
64
satisfaction level of non-migrants, implying that their life satisfaction would experience a
relative increase regardless of the change in status. Two facts suggest that this alternative
is incorrect. First, over 68% of work migrants experience a strict increase in occupational
status, a proportion that is significantly higher than that of non-migrants and non-work
migrants (40% and 52% respectively) (Table 3-6). Additionally, work migrants whose
status improves have a higher increase in life satisfaction than those whose status remains
the same or deteriorates. In fact, work migrants whose status deteriorates experience a
slight decrease in absolute life satisfaction, and no change in life satisfaction relative to
non-migrants. In contrast, work migrants whose status improves experience a significant
increase in life satisfaction above that of non-migrants (Table 3-6).
Table 3-6: Change in life satisfaction by migrant type and occupational trajectory
Work migrants
Life Satisfaction
N % total 1999 2009 99-09
Diff. from
non-migs
Occup status improved
(change 99-09>0) 147 68.4% 3.75 4.03 0.286 0.277***
No change in occup stat
(change 99-09=0) 54 25.1% 3.81 3.98 0.167 0.264*
Occup stat deteriorated
(change 99-09=0) 14 6.5% 4.00 3.86 -0.143 0.020
Total 215 100.0% 3.78 4.01 0.228 0.290
Non-work migrants
Life Satisfaction
N % total 1999 2009 99-09
Diff. from
non-migs
Occup status improved
(change 99-09>0) 169 52.3% 3.81 3.98 0.172 0.163*
No change in occup stat
(change 99-09=0) 128 39.6% 3.88 4.02 0.141 0.238**
Occup stat deteriorated
(change 99-09=0) 26 8.0% 4.04 3.92 -0.115 0.048
Total 323 100.0% 3.86 3.99 0.136 0.198
65
Table 3-6 (Continued)
Non-migrants
Life Satisfaction
N % total 1999 2009 99-09
Diff. in
change wrt
non-migs
Occup status improved
(change 99-09>0) 332 39.7% 3.94 3.95 0.009
reference
category
No change in occup stat
(change 99-09=0) 412 49.3% 4.08 3.99 -0.097
Occup stat deteriorated
(change 99-09=0) 92 11.0% 3.84 3.67 -0.163
Total 836 100.0% 4.00 3.94 -0.062
***significant at 1%; **significant at 5%; *significant at 10%
The previous findings highlight the importance of combining migration with
occupational improvements. Only work migrants whose occupational status improves or
remains equal experience a relative increase in life satisfaction; at the same time, the
increase in life satisfaction of those who experience an occupational improvement is
higher when it is accompanied by migration. Notice, that this result also provides support
for the finding that work-related migration is accompanied by an increase in life
satisfaction relative to what it would have been had the migrants not moved. As discussed
in the methods section, one of the possible differences between migrants and the
comparison group (non-migrants) could lay in the higher motivation levels of the
migrants. Since people with similar occupational trajectories are likely to have similar
levels of motivation, the comparison within respondents with comparable changes in
occupational status should partially account for different motivation levels. The results
therefore suggest that even within a group of highly-motivated individuals, that is people
whose occupational status improves, life satisfaction increases most for those who move.
66
Regression results that control for the correlates of migration further confirm this finding
(Table B-8).
Non-work migrants represent a different case: for them, the short to medium term
increase in life satisfaction associated with migration appears to be related mostly to
positive changes in the housing domain. Housing satisfaction of non-work migrants
displays improvements above those of non-migrants for both, those who moved less than,
and more than six years prior to 2009 (Table 3-5, Columns 4 and 6). However, the job
and financial domains do not appear to be significantly related to non-work migration. In
the job domain, neither occupational status nor satisfaction with occupation significantly
change as compared to non-migrants following the non-work related move (Table 3-3,
Columns 2-6, and 8-12). Regarding financial aspects, income levels – absolute and
relative – decline slightly, though generally not significantly, for non-work migrants as
compared to non-migrants in both the shorter and longer time periods considered (Table
3-4, Columns 4 and 10, and 6 and 12). At the same time no significant economic
satisfaction changes are experienced in either time period (Table 3-4, Columns 16 and
18).
The lack of an association between migration and the job and financial domains
for non-work migrants suggests that, unlike work migrants, those who move for reasons
other than work are not set on high-achieving career or income paths. However, they do
experience an increase in housing satisfaction soon after migration that persists over time
and is likely to be reflective of the motivation behind the non-work migrants’ move. This
housing improvement, however, is not accompanied by a long-lasting increase in life
67
satisfaction, which could be due to either long term adaptation to material circumstances
such as housing, or to the financial burdens that accompany dwelling improvements
suggested by previous authors (Nakazato, Schimmack and Oishi 2011).
3.6. Conclusions
Previous studies have found that changes in objective well-being following migration
may depend on characteristics of the migrants such as reason to move or age. In life
satisfaction analyses, however, little consideration has been given to the reason for
moving or other distinctive traits of the migrant group. The present analysis uses a
longitudinal approach to assess the change in life satisfaction that accompanies migration
of young adults, dividing the sample into those moving for work, and those moving for
other (non-work) reasons. Findings suggest that the change in life satisfaction following
internal migration is generally positive, but its persistence depends on the reason to
move. While both work and non-work migrants experience a significant improvement in
life satisfaction following moves within the past six years, only work migrants display a
significant increase in life satisfaction six to ten years after the move.
The difference between work and non-work migrants in the results for the long
term association between life satisfaction and migration may be explained by the life
domains that change following their moves. Those who migrate for work reasons
experience an improvement in occupational status which sets them on a relatively high-
achieving career path. This higher occupational status and its long term material
68
spillovers are accompanied by a persistent increase in life satisfaction. While an
improvement (or at least no change) in occupational status is necessary for the increase in
work-migrants’ life satisfaction to occur, it cannot fully account for this increase. Non-
work migrants experience an increase in housing satisfaction that is accompanied by life
satisfaction improvements in the shorter, but not in the longer, term. The lack of a long-
lasting relation between changes in housing and life satisfaction may be due to the high
costs associated with better housing suggested by previous studies or to adaptation to
material domains. The finding that the relationship between migration and long term
change in life satisfaction may depend on the reason to move could explain the mixed
results of earlier analyses which have typically combined all migrants regardless of
reason for moving.
69
CHAPTER 4. Explaining Well-Being Over the Life Cycle: A Look at Life
Transitions During Young Adulthood
4.1. Introduction
Young adulthood is a time of change. Leaving the parental household, finishing
education, getting a job, forming a relationship, becoming a parent – these are important
life transitions that most young adults go through between the late teens and early thirties.
With all these life shifting events occurring in just over a ten-year period, it shouldn’t be
surprising that well-being will also go through important changes during young adult
years. This paper’s main objective is to analyze the association between the major life
transitions occurring during young adulthood and the overall life satisfaction cycle
observed in those years. The transitions analyzed are the school-to-work transition (the
end of the formal education), changes in partnership status (the formation and dissolution
of relationships), and the parenting transition. The analysis is based on a panel of young
adults from Sweden interviewed three times between the years 1999 and 2009.
In studying the association between life transitions and life satisfaction two main
questions are addressed. First, what is the path followed by life satisfaction during the
young adult years? Second, what is the standard pattern of life transitions undergone in
those years, and to what extent can these transitions account for the overall life
satisfaction changes? To answer these questions, changes in overall satisfaction with life
are analyzed for various age intervals to define the life satisfaction cycle followed
70
between ages 22 and 40. Subsequently, an analysis of the timing and sequence of the
school-to-work transition, partnership formation, parenting, and partnership dissolution,
is carried out, and the relationship between the observed transition pattern and life
satisfaction is assessed. The last part of the study provides an introductory exploration of
the possible life domains (such as the financial, job, or family domains) mediating the
association between the transitions and life satisfaction by analyzing the association
between each transition and changes in these domains.
Previous work has considered life satisfaction over the life cycle and the
relationship between individual life transitions and changes in well-being. But a link
between the satisfaction cycle and the transitions followed over the young adult years has
rarely been drawn. The present study contributes to the current knowledge on well-being
by illustrating the degree to which the life satisfaction cycle may be explained by the
pattern of transitions followed during young adulthood. The analysis also provides
important information on the association between life satisfaction and the school-to-work
transition, partnership formation, the birth of the first child, and partnership dissolution.
While previous studies have considered the relationship between such individual
transitions and satisfaction, given their typically close timing, the effect attributed to one
of these transitions considered alone may be confused by the effects of the others. To
avoid this bias, the present analysis considers the school-to-work transition, partnership
formation, birth of a child, and partnership dissolution jointly. This joint analysis aims to
capture the association with life satisfaction of each of these transitions controlling for
71
the effects of the others, providing more accurate information on how transitions during
young adulthood relate to overall life satisfaction.
4.2. Literature review
The analysis of well-being and its association with young adult transitions is at the
intersection of literature in demography and economics. In demography, studies
discussing the transition into adulthood are generally situated in the context of life course
analysis. This literature provides a detailed description of events characterizing the young
adult years, as well as their association with cultural and social surroundings (Elder 1998;
Elder, Johnson and Crosnoe 2003; Shanahan 2000; Vogel, 2002). Analysis of the
standardization, and later individualization of the life course suggests that the transition
into adulthood, while following common patterns, has become more variable in recent
years with frequent deviations from the standard sequence (Shanahan 2000).
Given that the interest of the present study is mostly on the life satisfaction cycle
during young adulthood, a comprehensive analysis of the timing and sequence of the
transitions undergone by young adults (such as the one undertaken by demographic
research) is beyond the scope of the present analysis. Still, the methods used here borrow
on the demographer’s findings in two important ways. First, research from demography is
used to identify the main transitions undergone by young individuals in the process of
becoming an adult. These transitions typically include five life events – leaving the
parental household, completion of education, labor force entry, partnership formation,
72
and the birth of the first child (Marini 1984; Hogan and Aston 1986; Billari 2001).
Second, in the spirit of the life course literature, a holistic approach is taken in analyzing
the transitions occurring during young adulthood, considering their timing, interactions,
and the resulting impact of the common transition pattern on life satisfaction. While
recognizing that important variability may exist in transitions occurring during young
adulthood, this study focuses on the standard patterns of these transitions, leaving further
exploration of individual deviations from these patterns and their effects on well-being
for future analysis.
In economics, recent interest in subjective well-being as a measure of human
progress (Stiglitz, Sen and Fitoussi 2009), has been accompanied by an increasing
amount of research analyzing the relationship between various socio-economic events
and life satisfaction. Studies in this area that are most closely related to the present
analysis are those tracking life satisfaction over the life cycle (Baird, Lucas and
Donnellan 2010; Mroczek and Spiro 2005; Easterlin 2006), and those analyzing the
effects of critical life transitions – such as partnership formation or parenting – on life
satisfaction (Lucas et al. 2003; Myrskyla and Margolis 2012; Zimmermann and Easterlin
2006). Regarding the former line of research, an important distinction must be made
between studies that take a ceteris-paribus approach (holding factors other than age –
such as health – constant), and those that describe a general path of life satisfaction over
life cycle without controlling for the influence of other variables on well-being. The
findings of these two lines of research differ considerably, with the former finding a U-
shaped association between life satisfaction and age (Blanchflower and Oswald 2008;
73
Clark and Oswald 1994), and the latter generally finding the opposite, inverse U-shape
association (Easterlin 2006; Mroczek and Spiro 2005). These differences are likely due to
the potentially negative effects on well-being of the deterioration of health and economic
conditions at older ages. The ceteris-paribus approach holds these negative effects
constant, obtaining a positive association of well-being with age. The studies looking at
the general pattern of life satisfaction over the life cycle do not control for the negative
changes that accompany ageing, and capture the actual deterioration in life satisfaction
experienced by the older population.
The analysis presented here follows the approach of the latter line of studies,
considering the evolution of satisfaction over the young adult years without additional
control variables. Presently, not much economic research exists in this area. For the
United States, Mroczek and Spiro
24
(2005) and Easterlin (2006) observe a mostly flat
path of life satisfaction over the life cycle, displaying a slight inverse-U shape with a
decline at older ages. Specifically, in the United States happiness seems to increase
slightly during the midlife reaching a maximum around the age of 50, and declines
thereafter. Using a similar approach with data from Germany and Great Britain, Baird et
al. (2010) find that the life satisfaction trajectories over the life cycle are somewhat
different for these two countries. In Germany, life satisfaction is generally flat until the
age of 74, and declines strongly after this point. In Great Britain, life satisfaction declines
slightly from young adulthood until the mid-40s, after which it increases until the age of
70, and again declines sharply after this point. Combining the findings from the three
24
The sample used by Mroczek and Spiro is limited to veteran men after the age of 40.
74
studies described, one gets a picture of a generally flat trajectory of life satisfaction
throughout a long part of the adult life cycle, with trends in the mid-life that may depend
on country-specific circumstances, and a sharp decline at older ages, especially after the
age of 70. The specific trends of life satisfaction during the young adult years, however,
are not addressed in detail by the existing work.
In the area of economics analyzing the young adult transitions and life
satisfaction, a considerable amount of research has been carried out studying the well-
being effects of events such as marriage, divorce, and parenting. The school-to-work
transition has been studied less extensively in the subjective well-being literature (though
its economic outcomes have been addressed by previous work). Most of the studies
dealing with these transitions focus on a single life event considering, for example,
marriage but not parenting, or vice-versa. Their results are therefore relevant in creating
expectations as to what relationship may exist between each transition and life
satisfaction, but not in determining the life satisfaction cycle over the young adulthood,
since to do that a joint analysis of the transitions is necessary.
There exists a general consensus in the literature on subjective well-being that
marriage is associated with an increase in life satisfaction (Clark et al. 2008; Lucas et al.
2003, Zimmermann and Easterlin 2006). Though some disagreement exists on whether
this positive association is, or not, permanent, the general picture is that life satisfaction
for married couples remains above the baseline as measured prior to both marriage and
cohabitation (Zimmermann and Easterlin 2006). This general picture is reinforced by
further findings that cohabitation, as well as marriage, are associated with an increase in
75
life satisfaction, and that in the long run the effects of both types of partnerships are very
similar (Musick and Bumpass 2012). Conversely, divorce and partnership dissolution
have been found to be accompanied by a decrease in life satisfaction (Clark et al. 2008;
Lucas 2005).
On the association between parenting and subjective well-being a very extensive
literature has been developed using cross-sectional analyses. Its findings have been
mixed, though generally paint a bleak picture of the effects of parenthood with results
showing a predominantly negative association between having children and various well-
being measures, including life satisfaction (Aassve, Goisis and Sironi 2012; Hansen,
Slagvold and Moum 2009, Hansen 2012; McLanahan and Adams 1987). Recently,
longitudinal studies have challenged these findings arguing that the negative association
between having children and life satisfaction in the cross-sectional analyses is due to a
self-selection bias. These longitudinal analyses find that an increase in life satisfaction
takes place right before the birth of the first child (Clark et al. 2008). This increase, while
dissipating over time, has been found to persist for at least two years following the birth
of the child (Baranowska and Matysiak 2011; Myrskyla and Margolis 2012).
As previously mentioned, the life satisfaction changes during the school-to-work
transition have not been studied extensively. Perhaps most relevant to this topic is the
analysis carried out by Creed, Muller and Patton (2003), who study the changes in well-
being for young adults in Australia during the transition from high-school to both, work,
and post-secondary education. Their findings indicate that life satisfaction declines for
those who enter post-secondary education as well as for those who enter the labor market
76
but are unable to obtain full-time employment. For those who enter the labor market and
become fully employed, life satisfaction remains constant. A related study of the
Australian youth with post-secondary education (Dockery 2005) also finds that
employment status affects life satisfaction, with the effects of unemployment being
negative and of job quality being positive.
To summarize, while the trends in well-being that accompany marriage (or
partnership formation) and divorce (or partnership dissolution) seem clear – positive in
the first case, and negative in the second –life satisfaction changes after becoming a
parent and the school-to-work transition appear more complex. In the case of parenting, a
short-term increase in life satisfaction is observed after the birth of the first child, but this
increase may not persist over time. This finding suggest that in the analysis of well-being
changes parents should be divided into those for whom the parenting transition took place
recently, and those for whom it took place several years earlier. As to the life satisfaction
trends associated with the school-to-work transition, these may depend on the type of
occupation obtained after the transition, which should also be considered in the analysis
of well-being. What the above literature does not answer, however, is what overall life
satisfaction pattern would emerge for young adults given the variety of transitions
typically undergone in this period of life. This question is addressed in the present
analysis.
77
4.3. Data description
The main source of data used is the Young Adult Panel Study (YAPS) carried out by
demographer Eva Bernhardt from Stockholm University. The YAPS is a longitudinal
survey of three cohorts of Swedish young adults (born in 1968, 1972, and 1976), that
were interviewed at three points in time corresponding to the years 1999, 2003, and 2009.
Survey responses were linked with the Swedish Register record by researchers in charge
of data collection to complement the socio-demographic information provided by the
respondents. This final dataset includes a comprehensive set of variables related to a
person’s family life, and various demographic and economic characteristics. For the
purpose of the present analysis the sample from YAPS is restricted to those answering the
main questions of interest in all three survey years.
The main dependent variable of the study, life satisfaction, is measured as the
answer to the question: “Are you satisfied or dissatisfied with life in general right now?”.
Response categories are given on a scale from 1 to 5, with 1 meaning very dissatisfied
and 5 very satisfied. The additional dependent variables are satisfaction with different life
domains. Domains for which specific satisfaction questions were asked include the
financial, job, and housing domains. For each domain, the question asked measures how
satisfied a person is with their economic situation, what they are currently doing, and
their housing, respectively with responses ranging from 1 (very dissatisfied) to 5 (very
satisfied). In addition, questions on satisfaction with the relationship with partner,
mother, and father are used to approximate changes in family satisfaction.
78
The domain measures used are subject to several limitations. First, satisfaction
with what the person is currently doing may represent an imperfect measure of the job
domain as it measures satisfaction with any activity that the person is doing, which
should most often, but not always, be interpreted as occupation. At the same time the
analysis is unable to provide an accurate assessment of changes in the family domain as a
specific question on family satisfaction was not asked. The questions used to approximate
this domain, satisfaction with relationship with partner, mother, and father, are
insufficient as they do not capture one of the main changes in the family situation
experienced by young adults: satisfaction with children. Moreover, these questions are
subject to a serious problem of missing values, as a number of respondents (most likely
those single or whose parents had deceased) did not respond these questions. Given these
limitations, the domain analysis presented should be interpreted as introductory rather
than conclusive.
The main independent variables are those identifying people as going through the
school-to-work, partnership formation, parenting, and partnership dissolution transitions
in each of the periods under analysis (1999-2003 and 2003-2009). While leaving the
parental household may be an equally important transition, when the respondents are first
observed they are already 22 years of age and 93% of them are no longer living with their
parents. The small sample of young adults leaving the parental household observed
makes the analysis of the association of this transition with life satisfaction impossible.
Young adult are identified as going through the school-to-work transition if they
attain their highest level of education in between any two surveys. Those who interrupt
79
their education at any point, either due to spells of employment or inactivity, are
considered to go through the school-to-work transition only after they re-enter education
and graduate with their highest degree attained.
25
Partnership formation is defined as
entering a new marriage or cohabitation
26
during any of the two periods under analysis.
The parenting transition is considered to take place with the arrival of the first child
(either biological or adopted) into the respondent’s household. Finally, partnership
dissolution takes place if a person reports to be in a partnership (marriage or cohabitation)
during either 1999 or 2003, and to be single or divorced/widowed in the following survey
(2003 or 2009 respectively). For more information on these variables (including the exact
survey questions used in their construction), see Appendix C.1, Table C-1.
Other variables used in the analysis include work income, occupation, and the
child’s age. Income is provided at the individual level and adjusted for inflation. Given
that the data was collected in between March and May of each year, the satisfaction
levels reported during the survey are most likely to reflect past years’ income. Because of
this, the work income from the year previous to each survey is used. Occupational
categories are constructed by combining two survey questions: main activity, defining the
person’s employment status; and main occupation, defining the person’s production
sector. The final occupation categories used are: student, unemployed, inactive, goods
production, service production, assistant non-manual, intermediate non-manual,
25
This definition was used because of the high rates of young adults in Sweden that briefly
interrupt their education soon after high-school to engage in either work or leisure activities before re-
entering education at the post-secondary level (Cook and Furstenberg, 2002).
26
Given the similar positive effects of both marriage and cohabitation on life satisfaction
(Appendix C.2, Table C-2), the two partnership states were combined to increase the number of
observations.
80
farmer/self-employed, and professional/higher non-manual/executive. Child’s age was
calculated using the year of birth of each child as reported by the respondent. For more
information, see Appendix C.1.
As mentioned above, the analysis restricts the sample of YAPS respondents to
those answering all three surveys. This reduces the number of observations to
approximately 1,385 of the 2,820 young adults originally interviewed in 1999, some of
whom were dropped in the analysis due to missing data in one or more of the main
questions of interest (life satisfaction and the questions defining the transitions). Such
high attrition rates (of around 50%) are not uncommon in developed countries (Abraham,
Maitland and Bianchi 2006; Becketti et al. 1988). Attrition in the YAPS survey could
represent a problem to the present study if the non-responses were systematically related
to both the change in life satisfaction, and any of the specific life transitions under
analysis. That is, the results may be biased if a specific sub-group of the people going
through a life transition was both more likely to attrit and to experience a specific change
in life satisfaction (either an increase or a decrease).
While it is impossible to test whether or not attrition is associated with an increase
(or decrease) in life satisfaction for the people who leave the survey – as, by definition,
their life satisfaction levels are not observed after they leave – it is possible is to check
whether the life satisfaction change of people interviewed between 1999 and 2003 is
associated with their future attrition in the 2009 survey. Using information on life
satisfaction changes in 1999-2003, and on future attrition in 2009, a test of the
significance of attrition suggested by previous literature is used (Fitzgerald, Gottschalk
81
and Moffitt 1998). This test consists of regressing the main dependent variable (in this
case, life satisfaction change) on subsequent attrition. If attrition is in fact a problem, then
its coefficient in such a regression should be significant. Performing this test using the
YAPS data shows that attrition is not a significant determinant of life satisfaction
changes (Appendix C.2, Table C-2). Repeating the same analysis for the other dependent
variables (changes in the domain satisfaction variables), shows that attrition is not a
significant determinant of any of these. The results of this simple test are reassuring in
that the attrition bias should not represent a major problem for the present analysis.
A brief statistical description of the four age intervals constructed is provided in
Table 4-1. As could be expected, the younger age groups have a lower income, and are
less likely to hold high status occupations, such as being a professional, high level non-
manual worker, or an executive. At the same time they are more likely to be studying and
to be single. By the age 30/32 the percent of respondents still studying and not married or
cohabiting drops, and the percent of people parenting starts to increase (Table 4-1). As
will be further discussed in the results section, these changes roughly outline the
evolution undergone by young adults between ages 22 and 40: from mostly single, non-
parenting students at age 22, to predominantly married (or cohabiting) parents with
completed final education levels by age 40.
82
Table 4-1: Descriptive statistics, by age, beginning and end of each age interval
Age interval:
22-26
Age interval:
26-30/32
Age interval:
30-34/36
Age interval:
34-40
Age 22 Age 26 Age 26
Age
30/32 Age 30
Age
34/36 Age 34 Age 40
Income
(in ‘000 SEK) 69.27 140.78 134.79 198.21 177.29 222.52 211.28 270.67
Prof/higher
nm/exec 1.1% 12.7% 11.7% 20.5% 15.16% 20.29% 16.45% 25.90%
Active in
labor force* 42.2% 66.5% 66.0% 79.6% 74.11% 86.71% 80.16% 96.69%
Student 48.4% 21.8% 20.5% 6.5% 8.02% 4.48% 6.75% 1.04%
Single 61.6% 39.6% 40.1% 22.4% 25.12% 16.51% 20.78% 15.84%
Cohabiting/M
arried 38.2% 59.7% 59.3% 75.5% 72.52% 78.30% 74.81% 76.62%
Divorced/Wid
owed 0.2% 0.6% 0.6% 2.0% 2.36% 5.19% 4.42% 7.53%
Parenting
(with partner) 4.0% 17.4% 20.0% 50.5% 46.23% 65.80% 62.34% 66.75%
Parenting
(alone) 0.8% 1.3% 1.7% 3.9% 4.48% 8.14% 9.09% 13.51%
N children 0.07 0.24 0.31 0.87 0.83 1.37 1.28 1.49
Final educ =
postsec. 64.99% 60.32% 52.48% 48.83%
Male 41.93% 40.96% 41.86% 44.16%
Cohort 76 100.00% 50.75% 0.00% 0.00%
Cohort 72 0.00% 49.26% 54.60% 0.00%
Cohort 68 0.00% 0.00% 45.40% 100.00%
N obs. 477 940 848 385
*Includes all those actively employed, excluding students, and those unemployed, in participating
labor market programs (such as adult learning), on parent leave, and inactive
4.4. Methods
To identify the average life satisfaction path followed during young adulthood the
analysis focuses on changes in satisfaction over specific age intervals. The four intervals
considered cover ages 22 to 26, 26 to 30/32, 30 to 34/46, and 34 to 40. These intervals are
83
constructed pooling observations for different cohorts interviewed around the same age at
least twice. For example, respondents born in 1972 interviewed in 1999 and 2003 (at ages
26 and 30) were pooled with those born in 1976 interviewed in 2003 and 2009 (at ages 26
and 32) to construct the 26 to 30/32 age interval. Since a first difference analysis is used
throughout the study, the main criterion to select the cohorts pooled is that observations
for that cohort have to be available both at the beginning and end of the age interval
considered. (For more information on the age interval construction, see Appendix C.3.)
Life satisfaction of individual i at age a could be represented by the following:
(4.1) LS
ia
= α
a
+ δ
i
+ βX
ia
+ε
ia
Where, α
a
is the effect of age, δ
i
is an individual fixed effect, X
ia
is a vector of
covariates that are allowed to change over age, and ε
ia
is an error term. The individual
fixed effect in (4.1) includes all personal characteristics that are time-invariant, including
personality and cohort effects, among others. Applying a first difference to (1) provides:
(4.2) ΔLS
i
= Δα
a1-a0
+ βΔX
i
+Δε
i
where Δα
a1-a0
= (α
a1
- α
a0
)
. In specification (4.2) δ
i
, representing all time invariant traits,
is automatically subtracted from the equation. Previous studies have recognized the
importance of controlling for cohort effects – such as being born during a war or a
recession – when studying life satisfaction over the life cycle (Blanchflower and Oswald
84
2008; Easterlin 2006). By using first differences, the present analysis goes a step further
and eliminates not only the effects of the year of birth, but also those of any personal time
invariant traits – such as being an optimist – on life satisfaction.
In equation (4.2), Δα
a1-a0
captures the association between the age interval
starting at 0 and ending at 1 and overall life satisfaction. Notice that Δα
a1-a0
is age-
specific, which implies that different age intervals may have a different association with
life satisfaction. To capture the general path of life satisfaction during young adulthood, a
regression is run with individual life satisfaction change as the dependent variable, and
four age interval dummies as explanatory variables. The regression is run using an OLS
first-difference model without a constant
27
. This is methodologically equivalent to an
OLS first-difference regression that includes a constant but omits one of the age intervals
from the estimation
28
. Since no socio-economic control variables are used, the coefficient
on each age interval dummy represents the total life satisfaction change for the average
young adult over that interval.
27
Given the grouping into four age intervals, it may be econometrically appealing to cluster the
standard errors at the age interval level. This was not done in the main part of the analysis as it creates a
problem of estimation with few clusters (for a discussion see Cameron and Miller 2011). As a robustness
check, the main regressions of the analysis were re-run using standard errors clustered by age interval, and
applying the standard adjustment for few clusters implemented by Stata, which uses a T distribution
(instead of the normal) for inference. All the coefficients significant in the original results remained
significant after the clustering (results available upon request).
28
The model used was preferred for three reasons. First, from a theoretical perspective, a change
in life satisfaction holding age (and through it, time) perfectly constant, is implausible. Second, the model
used makes the interpretation of the results easier: the average change in life satisfaction undergone during
each age interval is captured by the coefficient of that specific interval. Finally, the econometric
controversy about using OLS without a constant revolves around the diagnostic measures, such as the R
squared, which are not the main interest of this study (Eisenhaur 2003). For robustness, all analyses were
repeated using the alternative specification with a constant, which was found to have little effect on the R
squared and (as expected) none on the coefficients (results available upon request).
85
The previous estimation is based on two assumptions. First, the change in life
satisfaction is assumed to depend only on characteristics related to a person’s age, but not
on external time trends. This is a sensible supposition as long as the socio-economic
conditions of the country under analysis remain stable. In the case of Sweden, while GDP
growth in the periods 1999 to 2003 and 2003 to 2009 was reasonably stable, the
unemployment rate did experience important fluctuations. To check for the importance of
time trends in influencing the main findings, a robustness test is carried out where a
control variable for the 2003 to 2009 time period is included into the regressions. Doing
so does not affect the results (Appendix C.2, Table C-4). The second assumption is that
all cohort effects are fixed and therefore disappear in the first difference equation. To
assure that the results do not depend on time-variant cohort effects, life satisfaction
change for each age interval is analyzed separately by cohort before running the pooled
regression in the results section. Additionally, pooled regressions are re-run using control
dummies for birth cohort (Appendix C.2, Table C-4). The results of both tests support the
assumption of fixed cohort effects.
Following the description of the life satisfaction path, the main transition pattern
for young adults ages 22 to 40 is identified. Four transitions that characterize young
adulthood are considered: partnership formation, school-to-work transition, parenting,
and partnership dissolution. For the parenting transition, the long time span (four to six
years) between surveys implies a considerable variance in the age of the child born over a
given interval at the time of the second survey. Because a child’s age may influence the
parenting experience, the new parents are sub-divided into those whose child is less than
86
two years old and those whose child is two years or older at the time of the survey.
Partnership dissolution may also represent a different process depending on whether a
child is involved or not, and therefore respondents going through this transition are sub-
divided into those with and without a child. Finally, young adults not going through any
transition at a given age interval are divided into those who have already gone through
the school-to-work and parenting transitions, and those who have not
29
. This is done to
capture the lasting effects of some transitions (such as parenting) on the life satisfaction
of young adults. To identify the common transition pattern, the percent of respondents
going through each transition is calculated for every age interval.
To estimate the degree to which transitions typical to young adults account for
their life satisfaction changes, regression (4.2) is run including the transitions undergone
by each individual as explanatory variables. Doing so leads to:
(4.3) ΔLS
i
= Δα
a1-a0
+ β
T
’Ti +Δε
i
where T
i
is a matrix of bivariate variables with value 1 if person i has gone through
transition T in the age interval from to and 0 otherwise. Matrix T
i
includes
partnership formation, the school-to-work transition, parenting (with child below age 2 at
time of interview), parenting (with child 2 years or older at time of interview),
partnership dissolution (with a child), and partnership dissolution (without a child), as
29
Since the proportion of young adults going through parenting before partnership formation is
very small (Table 4-1), it is assumed that all those that have gone through parenting have also gone through
partnership formation.
87
well as a dummy variable for those not currently going through any transitions but who
have already completed their school-to-work and parenting transitions. In this
specification, the age interval coefficients capture the association between change in age
and life satisfaction for the omitted category (those not going through any transition and
who have not yet completed the school-to-work and parenting transitions). Since all
transitions are included simultaneously, the coefficients β
T
capture the pure association
between each transition and life satisfaction, controlling for the effects of all other
transitions. This allows to separately identify the associations with life satisfaction for
transitions that may occur jointly, such as partnership formation and parenting. Notice,
again, that no socio-economic control variables (other than the transitions themselves) are
included in the model.
To assess whether the common transition pattern experienced by young adults can
account for their overall life satisfaction path, the life satisfaction change for each person
is predicted using only its association with the transitions occurring in that person’s life
(that is, using coefficients β
T
to predict the change in life satisfaction at the individual
level). Using these estimates, the average life satisfaction change for every age interval is
obtained. This average represents the change in life satisfaction that could be expected to
take place for a given age interval based only on the transitions common to that age.
Using these predictions allows to construct the estimated life satisfaction path over ages
22 to 40 as projected exclusively by the common young adult transition pattern. If the
estimated path accurately approximates the actual path followed by life satisfaction, this
88
result could be taken to imply that young adult transitions are an important determinant of
overall life satisfaction during this part of the life cycle.
The final step of the analysis aims to identify the impact of the four young adult
transitions on several aspects of a person’s life (commonly referred to as life domains).
The domains considered include the financial, work, and housing domains, as measured
by the financial, occupation and housing satisfaction respectively. Additionally, changes
in a respondents’ satisfaction with their relationships with partner, mother, and father, are
also analyzed. The impact of the young adult transitions on each domain is approximated
by the association between a given transition and each of the five satisfaction variables
estimated using specification (4.3) for all age intervals pooled. Though this analysis
provides some information on the relationship between young adult transitions and life
domains, its extent is limited as several domains of interest (such as family satisfaction)
are not available in the YAPS survey. Additionally, the question on satisfaction with
partner might not have been answered if the person was not currently in a stable
relationship. A more precise evaluation of the impact of young adult transitions on
different life domains is left for future analysis.
4.5. Results
4.5.1 Life satisfaction path during ages 22 to 40
89
The life satisfaction path followed between ages 22 and 40 displays a slight inverse U-
shape, with overall satisfaction increasing until 30/32 and decreasing thereafter. The
average change in life satisfaction for each age interval is captured by the coefficient of
the age interval’s dummy in a regression with observations for all cohorts pooled and
with change in life satisfaction as the dependent variable (Table 4-2, Column 4). Life
satisfaction increases between age 22 and 30/32, with the increase being steepest in the
second part of the decade, and decreases steadily in the following ten years.
Table 4-2: OLS regressions: change in life satisfaction as dependent variable,
age intervals as explanatory variables
LS change
1976
cohort
1972
cohort
1968
cohort all cohorts
Age 22 to 26 0.038 0.038
(0.83) (0.82)
Age 26 to 30/32 0.046 0.060 0.053
(1.02) (1.31) (1.62)
Age 30 to 34/36 -0.017 -0.049 -0.032
(0.38) (0.93) (0.92)
Age 34 to 40 -0.34 -0.034
(0.64) (0.66)
Observations 954 926 770 2650
R-squared 0.002 0.002 0.002 0.002
Absolute value of t statistics in parentheses;
+ significant at 10%; * significant at 5%; ** significant at 1%
The initial increase in life satisfaction between ages 22 to 26 and 26 to 30/32, and
the consequent decrease between 30 to 34/36 and 34 to 40, hold for all cohorts for which
observations at those ages are available (Table 4-2, Columns 1-3). The actual path of life
satisfaction observed for each cohort during the decade between the first and last surveys
is shown in the left panel of Figure 4-1. Though some cohort level differences are clear,
with the 1972 cohort appearing on average more satisfied than either the 1976 or the
90
1968 cohorts, the trends followed during overlapping age intervals are similar for all
cohorts (Figure 4-1). Additionally, the difference in the interval coefficients between
cohorts with overlapping age intervals is small and not significant, suggesting that the
general trends in life satisfaction by age do not differ depending on the cohort of birth
(Table 4-2, Columns 1-3).
Figure 4-1: Life Satisfaction by age, by cohort and pooling cohorts by age interval
Using the coefficients from the pooled regression, and adjusting life satisfaction at
age 22 to 0 to avoid cohort level effects, the average path of life satisfaction for young
adults ages 22 to 40 is illustrated in the right panel of Figure 4-1. Each life satisfaction
point of this path can be interpreted as the difference in life satisfaction from age 22
1976 cohort
1972 cohort
1968 cohort
3.75 3.85 3.95 4.05 4.15 4.25
Life Satisfaction (actual values)
20 22 24 26 28 30 32 34 36 38 40
age
By cohort, actual LS values
(LS age 22=0)
-.25 -.15 -.05
.05 .15 .25
Life Satisfaction (adjusted to 0 at age 22)
20 22 24 26 28 30 32 34 36 38 40
age
Cohorts pooled, adjusted LS values
91
expected at a given age for an average person
30
. Adding the first two segments of the
path, the overall increase in life satisfaction for the upward trend from 22 to 30/32 is
approximately 0.09 (Table 4-2, Column 4). Though this may seem small given the
satisfaction scale (1 to 5), previous findings indicate that over 30 year spans covering
ages 18 to 51 and 40 to 70 respectively average life satisfaction changes by about 0.1
points on a scale of 1-3 in the first case, and 1 point on a scale of 1-11 in the second
(findings for the American population, Easterlin 2006, and Mroczek and Spiro 2005).
Given that the time span used here is a decade, one third of that analyzed for the
American population, the change in life satisfaction between ages 22 and 30/32, though
small, may be considered relevant. During the following decade covering ages 30 to 40
life satisfaction shifts directions decreasing steadily by about 0.066 overall points (Table
4-2, Column 4). Though small, this decrease is consistent for both the 1972 and the 1968
birth cohorts.
4.5.2 Common transition pattern and its association with life satisfaction
The main transition pattern between ages 22 and 40 is characterized by young adults
typically going through partnership formation and the school-to-work transition before
age 30/32, and then through parenting between ages 26 and 34/36 (Table 4-3). Though
partnership dissolution does not represent a common transition for the majority of young
30
Notice that, since life satisfaction is adjusted to 0 at age 22, the right panel of Figure 4-1 says
nothing about satisfaction levels, which should be set by personal fixed circumstances, such as birth cohort
or personality traits. Regardless of the level, this figure illustrates the life satisfaction path expected over
the ages 22 to 40 for the average young adult.
92
adults at any age before 40, after 30 the proportion of couples that dissolve their
partnership starts to steadily increase. Most young adults go through only one transition
(if any) at a time, with partnership formation and the school-to-work transition between
ages 22 and 26 being the only transitions that occur jointly for more than 10% of the
sample (Table 4-3). A visual representation of the main transitions undergone, by age
interval, is given in Figure 4-2. During the youngest age interval, 22 to 26, over 50% of
the young adults observed go through either partnership formation or the school-to-work
transition, or both. After age 26 and before 30/32 parenting represents the most common
transition, with 32% of the respondents having a first child born in this age interval.
Partnership formation and the school-to-work transition between ages 26 and 30/32 are
also observed for important proportions of the sample (22% and 25% respectively). After
the age of 30/32, parenting remains as the only transition that still occurs for over 15% of
all young adults, and by the final age interval, 34 through 40, most young adults are no
longer undergoing any transitions.
Table 4-3: Life transitions undergone, by age group
Age group:
22-26
Age group:
26-30/32
Age group:
30-34/36
Age group:
34-40
Transitions occurring
individually: N % N % N % N %
Partnership formation 61 12.8 88 9.4 56 6.6 22 5.7
School-work 100 21.0 113 12.0 65 7.7 30 7.8
Parenting (all) 31 6.5 183 19.5 128 15.1 19 4.9
Parenting (child 1 year less) 18 3.8 66 7.0 45 5.3 4 1.0
Parenting (child 2 yrs more) 13 2.7 117 12.4 83 9.8 15 3.9
Partnership dissolution (all) 23 4.8 34 3.6 46 5.4 27 7.0
Partnership diss (with child) 2 0.4 19 2.0 27 3.2 21 5.5
Partnership diss (no child) 21 4.4 15 1.6 19 2.2 6 1.6
Transitions occurring
jointly:
Partnership form+school-work 54 11.3 44 4.7 11 1.3 3 0.8
93
Table 4-3 (Continued)
Partnership form + parenting 19 4.0 56 6.0 35 4.1 12 3.1
School-work + parenting 11 2.3 43 4.6 12 1.4 1 0.3
Partnership diss + school-work 11 2.3 12 1.3 11 1.3 3 0.8
Partnership diss + parenting 0 0.0 6 0.6 9 1.1 1 0.3
Partner. form+school-
work+parenting 4 0.8 18 1.9 12 1.4 1 0.3
Partnership diss+school-
work+parenting 1 0.2 2 0.2 1 0.1 0 0.0
No transition (all) 162 34.0 341 36.3 462 54.5 266 69.1
No trans (trans incomplete) 151 31.7 217 23.1 151 17.8 46 11.9
No trans (comp majort trans) 11 2.3 124 13.2 311 36.7 220 57.1
Total age group 477 100.0 940 100.0 848 100.0 385 100.0
Figure 4-2: Main life transitions, by age interval, age 22 to 40
While at younger ages, partnership dissolution generally occurs for those without
children, by ages 34 to 40 almost 75% of those going through partnership dissolution do
so after having a child (Table 4-3). By age 30 a shift also occurs for those not going
through any transitions: while at younger ages this group is mostly composed by childless
people who are still studying, after age 30 the majority of this group has completed the
94
main young adult transitions, such as school-to-work and parenting. Finally, those who
become new parents are more likely to have children two years or older (rather than one
years or younger) at all age intervals, except for the youngest – 22 through 26 – when
parenting is still uncommon.
To summarize, the common transition pattern is that of young adults going
through partnership formation and school-to-work transitions before age 30/32, parenting
(mostly with children two years or older at time of interview) between 26 and 34/36, and
no more transitions, having completed the main ones, between ages 34 and 40.
Partnership dissolution after having children, though still uncommon, increases after the
age of 34. As a caveat, it is important to mention that though this pattern represents the
most common transitions as followed by the majority of young adults in the sample,
important deviations from it may exist. Although the analysis of these individual
variations in the patterns followed is beyond the scope of the present study, it represents
an area of interest for future exploration.
How does the common transition pattern relate to the life satisfaction path during
ages 22 to 40? To address this question, the association between each transition and life
satisfaction is assessed using regression analysis (Table 4-4). The transition coefficients
in these regressions may be interpreted as the change in life satisfaction for those going
through a given transition relative to those not going through any transitions and who
have not yet completed the major life transitions such as school-to-work and parenting.
Since all transitions are included simultaneously, the associations with life satisfaction for
transitions that may occur jointly are identified separately.
95
Table 4-4: OLS regressions: change in life satisfaction as dependent variable, main
life transitions and age intervals as explanatory variables
All age
intervals
pooled
Age
interval
22-26
Age
interval
26-30/32
Age
interval
30-34/36
Age
interval
34-40
Partnership formation 0.292 0.142 0.273 0.474 0.33
(5.58)** (1.37) (3.41)** (4.50)** (1.62)
School-to-work -0.023 -0.088 0.071 -0.066 -0.181
(0.45) (0.91) (0.92) (0.62) (0.89)
Parenting (child 1yr-) 0.245 0.282 0.268 0.169 0.296
(3.49)** (1.71)+ (2.73)** (1.27) (0.76)
Parenting (child 2yrs+) -0.129 -0.244 -0.09 -0.081 -0.582
(2.13)* (1.17) (1.05) (0.78) (2.52)*
Partnership diss (with child) -0.343 -0.59 -0.075 -0.292 -0.75
(3.25)** (1.01) (0.4) (1.82)+ (3.06)**
Partnership diss (no child) -0.099 -0.254 -0.12 0.029 0.005
(0.9) (1.35) (0.59) (0.14) (0.01)
No trans (major completed) -0.135 0.153 -0.068 -0.167 -0.232
(2.38)* (0.59) (0.65) (1.89)+ (1.51)
Age 22-26 -0.041 0.034
(0.75) (0.46)
Age 26-30/32 0.018 -0.029
(0.39) (0.5)
Age 30-34/36 0.002 -0.013
(0.04) (0.19)
Age 34-40 0.044 0.164
(0.68) (1.2)
Observations 2650 477 940 848 385
R-squared 0.03 0.03 0.03 0.05 0.06
As could be expected, partnership formation is accompanied by an increase in life
satisfaction that is significant for all age intervals pooled, as well as for most of the age
intervals during which partnership formation is common
31
. The school-to-work transition,
however, does not display a significant association with life satisfaction change for any of
the age intervals, nor for all pooled (Table 4-4). Previous literature has shown that the
31
The lack of a significant association of partnership formation with change in life satisfaction
during the first age interval could be representative of that partnerships formed earlier in life are perceived
as less important. Though further exploration may represent an area of interest, due to data limitations, such
an exploration is left for future studies.
96
effects of the school-to-work transition on well-being may depend on personal
circumstances. Given this, additional specifications were run in which those going
through the school-to-work transition were divided by type of occupation after education
completion, and by level of final education. Since no significant association with life
satisfaction change was found for any of the groups considered (Appendix C.2, Tables
C-5 and C-6), the original specification (using only one group for the school-to-work
transition) is used in the main analysis.
The association of parenting with life satisfaction clearly depends on the age of
the child during the time of the interview. For the new parents whose child is one year old
or younger parenting shows a clear positive and significant association with life
satisfaction change for all age intervals pooled, and for each interval separately (though
losing its significance in some cases, probably due to low number of observations) (Table
4-4). On the contrary, for the parents with children two years or older, this association is
negative, though only significant for the pooled regression and for the last age interval.
This finding is in accordance with previous literature, which has shown positive but
decreasing changes in well-being in the years following the birth of a first child. The
negative coefficients on parenting for those with children two years or older may imply
that the positive association between the birth of a first child and well-being may not only
be short lived, but in fact, may become reversed in the long run.
Partnership dissolution is always accompanied by a decrease in life satisfaction.
This decrease, however, is only significant for those going through partnership
dissolution with children for all age intervals pooled and for the older age intervals, when
97
this transition becomes more common (Table 4-4). The negative coefficient for
partnership dissolution without children is never significant, though this may be due to
the small number of people going through this transition before age 40. The stronger
decrease in life satisfaction for those going through partnership dissolution with, rather
than without children, confirms the study’s expectations and shows the importance of
considering personal circumstances during this transition. Finally, those who have
already completed their parenting and school-to-work transitions and are not going
through any more transitions at a given age interval usually experience a decrease in life
satisfaction. This negative association holds for all age intervals but the first (during
which this is group is very small), but is only significant for all intervals pooled and for
those ages 30 to 34/36 (Table 4-4). This finding could be indicative of mounting
pressures, possibly related to parenting, during the later stages of young adulthood.
Can the common transition pattern transitions account for the well-being path
during young adulthood? Recall that life satisfaction displays a slight inverse U-shape,
increasing in the age intervals 22 through 26, and 26 through 30/32, and decreasing in the
two consecutive intervals – 30 through 34/36 and 34 through 40. In the first part of this
cycle while life satisfaction is increasing, people are mostly going through partnership
formation and the school-to-work transition at ages 22 to 26, and through parenting in
addition to the previous two transitions at ages 26 to 30/32. Given the positive
relationship with life satisfaction of partnership formation and of parenting young
children, and the lack of significance of the school-to-work transition, these transitions
could potentially explain the increasing life satisfaction trend. After the age of 30, people
98
go through parenting with mostly older children, or through no transitions, having no
more transitions pending. A slight proportion of the sample also goes through partnership
dissolution after having a child. All of these transitions undergone after 30 have a
negative (though not always significant) association with life satisfaction and so could
possibly account for the slight downward trend in life satisfaction between ages 30 and
40.
To formalize this reasoning, a prediction of life satisfaction change is estimated
using the coefficients of the regression for all age intervals pooled (Table 4-4, Column 1).
The average of this prediction for each age interval represents the change in life
satisfaction that could be expected during that interval based only on the proportion of
people going through each of the transitions considered. Comparing these average
predictions to the actual life satisfaction changes taking place during each age interval, it
is found that the transitions can in fact predict the slight inverse U-shape of the overall
life satisfaction path (bottom two rows of Table 4-4). The predicted path displays slightly
bigger changes in life satisfaction, especially at the beginning and end of the period, but
the age at which the maximum is reached and life satisfaction starts decreasing is
predicted correctly by the transitions alone (Figure 4-3). Taking into account that this
prediction excludes all of the variables typically considered as associated with a person’s
life satisfaction trend – such as changes in income, health or job status – the capacity of
the transitions alone to predict the well-being path is striking. Young adult transitions are
clearly important factors contributing to the life satisfaction path followed during this part
of the life cycle.
99
Figure 4-3: Life Satisfaction by age, actual and predicted, adj. values (LS age 22=0)
4.5.3 A glimpse into the life domain changes between ages 22 and 40
What are the life domains mediating the association between young adult transitions and
life satisfaction? To answer this question, ideally a comprehensive analysis of changes in
the various life domains impacting well-being should be carried out. Due to data
restrictions, however, the present analysis is limited to three life domains, representing
satisfaction with the financial, housing, and job situation. The association between the
change in each of these three domains and young adult transitions is analyzed.
Additionally, measures of satisfaction with relationships with one’s partner, mother, and
predicted
actual
-.25 -.15 -.05
.05 .15 .25
Life Satisfaction (adjusted to 0 at age 22)
20 22 24 26 28 30 32 34 36 38 40
age
100
father are used as auxiliary variables to approximate satisfaction in the family domain.
Still, since satisfaction with children is not measured, this family domain analysis is
incomplete. Recall also that the job domain is measured with satisfaction with what the
person is currently doing, which may not always correspond to one’s job, and that the
measures on satisfaction with one’s relationships are imperfect due to high non-response
rates (probably by those who do not have a partner, or whose parents have deceased).
Given these limitations, the present analysis is very preliminary, and its findings should
be interpreted as suggestive, not conclusive.
The domains most affected by the young adult transitions are the financial, and
family domains. Before age 30, transitions such as partnership formation, school-to-work,
and parenting younger children are associated with positive changes in the financial,
housing, and family domains. After 30, transitions such as partnership dissolution and
parenting older children are accompanied by financial pressures, strains on relationships
with family members, and a general decrease in satisfaction with what the person is
doing.
Not surprisingly, partnership formation has a strong positive association with
satisfaction with partner. Interestingly, forming a partnership is also related positively
with financial satisfaction, and negatively with satisfaction with occupation. The positive
relationship for the financial domain could be due to the effects marriage and
cohabitation may have on increasing household income (Korenman and Neumark 1991;
Waite 1995). The negative association in the job domain is more of a puzzle, but could
101
possibly be explained if for some (especially women) forming a partnership requires an
occupational shift that is perceived as unpleasant.
32
As one might expect, the school-to-work transition has a strong positive
relationship with financial satisfaction (Table 4-5). However, its association with the job
domain is negative and also significant. This finding may be due to the way the question
was asked, which aims to reflect satisfaction with what the person is currently doing,
rather than with a job itself.
Table 4-5: OLS regressions: change in life domains as dependent variables, main life
transitions as explanatory variables - all age intervals pooled
Financial
sat.
Sat. with
occup.
Housing
sat.
Sat. with
partner
Sat. with
mother
Sat. with
father
Partnership 0.222 -0.166 0.148 0.38 0.103 0.113
formation (3.65)** (2.41)* (2.24)* (5.73)** (2.19)* (2.04)*
School-to-work 0.599 -0.195 0.028 0.083 -0.049 -0.01
(10.24)** (2.94)** (0.43) (1.36) (1.09) (0.18)
Parenting -0.069 0.093 0.15 -0.029 -0.137 -0.034
(child 1yr-) (0.85) (1.01) (1.70)+ (0.38) (2.20)* (0.46)
Parenting -0.171 0.027 -0.073 -0.181 -0.161 -0.141
(child 2yrs+) (2.43)* -0.34 (0.96) (2.64)** (2.93)** (2.17)*
Partner. diss -0.443 -0.297 -0.208 -0.157 0.059 0.202
(with child) (3.62)** (2.15)* (1.55) (1.28) (0.63) (1.76)+
Partners. diss -0.216 -0.085 -0.221 -0.403 0.065 -0.023
(without child) (1.69)+ (0.59) (1.6) (2.89)** (0.66) (0.2)
No trans (major 0.099 -0.204 -0.01 0.008 -0.045 -0.015
completed) (1.51) (2.75)** (0.15) (0.12) (0.87) (0.25)
Age 22-26 -0.108 0.127 0.012 -0.065 -0.108 -0.113
(1.74)+ (1.81)+ (0.17) (0.92) (2.25)* (2.02)*
Age 26-30/32 0.133 0.137 0.233 -0.09 -0.056 -0.127
(2.57)* (2.34)* (4.15)** (1.61) (1.42) (2.69)**
Age 30-34/36 0.111 0.236 0.115 -0.161 -0.091 -0.124
(1.99)* (3.74)** (1.89)+ (2.72)** (2.10)* (2.38)*
32
A separate analysis for men and women is left for future research. Such an analysis could
potentially reveal important gender differences in the association of young adult transitions with changes in
occupational and other domains. To assure that gender differences do not affect the main results of the
study, life satisfaction regressions were run separately for men and women revealing no major gender
differences (results available upon request).
102
Table 4-5 (Continued)
Age 34-40 0.113 0.284 0.12 -0.246 -0.059 -0.03
(1.52) (3.36)** (1.48) (3.24)** (1) (0.42)
Observations 2632 2579 2629 1933 2500 2301
R-squared 0.09 0.01 0.03 0.05 0.02 0.02
Absolute value of t statistics in parentheses;
+ significant at 10%; * significant at 5%; ** significant at 1%
Life domain changes following the transition into parenting depend on the age of
the child at the time of the survey. New parents with children one year or younger report
a significant increase in satisfaction with housing, and a decrease in satisfaction with their
mothers (Table 4-5). The decrease in satisfaction with mothers, however, is likely to be
counteracted by a strong increase in satisfaction with children following the recent birth
of a first child, leading to a likely increase in overall family satisfaction. New parents
with children two years or older represent a different case, experiencing a significant
decrease in financial satisfaction, as well as satisfaction with all family members (partner,
mother, and father) in general (Table 4-5). Combined, these findings suggest that after an
initial period of increased satisfaction of having a new child, the economic burdens and
strains on the relationship associated with having older children may become strong
enough to outweigh the pleasures of parenting
33
.
Partnership dissolution for those with a child is accompanied by strong negative
changes in the financial, and job domains, that are much weaker, or not significant, than
for those without a child (Table 4-5). Though the negative relationship between
partnership dissolution and partner satisfaction is only significant for those going through
33
The possibility of children having a negative impact on financial satisfaction has been
previously suggested by Zimmermann and Easterlin 2006.
103
this transition without a child, this should be interpreted cautiously as it is likely due to
the limited number of respondents reporting satisfaction with partner following
partnership dissolution (when divorced or single). Finally, those not going through any
transitions after having completed the school-to-work and parenting transitions
experience a decrease in satisfaction with what they are currently doing and insignificant
changes in other life domains (Table 4-5).
The results of the domain analysis suggests an explanation for the increase, and
subsequent decrease in life satisfaction between ages 22 and 40. The transitions most
common to the younger age intervals (partnership formation, the school-to-work
transition, and parenting younger children) are accompanied by positive changes in the
financial, housing, and family domains, which may explain the increase in life
satisfaction before age 30. Conversely, transitions more common after the age of 30
(partnership dissolution and parenting older children) are accompanied by financial and
family burdens, leading to a decrease in overall life satisfaction. Still, as previously
mentioned, due to its numerous limitations this analysis is very preliminary, and further
exploration of the life domain changes that accompany young adult transitions is needed
to obtain more conclusive results.
4.6. Conclusions
During young adulthood a number of important changes affecting well-being are
experienced. This study focused on the life satisfaction path followed between ages 22
104
and 40, and on its association with the main young adult transitions, such as partnership
formation, school-to-work transition, parenting, and partnership dissolution. For the
Swedish young adults analyzed, life satisfaction follows a slight inverse U-shape
increasing between ages 22 and 30/32, and decreasing after 30. At the same time, these
young adults experience various life transitions that take them from being predominantly
single students at age 22, to being in a marriage or cohabitation, working, and parenting
by the age of 40. The common transition pattern observed starts with partnership
formation and the school-to-work transition before the age of 30/32. Parenting begins to
be common between ages 26 and 30/32, and becomes the most common transition during
the age interval from 30 to 34/36. After the age of 34, the majority of young adults no
longer go through any of the transitions described, having completed all of them. A small
but growing proportion of the sample also begins to experience partnership dissolution.
This common transition pattern is found to account in large part for the life
satisfaction path. The early young adult transitions exert positive effects on life
satisfaction mostly through the strong and positive association between partnership
formation and life satisfaction. Later in life, parenting of older children and partnership
dissolution, both of which are accompanied by negative (though not always significant)
changes in life satisfaction, are partially responsible for the downward trend in well-
being. Predicting life satisfaction changes with only the transitions undergone at each age
interval produces an estimated life satisfaction path that recreates the inverse U-shape
actually observed between ages 22 and 40. The close resemblance between the life
satisfaction path as predicted by the transitions alone, and as actually observed, is
105
especially striking given that none of the variables typically associated with well-being –
such as income – are included in the prediction.
To explain the strong association between young adult transitions and life
satisfaction changes, a domain analysis is used. Though the domains available are
limited, the findings suggest a possible explanation for the increase, and subsequent
decrease in well-being. The initial upward trend in life satisfaction is accompanied by
positive changes in financial and partner satisfaction associated with the school-to-work
transition and partnership formation. The later downward trend in life satisfaction comes
with strains to the financial conditions and to the relationship with partner associated with
parenting older children and partnership dissolution. Future exploration of life domain
changes following young adult transitions could include assessment of gender
differences, and a detailed analysis of how family pressures interact with the job and
financial domains.
The present results imply that the life satisfaction path followed during young
adulthood is to a big extent a reflection of the transitions undergone. Helping young
adults handle the pressures created by these transitions may therefore have important
well-being effects. In particular, this analysis suggests that even in a country with a
strong welfare support system such as Sweden, young parents are exposed to
accumulating financial strains that lead to a decrease in overall life satisfaction. These
strains become especially taxing for single parents following partnership dissolution.
Policies aimed at aiding new parents with older children, such as extended day care
programs, could be used to alleviate these pressures leading to potential well-being
106
improvements. In designing these policies results from a further life domains analysis
may be used to identify the life aspects associated with each transition.
107
CHAPTER 5. Summary and Conclusions
The analysis carried out in this dissertation provides new knowledge on the association
between important social conditions and life satisfaction. In the past decades there has
been a growing interest among economists in the study of subjective well-being
measures, such as life satisfaction, and a shift towards the use of these measures for
policy design has been suggested. The three studies presented here analyze how living in
a rural or urban area, internal migration, and life transitions, are each associated with life
satisfaction. The results are based mostly on panel-survey data, minimizing the
personality and cultural biases that are often problematic in subjective well-being
analyses. This dissertation represents a response to the ongoing shift towards subjective
well-being measures, and provides information that could be used to design policies
aimed at improving the overall well-being of the population of a country and not just its
economic growth.
In chapter two, two questions are asked: how does life satisfaction vary by size-
of-place, and what can account for these variations. The focus is on fourteen Latin
American countries in the time period covering 2003 to 2006. Findings show that the
fourteen countries studied may be divided into two groups: those where people in rural
areas are happier than people in urban areas (“rural-happy countries”), and those where
they are less happy (“urban-happy countries”). Four possible explanations for the
different relative well-being patterns in rural-happy and urban-happy countries are
explored. The first two, economic development and public social spending, focus on
108
country level variables that could affect those living rural places differently from than
those in urban places. The third and fourth, social values and indigenous origins, consider
individual level traits that may differ between the rural and urban population. These four
explanations are examined singly as well as in multivariate analysis. Considered singly,
economic development and public social spending are both found to be important in
explaining the difference between rural-happy and urban-happy countries by positively
affecting rural, relative to urban, life satisfaction. A multivariate analysis, however,
shows that public social spending, is the main determinant of the rural-urban life
satisfaction variations through its positive relative effect on rural well-being. Social
values and indigenous origins explanations do not come up as significant determinants of
the rural-urban life satisfaction patterns in either separate, or joint analysis. These
findings suggest that if policy makers are interested in helping to promote higher levels of
well-being in rural areas, special attention should be given to public spending on social
services such as health care and proper housing.
The third chapter explores the change in life satisfaction following an internal
move for work and non-work migrants, and the aspects of life that could affect the
change in well-being for these two groups of movers. Results show that while life
satisfaction generally increases after a move, the persistence of this increase depends on
the reason for moving. Work migrants experience a significant increase in life
satisfaction that holds for those who moved within the past six years, as well as for those
for whom more than six year have gone by since the move. In contrast, although life
satisfaction of non-work migrants increases, this increase is only significant for those
109
who moved recently, not for those for whom more than six years have gone by since the
move. The finding that life satisfaction remains significantly higher in the long term for
work-migrants, but not for non-work migrants, is explained by the aspects of life that
change following the move. For work migrants, migration is conducive to a high-
achieving career path associated with improvements in long term relative status and
financial situation. The case of non-work migrants is different, as for this group migration
is associated with an increase in housing satisfaction but no changes in other aspects of
life, such as relative status. These findings suggest that policies facilitating work related
migration could increase overall satisfaction with life as a whole. Costs to labor mobility,
such those created by excessive regulation of the housing market, could create negative
repercussions in terms of overall well-being.
The last study of this dissertation (chapter four) analyzes the life satisfaction path
followed during young adulthood, and the degree to which young adult transitions alone
may account for this well-being path. Life satisfaction between ages 22 and 40 is found to
follow a slight inverse U-shape increasing until the age of 30, and decreasing thereafter.
During this same age period, most young adults undergo a similar pattern of life
transitions. Before the age of 30 the most typical transitions are partnership formation,
the school-to-work transition, and parenting of young children, all of which are
associated with an increase in life satisfaction. Between age 30 and 40, young adults
continue parenting children that are now older (age two or higher) and an increasing rate
of partnerships start dissolving. Both parenting older children and partnership dissolution
are negatively associated with life satisfaction. The typical pattern of life transitions
110
followed by young adults is found to largely account for the life satisfaction path between
ages 22 and 40. These results demonstrate the importance of life transitions in
determining the well-being of young adults. They suggest that, even in a country with a
strong welfare support system such as Sweden, additional policies helping young adults
to handle the job, family, and financial pressures accompanying life transitions may be
useful to maintain or improve long-term well-being.
111
REFERENCES
Aassve, Arnstein, Alice Goisis, and Maria Sironi. 2012. “Happiness and Childbearing
Across Europe.” Social Indicators Research 108(1): 65-86.
Abraham, Katherine G., Aaron Maitland, and Suzanne M. Bianchi. 2006. “Nonresponse
in the American time Use Survey: Who Is Missing from the Data and How Much
Does it Matter?” Public Opinion Quarterly 70(Special Issue No. 5): 676-703.
Ambler, Gareth, and Rumana Z. Omar. 2007. “A comparison of imputation techniques
for handling missing predictor values in a risk model with a binary outcome.”
Statistical Methods in Medical Research 16: 277-298.
Arriagada, Camilo. 2000. “Pobreza en America Latina: Nuevos escenarios y desafíos de
políticas para el hábitat urbano.” United Nations, ECLAC Series Medio Ambiente
y Desarrollo, No. 27.
Baird, Brendan M., Richard E. Lucas, and M. Brent Donnellan. 2010. “Life Satisfaction
Across the Lifespan: Findings from Two Nationally Representative Panel
Studies.” Social Indicators Research 99(2): 183-203.
Baranowska, Anna and Anna Matysiak. 2011. “Does parenthood increase happiness?
Evidence for Poland.” Vienna Yearbook of Population Research 9: 307-325.
Barcus, Holly. 2004. “Urban-Rural Migration in the USA: An Analysis of Residential
Satisfaction.” Regional Studies 38(6): 643-657.
Bartel, Ann P. 1979. “The Migration Decision: What Role Does Job Mobility Play?” The
American Economic Review 69(5):775-786.
Bartram, David. 2011. “Economic Migration and Happiness: Comparing Immigrants’ and
Natives’ Happiness Gains From Income.” Social Indicators Research 103(1): 57-
76.
Becketti, Sean, William Gould, Lee Lillard and Finis Welch. 1988. “The Panel Study of
Income Dynamics after Fourteen Years: An Evaluation.” Journal of Labor
Economics 6(4): 472-492.
Berry, Brian J.L. and Adam Okulicz-Kozaryn. 2009. “Dissatisfaction with city life: A
new look at some old questions.” Cities 26, pp. 117-124.
112
Billari, Francesco C. 2001. “The Analysis of Early Life Courses: Complex Descriptions
of the Transition to Adulthood.” Journal of Population Research 18(2): 119-142.
Blackburn, McKinley L. 2009. “Internal migration and the earnings of married couples in
the United States.” Journal of Economic Geography 10: 87-111.
Blanchflower, David G. and Andrew J. Oswald. 2008. “Is well-being U-shaped over the
life cycle?” Social Science & Medicine 66: 1733-1749.
Böheim, René and Mark P. Taylor. 2007. “From the dark end of the street to the bright
side of the road? The wage returns to migration in Britain.” Labour Economics
14: 99-117.
Cameron, A. Colin, Douglas L. Miller. 2011. “Robust Inference with Clustered Data.” In
Handbook of Empirical Economics and Finance, edited by Arman Ullah and
David E.A. Giles, 1-28. Boca Raton, FL., USA: Taylor and Francis.
Chen, Yong and Stuart S. Rosenthal. 2008. “Local amenities and life-cycle migration: Do
people move for jobs or fun?” Journal of Urban Economics 64: 519-537.
Clark, Andrew E., Ed Diener, Yannis Georgellis, and Richard E. Lucas. 2008. “Lags and
Leads in Life Satisfaction: A Test of the Baseline Hypothesis.” The Economic
Journal 118: F222 – F243.
Clark, Andrew E. and Andrew J. Oswald. 1994. “Unhappiness and Unemployment.” The
Economic Journal 104 (May): 648-659.
Clark, Andrew E. and Andrew J. Oswald. 1996. “Satisfaction with comparison income.”
Journal of Public Economics 61: 359-381.
Cooke, Thomas J. and Adrian J. Bailey. 1996. “Family Migration and the Employment of
Married Women and Men.” Economic Geography 72(1): 38-48.
Creed, Peter A., Juanita Muller, and Wendy Patton. 2003. “Leaving high school: the
influence and consequences for psychological well-being and career-related
confidence.” Journal of Adolescence 26: 295-311.
De Jong, Gordon F., Aphichat Chamratrithirong and Quynh-Giang Tran. 2002. “For
Better, for Worse: Life Satisfaction Consequences of Migration.” International
Migration Review 36(3): 838-863.
113
De Jong, Gordon F., Ricardo G. Abad, Fred Arnold, Benjamin V. Carino, James T.
Fawcett and Robert W. Gardner. 1983. “International and Internal Migration
Decision Making: A Value-Expectancy Based Analytical Framework of
Intentions to move from a Rural Philippine Province.” International Migration
Review 17(3): 470-484.
Department of Commerce: Bureau of the Census. 2002. “Qualifying Urban Areas for
Census 2000; Notice.” Federal Register 67(84): 21962-21967.
Diaz-Serrano, Luis and Alexandrina P. Stoyanova. 2009. “Mobility and housing
satisfaction: an empirical analysis for 12 EU countries.” Journal of Economic
Geography: 1-23.
DiTella, Rafael, Robert J. MacCulloch Andrew J. and Oswald. 2003. “The
Macroeconomics of Happiness.” The Review of Economics and Statistics 85(4):
809-827.
Dockery, Alfred M. 2005. “The Happiness of Young Australians: Empirical Evidence on
the Role of Labour Market Experience.” The Economic Record 81(255): 322-335.
Easterlin, Richard A. 2001a. “Income and Happiness: Towards a Unified Theory.” The
Economic Journal 111(473): 465-484.
Easterlin, Richard A. 2001b. “Life cycle welfare: Trends and differences.” Journal of
Happiness Studies 2: 1-12.
Easterlin, Richard A. 2005. “Building a Better Theory of Well-Being.” In Economics &
Happiness: Framing the Analysis, edited by Luigino Bruni and Pier L. Porta, 29-
64. New York: Oxford University Press.
Easterlin, Richard A. 2006. “Life cycle happiness and its sources: Intersections of
psychology, economics, and demography.” Journal of Economic Psychology 27:
463-482.
Easterlin, Richard A. and Laura Angelescu. 2009. “Happiness and Growth the World
Over: Time Series Evidence on the Happiness-Income Paradox.” IZA Discussion
Paper No. 4060.
Easterlin, Richard A., Laura Angelescu and Jacqueline S. Zweig. 2011. “The Impact of
Modern Economic Growth on Urban-Rural Differences in Subjective Well-
Being.” World Development 39(12): 2187-2198.
114
Easterlin, Richard A. and Onnicha Sawangfa. 2009. “Happiness and Domain Satisfaction:
New Directions for the Economics of Happiness.” In Happiness, Economics, and
Politics: Towards a Multi-Disciplinary Approach, edited by Amitava K. Dutt and
Benjamin Radcliff, 70-94. Northampton, MA.: Edward Elgar.
Economic Commission for Latin America and the Caribbean. 2006. Social Panorama of
Latin America, 2005. Santiago de Chile: United Nations.
Economic Commission for Latin America and the Caribbean. 2010a. “Estadisticas de
America Latina y el Caribe (CEPALSTAT).” Accessed May 2010.
http://websie.eclac.cl/sisgen/ConsultaIntegrada.asp
Economic Commission for Latin America and the Caribbean. 2010b. “Sistema de
Indicadores Sociodemograficos de Poblaciones y Pueblos Indígenas.” Accessed
May 2010. http://celade.cepal.org/ redatam/PRYESP/SISPPI/
Eisenhaur, Joseph G. 2003. “Regression through the Origin.” Teaching Statistics 25(3):
76-80.
Elder, Glen H. 1998.“The Life Course as Developmental Theory.” Child Development
69(1): 1-12.
Elder, Glen H., Monica K. Johnson and Robert Crosnoe. 2003. “The Emergence and
Development of Life Course Theory.” In Handbook of the Life Course, edited by
Jeylan T. Mortimer and Michael J. Shanahan, 3-19. New York, USA: Kluwer
Academic Publishers.
Ferrer-i-Carbonell, Ada and Paul Frijters. 2004. “How important is methodology in the
estimates of the determinants of happiness?” The Economic Journal 115(497):
641-659.
Findlay, Allan M. and Beata Nowok. 2012. “The uneven impact of different life domains
on the wellbeing of migrants.” Centre for Population Change Working Paper No.
26.
Finnie, Ross. 1999. “Inter-Provincial Migration in Canada: A Longitudinal Analysis of
Movers and Stayers and the Associated Income Dynamics.” Canadian Journal of
Regional Science 22(3): 227-262.
Fitzgerald, John, Peter Gottschalk and Robert Moffitt. 1998. “An Analysis of Sample
Attrition in Panel Data: The Michigan Panel Study of Income Dynamics.” NBER
Technical Working Paper No. 220.
115
Frey, William H. (1992). “Perspectives on Recent Demographic Change in Metropolitan
and Nonmetropolitan America.” Population Studies Centre of University of
Michigan Research Report No. 92-248.
Frijters, Paul, John P. Haisken-DeNew and Michael A. Shields. 2004. “Income Does
Matter! Evidence from Increasing Real Income and Life Satisfaction in East
Germany Following Reunification.” The American Economic Review 94(3): 730-
740.
Frijters Paul, Ingo Geisheckek, John P. Haisken-DeNew and Michael A. Shields. 2006.
“Can the Large Swings in Russian Life Satisfaction be Explained by Ups and
Downs in Real Incomes?” The Scandinavian Journal of Economics 108(3): 433-
458.
Ganzeboom, Harry B.G. and Donald J. Treiman. 1996. “Internationally Comparable
Measures of Occupational Status for the 1988 International Standard
Classification of Occupations.” Social Science Research 25: 201-239.
Ghatak, S., Levine, P., and Price, S.W. (1996). Migration Theories and Evidence: An
Assessment. Journal of Economic Surveys, 10(2): 159-198.
Groves Robert M. 2006. “Nonresponse Rates and Nonresponse Bias in Household
Surveys.” Public Opinion Quarterly 70(Special Issue No. 5): 646-675.
Hansen, Thomas, Britt Slagsvold and Torbjørn Moum. 2009. “Childlessness and
Psychological Well-Being in Midlife and Old Age: An Examination of Parental
Status Effects Across a Range of Outcomes.” Social Indicators Research 94: 343-
362.
Hansen, Thomas. 2012. “Parenthood and Happiness: a Review of Fold Theories Versus
Empirical Evidence.” Social Indicators Research 108: 29-64.
Harris, John R. and Michael P. Todaro. 1970. “Migration, Unemployment and
Development: A Two-Sector Analysis.” The American Economic Review 60(1):
126-142.
Hausman Jerry A. and David A. Wise. 1979. “Attrition Bias in Experimental and Panel
Data: The Gary Income Maintenance Experiment.” Econometrica 47(2): 455-473.
Hogan, Dennis P. and Nan M. Aston. 1986. “The Transition to Adulthood.” Annual
Review of Sociology 12: 109-130.
116
Knight, John and Ramani Gunatilaka. 2010. “Great Expectations? The Subjective Well-
Being of Rural-Urban Migrants in China.” World Development 38(1): 113-124.
Korenman, Sanders and David Neumark. 1991. “Does Marriage Really Make Men More
Productive?” The Journal of Human Resources 26(2): 282-307.
Lansing, John B. and James N. Morgan. 1967. “The Effects of Geographical Mobility on
Income.” The Journal of Human Resources, 2(4): 449-460.
Latinobarometro Corporation. “Latinobarometro Survey 2003-2006.” Accessed July
2009. http://www.latinobarometro.org.
Lucas, Richard E. 2005. “Time Does Not Heal All Wounds: A Longitudinal Study of
Reaction and Adaptation to Divorce.” Psychological Science 26(12): 945-950.
Lucas, Richard E., Andrew E. Clark, Yannis Georgellis and Ed Diener. 2003.
“Reexamining Adaptation and the Set Point Model of Happiness: Reactions to
Changes in Marital Status.” Journal of Personality and Social Psychology 84(3):
527-539.
Machinea, José L. and Martín Hopenhayn. 2005. “La esquiva equidad en el desarrollo
latinoamericana: Una vision structural, una aproximacion multifacética.” United
Nations, ECLAC Series Informes y Estudios Especiales, No. 14.
Marini, Margaret M. 1984. “The Order of Events in the Transition to Adulthood.”
Sociology of Education 57(2): 63-84.
McLanahan, Sara and Julia Adams. 1987. “Parenthood and Psychological Well-Being.”
Annual Review of Sociology 13: 237-257.
Melzer, Silvia M. 2011. “Does Migration Make You Happy? The Influence of Migration
on Subjective Well-Being.” Journal of Social Research and Policy 2(2): 73-92.
Molloy, Raven, Christopher L. Smith and Abigail K. Wozniak. 2011. “Internal migration
in the United States.” NBER Working Paper No. 17307.
Mroczek, Daniel K. and Avron Spiro. 2005. “Change in Life Satisfaction During
Adulthood: Findings From the Veterans Affairs Normative Aging Study.” Journal
of Personality and Social Psychology 88(1): 189-202.
Munshi, Kaivan. 2003. “Networks in the Modern Economy: Mexican Migrants in the
U.S. Labor Market.” The Quarterly Journal of Economics 118(2): 549-599.
117
Musick, Kelly and Larry Bumpass. 2012. “Reexamining the Case for Marriage: Union
Formation and Change in Well-Being.” Journal of Marriage and Family, 74(1):
1-18.
Myrskyla, Mikko and Rachel Margolis. 2012. “Happiness: Before and After Kids.”
MPIDR Working Paper WP 2012-013.
Nakazato, Naoki, Ulrich Schimmack and Shigehiro Oishi. 2011. “Effect of Changes in
Living Conditions on Well-Being: A Prospective Top-Down Bottom-Up Model.”
Social Indicators Research, 100: 115-135.
Nowok, Beata, Maarten van Ham, Allan M. Findlay and Vernon Gayle. 2013. “Does
Migration Make You Happy? A Longitudinal Study of Internal Migration and
Subjective Well-Being.” Environment and Planning 45(4): 986-1002.
Oswald, Andrew J. 1997. “Happiness and Economic Performance.” The Economic
Journal 107(445): 1815-1831.
Pekkala, Sari and Hannu Tervo. 2002. “Unemployment and Migration: Does Moving
Help?” The Scandinavian Journal of Economics 104(4): 621-639.
Rabe, Birgitta and Mark Taylor. 2010. “Residential mobility, quality of neighbourhood
and life course events.” Journal of the Royal Statistical Society 173(3): 531-555.
Rätzel, Steffen. 2012. “Labour Supply, Life Satisfaction, and the (Dis)Utility of Work.”
The Scandinavian Journal of Economics 114(4): 1160-1181.
Rojas, Mariano. 2006. “Life Satisfaction and Satisfaction in Domains of Life: Is it a
Simple Relationship?” Journal of Happiness Studies, 7: 467-497.
Scheffer, Judi. 2002. “Dealing with Missing Data.” Research Letters in the Information
and Mathematical Sciences 3: 153-160.
Shanahan, Michael J. 2000. “Pathways to Adulthood in Changing Societies: Variability
and Mechanisms in Life Course Perspective.” Annual Review of Sociology 26:
667-692.
Pichler, Florian, Mark Shucksmith, Stuart Cameron and Tanya Merridew. 2006. First
European Quality of Life Survey: Urban-rural differences. Luxembourg: Office
for Official Publications of the European Communities.
Sjaastad, Larry A. 1962. “The Costs and Returns of Human Migration.” Journal of
Political Economy 70(5): 80-93.
118
Speare, Alden. 1974. “Residential satisfaction as an intervening variable in residential
mobility.” Demography 11(2): 173-188.
Spellerberg, Annette, Denis Huschka and Roland Habich. 2007 “Quality of life in rural
areas: processes of divergence and convergence.” Social Indicators Research
83(2), pp. 283-307.
Statistics Sweden Register data compilations: the Register of Education, the Register of
Income and Taxation, and the Total Population Register.
Statistics Sweden Database. Economic Statistics. “Prices and consumption.” Accessed
March 2011. http://www.scb.se/Pages/SubjectArea____11537.aspx.
Statistics Sweden Database. Environment. “Land and water areas.” Accessed March
2011. http://www.scb.se/Pages/ProductTables____12833.aspx.
Stiglitz, Joseph E., Amartya Sen and Jean-Paul Fitouss. 2009. “Report of the commission
on the measurement of economic performance and social progress.” Accessed
March 2010. http://www.stiglitz-sen-fitoussi.fr/en/index.htm.
Thomas, Duncan, Elizabeth Frankenberg E., James P. Smith. 2001. “Lost but Not
Forgotten: Attrition and Follow-up in the Indonesia Family Life Survey.” The
Journal of Human Resources 36(3): 556-592.
Thomas, Duncan, Firman Witoelar, Elizabeth Frankenberg., Bondan Sikoki, John
Strauss, Cecep Sumantri and Wayan Suriastini. 2012. “Cutting the costs of
attrition: Results from the Indonesia Family Life Survey.” Journal of
Development Economics 98(1): 108-123.
United Nations. 2008. “World Urbanization Prospects: The 2007 Revision Population
Database.” Accessed May 2010. http://esa.un.org/unup/.
Veenhoven, Ruut. 1994. “How satisfying is rural life? Fact and value.” In Changing
values and attitudes in family households. Implications for institutional transition
in East and West FAA report 296, edited by James Cecora, 41-51. Bonn
Germany: Society for agricultural policy research in rural society.
Vogel, Joachim. 2002. “European Welfare regimes and the transition to adulthood: A
comparative and longitudinal perspective.” Social Indicators Research 59(3):
275-299.
von Hippel, Paul T. 2007. “Regression with Missing Y’s: an Improved Strategy for
Analyzing Multiply Imputed Data.” Sociological Methodology 37(1): 83-117.
119
Waite, Linda J. 1995. “Does Marriage Matter?” Demography 32(4): 483-507.
Weiss, Leonard and Jeffrey G. Williamson. 1972. “Black Education, Earnings, and Inter-
regional Migration: Some New Evidence.” The American Economic Review,
62(3): 372-383.
Wooldridge, Jeffrey M. (2002). Econometric Analysis of Cross Section and Panel Data.
United States: Massachusetts Institute of Technology.
World Bank. 2010. “World Development Indicators & Global Development Finance.”
Accessed May 2010. http://databank.worldbank.org/ddp/home.do?Step=12&id=
4&CNO=2.
Zabel Jeffrey E. 1998. “An Analysis of Attrition in the Panel Study of Income Dynamics
and the Survey of Income and Program Participation with and Application to a
Model of Labor Market Behavior.” The Journal of Human Resources 33(2): 479-
506.
Zimmermann, Anke C. and Easterlin, Richard A. 2006. “Happily Ever after?
Cohabitation, Marriage, Divorce, and Happiness in Germany.” Population and
Development Review 32(3): 511-528.
120
APPENDICES
Appendix A: Additional analysis for Chapter 2
A.1: Methods
As mentioned in the text, certain issues with the Latinobarometro Survey created
limitations concerning the data used. Firstly, as to the year coverage, even though the
Latinobarometro was carried out since 1995, the “size of town” variable used in this
study was not collected until 2001. In 1997 and 1998 there was information available
about the size of the town in which the person lived, however, it was collected using a
different classification than the one employed in this study and is therefore not
compatible with the analysis performed. Moreover, the amount, order, and formulation of
the questions asked changed in several years. Therefore, for consistency, the year
coverage had to be restricted to 2003 to 2006
as before and after these years at least one
of the main variables of interest was not available.
The second data problem faced had to do with the data coverage of each of the
countries surveyed by the Latinobarometro, the main issue being that for certain countries
the rural population was greatly under sampled in some (or all) years. To assure that the
coverage of population was at least roughly representative at a national scale, a
comparison was performed of the percent of people living in areas with less than 5,000 or
less than 10,000 population – weighted values used – surveyed by the Latinobarometro
121
(Table A-1) versus the percent of people living in rural areas as reported by the United
Nations (United Nations 2008) for each country. Based on this analysis the following
decisions were made: Guatemala and Nicaragua were completely excluded from the
study due to their almost complete lack of coverage of cities with less than 5,000 or less
than 10,000 inhabitants. Moreover, for some countries (Costa Rica, Chile, Ecuador,
Mexico, Panama, Peru and Venezuela) years in which the rural population interviewed
seemed not to be representative of the national values were also omitted. For Panama,
where the percent of urban population as reported by the UN is not available, the decision
to omit 2006 was based on the big jump observed in that year for the number of people
surveyed in cities with less than 5,000 inhabitants. Also, even though for Costa Rica the
under sampling of rural cities was not a problem, the year 2006 was still left out since in
that year no city with a population of over 100,000 people was surveyed.
Table A-1: Percent of urban population (living in cities of 5,000/10,000 or more
inhabitants) as surveyed by the Latinobarometro (weighted values used)
Argentina Bolivia
year
% in cities
>5k (LB)
% in cities
>10k (LB) Year
% in cities
>5k (LB)
% in cities
>10k (LB)
1985 1985
1990 1990
2003 79 79 2003 58 54
2004 82 80 2004 54 51
2005 90 82 2005 53 52
2006 85 85 2006 99 94
Brazil Colombia
year
% in cities
>5k (LB)
% in cities
>10k (LB) Year
% in cities
>5k (LB)
% in cities
>10k (LB)
1985 1985
1990 1990
2003 96 95 2003 94 89
2004 97 91 2004 94 89
2005 97 88 2005 95 91
2006 94 87 2006 96 92
122
Table A-1 (Continued)
Costa
Rica Chile
year
% in cities
>5k (LB)
% in cities
>10k (LB) year
% in cities
>5k (LB)
% in cities
>10k (LB)
1985 1985
1990 1990
2003 88 71 2003 100 100
2004 87 61 2004 100 100
2005 86 67 2005 100 100
2006 86 62 2006 77 73
Ecuador
El
Salvador
year
% in cities
>5k (LB)
% in cities
>10k (LB) year
% in cities
>5k (LB)
% in cities
>10k (LB)
1985 1985
1990 1990
2003 75 64 2003 54 47
2004 74 61 2004 53 44
2005 77 67 2005 51 42
2006 100 99 2006 48 44
Guatemala Honduras
year
% in cities
>5k (LB)
% in cities
>10k (LB) year
% in cities
>5k (LB)
% in cities
>10k (LB)
1985 1985
1990 1990
2003 100 93 2003 100 85
2004 100 98 2004 93 82
2005 100 96 2005 98 84
2006 100 94 2006 97 89
Table A-1 Continued
Mexico Nicaragua
year
% in cities
>5k (LB)
% in cities
>10k (LB) year
% in cities
>5k (LB)
% in cities
>10k (LB)
1985 1985
1990 1990
2003 100 100 2003 100 97
2004 72 67 2004 98 96
2005 69 64 2005 98 96
2006 75 74 2006 100 98
123
Table A-1 (Continued)
Panama Paraguay
year
% in cities
>5k (LB)
% in cities
>10k (LB) year
% in cities
>5k (LB)
% in cities
>10k (LB)
1985 1985
1990 1990
2003 74 66 2003 100 100
2004 67 56 2004 100 100
2005 73 63 2005 83 69
2006 46 46 2006 76 66
Peru Uruguay
year
% in cities
>5k (LB)
% in cities
>10k (LB) year
% in cities
>5k (LB)
% in cities
>10k (LB)
1985 1985
1990 1990
2003 39 32 2003 82 78
2004 62 53 2004 80 74
2005 65 62 2005 80 74
2006 99 99 2006 81 75
Venezuela
year
% in cities
>5k (LB)
% in cities
>10k (LB)
1985
1990
2003 93 87
2004 100 96
2005 88 78
2006 89 81
Finally, Paraguay had to be excluded from the analysis due to a problem with the
response ratios to the question used to identify the indigenous population. Since the
question used to detect the indigenous origins of a person was the mother tongue as
reported by the individual, one could expect, if anything, an under-sampling of the
individuals identified as indigenous at the national level. This is because, if a person
living in an indigenous community reported Spanish (or Portuguese) as her mother
tongue, which is possible in a country where Spanish (Portuguese) represents the main
124
language, they would not be classified as having indigenous origins. In fact, in most
countries, the percent of the population identified as indigenous by the survey (Table A-
2) is below the national percentage of indigenous population as reported by the Economic
Commission for Latin American and Caribbean Association (2010b). However,
Paraguay stands out alone in having a much higher percent of people reporting an
indigenous mother tongue in the survey than what is to be expected based on the nation’s
ratio of indigenous population; based on their mother tongue, over one half of the
population in Paraguay could be considered indigenous. This abnormality is consistent
for all years of the survey used for Paraguay. Due to this, and since indigenous origins is
one of the main variables of interest in this study, Paraguay was left out of the analysis.
Table A-2: Percent of people classified as having indigenous origins using the
Latinobarometro data (weighted values)
Country
% People reporting indigenous language
as their mother tongue (LB)
Argentina 0.78
Bolivia 33.85
Brazil 0.95
Chile 0.02
Colombia 0.64
Costa Rica 0.05
Ecuador 1.66
El Salvador 0.04
Honduras 0.29
Mexico 1.33
Panama 1.86
Paraguay 58.33
Peru 10.63
Uruguay 0.23
Venezuela 0.24
125
A.2: Description of countries in the analysis
Table A-3: Number of observations used for analysis, per year and country
Year Argentina Bolivia Brazil Colombia
Costa
Rica Chile Ecuador
2003 1193 1200 1195 1195 999 1199
2004 1198 1179 1189 1192 1000 1204
2005 1193 1193 1200 1193 995 1196
2006 1197 1191 1202 1198 1196
Total 4781 4763 4786 4778 2994 1196 3599
El
Salvador Honduras Mexico Panama Peru Uruguay Venezuela
2003 1000 1003 993 1195 1197
2004 1003 1000 1200 1000 1193 1188
2005 990 997 1188 1002 1187 1196 1195
2006 1020 997 1194 1195 1194
Total 4012 3997 3582 2995 2380 4774 3586
Table A-4: Mean life satisfaction by size-of-place and country
Argentina Bolivia
Location N %
Mean
LS Location N %
Mean
LS
<=5k 760 16 2.93 <=5k 1622 34 2.51
5k-40k 675 14 2.97 5k-40k 730 15 2.65
40k-100k 353 7 3.08 40k-100k 275 6 2.76
>100k/capital 2992 63 2.92 >100k/capital 2135 45 2.60
Total 4781 100 2.94 Total 4763 100 2.59
Brazil Colombia
Location N %
Mean
LS Location N %
Mean
LS
<=5k 200 4 2.74 <=5k 178 4 3.06
5k-40k 1392 29 2.81 5k-40k 701 15 3.24
40k-100k 1065 22 2.72 40k-100k 814 17 3.03
>100k/capital 2129 44 2.68 >100k/capital 3085 65 3.20
Total 4786 100 2.73 Total 4778 100 3.17
126
Table A-4 (Continued)
Costa Rica Chile
Location N %
Mean
LS Location N %
Mean
LS
<=5k 391 13 3.38 <=5k 279 23 2.84
5k-40k 2160 72 3.36 5k-40k 75 6 2.87
40k-100k 173 6 3.34 40k-100k 128 11 2.80
>100k/capital 270 9 3.38 >100k/capital 714 60 2.94
Total 2994 100 3.36 Total 1196 100 2.90
Ecuador El Salvador
Location N %
Mean
LS Location N %
Mean
LS
<=5k 902 25 2.70 <=5k 1947 49 2.97
5k-40k 948 26 2.61 5k-40k 690 17 2.98
40k-100k 220 6 2.78 40k-100k 298 7 3.01
>100k/capital 1529 42 2.82 >100k/capital 1077 27 3.06
Total 3599 100 2.73 Total 4012 100 3.00
Honduras Mexico
Location N %
Mean
LS Location N %
Mean
LS
<=5k 115 3 3.14 <=5k 999 28 2.90
5k-40k 1819 46 3.18 5k-40k 522 15 2.98
40k-100k 602 15 3.19 40k-100k 269 8 3.09
>100k/capital 1460 37 3.12 >100k/capital 1791 50 3.09
Total 3997 100 3.16 Total 3582 100 3.02
Panama Peru
Location N %
Mean
LS Location N %
Mean
LS
<=5k 824 28 3.07 <=5k 869 37 2.48
5k-40k 993 33 3.17 5k-40k 327 14 2.53
40k-100k 258 9 3.34 40k-100k 135 6 2.53
>100k/capital 920 31 3.20 >100k/capital 1049 44 2.46
Total 2995 100 3.17 Total 2380 100 2.48
Uruguay Venezuela
Location N %
Mean
LS Location N %
Mean
LS
<=5k 848 18 2.88 <=5k 351 10 3.52
5k-40k 1198 25 2.93 5k-40k 1224 34 3.39
40k-100k 679 14 2.74 40k-100k 1002 28 3.41
>100k/capital 2049 43 2.79 >100k/capital 1010 28 3.40
Total 4774 100 2.83 Total 3586 100 3.41
127
A.3: Description of variables in the analysis
Table A-5: Questions and response categories for each of the explanations analyzed
Hypothesis
tested Specific variable Question asked in Latinobarometro
Income/
development
Ownership of goods
A sum of the ownership of following goods:
colored tv, refrigerator, home, computer, washer,
phone, car and holiday home (scale 0-8) -- the
exact question says: Possession of household
goods, response categories: yes/no
Rating of personal
economic situation
In general, how would you describe your present
economic situation and that of your family? Would
you say that it is very good, good, about average,
bad or very bad? (scale 1-5)
Relationship of income
to needs
Does you salary and the total of your family's salary
allow you to satisfactorily cover your needs? Which
of the following situations do you find yourself in?
Covers them well (can save), covers alright, does
not cover (there are difficulties), does not cover
(there are great difficulties). (scale 1-4)
Combined
employment/occupation
What is the current employment situation
or kind of work performed? Possible
categories: Professional/manager, other
public worker, other private worker, owner,
informal worker, farmer/fisher, unemployed,
retired, student, homemaker
Social values
Marital status
Classification into single, married or
separated/divorced/widowed -- Exact
question: marital status with above categories
Crime
Have you, or someone in your family, been
assaulted, attacked, or been the victim of a
crime in the last 12 months? (yes/no)
Corruption
Have you or someone in your family been
aware of an act of corruption in the last
12 months? (yes/no)
Religiosity
How would you describe yourself? Very devout,
devout, not very devout, or not devout at all?
128
Table A-5 (Continued)
Welfare State
Access to education
Would you say that you are very satisfied, rather
satisfied, not very satisfied or not at all satisfied
with the education to which you have access?
Access to health
Would you say that you are very satisfied, rather
satisfied, not very satisfied or not at all satisfied
with the health care to which you have access?
Basic services
A sum of the availability of following services
in the household: water, hot water, sewage
(scale 0-3)
Indigenous
origins Mother tongue
What is your mother tongue? Response categories:
Spanish, Portuguese, Indigenous language, other.
Using this question all people who responded that
indigenous language was their mother tongue were
classified as having indigenous origins
Appendix B: Additional analysis for Chapter 3
B.1: Attrition in the Young Adult Panel Study
Given its longitudinal nature, the YAPS survey faces the inevitable problem of attrition.
Of the 2820 individuals interviewed in 1999, 1575 were re-interviewed in 2009. This
generated an attrition rate of 44% over the 10 year period, which is similar to the rates
typically observed in longitudinal surveys from other developed countries (Becketti et al.
1988, Abraham, Maitland and Bianchi 2006). The high non-response in YAPS creates
concerns about the existence of an attrition bias. In what follows, first, the main
characteristics at baseline of those who attrit (not interviewed in 2009) and who do not
attrit are compared. Then, two main problems related to attrition are discussed: selection
129
on migration, and selection on unobserved time-varying characteristics related to the
dependent variables of the study.
At baseline, attritors have generally lower income
34
, lower economic satisfaction,
and less years of education, then people who are interviewed in both 1999 and 2009.
Attritors are more likely to be male, young, and have Swedish background (Table B-1).
The characteristics related to income and education are unlike those observed in previous
studies in developing (Thomas, Frankenberg and Smith 2001 and 2012) and developed
countries (Hausman and Wise 1979, Becketti et al. 1988), where attrition was found to
have a positive association with income and education. This may be difference due to the
design of the YAPS survey which targets young adults (ages 22 to 30 in 1999), and
therefore has a high proportion of student respondents (with low income) at the time of
the first survey. Given that young people are more likely to leave the survey, a higher
percent of attritors has not achieved their final levels of education in 1999, lowering the
average education level of this group and their income and economic satisfaction.
Table B-1: Comparison of the characteristics at baseline (1999) of respondents who
consequently attrit (not interviewed in 2009) and do not attrit (interviewed in 2009)
Complete sample Non-attritors Attritors
N Mean N Mean N Mean
Life satisfaction 2785 3.91 1560 3.92 1225 3.9
Self reported income
(in 1000 SEK)***
2800 101 1573 104 1227 97
Economic satisfaction 2789 3.05 1564 3.11 1225 2.97
Satisfaction with housing 2776 3.7 1556 3.69 1220 3.73
34
The income variable used here is self-reported income in 1999, and is different from the
Register data used in the study. The Register data could not be used to analyze the problem of attrition, as it
is only available for the people who are interviewed in 2009 – consequently, it is only available for non-
attritors.
130
Table B-1 (Continued)
Satisfaction with partner 2075 4.47 1159 4.45 916 4.49
Satisfaction with occup. 2751 3.78 1551 3.81 1200 3.76
Educ level 1999** 2782 11.98 1565 12.19 1217 11.71
Hours worked per week 2014 37.47 1132 37.79 882 37.06
% Male 1320 46.80% 702 44.57% 618 49.64%
% Studying 208 7.71% 121 7.94% 87 7.40%
% Cohort 1976 (age 22) 1107 39.30% 589 37.40% 518 41.60%
% Cohort 1972 (age 26) 973 34.50% 543 34.50% 430 34.50%
% Cohort 1968 (age 30) 740 26.20% 443 28.10% 297 23.90%
% Married 393 14% 208 13.20% 185 15.10%
% Swedish background 2283 80.96% 1336 84.83% 947 76.06%
%Polish/Turkish
background
537 19.04% 239 15.17% 298 23.94%
Bold values imply that the mean or % for attritors and non-attritors are statistically different at
5% significance level.
** information reported in 1999; different from Register information used in study
The relationship between the birth cohort and attrition is similar to that observed
in previous literature, with younger cohorts being more likely to attrit in subsequent
interviews. The difference in the attrition rates of people with Swedish and non-Swedish
background may be related to previous findings that early life experience and parent
characteristics are related to attrition (Thomas et al. 2012). Interestingly, higher levels of
attrition are not associated with more hours worked per week, as could be expected if
busy people were less likely to be re-interviewed. Previous studies conducted with
surveys from the United States have found that non-contact is in fact associated with
longer work times, though the same did not hold for refusals, with refusal rates showing
no association with work time (Abraham, Maitland and Bianchi 2006).
Attrition in the YAPS survey could represent a major problem if it was selective
on migration given that the main focus of the present study is on comparisons of migrants
and non-migrants. Past research has found that attrition in longitudinal surveys may, in
131
fact, be selective on migration. This problem arises especially in the case of surveys
performed in developing countries (Thomas, Frankenberg and Smith 2001; Thomas et al.
2012), as in developed countries non-response rates in surveys are mostly associated with
refusals as opposed to failure to contact the respondents. Still, Abraham, Maitland and
Bianchi (2006) find that non-contact rates may also be high in developed countries, as
documented by their observations about the American Time Use Survey.
The problem of attrition due to migration should be lessened in the YAPS due to
the access of the employees of Statistics Sweden, who were in charge of the data
collection, to the Swedish Register records. The Register consists of data collected by the
Swedish Tax Agency and includes specific information about current place of residence
for all individuals. Access to this information should potentially make the task of
following migrants considerably easier than in countries with less precise demographic
information on their inhabitants.
A comparison of non-contact versus refusal rates in the YAPS could be
informative, as non-response associated with non-contact may be more related to trouble
finding a person who has moved. Unfortunately, the YAPS survey was performed by
mail, and no information of non-contact versus refusal rates was collected. Still, because
attrition is generally associated with similar demographic characteristics across different
surveys (Zabel 1998), a comparison of the characteristics of attritors in the YAPS to the
characteristics of attritors due to non-contact in other surveys could provide insight into
this problem. In developed countries such as the United States, non-contact is typically
associated with being single, working longer hours, and being a high school graduate
132
(Abraham, Maitland and Bianchi 2006). In the YAPS, the proportion of people married
and the hours worked at baseline are not statistically different for attritors and non-
attritors. Moreover, attritors have significantly less years of education, which is the
opposite of the association between education and non-contact found by Abraham,
Maitland and Bianchi. If the same associations between non-contact and demographic
characteristics hold for Sweden as for United States, this could imply that a big
proportion of attrition in the YAPS is due to refusal. Still, it is not clear that Swedish
attrition should follow the same patterns as those observed in studies from other
countries, and so the previous implication may be considered inconclusive.
An additional indirect test of selection on attrition used by previous literature
consists of comparing characteristics of interest of the observed survey sample to those of
a similar sample of the general population (Groves 2006). Using this method, a test of
attrition selective on migration in the YAPS is performed comparing rates of mobility by
cohort of survey respondents interviewed in both years to those of the general population
of Sweden (Table B-2). For every cohort, the mobility of the general population is
slightly above that of the non-attritors from YAPS, with the difference between the two
populations being highest for the 1976 cohort. For all cohorts combined, the difference in
the migration proportions between the general population and the YAPS is 3% (44% for
general population and 41% for YAPS). This difference implies that, though selection on
migration might have certainly taken place in the YAPS survey, the magnitude of this
selection appears small.
133
Table B-2: Mobility by cohort: general population vs. YAPS non-attritors
% Migrants
Cohort
General Pop.
(Register)
YAPS
non-attritors
Difference
1968 31.16% 29.35% 1.81%
1972 44.04% 38.67% 5.37%
1976 57.65% 52.98% 4.67%
Total 43.63% 41.39% 2.24%
The second reason why attrition could bias the results is if it was selective on
time-varying characteristics associated with either changes in life satisfaction or any of
the other dependent variables used. Based on the analysis of baseline characteristics it
appears that, in levels, attrition is not highly associated with most of the dependent
variables used, with income and economic satisfaction being the two exceptions (Table
B-1). To analyze the issue of selection on unobservables, a test from previous literature
(Fitzgerald, Gottschalk and Moffitt 1998) is used. The test checks the significance of
attrition by employing regressions of the main dependent variables at baseline on
subsequent attrition and control variables. If attrition is in fact a problem, then its
coefficient in such regressions should be significant.
Attrition is not significant, both with and without additional control variables, for
life satisfaction, satisfaction with housing, and satisfaction with occupation. This
indicates that, most likely, attrition is not selective on these variables. Controlling for the
personal characteristics that are accounted for in the main regressions
35
, attrition becomes
35
The control variables are chosen based on fixed characteristics that would be accounted for in a
first difference regression (i.e. gender, and nationality), and additional controls similar to those used in the
main regressions in the study (i.e. cohort of birth, marital status, a dummy for being a student – not having
completed education – and a dummy for having a child in the household.
134
insignificant in the income regression, and reduces its significance levels in both the
economic satisfaction, and occupational status regressions (Table B-3).
Table B-3: OLS regressions: variables of interest (in levels) on future attrition
Life sat. Self-reported income Economic sat.
attrit99_09 -0.028 -0.005 -7.711 -2.251 -0.139 -0.103
-0.79 -0.15 (2.57)** -0.9 (3.19)*** (2.34)**
male
-0.123
21.784
0.15
(3.47)***
(8.67)***
(3.40)***
swedish
0.131
5.519
-0.079
(2.67)***
(1.76)*
-1.27
1972 c.
0.043
60.499
0.199
-1
(20.4)***
(3.65)***
1968 c.
-0.018
99.798
0.364
-0.32
(24.7)***
(5.43)***
married
0.193
0.889
0.316
(3.45)***
-0.22
(4.67)***
div/wid
-0.204
-27.811
-0.462
-1.45
(3.44)***
(2.45)**
student
-0.019
-35.249
-0.35
-0.3
(11.4)***
(4.13)***
child in hh
0.186
12.569
-0.236
(3.81)***
(3.58)***
(3.99)***
Constant 3.924 3.798 104.27 40.809 3.111 3
(171)*** (69.5)*** (51.4)*** (12.4)*** (110)*** (44.6)***
Observations 2785 2631 2800 2659 2789 2634
R-squared 0 0.03 0 0.38 0 0.04
Occupational status Sat. with occupation Sat. with housing
attrit99_09 -0.097 -0.057 -0.053 -0.044 0.043 0.055
(2.83)*** (1.84)* -1.22 -1.01 -1 -1.24
male
0.103
-0.021
-0.119
(3.24)***
-0.49
(2.72)***
swedish
0.005
-0.023
0.081
-0.12
-0.4
-1.29
1972 c.
0.492
0.135
0.086
(12.8)***
(2.53)**
-1.56
1968 c.
0.752
0.11
0.141
(15.7)***
-1.64
(2.15)**
married
0.151
0.077
0.115
(3.03)***
-1.07
(1.74)*
div/wid
-0.15
-0.307
-0.066
-1.21
-1.54
-0.37
student
-0.834
0.462
-0.143
(32.4)***
(6.74)***
(1.75)*
135
Table B-3 (Continued)
child in hh
-0.172
-0.106
0.132
(4.13)***
(1.72)*
(2.21)**
Constant 2.065 1.729 3.808 3.764 3.686 3.568
(87.6)*** (39.6)*** (138)*** (59.1)*** (133)*** (51.6)***
Observations 2736 2660 2751 2605 2776 2624
R-squared 0 0.22 0 0.02 0 0.02
Robust t statistics in parentheses
* significant at 10%; ** significant at 5%; *** significant at 1%
The dependent variables used in the previous regressions are in levels, whereas
those used in the main part of the analysis are all first difference dependent variables.
The first difference variables should be more robust to possible selection problems, as
they implicitly control for any fixed characteristics of the respondents that could be
related to their subsequent non-response. Still, previous research has shown that attrition
could also be related to time-varying unobserved characteristics that could bias the results
of a first-difference regression (Thomas et al. 2012).
Since attritors are not interviewed in 2009, it is impossible to check whether the
changes in the variables of interest over the period under analysis (99-09) differ
depending on whether a person drops out of the survey or not. However, two additional
tests may be carried out using first difference, as opposed to level, variables to
approximate the methods used in the study. First, even though the attritors are not
observed in 2009, some of them did participate in an intermediate survey performed in
2003. Using these 2003 responses, a comparison of the 99-03 changes in the main
variables of interest may be performed between people who remain in the survey in 2009
and those who eventually drop out (the attritors). Using these 99-03 first difference
136
variables, regressions on future attrition alone, and with the available control variables
36
are run. Attrition is not significant in any of these regressions (Table B-4), indicating that
attrition is unlikely to be selective on the first difference variables used in the main
analysis.
Table B-4: OLS regressions: variables of interest (in 99-03 changes) on attrition
Life
satisfaction
Self-reported
income
Economic
satisfaction
attrit99_09 -0.05 -0.07 -4.536 -5.005 -0.078 -0.055
-1.06 -1.44 -0.94 -0.99 -1.48 -1.09
1972 c.
-0.024
-4.623
-0.198
-0.41
-0.81
(3.28)***
1968 c.
-0.1
-15.521
-0.354
(1.68)*
(2.69)***
(5.92)***
married FD
0.02
22.758
0.037
-0.31
(2.92)***
-0.51
div/wid FD
0.112
44.704
0.033
-0.47
(2.24)**
-0.16
student FD
-0.056
7.369
-1.106
-0.62
-0.99
(12.2)***
child born
0.164
14.162
-0.052
(2.82)***
(2.55)**
-0.88
Constant 0.023 0.023 70 71.688 0.397 0.509
-0.83 -0.49 (25.4)*** (17.2)*** (12.8)*** (10.2)***
Observations 2049 1912 2086 1942 2002 1945
R-squared 0 0.01 0 0.02 0 0.1
Occupational
status
Sat with
occupation
Satisfaction with
housing
attrit99_09 0.043 0.035 -0.065 -0.091 0.067 0.049
-0.75 -0.59 -1.08 -1.43 -1.05 -0.76
1972 c.
0.072
0.116
-0.07
-1.03
-1.53
-0.95
1968 c.
0.049
0.084
0.069
-0.72
-1.1
-0.86
married FD
0.181
0.191
0.127
(2.21)**
(2.05)**
-1.46
36
The main difference in the control variables used here and in the main analysis is that, for
attritors, the education completion variable could not be constructed (because of absent register data), and
so it is not used in Table B-4. Instead, a student dummy first difference variable is used to proxy for
education completion.
137
Table B-4 (Continued)
div/wid FD
-0.307
0.184
-0.321
-1.45
-0.83
-1.22
student FD
-0.447
-0.173
0.451
(3.97)***
-1.49
(4.22)***
child born
-0.191
0.092
0.033
(2.96)***
-1.21
-0.44
Constant 0.122 0.084 0.103 -0.005 -0.003 0.011
(3.70)*** -1.47 (2.93)*** -0.07 -0.08 -0.18
Observations 2052 1916 2046 1911 2004 1876
R-squared 0 0.02 0 0.01 0 0.01
Robust t statistics in parentheses
* significant at 10%; ** significant at 5%; *** significant at 1%
The second test performed using first difference variables consists of comparing
the changes in a clue variable for the sample of respondents from the YAPS interviewed
in both 1999 and 2009, to the changes in the same variable for the general population.
This comparison is carried out for income changes (Table B-5). There are two main
reasons to use income for this test. First, disposable income is readily available from the
Statistics Sweden for both, the YAPS sample, and the general population. Second,
attrition has been specifically found to be selective on changes in returns to human
capital, such as education (Thomas et al. 2012), which could possibly be reflected in
changes in disposable income.
Table B-5: Mean disposable income (hundreds of SEK), whole population (1968,
1972 and 1976 cohorts) and YAPS (non-attritors), by migration status, by year
Population YAPS
Period 1998 2007 Change 1998 2007 Change
Migrant 1052 2262 1209 1075 2382 1307
Non-Migrant 1132 2076 944 1138 2227 1089
Both migrants and
non-migrants
1097 2157 1060 1112 2290 1178
Difference migrants
- non-migrants
-80 186 265 -63 155 218
138
For both migrants and non-migrants observed in the YAPS survey in 1999 and
2009, the changes in disposable income are slightly above those of the general
population.
37
Because the present study is based on the comparison of migrants versus
non-migrants, one may be especially interested in comparing the difference in changes in
income for these two groups for the YAPS sample and the general population. For the
sample of non-attritors from YAPS, the difference between changes in income for
migrants and non-migrants is 21800 SEK; the difference between the migrant groups for
the general population is 26500 SEK (Table B-5). The closeness between these two
differences is reassuring.
Because of the high levels of attrition in the YAPS survey, concerns with possible
bias may certainly arise. Given the previous analysis, selective attrition on migration,
though possible, appears to be generally small in magnitude. The first-difference
regression analysis used in the study allows to control for all time invariant unobserved
characteristics that could be related to both attrition and the variables of interest. Though
the possibility of time varying unobserved characteristics related to attrition remains, the
two additional tests performed (using first difference variables over 99-03 and a
comparison of the changes in income for migrants and non-migrants for the YAPS
sample and the general population) both provide results indicating that the first difference
variables do not appear to be selective on attrition. In conclusion, the results of the
analysis performed in this section provide reassurance that the possible attirition bias in
the survey should not have a strong effect on the main results of the study.
37
The general population encompasses all inhabitants of Sweden born in the 1968, 1972 and 1976
cohorts for whom Register information was available in 1999 and 2009.
139
B2: Description of variables used in the study
Table B-6: Number of people surveyed answering each question in both 99 and 09,
by migration status and reason to move
Work
migrants
Non-
work
migrants
All
migrants
Non-
migrants
Total
Life satisfaction 218 338 630 911 1541
Economic satisfaction 220 340 636 919 1555
Satisfaction with house 219 341 632 912 1544
Satisfaction with occupation 222 334 629 893 1522
Satisfaction with partner 121 244 415 642 1057
Occupation group 215 326 609 860 1469
Civil status 222 344 643 930 1573
Education 221 343 641 923 1564
Work Income 222 344 643 930 1573
Disposable Income 222 344 643 930 1573
Table B-7: Description of original survey questions used in the analysis
Variable Question asked Response categories
child born
(year/month)
Self-reported year in which children 1-
5 were born
Year and month recorded
separately
economic
satisfaction
Are you satisfied or dissatisfied with
your economic situation?
Scale: 1= very dissatisfied to
5= very satisfied
life satisfaction
Are you satisfied or dissatisfied with
life in general right now?
Scale: 1= very dissatisfied to
5= very satisfied
long distance move
(year/month)
When did you last make a long
distance move? (year and month)
Year and month recorded
separately
main activity What is your current main activity?
Response grouped as:
1. permanent emp; 2. casual/
limited emp ; 3. self-emp;
4. studies; 5. kunskapslyftetet;
6. employment measures;
7. unemp >= 6 months;
8. unemp < 6 months; 9.
parental leave; 10.
housekeeping; 11. military; 12.
retired; 13. on long term sick
leave; 14. doctoral student; 15.
on leave from work; 16. other
140
Table B-7 (Continued)
Occupation
What is your main occupation? What
are your main tasks at work?
Open ended response from
survey regrouped as:
1.unskilled in good production;
2.unskilled in service
production; 3.skilled in goods
production; 4.skilled in service
production; 5.assistant non-
manual, lower level i;
6.assistant non-manual, lower
level ii; 7.intermediate non-
manual; 8.professionals and
other higher non-manual;
9. upper-level executives;
10. self-employed
professionals;
11. entrepreneurs; 12. farmers
reason_move
What was the most important reason
for you to move?
My work/studies ; My partners
work/studies ; I wanted to move
to my partner; I wanted to come
closer to friends and family; I
wanted a change of
environment; I wanted to move
back to where I grew up; My
partner wanted to move; Other,
namely...
satisfaction with
housing
Are you satisfied or dissatisfied with
your housing situation?
Scale: 1= very dissatisfied to
5= very satisfied
satisfaction with
partner
Are you satisfied or dissatisfied with
your relationship with your partner?
Scale: 1= very dissatisfied to
5= very satisfied
satisfaction with
what the person is
doing
Are you satisfied or dissatisfied with
what you are currently doing?
Scale: 1= very dissatisfied to
5= very satisfied
141
B.3: Additional regression results
Table B-8: OLS regressions: results dividing respondents by occupational status
trajectory and reason for moving
Occupational status improved No change in occupational status
All
More
recent
Less recent All
More
recent
Less recent
work 0.236 0.236 0.238 0.332 0.423 0.247
migrants (2.04)** -1.46 (1.83)* (2.46)** -1.67 (1.88)*
non-work 0.125 0.155 0.078 0.191 0.224 0.178
migrants -1.12 -1.52 -0.47 (1.97)* -1.48 -1.56
married_fd 0.036 0.002 -0.006 -0.092 -0.122 -0.077
-0.44 -0.02 -0.06 -1.15 -1.23 -0.72
div/wid_fd -0.166 -0.271 -0.105 0.099 0.292 0.027
-0.65 -1.17 -0.27 -0.44 -1.37 -0.1
educ -0.043 0.068 -0.107 0.241 0.206 0.289
completion -0.58 -0.79 -1.47 (2.45)** (1.84)* (2.62)**
child birth 0.006 -0.024 0.078 -0.049 -0.087 -0.056
-0.08 -0.31 -0.73 -0.48 -0.64 -0.52
final educ 0.029 -0.001 0.004 0.002 0.041 -0.019
level -0.69 -0.02 -0.07 -0.06 -1.07 -0.51
Constant -0.018 0.11 0.227 -0.107 -0.228 -0.015
-0.08 -0.47 -0.93 -0.7 -1.25 -0.1
Obs 676 507 496 617 505 512
Occupational status deteriorated
All
More
recent
Less recent
work 0.068 0.479 -0.549
migrants -0.22 -1.05 -1.04
non-work 0.038 0.203 -0.089
migrants -0.23 -0.9 -0.39
married_fd 0.078 -0.127 0.123
-0.31 -0.41 -0.43
div/wid_fd -0.689 -0.582 -0.621
(2.53)** (1.89)* (2.35)**
educ -0.648 -0.716 -0.999
completion -1.42 -1.2 (2.09)**
child birth 0.308 0.396 0.309
-1.35 -1.59 -1.14
final educ 0.058 0.039 0.133
-0.66 -0.32 -1.71
Constant -0.42 -0.464 -0.867
level -0.78 -0.63 (1.96)*
142
Table B-8 (Continued)
Obs 141 115 115
t-statistics in parentheses, standard errors clustered at change in county level
* significant at 10%; ** significant at 5%; *** significant at 1%
Note: additional control variables for all regressions include cohort of birth and county of origin
Appendix C: Additional analysis for Chapter 4
C.1: Description of variables in the analysis
Table C-1: Description of all variables used in the analysis
Variable Question asked Response categories
Satisfaction variables
life
satisfaction
Answer to: "Are you satisfied or dissatisfied with
life in general right now?"
Scale: 1 = very dissatisfied, to
5 = very satisfied
financial
satisfaction
Answer to: "Are you satisfied or dissatisfied with
economic situation in general right now?"
Scale: 1 = very dissatisfied, to
5 = very satisfied
sat. with
occupation
Answer to: "Are you satisfied or dissatisfied with
what you are currently doing?"
Scale: 1 = very dissatisfied, to
5 = very satisfied
sat. with
housing
Answer to: "Are you satisfied or dissatisfied with
your housing situation?"
Scale: 1 = very dissatisfied, to
5 = very satisfied
sat. with
partner
Answer to: "Are you satisfied or dissatisfied with
your relationship with your partner?"
Scale: 1 = very dissatisfied, to
5 = very satisfied
sat. with
mother
Answer to: "Are you satisfied or dissatisfied with
your relationship with your mother?"
Scale: 1 = very dissatisfied, to
5 = very satisfied
satisfaction
with father
Answer to: "Are you satisfied or dissatisfied with
your relationship with your father?"
Scale: 1 = very dissatisfied, to
5 = very satisfied
Transition variables and variables used for their construction
school-to-
work
transition
Dummy variable taking the value the 1 if the
respondent achieved his or her highest level of
education in between 1999 and 2003 (2003 and
2009)
0 - other
1 - completed school-to-work
transition in 99-03 (03-09)
partnership
formation
Dummy variable taking the value 1 if the
respondent changed partnership categories from
being single living alone in 1999 (2003) to being
in a marriage or cohabitation in 2003 (2009)
0 - other
1 - experienced partnership
formation in 99-03 (03-09)
143
Table C-1 (Continued)
parenting
transition
Dummy variable taking the value 1 if the
respondent had the first biological or adoptive
child born in 1999-2003 (2003-2009) (see
below for parent status); and 0 otherwise
0 - other
1 - completed parenting
transition in 99-03 (03-09)
partnership
dissolution
Dummy variable taking the value 1 if the
respondent changed partnership categories from
being in a marriage or cohabitation in 1999 (or
2003) to being single living alone, divorced or
widowed (including divorced/widowed
cohabiting) in 2003 (or 2009)
0 - other
1 - experienced partnership
dissolution in 99-03 (03-09)
marital
status
Marital status from Swedish register which
classifies people as single, married, divorced or
widowed, combined with self-reported
information on cohabiting obtained from each
survey year.
1. single living alone;
2. single cohabiting ;
3. married; 4. divorced or
widowed living alone;
5. divorced or widowed
cohabiting
education
level
Education level from the Swedish register data
compulsory 9 years;
secondary <3 years;
secondary 3 years; post-
secondary <3 years; post-
secondary >=3
years/postgraduate
parent
status
1999 Survey:
Q36a_1-Q36c_1: Year of birth, biological child
1-3; Q36a_4-Q36c_4: Year of birth, other child
1-3
2003 Survey:
Q37a_1-Q37d_1: Year of birth, child 1-4 living
in household; Q37a_4-Q37d_4: Child 1-4 is:
respondent's and partner's child, respondent's
but not partner's child, partner's but not
respondent's child, adoptive child, foster-child
2009 Survey:
F20a1_ar_IP-F20a5_ar_IP: Year of birth,
biological or adoptive child 1-5; F20d1_IP-
F20d5_IP: Does the child live with you?
Based on the answers to the
YAPS questions
respondents were classified
in the following parent
categories:
1. Non-parents (no children
born by 2009)
2. Parents in 2009 (first child
born in 2003-2009)
3. Parents in 2003 and 2009
(first child born in period
1999 to 2003, more children
born in period 2003 to 2009)
4. Parents in 2003 not 2009
(first child born in 1999-
2003, no children born in
2003-2009)
cohort Register data for year person was born 1968, 1972, or 1976
gender Register data for gender of person surveyed male or female
work
income
Register information on "income from work
before tax" for the years 1998, 2002, and 2008
(in thousands of SEK)
Real thousands of SEK,
adjusted for inflation in
2002 and 2008 using the CPI
index from Sweden Statistics
144
Table C-1 (Continued)
occupation
Classification constructed from two questions
asked in the YAPS survey:
1 - What is your main occupation? What are
your main tasks at work?
2 - What is your current main activity?
Occupation categories used in
the study are divided into
following groups:
1) Student; 2) Unemployed;
3) Inactive (including
military service, parental
leave, housekeeping and
those participating in an
active labor market program
such as adult learning);
4) Goods production; 5)
Service production; 6)
Assistant non-manual
7) Intermediate non-manual;
8)Farmer/self-employed non-
professional;
9) Professional/higher
manual/executive
partner's
occupation
What is your partner's occupation at the
moment?
1) Permanent work;
2) Casual work; 3) Own
business; 4) Studies; 5) Adult
learning;
6) Employment measures;
7) Unemployed >6months;
8) Unemployed <6months;
9) Parental leave;
10) Housekeeping;
11) Military service;
12) Other
child's age
Constructed based on the YAPS questions about
the year in which each child was born. The
child's age was classified as follows:
In 2003: 1 year or less: If child was born in
2003 or 2002, 2 years or more: if child was born
in 2001 or earlier;
In 2009: 1 year or less: If child was born in
2009 or 2008, 2 years or more: if child was born
in 2007 or earlier
Child age categories used: 1
year or less, 2 years or more
parental
leave
Dummy variable taking the value 1 if the
respondent reported his or her main activity to
be "parental leave" during the time of the survey
0 - other
1 - person currently on
parental leave
145
Table C-1 (Continued)
respondent's
age
Age based on the register data for birth cohort
and on the year survey was conducted
Age assigned as follows for
each birth cohort: 1968
cohort: 30 in 1999, 34 in
2003, 40 in 2009; 1972
cohort: 26 in 1999, 30 in
2003, 36 in 2009; 1976
cohort: 22 in 1999, 26 in
2003, 30 in 2009
C.2: Additional statistical analysis and regressions
Table C-2: Life satisfaction changes associated with marriage and cohabitation
Age 22-26 Age 26-(30/32)
Partnership
formed: N
LS age
22
LS age
26
Change
in LS N
LS age
26
LS age
30/32
Change
in LS
Cohab 53 3.79 3.91 0.11 69 3.64 4.01 0.38
Marriage 8 4.13 3.75 -0.38 20 4.15 4.20 0.05
Age 30-(34/36) Age 34-40
Partnership
formed: N
LS age
30
LS age
34/36
Change
in LS N
LS age
34
LS age
40
Change
in LS
Cohab 46 3.41 3.87 0.46 16 3.69 4.38 0.69
Marriage 11 3.82 4.00 0.18 6 4.00 4.17 0.17
Table C-3: OLS regressions: variables of interest (in changes) on future attrition
Life sat Sat econ
Sat
house
Sat
occup
Sat
partner
Sat
mother
Sat
father
Attrition -0.05 0.043 -0.065 0.067 -0.057 0.038 0.032
09 (1.06) (0.75) (1.07) (1.06) (0.97) (0.93) (0.66)
Constant 0.023 0.122 0.103 -0.003 -0.048 -0.115 -0.147
-0.83 (3.70)** (2.94)** -0.08 -1.4 (4.82)** (5.23)**
Obs 2049 2052 2046 2004 1414 1983 1857
R-sq 0 0 0 0 0 0 0
Absolute value of t statistics in parentheses
+ significant at 10%; * significant at 5%; ** significant at 1%
146
Table C-4: OLS regressions: specification including time trends and cohort
dummies
Including time trend Including cohort dummies
1 2 3 4
Age 22-26 0.038 -0.04 0.038 -0.041
(0.82) (1.2) (0.82) (1.2)
Age 26-30/32 0.049 0.013 0.046 0.01
(1.21) (0.48) (1) (0.29)
Age 30-34/36 -0.036 -0.003 -0.032 0.004
(0.83) (0.13) (0.39) (0.1)
Age 34-40 -0.041 0.034 -0.016 0.066
(0.59) (2.19) (0.15) (1.61)
Partnership formation 0.292 0.292
(4.24)* (4.24)*
School-to-work -0.023 -0.022
(0.44) (0.43)
Parenting (child 1yr-) 0.245 0.246
(8.56)** (8.74)**
Parenting (child 2yrs+) -0.131 -0.13
(2.61)+ (2.56)+
Partnership diss (with child) -0.343 -0.344
(2.46)+ (2.48)+
Partnership diss (no child) -0.099 -0.097
(1.4) (1.35)
No trans (major completed) -0.135 -0.135
(3.29)* (3.29)*
2003-2009 time trend 0.008 0.01
(0.16) (0.48)
1976 birth cohort 0.014 0.015
(0.22) (1.86)
1968 birth cohort -0.018 -0.022
(0.19) (1.59)
Observations 2650 2650 2650 2650
R-squared 0 0.03 0 0.03
Absolute value of t statistics in parentheses;
+ significant at 10%; * significant at 5%; ** significant at 1%
147
Table C-5: OLS regressions: specification dividing those going through school-to-
work transition by occupation type after education completion
All age
intervals
pooled
Age
interval
22-26
Age
interval
26-30/32
Age
interval
30-34/36
Age
interval
34-40
Partnership formation 0.298 0.128 0.28 0.468 0.418
(5.66)** (1.23) (3.47)** (4.42)** (2.03)*
School-to-work -0.11 -0.304 0.062 -0.357 no obs.
(low level occup) (0.82) (1.23) (0.34) (1.07)
School-to-work 0.008 0.188 0.15 -0.217 -0.349
(med level occup) (0.08) (1.07) (0.91) (1.08) (1.25)
School-to-work -0.023 -0.135 0.04 0.011 0.078
(high level occup) (0.38) (1.22) (0.45) (0.09) (0.29)
Parenting (child 1yr-) 0.255 0.336 0.27 0.173 0.285
(3.56)** (1.93)+ (2.69)** (1.3) (0.74)
Parenting (child 2yrs+) -0.132 -0.231 -0.095 -0.093 -0.584
(2.16)* (1.11) (1.09) (0.9) (2.54)*
Partnership diss (with child) -0.342 -0.58 -0.074 -0.295 -0.702
(3.24)** (0.99) (0.39) (1.83)+ (2.86)**
Partnership diss (no child) -0.1 -0.304 -0.121 0.015 0.062
(0.91) (1.6) (0.59) (0.07) (0.15)
No trans (major completed) -0.133 0.152 -0.067 -0.172 -0.199
(2.34)* (0.59) (0.64) (1.94)+ (1.3)
Age 22-26 -0.041 0.035
(0.77) (0.48)
Age 26-30/32 0.014 -0.029
(0.32) (0.51)
Age 30-34/36 -0.001 -0.008
(0.02) (0.12)
Age 34-40 0.044 0.131
(0.68) (0.96)
Observations 2637 476 935 845 381
R-squared 0.03 0.03 0.03 0.05 0.07
Absolute value of t statistics in parentheses;
+ significant at 10%; * significant at 5%; ** significant at 1%
148
Table C-6: OLS regressions: specification dividing those going through school-to-
work transition by level of final education (postsecondary, or secondary and less)
All age
intervals
pooled
Age
interval
22-26
Age
interval
26-30/32
Age
interval
30-34/36
Age
interval
34-40
Partnership formation 0.293 0.14 0.274 0.478 0.33
(5.60)** (1.35) (3.42)** (4.53)** (1.62)
School-to-work (sec or less) 0.144 0.128 0.245 0.102 -0.125
(1.33) (0.58) (1.47) (0.45) (0.38)
School-to-work (postsec) -0.055 -0.118 0.038 -0.101 -0.204
(1.02) (1.17) (0.47) (0.89) (0.89)
Parenting (child 1yr-) 0.249 0.279 0.272 0.173 0.296
(3.54)** (1.68)+ (2.78)** (1.3) (0.76)
Parenting (child 2yrs+) -0.127 -0.238 -0.091 -0.078 -0.579
(2.09)* (1.14) (1.05) (0.75) (2.50)*
Partnership diss (with child) -0.345 -0.583 -0.078 -0.295 -0.751
(3.27)** (1) (0.41) (1.84)+ (3.06)**
Partnership diss (no child) -0.102 -0.252 -0.131 0.023 0.008
(0.93) (1.34) (0.65) (0.11) (0.02)
No trans (major completed) -0.132 0.153 -0.068 -0.165 -0.232
(2.33)* (0.59) (0.65) (1.87)+ (1.51)
Age 22-26 -0.038 0.034
(0.71) (0.47)
Age 26-30/32 0.016 -0.029
(0.37) (0.51)
Age 30-34/36 -0.0002 -0.015
(0) (0.21)
Age 34-40 0.04 0.164
(0.61) (1.2)
Observations 2650 477 940 848 385
R-squared 0.03 0.03 0.03 0.05 0.06
Absolute value of t statistics in parentheses;
+ significant at 10%; * significant at 5%; ** significant at 1%
C.3: Further details on age interval construction
Given the focus on changes in life satisfaction and life transitions over time, the first step
of the analysis is to identify the age intervals for which these changes are described. This
149
is done using observations for young adults from three different cohorts at three points in
time and pooling respondents from different cohorts observed at the same (or similar) age
(Figure C-1). Since a first difference analysis is used throughout the study, the main
criterion used to pool the cohorts in constructing the age intervals is that observations
have to be available both at the beginning and end of each age period. Given this criterion
four age intervals are constructed: age 22-26 (1976 cohort in 99 and 03), age 26-30/32
(1976 cohort in 03 and 09 and 1972 cohort in 99 and 03), age 30-34/36 (1972 cohort in
03 and 09, and 1968 cohort in 99 and 03), and age 34-40 (1968 cohort in 03 and 09). In
Figure C-1, these age intervals correspond to arrows numbered one (22-26), two (26-
30/32), three (30-34/36), and four (34-40), respectively. Notice that while the 1968 cohort
is observed at age 30, it is not included in the age interval 26-30/32, because observations
at the beginning of this period are not available. Because the young adults were surveyed
at three point in time, each respondent is included in two consecutive age intervals,
covering periods 99 through 03, and 03 through 09 respectively. This implies that the
respondents born in 1976, for example, are included first in the 22 through 26 age span
(covering 99-03), and second in the 26 through 30/32 period (covering 03-09) (Figure C-
1).
150
Figure C-1: Method of pooling of cohorts for age interval creation
Abstract (if available)
Abstract
The present work provides new knowledge on the relation between various life circumstances and quality of life, as measured by life satisfaction. Chapter 2 analyzes rural‐urban life satisfaction differences in 14 Latin American countries using Latinobarometer survey data for four years. The countries divide into two groups: those where rural people are happier than urban people, and those where they are less happy. In a multivariate analysis higher public social spending is found to increase rural relative to urban life satisfaction through its positive influence on access to services such as health. Chapter 3 examines life satisfaction changes following internal migration for young adults in Sweden, using panel data for years 1999 and 2009. Life satisfaction increases following an internal move, but the persistence of this increase depends on the reason for moving. Migrants who move for work‐related reasons experience a long term increase in life satisfaction. Those who move for reasons other than work (such as housing), experience a medium term, but no long term, well‐being improvement. Occupational status improvements are the main reason behind the long lasting increase in life satisfaction of work migrants. Chapter 4 explores the life satisfaction path during young adulthood, again using the panel data for Sweden. Life satisfaction increases slightly until the age of 30 and then turns downward. This pattern is largely explained by major transitions in life circumstances. Before 30 partnership formation, the school‐to‐work transition, and having babies all tend to increase life satisfaction. After 30 life satisfaction declines as children grow older and the breakup of unions becomes more prevalent.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
What leads to a happy life? Subjective well-being in Alaska, China, and Australia
PDF
Four essays on how policy, the labor market, and age relate to subjective well-being
PDF
Essays on the economics of subjective well-being in transition countries
PDF
Essays on health and well-being
PDF
Three essays on social policy: institutional development, and subjective well-being as a cause and consequence of labor market outcomes
PDF
Adaptation, assets, and aspiration. Three essays on the economics of subjective well-being
PDF
Essays on the new economic history of the Middle East
PDF
Beyond revealed preferences: how gender, relative socioeconomic status, and social norms drive happiness and behavior
PDF
Essays on education: from Indonesia to Los Angeles, CA
PDF
Selection and impacts of early life events on later life outcomes
PDF
Essays in international economics
PDF
Later life success of former college student-athletes as a function of retirement from sport and participant characteristics
PDF
Essays on health and aging with focus on the spillover of human capital
PDF
Essays on health economics
PDF
Feeling good and living well: on the nature of pleasure and its role in well-being
PDF
Three essays in health economics
PDF
Essays on development and health economics
PDF
Essays in panel data analysis
PDF
Essays on behavioral decision making and perceptions: the role of information, aspirations and reference groups on persuasion, risk seeking and life satisfaction
PDF
Essays on education and institutions in developing countries
Asset Metadata
Creator
Switek, Malgorzata A.
(author)
Core Title
Explaining well-being: essays on the socio-economic factors accompanying life satisfaction
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Economics
Publication Date
04/02/2014
Defense Date
03/10/2014
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
life satisfaction,OAI-PMH Harvest,well-being
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Easterlin, Richard A. (
committee chair
), Bechara, Antoine (
committee member
), Nugent, Jeffrey B. (
committee member
)
Creator Email
mac.switek@gmail.com,switek@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-372250
Unique identifier
UC11296664
Identifier
etd-SwitekMalg-2316.pdf (filename),usctheses-c3-372250 (legacy record id)
Legacy Identifier
etd-SwitekMalg-2316.pdf
Dmrecord
372250
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Switek, Malgorzata A.
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
life satisfaction
well-being