C.uOOL - World Database of Happiness

World Database of Happiness
'C.uOOL
2..ÖI
"2
author's code
\2 \.{. cl
author's
ti tIe
Cc-
1./
I (:) lè-,
k
kJ
r'
source
f
4.
f { v,
c) "'"
4:
r
k
'k " v-.e cc
U
o
excerpted
by
o
excerpt checked
by
o
excerpt send to
author
dd
excerpt returned by
author
dd
final con trol
by
o
reprint in Databook files
o
book/report available at
'-t.
c' U
0
'-"<.
-f'
Jl.
e v S ( ' /; ')
'101 '2
ISBNIISSN:
ISSN E:
PDF: ja/neen
o
') '" (
IJ.. Ij! \{' " VL.e 7 C;
URLIDOI:
o
S
Can a Woman's Income Buy Marital Happiness?
Life Events, Adaptation, and Gender in South Korea
By Robert Rudo lf'
March 20 12
a
Division of Intern ational Stud ies, Korea Uni versity
Anam-dong, Seongbuk-gu, Seoul 136-7
outh
Tel. +82 -2-3290-24 1 , rrudo lf@ko rea.ac.kr
ca
A bstract:
Us in g detail ed longi tudi nal data from the Korean Labor and Income Study (K LlPS) from 1998 to
2008, thi s paper ana lyzes gender-spec ifi c impacts as weil as adaptation to major life and labor market
events in Korea. The results show substa nti al gender-specific di ffe rences pat1i cul arly regardin g the
impacts of events related to marital status change. Husba nds remain on a hi gher happiness level
th roughout marriage whi le women return to their base line happiness with in only two years . Men suffer
more fro m divorce and widowhood. The fo und gender happiness gap is eq ui va lent to a (husband-only)
increase of annual per-capita household income of approx im ate ly US$ 17,800 during marriage.
Women who contri bute relati vely more to total househo ld income are able to close thi s gap. Moreover,
gender roles matter: a non-com pl iance with socially prescribed roles in the form of a too hi gh share of
wives' earni ngs leads to dissatisfact ion and thus restricts women in thei r life choices.
JEL: 0 13, 112, J71.
Keywords: Subj ective we ll-being; Adaptation; Intra-house hold barga ining; Gender ea rnings gap;
Gender ident ity.
1. Introduction
Recent work in subj ecti ve well-being has shown that individuals' perception of happi ness
tends to adapt to most life and labor market events (e.g. Jonathan Gardner and Andrew
Oswald, 2006; StutzeI' and Frey, 2006; Andrew Clark et al. , 2008; Andrew Clark and Ymmi s
George lli s, 20 10; Paul Frijters, David Johnston, and Mi chael Shields, 2008; Andrew Oswald
and Nattavudh Powdthavee, 2008). These papers argue in favor of the baseline hypolhesis,
that individllals are equi pped with a certain fo rm of baseline happiness whi ch can main ly be
explained by quas i-fix fac tors such as genes, early childhood educati on, and persistent
psychological traits. Only a sm all share of an indi vidual's happiness is attri buted to
demographic and socio-economic factors and these rather short-term variabl es are assumed to
indllce fl uctllation arollnd the base line level. I The argument is that changes in most
demographic and soc io-economic variables have no long-run effect on an indi vidual's leve l of
happiness or life satisfaction, and so indi viduals rather experi ence both anticipation of and
adaptation to most life events. Thi s has recently changed our understanding of the
intertemporal dimension of we ll-being and requires flIrther research. !-Iow do shocks, positi ve
or negati ve, effect incli vidllals over time? Do people fully recllperate after a particul ar shock,
or is there a lasting shift in happiness levels? And if so, which are the fac tors that do not only
temporaril y, but permanently shift an indi viclual's we ll-being? Do differences exist in the
impact of, and adaptati on to shocks bet ween women anclmen?
Looking at major li fe events li ke marriage, or labor market events Iike unempl oyment calls
fo r an in-depth study of gender differences in li fe satisfact ion responses over time .
I Lykke n and Te ll ege n ( 1996) show th at dcmog rap hic and soc io-economie factors acco unt fo r o nly a smal! pan
ofthe va ri ance in subjective well-be ing measures. Th ey estim ate th e contribut ion of the Slab je component of
subject ive we ll-being, which they ascribe to gcncs and pers istent psychological lrails, to exp lai n as much as 80
percent.
2
Particularly in countries with rather traditional gender roles, both the magnitude of a certain
event 's impact on individual well-being as weil as adaptation to it might differ significantly
between women and men. While some related studi es have found th at women are more likely
to adapt to unemp loy ment than men , to the best of our know ledge no study has ever found
significant gender differences for a number of events yet. Thi s might be due to the spatiall y
selecti ve nature of happiness research in the past. Most studies in thi s fi e ld concentrate on
"Western" OECD countries with comparatively low leve ls of gender inequality. In their study
of Gernlan panel data (GSOEP), C lark et a l. (2008) estimate intertempora l changes in life
sati sfaction due to marriage, divorce, widowhood, birth of child, unemployment and layoff
separately for the two sex es. T hey find full or partial adaptation 10 most events and concl ude
that the anticipation and adaptation patterns due to life events are " remarkably similar
between men and women". In a follow-up paper, Clark and Georgelli s (2010) confirnl these
findings for British panel data (BH PS) and conclude that adaptation might be a "genera I
phenomenon". Frijters, Johnston and Shields (2008) use quarterly life event data from the
Household, Income and Labour Dynamics in Australia (HILDA) survey fo r the study of ten
events, in which they ma inl y show that life events are not randomly di stributed and thus stress
lhe importance of fixed-effects estimation.
In this study, we will perform a s imilar analysis, yet for a country that has maintained
relatively strong traditional gender roles. Within such a society , preferential access to labor
markets and income mi ght have important impacts on intra-marriage and work pl ace utility
outcomes. Intra-marriage bargaining might be skewed in the sense th at women's financial
dependence might lead to a gender gap in deci sion-making and eventual utili ty outcomes. The
recent ri se of subj ective well-being measures in econom ics all ows us to di rectly measure
perceived we ll-being of indi vi duals.
3
For our analysis, we use eleven waves of the Korean Labor and Income Panel Study (KLIPS)
from 1998 to 2008 to study gender-specific impacts of events as wei l as their related
anticipation and adaptation patterns. We wil! estimate the impact of the fo l!owing six major
life and labor market events on individual happiness: marriage, divorce, widowhood,
unemployment, first job entry, and working day reduction/introdllction of the five-day
working week. Determinants of potential!y arising gender differences wi l! then be analyzed in
detail.
Psychology literature suggests that the wife's absolute and relative earnings play important
roles for intra-marriage wel!-being. Higher female employment and income can have positive
effects for married women if other roles such as being a housewife are less satisfy ing (Ross,
Mirowsky and Go ldsteen, 1990). For men, the effect mainly depend s on prevailing gender
roles and the level of econom ic hardship. In case of more liberal ro les or more severe
economie hardship, men wil! appreciate higher contributions of their wives in broadening the
hOllsehold 's total economie resources. In contrast, if gender roles are rather traditional and the
fami ly does not necessarily depend on the wife 's earnings, then a higher share of female
earnings might give the impression that the man is unable to fulfil! the breadwinner role and
this can have negative effects on the husband 's wel!-being (Arl ie Hochschild , 1989; Rogers
and DeBoer, 200 I). Higher earnings of wives improve fema le bargain ing in the household,
and Ihus can lead 10 changes in the househo ld divi sion of labor, and in spousa l roles more
general!y. Thi s can have beneficial effects for the cOllple if more liberal roles are prevalent in
society. However, it maya lso increase the risk of fema le-sided marital breaklIp.
There are two related theoretical threads of literature in the field of economics. First, there is
an establi shed li terature on intra-household bargaining and deci sion-making (e.g. Lawrence
Haddad, John I-Ioddinott and Harold Alderman, 1997). These studies basicaJl y argue against
4
mcome poo ling withi n a household as proposed by Gary Becker ( 198 1) in his a ltruistic
di ctator model and rather cla im that it is important for intra-househo ld deci s ion-making who
earns the money. It has been shown in several studies that the hi gher the relative share of
income earned by a woman (alternati vely the share of assets brought into marriage), the
higher her relati ve bargaining power (e.g. Phipps and Burton, 1998; Phili p Brown, 2009; Lise
and Seitz, 20 11) . Related to hi gher female income is an increase in the credibility of her
potential di vorce threat, eguipping a woman with a viabie alternative in case of an
un satisfy ing marriage. Women who are fin anciall y more dependent mi ght not have thi s
alternati ve at hand, since ex pected income after di vorce will be rather low. In the context of
happiness modeis, we are the fi rst to propose to measure the relati ve barga ining pos ition of
women by their relative income contribution. A secOlld related thread of literature is rather
new and starts with George AkerIof and Rachel Kranton (2000) who introduce the concept of
"identity" into economic mode Is of behavior. They argue that own utility does not only
depend on own actions, but also on others' acti ons, own identity, and the degree of
accordance of own actions with others acti ons and societal presc riptions fo r our identity. Thus,
if in a society with traditional gender ro les men are expected to be mai n breadwinners, those
who do not compl y with th is prescripti on may suffer negati ve utility effects. We will include
thi s concept in oW" model mainly by relating wives ' contri buti on to total household income to
the same vari abl e's regiona l average .
We contribute to the literature in the following way . First, we are the first to study welfare
impacts of maj or li fe and labor market events as we il as their re lated intertemporal effects in
the context of hi gh gender ineguali ty . We will show how gender discrimination in the
household and in the labor market aftècts gender-spec ifi c utili ty outcomes. Second, by
exp laini ng gender differe nces in happiness during marriage we are able to test and confirm
theoretical pred ictions deri ved from intra-household bargaining and gender identi ty models
5
while using a direct measure of well-being. Third, we wi ll expand the happiness literature to
include East Asia. Onl y few studies have recentl y started to analyze subj ecti ve we ll-being
questions in the East Asian context (e.g. Kang, 20 10; Rudo lf aIld e ho, 20 11 ) and happiness
economics is still very much focused on high-income " Western" economies. Fourth, we will
make use of a uni que pane l data set and state-of-the art fixed-effects estimation techniques.
Our results prov ide a number of new and interesting findings. First, evidence for partial and
full adaptation to most events is confinned. However, resul ts suggest larger absolute initial
im pacts for men and a somewhat faster and more complete adj ustment for women. Second,
results indicate strong gender-specific impacts of vari ous events, particul arly of changes in
marital status: Whil e marriage has astro ng and lasting positi ve effect on ma le happiness, the
average fe male happiness gain is limited to not more than two years. Third , men suffe r more
than women from di vorce and widowhood, while women show no negati ve effects in the case
of their spouse's death. Fourth, the intra-marriage gender happiness gap can be explained with
women' s low contribution to total household income and with soci al restrictions leadi ng to
disutility for couples in which wives earn more than the socia lly to lerated level, providing
evidence fo r both intra-house hold bargaining and the gender identity hypothesis. Fifth,
concerning labor market events, our fi ndings also point towards a gender-segregated working
cu lture.
The rema inder of the paper is organized as fo llows: Secti on 2 discusses gender inequali ty in
Korea in more detail. Secti on 3 introduces the data set and methodology used fo r the analysis.
The methodologica l part fi rst introduces the adaptation mode l and then the intra-marriage
gender happiness model. Secti on 4 then di sc usses regression results. Whil e the fiTst two
subsecti ons cover marri age-related and labor market events, the third subsection presents the
6
determinants of the intra-marriage gender happiness gap. Finally, the intra-marriage gender
gap is analyzed in greater detail in the last subsection. Section 5 concludes.
2. Gender Gap in Korea
The Republic of Korea presents a very interesting case of economic and socio-cultural change.
The Korean economy has shown spectacular growth rates over the past fi fty years. Related
rapid socio-econom ic development all owed the country to join the group of hi gh-income
2
OECD countries in 1996. At the begimling of the 2 1SI century, Korea is an established
member of the rich world. However, traditi onal gender roles show them se lves to be much
more resistant to change than political or eco nomic variables. Until today, Korea has
maintained strong trad itional values, particularly when socia l and fami ly relations are
concerned 3 In 20 10, Korea had the third lowest fema le labor force participation rate (25 -54
years of age) among the 34 OECD countri es. Lower rates were only observed in Turkey and
Mexico. In 2005, the UNDP dedicated an entire report to the issue. The " Korean Human
Deve lopment Report on Gender" points out that "women's participation in political and
economic sectors, especially in deci sion-ranking positi ons is very low in Korea, despite
marked growth in the country 's economy over the past decades". While Korean wome n, when
compared internationally, are high ly educated, they still face a lllllnber of obstacles preventing
them fro m engaging in the labor market in a simil ar manner as men do. The most prominent
2
In term s ofGD P per cap ita in purchas ing power parity units, Korea is expected ta reac h the leve l of Germany
Wilhin the next decade.
3
For an introduction
10
Korean gender relati ons, see e.g. Kim (2007) or Renate Clasen and You - Kyoung Maan
(20 10).
7
of these are strong gender gaps in earnings, very long working hours, the absence of part-time
work options outside the low-skilled service sector, and defici ts in ch ild care supply.
Table I presents the gender gap in Korea in a cross-country comparison. Korea ranks
particularly low in indicators that strictly focus on the gender gap measured in terms of
female-to-ma le ratios of education, health, economie and politica l empowerment (Gender
Empowerment Measure (GEM) and Global Gender Gap Index (GGG». The presumably most
comprehensive attempt to measure the gender gap in a soc iety is the GGG index whi ch has
been published by the World Econom ic Forum since 2006. Acco rding to this index, Korea
ranks only 104,h out of 134 countries in 20 10, suggesting very hi gh leve ls of gender inequality
in Korea. Korea's low rank is not at last due to low female -to-male ratios in the following
sub-indexes: " wage ineq uality for similar work", "professional and technical workers",
" Iegislators, semor offic ials, and managers", "tertiary enrolment ratio" and " seats in
parliament" . Traditional gender roles in Korea have been further acknowledges in a recent
OECD "Society at a Glance" report (20 11 ) comparing male and female shares in housework
across OECD countries. While accordi ng to the OECD country average, women do 2 .13 times
the housework that men do, in Korea th is ratio amounts to 5.05 . For comparison , this ratio is
1.48 in Norway, 2 .75 in Spain, and 3.24 in Turkey. Moreover, comparing Korea to Singapore,
another Asian growth miracJe, further suggests that high gender inequality is not inevitable in
4
countries with Chinese cultural heritage. While Korea and Singapore take up simil ar ranks in
the human development index (HOI ), Korea ranks much lower in gender gap measures.
Rudolfand e ho (201 1) shows fo r marri ed Korean coup les that even ifa woman is doing most
of the market work, she still takes care of about 70 percent of the housewo rk . The same articJe
, Chinese cu ltura l understanding of a proper fema le life is main ly innuenced by Confuc ian and Daoist tho ugh!.
Ba th phil osophies suggest gender segregation afwork, ye t Confucian isl11 stresses even stronger thc
subordi nati on o fwi ves to their husband (Josep h Ad ler. 2006).
8
also argues that very long wo rking hours and the absence of high- ski lied part-time work pose
important restrictions on furth er femal e engagement in the labor market. Lee ( 1998) argues
th at many Korean girls only stri ve for higher education in order to increase the likelihood to
find a we ll-educated husband. This might be a rational strategy under the present
circumstances, given that marital sorting is relatively high in Korea (Lee, 2008). It might thus
be a good sign th at gender roles are slowly starting to change in Korea and so does marital
labor-sharing. Young generations of husband s al ready show sli ghtl y higher engagement in
housework and labor force participation of women in their prime motherhood years has risen
from 55.8 to 62.3 percent between 1998 and 20 10 (Rudolfand Cho, 20 11 ; OECD, 2011). Vet,
to sum up, it becomes evident that Korea, in contrast to its economic and human development
achievements, is stilllacking behind substantiall y in tenTIS of gender equality.
[Tab Ie I about here]
3. Data and Methodology
3.1 Data
Data for our analys is comes from the Korean Labor and Income Panel Study (KLIPS) for the
years 1998 to 2008. KLIPS is a nationally representative longitudina l study .of urban Korean
.
households, modeled after the US National Long itudinal Surveys (NLS) and Panel Study of
Income Dynamics (PSID). It is conducted annllally by the Korea Labor Institute, a
government-sponsored research institute. The study star1ed in 1998 with 5,000 hOllseholds
and 13,783 individllal s aged 15 years or older. KLIPS collects a wide range ofinformation on
9
indi viduals, such as earnings, fami ly, education, employment backgrounds, and demographic
characteristics. In addition, it offers broad in format ion on various ind icators of life and job
satisfaction. The data quality KLIPS provides satisfies highest international standards. The
panel maintains 76.5% of the original sample throughout all waves, which is comparable to
lhe US PSID (78%) ; the German Socio-economie Panel (GSOEP, 79%); and the British
Household Panel Survey (BI-IPS 77%). Kang (20 I 0) shows lhat potenlial bias produced by
altrition is negligible in KLIPS data.
We restriet our sample to the age group of 16 to 60 year olds which yields a tolal number of
55,447 person-year observations for fema\es and 57,574 for mal es. For the analysis of
widowhood, we extend the upper age limit to 80 years, resulting in a sample of 66,592
person-year observat ions for fema\es anel 65,986 for males. The panel is unbalanced in lhat
not all individuals are present in all waves. Thus, our minimal requiremenl is lhal an
individual was observed at least once before and after the event, thus exc\uding lefl-censored
spe lis.
The central variab ie for our analysis IS overall life satisfaction. In KLIPS ' individual
queslionnaire, the question on overall life satisfaction is precedeel by a sel of delailed
questions on the satisfaction with different aspecls of life: household income, leisure life,
housing envirorunent, family relalions, relalions with relatives, and social relations . The exact
word ing of the overalllife satisfaction question is then: "Overall,
110W
salis(ied Ol' dissalisfir;d
are you wilh your lire?" Individuals are asked to respond according to a scale ranging from I
("very satisfied") to 5 ("very dissatisfied"). For the sake of eas ier interpretation, we recoded
the scale so lh at higher numbers cOITespOild 10 higher levels of salisfaction.
Table 2 shows the distribulion of life satisfaction for the sample of 16-60 year olds separately
for men and women. About half of all women and men report to be " neither satisfied nor
c>-çLv.-I\.Ar~~ JV (sjo.
dissatisfied ". More people report to be "satisfied" than "dissati sfied". It might be a cultural
particularity th at Koreans tend to avo id the extreme categories "very satisfied" or "very
dissatisfied". This contrasts with studi es on e.g. Germany or Great Britain, where usually
around 10 percent of the sampl e chooses the highest category of satisfaction.5 Compared to
their male counterparts, fe males in Korea report about the same if not a slightly higher
average life satisfaction.
[Tabie 2 about here]
In order to exam ine the intertemjloral change in satisfaction levels due to major life and labor
market events and potential gender differences that might arise, we wi ll analyze the following
six major events: marriage, divorce, widowhood, unemployment, first j ob entry, and
introduction of the five-day working week. Since we wou ld like to avo id potential bias
through habituation to events, we only take into considerati on the first event of its kind for
each individual during the sample period. Th us, observati ons are right-censored in case of e.g.
second unemployment spe lIs or remarriages.
For changes in marital status and first job entry, the questionnaire explicitly asks for !he exact
year and month of the change or entry. This allows us to calculate for each person-year
observation the years passed since the event, or the years from a particular wave unlil the
event if it has yet to occur. 6 COl11pared
10
simpl y look ing at mari tal status changes between
two interviews, thi s has the advantage to be able to identi fy "quick remarriages", where
' Com pare C lark et al. (2008); Clark a nd George ll is (20 I 0). 11 is ral her un likely lhal
50
rew people are "very
sat isfi ed" in Korea, as comparcd 10 We stern European co untries. In stcad , we be li eve thaI thi s has to do wit h
cu ltural-specifïc behavior: a modest and hum bie lI se of language is often required
by Ko rean soc ial norm s.
Espec ialty whcn talkin g to an unkno wll interviewe r ahollt yaur perso nal happiness, Koreans might be incl ined to
respond in a more reserved way.
A few observations only report ed the year ofthe event, but not the exact month. Here, the rniddl e of th e year is
used to calculate the tim e elapsed since th e event (e.g. 1998 +0.5 years). Addit ionally, we check for th e person's
marita1 statu s at the time ofthe even t and the year preceding it.
6
11
individuals become divorced or widowed and then remarry within only two survey waves.
Moreover, we can better identify the exact date of change in case there is a gap of two or
more years between two interviews. We then coded lead and lag dummies for each year
reaching from "3 to 4 years before the event" to "5 or more years after the event". For
example, the latter dummy in the case of the event "marriage" would take the va lue I if at
least five years have passed since a person's first marriage and if she has not had a change in
marital status in the meantime. It would take the value 0 otherwise.
For unemployment we use a simi lar methodology as Clark et al. (2008) who use the
occupational status of individual i in each period, 1= 1 to T, to calcu late if and for how long an
individual currently is and has been unemployed. And if currently working but unemployed in
the future, it is calculated how far she is away from future unemployment. In 2004, following
a change in national legislation, the official six-day working week was replaced by a five-day
working week. To model thi s empirically, we observe the date of the first change from above
tive to five working days for each individual. We demand for an indi vidual to have worked on
average at least 5.5 days a week during two years, and to then have adopted and maintained a
five-day wo rking week (4.5:'ó ave rage weekly working days <5.5) fo r at least another two
periods.
Table 3 shows ave rage satisfaction of leads and lags byevent and sex. We can see that for
al most all events there was a considerable upward or downward change in life satisfaction
after the event took place. Most yet not all effects are as expected. In the case of marriage we
see the expected increase in ave rage satisfaction by 0.35 for fe ma les and 0.34 for males in the
year following the wedding as compared to the year preceding the wedding. Looking only at
descriptive data, both females and males manage to stay on a higher average leve l of li fe
satisfaction even after more than five years. While women after two years then seem to
12
partially return to their baseline leve l of happiness, it loo ks as ij' men are do slightly better in
remaining on a hi gher level after marriage. In the case of di vorce, both women and men are
already relatively unhappy before the event. Relative to their baseline level, women
experi ence a brief decline and a fast recovery whil e men seem to experience a long-Iasting
negatÎ\/e effect of being and stay ing divorced. Particularly su rpri sing are the gender-specifi c
results for widowhood . Fo r this on average e lder age cohort, we would expect gender roles to
be most traditional (compare Rudolf and Cho, 20 11 ). Men are happier than women by about
0.3 units in the years preceding the event. After becoming widowed, women's life sati sfaction
increases whil e that of men suffers a brief decline, but then seems to recover rather fast. Thus,
we do not observe seriou s negative effects of widowhood, and if so, onl y for men.
[Tabie 3 about here]
Next, we are interested in labor market events. Considering unemployment effects, it seems
that there is some anticipation of unemployment in the preceding year both for women and
men . When getting unemployed, average life satisfaction of women drops by 0.15 and that of
men by 0.27 as compared to one to two years before the event. Women seem to quickly adapt
to unemployment, return ing to the old level when being unempl oyed for one or more years.
Men, on the other hand , recover only partially eve n after two or more consecutive years of
unemployment. 7 The next two co lumns of Table 3 refer to the time around a graduate's first
job entry. In the case of females , we see a gradual increase in life satisfaction after leaving
school, college or university and entering their tirst job. Growi ng experience, a bette I' stand ing
in lhe company, and hi gher income might be potential explanations. The increase observed
7
The extremely low leve ls ofunemp loymcnt in general and long-term unemployment in particular found in Dur
data restriet the analys is of the effect of long-term uncmploymel1l on life sat israction.
13
for mal es is comparable; however, their initia I happiness increase in the year of starting the
first job is stronger. Thi s mi ght be due to hi gher performance expectati ons and related social
pressure on men as future main breadwinners. 8
The last two co lumns of Table 3 show the effect of the red ucti on in working days on
satisfaction with working hours. We see that while there is a genera I upward trend in hours
sat isfaction due to gradual work ing hours reductions that took place in Korea during the ti me
of our study, there is a particularly strong effect when work ing days are reduced from above
five to fi ve days a week. 9
3.2 Anticipation and Adaptation Model
In order to capture the intertemporal effects of the discussed events in a multivariate
regression framework, we estimate the fo ll owing empirical model. We use life and hours
satis tàction as our main response variab les. As Ada FeITer-i-Carbonnel and Paul Frijters
(2004) point out, assum ing cardinality or ordi nality of the satisfaction measure does produce
very similar results. In order to be able to bet1er interpret the magn itude of the impact of, and
adaptation to the events fo ll owed in thi s study, and to assign monetary values, we wi ll use
linear fi xed-effects estimation as our main technique and use fixed-effect s ordered logit
estimators fo r consistency checks. Controlling for fixed-effects is essential in sati sfaction
models si nce unobserved personality traits are li ke ly to be correlated with certain decisions
that appear on the right hand side of the equation, such as employment and marital status
(C lark, 2003; Stutzer and Frey, 2006).
Satisfaction Sof ind ividual i in period { is modeled as fo llows:
8
Whi le young men are supposed 10 have a secure job and income before being able 10 marry. wamen often have
give up th eir career ror family duties after marriage (Lee et al. , 2008).
10
9
For more informat ion on work ing haurs reduction sec Rudolf and e ha (20 1 I ).
14
Sit = LEitfJ
+ Xit Y + Ui + 1] t + fit,
i = 1, .. . , N t = 1, ... , T
( 1)
where LEit is a vector contain ing a set of binary lead and lag variables to control for the
intertemporal effect of a certai n life or labor market evenl.
Xit is a vector of standard control vari ables in satisfacti on mode Is, including individual and
household demographi c and socio-econom ic variables. 1o Ui is indi vidua l i's fi xed effect, 1]t
control s for year effects, and fit is an i.i .d. error term. Since it is very like ly that ECu"Xit ) *0
and ECu" LElt) *0, we should estimate this modeillsing a fi xed-effects estimator in order to
yield consistent estimates of the model parameters fJ and y. In order to yie ld ge nder-specific
estimates we will run the regressions separately for men and women.
Wh ile the basic model is estimated by linear fixed-effects, for consi stency checks we will
then also use lhe FF-estimator (Ferrer-i- Carbonell and Frijters, 2004) and the BUC-estimator
(G regori BaetschmanI1, Kevin Stallb, and Rainer Winkelm3lm, 201 1) which eSlimate fi xedeffects in the presence of an ordinal dependent variabie. The FF-estimator was recently lIsed
by Ali son BOO lh and Jan va n Ours (2008; 2009) for their study of working hours and life
sati sfaction in Germany and Australia. We willllse the " mean vers ion" of the estimator in our
analys is. Very recently, Baetschmann , Staub and Winkelm3l1l1 (2011) proposed the BUCest imator and they show in Monte-Car1o simlliations that it performs best among a set of
fixed -effects ordered logit esti mators.
3.3 Intra-Marri age Gender Happiness Model
10
For furthe r information on all va riab les used in th is paper sec a detailed description in Table A I.
15
In order to examine c10ser the dri vers of a potential gender-specific impact of marriage, we
will test both the predi ctions of the intra-household bargaining theory and o f Akerlof and
Kranton 's gender identity hypothesis in one model. According to models of intra-household
bargaining, women are able to increase their re lati ve bargai ning power within the household
as the ir share in total earnings rises (Haddad , Hoddinott, and A lderman, 1997). Related to
higher fe male income is a more credible di vorce threat, leav ing a woman with a viabie
alternati ve in case of an llnsati sfy ing marriage.
Thus we would expect the utility difference, Udif/j , between husband and wife ofcouplej to
. h the WI·ti·
. tota I coup Ie earnmgs:
.
decrease Wit
e' s s I1are Tj m
aUd i ffj
arj
< 0.
Another related hypothesis is that of Akerlof and Kranton (2000) who introduce the
psychological-sociologica l concept of " identity" into an economic model of behavior. In their
model, individual i' s identity or self image, I;, depends on own actions, a ;, other' s actions,
a_;, the ass igned gender, sexi , own given characteri stics, Ei, and the social ideal of the
assigned gender (gender prescription), P.
Individual we ll-being, U;, then depends on own actions, a;, olher's actions, a_i , and own
identity .
Compl ying wilh expected gender behavior P is rewarded , while non-compliance can have a
negative effect on own utility.
Thi s leads us 10 the following regression model where we regress indi vidllal sati sfaction of
husband and wife, as weil as the Slim of the two and their difference in separate regressions on
own occupation, spouse's occupation, and a vector of bargaining variables, BARGit , and a
16
vector of soc ial prescriptions, PRESCRit . As proxies for relati ve bargaining in the household,
we include the share of women in total coup Ie earnings, as weil as the hllsband-wife age and
education (years of schoo ling) differences. From the theoretical model , we wOll ld ex peet
women to increase thei r utility with increases in relative bargaining. Social prescriptions enter
the model throllgh ave rage regional per-capita hOllseho ld income and the ratio of the
hOllsehold 's share in tota l cOllple earnings to this variable's regional average . Thus, following
Akerlof and Kranton (2000) , we acco unt for the fact that indi vidllals compare them selves to
surrounding peer groups. We add a number of standard household demographic and soc ioeconomie controls, as weil as cross-partner variables controlling for the partner's occupation.
We wOll ld expect soc ial prescriptions to reward men relatively more when these are mai n
breadwinners and have a higher share of earnings." In an attempt to di sentangle the effects of
intra-household bargaining and gender identity on well -being, we estimate the fo llowing
model :
Si'
= OWNOCCit {3 + SPOUSEOCCi,y + BARGi,o + PRESCR it 1/! + Xi,À + Ui + 17, + Ei'
(3)
Where Sit is life satisfaction, ownoCCit and sp ouseoCCit are own and spouse's occupation for
partner i in period I , Xit is a vector of further control s, Ui are indi vidllal fi xed effects, 17, are
year dummies, and Ei' is an idiosyncratic error term. The model in equation (3) will be
estimated separately for wives, hu sbands, for their joint cOllple happiness, and the husbandwife happiness difference lIsing linear fixed-effects estimation.
4. Regression Results
11
Further research should also include the erfect ofassets brought into marriage onthe gender happiness
differential. Traditional practice in Korea requests the groom's parCllt S la prov ide the house for the ncw couplet
while th e bridc's parcilts should pro vide lhe necessary house equipment. As th e reJative price afhollsing has
been constantly on th e rise, thi s lIsually entails a higher share orasscts brought inta marriage by the groom.
17
Table 4 shows the results of the linear fi xed-effects regressions byevent and sex. Anticipation
only plays a minor ro le in most events. It is only sign ificant in the case of both male and
fe male unemployment and female transition to the five-day working week. Conceming
impact and adaptation, our data prodllces interesting findin gs. Most events show substantial
gender-specific impacts on li fe sati sfaction. The esti mates indicate fu ll adaptation to some
events, particularly for women. However, to others we see no or only partial habituation.
Table 5 presents determinants of the observed intra-marriage gender happiness gap.
4.1 Marriage-Status-Related Events
Co lumns ( I) to (6) of Table 4 show the effects of diftèrent changes in mari tal status on life
satisfacli on. Fema les experience a slrong positi ve and signifi cant effect in lhe firsl year after
marriage. Yet, th is positive effect halves one year laler and di sappears completely after !wo
years fro m marriage, shi fti ng women back to their base line level of happiness. [n comparison,
males also experience the strongesl positive effect in the year of marriage and this effect more
than halves in the secOlld year. For them, ho wever, the effect regains in strength three years
after marriage and is then susta ined in lhe long run. The resltlts sllggest that men mi ght
benefit more fro m marri age than wo men. Full results for the event "marriage" inclllding all
covariates can be seen in Table A2 in the appendix . [-lere, the main findings are confirmed for
the two ordered logit estimators (FF and BUC). The results provide addili onal evidence fo r
the fact that life events are not randomly di stribliled (Stutzer and Frey , 2006). Com paring lead
effecls of sim ple ordered logil estimales in col umns ( 1) and (5) with the fixed -effects
est imates, il can be seen that among the sti ll unmarried individuals of the same age cohort
those that eventuall y get married are already happier several years be/are marri age. Thus,
18
those with happier personality traits are more li kely to marry .12 While Ihis eonfinns onee
more the importanee to con trol for fixed effects to avoid potential se lection bias, the
comparison of different estimators also indicates that the choiee of the fixed-effects esti mator
does not change the mai n results.
[TabIe 4 about here]
Bearing in mind the additional marital benefits for men, results for divorce and widowhood
can be berter understood. In general, reslilts suggest that men suffer much more from marital
dissolution. When getting divorced, women experience a short negative effect in the first year
following di vo rce but qui ckly manage to recover.
IJ
Men experience a twice as strong negative
initial effect, and when stay ing divorced never return to their maritallevel of happiness withi n
the observed time fra me. Results on widowhood also di ffer between men and women. While
widowed men show a transient negative effect and ruil adaptation thereafter, widowed women
on ave rage show no negative effect after their spouses pass away. They experi ence, if
anything, small positive effects between the second and the fourth year after the event.
14
These results are particu larly interesting when comparing them to the findings of Clark et al.
(2008) and Clark and George lli s (20 10) lor German and British wo men and men. In their
analyses of more gender-equal societies, they do not find signi ficant gender differences in
12
Fu ll results arc only displayed ror the event "marriage" here because of space limitations. They
C3 n
be
obtained fo r all other events from th e au th ors on request. Exam ining potential se lect ioll effecls ror th e case of
divorce, we find th at unhappier female persona lity trai ts are more likely to get divorced, a finding th at does not
ho ld ror men in our data.
13 Note that jfthe woman has been a hOllsew ife throllghollt marriage, she is entitled 10 receive 30 percent ofthe
wea lth accumu lated during marriage \Vhen divorce takes place within thc first 10 yea rs, and 40 -50 percent
thereafter. Wea lth that was already fornled before marriage goes to thc partner who brought it into marriage. ln
the case of female widowhood, she is treated with a factor of 1.5, and each child of I . Thu s th e woman receives
about 43 percent of total inheritance in the case of one daughter and one son.
14 Of course, the laner results of widowhood focus on the eldest generation and cannot be generalized for
yo unge r Olarriages. Gender inequality is tikel y la be highest in the e ldest generation (Rudolf and eho, 20 11 ).
19
adaptation to marriage-re lated events. Both women and men adjust to the positive (negative)
effects of marriage (widowhood) rather quickly and return to their baseline happiness within a
similar time horizon as th at of our paper. On ly in the long run (five or more years) do they
find significant negative effects for women. In the case of divorce, German and British
women and men show strong negative effects before getting di vorced and then increase their
happiness gradually aftel'. Here also, no specific gender effects were established. Therefore,
Korean results do significantly differ fro m what has been found so far in the li terature.
In a next step, we want to visualize the magnitude of the gen der happiness gap. Thus, we
assign monetary values to the effect, as has been done, for example, by Oswald and
Powdthavee (2008) in the case of calculating compensati on payments for disabled people.
When looking at column (8) of Table A2 in the appendix, we see that the average long-run
happiness sh ift for men is at around 0.15. This conservative calcu lation is almost double the
effect of living in one's own house (.083), and it is equi valent to approximately a 300 percent
increase in per-capita household income (assuming that we could raise only male per-capita
household income). Since the mean of log annual household per-capita income in our sample
is equivalent to 6.08 million Korean Won (KR W) in real 2005 value, an increase of 300
percent is equivalent to an increase of about 18.2 Million KR W. Applying an average yearly
exchange rate of 1,024 KR W/$US in 2005, thi s is equivalent to the happiness effect
(exclusively for the husband) of raising yearl y per-capita household income by approximately
US$ 17,800.
4.2 Labor Market Events
Co lumns (7) to (12) ofTable 4 show estimation results for effects of unemployment, first job
entry, and the introd uction ofthe five-day working week on happiness. Unemployment and its
20
long-run effects can on ly partiall y be analyzed with the help of Korean data. The dynamic
nature of Korea's labor market and the virtual absence of public unemployment insuraJlce
both result in a very low number of long-term unemployed . Thus we had to reduce the
number of lag categories in our model. Still, the results suggest partial adj ustment for women
al ready after more than one year of unemployment. Men seem to not adjust over the limited
time horizon. Their negati ve satisfaction response does not diminish in size after two or more
years of continuous unemployment, it rather increases sli ghtly. Thi s is in line with Soong-Nang
Jang et al. (2009) who show that Korean men suffer much more from depressive symptoms
than their female cou nterparts when unemployed, early retired or out-of-Iabor-force.
Entering one' s first j ob does not show significant effects on happiness, on ce we control for
income effects. If anything, one might notice that men have rather positive coefficients and
women on ly negative ones following firstjob entry . In the long-run we can observe a negative
effect for women who stay in their first job.
15
Finall y, we look at what happens when indi viduals move from a six- to seven-day to a fiveday working week. Effects on li fe satisfacti on are in genera l rather weak. The estimates
suggest slightly positive effects fo r men in the years after the reduction in working days.
Women do not show significant effects after the reduction . Table A3 in the appendi x
exam ines this sati sfaction response in more detail. It reports the results of the same model;
on ly now, results are add itionall y estimated with hours and job satisfaction as dependent
variables. 16 The resu lts show that while women experience no effect on either of the three
satisfaction measures, men show positive responses in both their sati sfaction with working
15
Direct gender discrirninat ion in the Korean labor market not only takes the fcrm ofsignifïcant gen der wage
differentials but al50 manifests itselfthrough unequal promotion chances (Lee et al.. 2008).
16 Thc job satisfaction question is "Overall, /lOW satisjied or clissCirisfied are yOl/wir" you,. mailljob? ", wh ile
hours sat isfaction is derived from the answer given on the aspecl"Working hours" following the question "HolV
satisfied or dissatisjied are yau wi/h regcwd [0 YO II,. mail1job on {he Jol/owing CIspeels? ". Both variables use th e
same sca le as life satisfaction, i.e. from I ("very dissatisfied") to 5 ("very satisfied") in our ana lys is.
21
hours and their overall job satisfaction. The laner indicators show positive and significant
effects starting one to two years before the actua l reduction ; these then last in the long-run.
Thus, we cmmot reject non-adaptation to the observed working days reduction for men. 17 lt
seems that men, who, on ave rage, show higher labor force participation and work more days
and hours than women, benefited more from the introduction of one additional leisure day per
week.
To sum up, the findings on labor market events show that men are both harder hit by
unemployment and benefit more from a reduction in working days.
4.3 Determinants of Intra-Marriage Gender Happiness Gap
The finding of a significant intra-marriage gender happiness gap in the preceding subsections
leads us to the analysis of the factors that detem1ine this gap. The results of estimating
equation (3) that present empirical tests for the two hypotheses of intra-househo ld bargaining
and gender identity are shown in Table 5. As expected from bargaining mode Is, a higher
wife's share in eamings sign ificantly reduces the husband-wife happiness gap (colunm (1)).18
Women seem to be able 10 improve their relative bargaining position when improving their
relative income position. This effect works through an increase in the wife's individual life
satisfaction on ly (compare co lumn (3) and (4)). The COllple'S total life satisfaction (column
(2)) thus increases, however thi s effect is not statistica ll y signi ficant in our model. Conceming
the other two bargaining variables, women seem 10 be happier when having an o lder partner.
Differences in ed llcation between the partners seem 10 have no significant effect. Thus,
17
Note th at the result ofnon-ad ap tati on does not change after introducing working ha urs co ntrol s. Thu s, the
mere shifting afwork ing tim e from six or seven days te onl y five days a week appears ta have lasting positive
effects.
18 Note that the female share in marri ed co up Ie earnings in our sa mple rose from 1998 (0.18) to 1999 (0.22), but
then stagnated until 2008 (0.23).
22
predictions suggested by intra-household bargaining mode Is are confirmed by our results:
Women are able to significantly increase their re/arive utility outcome in the household wh en
contributing more to household income.
[Tabie 5 about here)
In order to test the predictions of Akerlof and Kranton (2000) , our ma in variabi e of interest is
"Wife 's share in earnings (hh/ regiona l average)" . This variabie measures the ratio of a
household 's share of earnings by the wife relative to thi s variabl e's average value in the
province of residence. We can interpret the average share of wives ' earnings in total
household earnings in the province of residence as the social ideal or social prescription that
households face. Moreover, we can assume that wives contTibut ing less than the socia lly
prescribed level does not lead to di sutility, since women being housewives still represent a
large share of the society and are th us fully accepted (Rudolf and eho, 20 11 ). However, when
wives' (husbands') income shares become too large (smalI) relative to social prescriptions,
both wife and husband might experience negative effects. Resllits in Table 5 confirm these
hypotheses. We see that a large earni ngs shares of wives relative to the provincial average
decrease both husbands' and wives' litè satistàction , and th us also joint cOllple sati sfaction. 19
If hllsbands do not comply with their soc ia lly ass igned ro le as the main breadwinner, then
both husband and wife experience a negative impact on their perceived well-being.
19
The coefficient shows the effects "at lhe mean", Since the mean ofthe variab ie "Wife's share in earnings
(hh/regiona l average)" is ene, i.e. the leve l where household and reg iona l ave rage shares ofwives earnings are
equal, the imerpretation ofthe coeffic ient is straightforward.
23
Thus, while femal e earnings shares provide an important explanati on for the intra-household
gen der happiness gap, social presc riptions pl ay an important ro le in restricting women from
freely opting for a career.
5. Conclusion
The main ai m of thi s paper was to provide evidence for the base line hypothesis for a soc iety
with relati ve ly strong traditional gende r roles. Korea presents an ideal case study, since
compared to its economie and human deve lopment achievements, the country still ranks very
low in ternlS of gender equi ty . Our analys is revea led that the inter-temporal nature of maj or
life and labor market events matters. Whil e partial and full adaptation can be observed for
most events, anticipation onl y pl ays a mi nor ro le. We find important gender-related
differences in both impact of, and adaptation to major events. On the one hand , all events
have a lower abso lute initial impact on women 's happiness. On the olher hand , wo men show
relative ly faste r adaptation and thus return faster to their base line level of happiness.
In parti cular the events related to changes in marital status provide evidence fo r a significant
gender gap during marriage in Korea. While marri age has astrong and lasting positi ve effect
on male happiness, female happiness ga ins are on ave rage limited to not more than the first
two years after marriage . We find that the long-run male happ iness shift due to marriage is
equi valent to twice the effect of li vi ng in one' s own house or to an increase of yearly percapita household income by approx imately US$ 17,800. Furthermore, men suffer relati vely
more than women from divorce anel wielowhood. Female widows do not show any negative
effects after their partner passes away . If anything, rather sma ll positi ve effe cts can be
observed. Examining potential determinants of the intra-marriage gender happiness gap
reveals that low female earnings shares prov ide an important explanation for the happiness
24
gap. Financial dependenee lowers relative female bargaining power in the household , which
then has impacts on the intra-marital happiness distribution. ThllS, when women contribute
relatively more to total hOllsehold income, they are able to increase their relati ve well -being.
We are also able to contirm predictions of the gender identity hypothesis: a non-compliance
with societal role prescriptions is shown to lower both hllsband 's and wife's well -being.
With respect to unemployment , our finding s sllggest partial adaptation for women but no
adaptation for men to unemployment in the short-run. The reduction of working days and the
move towards a five-day work ing week shows positive effects onl y for men. Thus, labor
market events also point towards the importanee of gendered ro les. Men's role as the main
breadwinner might explain stronger male respon ses to labor market events.
Although gender roles are changing in Korea today, the fact that the married women's share
in couple earnings has stagnated between 1999 (0 .22) and 2008 (0.23) indicates th at past
reforms have not led to significant improvements for married women yet. As traditional
gender roles and a highl y male-dominated labor market continue to be major obstacles to the
pursuit of gender equity in happiness in Korea, refonns in this area sholild be continued and
reqllire critica I evaluation. FlIture reforms sholild ensllre equal pay and equal promotion
chances for women, as weil as the creation of family -friendly job opportunities, particul arly in
high-ski lied employmenl. In order to beller separate gender-specific impacts of the event from
gender-spec ific adaptation patterns, fUIther research is needed.
25
References
Ad ler, Joseph. 2006. Daughter/ Wife/Mother or Sage/lmm ortal/Bodhisattva? Wom en in the Teaching
of Chinese Rel igions. ASIANetwork Excha nge, 14(2).
Akerlof, George A. and Rachel E. Kranton. 2000. Econom ics and identity. Quarterly Journalof
Eeonomies, 115, 7 15-53 .
Beeker, Gary S. 1981. A Treatise on the Fami ly. Cambridge, Mass.: I-Iarvard University Press.
Baetsehmann, Gregori, Kevin E. Staub, and Rainer Winke lmann. 2011 . Consistent estimation of the
fi xed effects ordered logitmode l. University of Zurieh Working Paper Series.
Booth, Alison L., and Jan C. van Ours. 2008. Job Sati sfaetion and Family Happiness: The Part-time
Work Puzzle. Economic Journal, 118, 77-99.
Booth, Ali son L. , and Jan C. van Ours. 2009. I-Iours of Work and Gender Identity: Does Part-time
Work Make the Fam ily Happier? Economiea, 76, 176-196.
Brown, Phil ip H. 2009. Dowry and Intrahouseho ld Barga ining. Evidenee from China. The Journalof
Human Resources, 44( 1),25-46 .
Clark, Andrew E. 2003. Unemployment and social norms: psychologieal ev idenee from panel data.
Journal of Labor Economies, 21,323- 51.
Clark, And rew E., Ed Diener, Yannis Georgelli s, and Riehard E. Lucas . 2008. Lags and Leads in Life
Satisfaction: A Test of the Base line I-Iypothesis. The Eeonom ic Journal, 118(June), F222 -F243.
26
Cl ark, Andrew E. and Yannis GeorgelJ is. 20 10. Back to Baseline in Britain: Adaptat ion in the BHPS.
Clasen, Renate and You-Kyoung Moon. 20 IO. Fam ily Status, Working EXJlerience, And Korean
Women 's Attitude to Feminism. Korean Studi es Forum , 4, 245-266.
Ferrer-i-Carbonel l, Ada and Paul Frijters. 2004 . How Important is Methodo logy for the Estimates of
the Detenninants of Happi ness? Econom ie Journal, I 14,64 I-659.
Frijters, Paul , David W. Johnsto n, and Michael A. Shields. 2008. Happiness Dynamics with Quanerly
Life Eve nt Data. IZA DP No. 3604. (fonhcoming in Sca nd inavian Journa l of Economics)
Gardner, Jonathan and Andrew J. Oswald. 2006. Do divorcing couples become happ ier by break ing up?
Journalof the Roya l Stati stica I Society A, 169(2), 3 I 9-336.
Haddad, Lawrence, John Hoddinot!, and Harold Alderman (Eds .). 1997. Intrahousehold Resource
AlJocat ion in Developing Countries: Modeis, Methods and Policy. Baltim ore: Johns Hopkins
University Press.
Hochsc hild , Arl ie R. 1989. The second shi ft. New Vork : Viking.
Jang, Soong-Nang, Sung-J1 Cho, Jiyeun Chang, Kachung Boo, Hyun-Goo Shin, Hyejung Lee, and
Lisa F. Berkman. 2009. Employment status and depressive symptoms in Koreans: results from a
baseline survey of the Ko rean longitudin al study of ag ing. Journalof Gerontology: Social Sc iences,
64B(5), 677-683.
27
/
Kang, Sllng-Jin. 20 IO. The Analys is on the Detenninants of Life Satisfaction in Korea. Kyongjehak
Yongu, 58(1), 5-36. (DODO 00000 DODO DO, in Korean)
Kim , Sellng-Kyung. 2007. Family, Gender, and Sexual Inequality. In: Kim, H.-R. and Song, B. 2007:
Modern Korean Soc iety: its development and prospect. 131 -1 57. U California, Berke ley: Insti tllte of
East Asian Studies.
Lee, Mijeong. 1998. Women's Education, Work , and Marriage in Korea: Women 's Li ves Under
Institlltional Conflict. Doctoral Thesis, Seoul: Seoul Nationa l University Press.
Lee, Soo-Hyung. 2008. Preferences and Choice Constraints in Marital SOI·ting: Evidence from Korea.
Stanford Institute for Economic Policy Reasearch Discussion Paper 07-042.
Lee, Bun Song, Soomyung Jang, and Jayanta Sarkar. 2008. Women's labor force partic ipati on and
marriage: The case of Korea. JOllrnal of Asian Economics, 19, 138- 154.
Lise, Jeremy and Shannon Se itz. 2011. Consumption Inequality and Intra-hollsehold Allocations.
Review of Economic Studies, 78, 328-355.
Lykken , David and Allke Tellegen. 1996. Happiness is a Stochastic Phenomenon. Psychological
Science, 7(3), 186-1 89.
OECD. 2011. Society at a Glance 2011: OECD Social lndicators , OECD Pllbli shing.
http://dx .doi.orgIl0.1787/soc_glance-201 I-en
OECD. 201 1. OECD Statistics on line (stats.oecd.org).
28
Oswald , And rew J. and Nattavudh Powdthavee. 2008. Ooes happiness adapt? A longitud inal study of
disability with im plicat ions for economists and judges. Journa l of Publ ic Economics, 92 , 1061 -1 077.
Phipps, She ll ey A. and Peter S. Burton. 1998. What's Mine is Yours? The In fluence of Ma le and
Female Incomes on Patte rns of Household Expendit ure. Econom ica, 65, 599-6 13.
Rogers, Stacy J. , OeBoer, Oanelle O. 200 1. Changes in Wives ' Income: Effects on Ma ri tal Happiness,
Psychological We ll- Being, and the Risk ofO ivorce. Journal of Marri age and Family, 63(2), 458-472.
Ross, Catherine E., John Mirowsky, and Karen Goldsteen. 1990. The im pact of fam ily on health: The
decade in review. Journal of Marriage and the Fami ly, 52, 1059- 1078.
Rudo lf, Robert and Seo-You ng Cho. 20 11. The Gender-Specific Effect of Working Hours on Fam ily
/
Happiness in South Korea. CRC-P EG Oiscuss ion Paper No. 77, University ofGöttingen.
StutzeI', Alois and Bruno Frey. 2006. Ooes marriage make people happy, or do happy people get
marri ed? The Journalof Socio-Economics, 35, 326-347.
UNOP. 2005. Korean Human Oevelopm ent Report on Gender.
UNOP. various years. Human Oevelopment Report.
World Economie Forum. 2010. The Global Gender Gap Report 20 10.
29
Tables and Figures
T a bl e I : G ender Ga E in Ko rea in C ross -Co unlr~ COI11Ea rison
No rwa~
Germany
Korea
S ~ail1
UK
Si n ga~ rc
Mexico
Tu rkc~'
la
12
20
26
27
56
83
35.308
295 18
29 ,66 1
35 ,087
48,893
13 ,97 1
13,359
Ranking by indicator
HDI ran k 20 10 (ouI of 169 counlri es)
GNI pcr capila (PPP 2008 $)
58,8 10
7
20
14
32
la
68
77
9
61
11
15
16
39
101
2
13
104
11
15
56
91
126
2 .67
9.7
6 1.7
12
20 .7
27 .3
66
101.3
Poplilation with at least secondary educ3 ti on (25 or older)
1.00
.98
.87
.94
1.01
.88
.9 1
.58
TCrliary cnroJlmcnt mte
1.62
I
.69
1.24
1.40
.98
.78
Labor force part icipati on mte
.94
.87
.73
.77
.84
.74
.55
.35
\Vagc equalily for si milar work
.75
.6 1
.52
.52
.67
.80
.54
.57
Lcgislators. senior orti cials, and managers
.46
.6 1
.11
.48
.53
A6
.44
.11
Professional and technical workers
1.06
1.01
.69
.98
.90
.82
.70
.54
Scats in parliament
.66
.49
.17
.58
.28
.3 1
.36
.10
'l ï me S ~CJ1t in houscwork
I A8
1.64
5.05
2 .75
1.82
3.3 1
3.24
Gil rank 2008 (of 169)
GEM rank 2009 (of 109)
GGG ra nk2010 (of 134)
Avcrage rank of thrcc gender indicators
5
Ratio fel/w ie 10 male (20 l a da{(l)
Sourees: \VEF 20 10: UNDP 2009, 20 10: OECD 20 11.
Table 2: Di stribution of Life Sali sfaclion by Sex
FeIlmies
Males
11 of obs
Percentage
11 of o bs
798
1.5
867
1.7
2 (dissalisfïcd )
6,28 1
11.9
6, 197
12.0
3 (ncilhcr satisfïcd no r dissa li sfï cd)
29,572
55.8
29.243
56.5
4 (salisfïcd)
15 ,928
30.1
15 ,070
29.1
426
.8
429
.8
TOlal
53 ,005
100
5 1,806
100
Mcan
3. 17
3. 15
S.o.
.70
.70
I (vcr)' dissali sfï cd)
5 (vcr)' sali sfïcd)
NOIC: Statistic!> calculatcd for thase agcd 16-60. Dala: KLiPS 1998-2008
Percentage
Table 3: Ave rage Life Sali sfacli o n o f Leads a nd Lags b~ Eve nl a nd Sex
M arri agc
Divorcc
Wi dowhood
Feuwles
Males
Fel1 /Gles
3-4 ycars hence
3.2 1 (314)
3.08 (289)
2.96 (69)
2.89 (72) 2.76 (180) 3.05 (39) 2.98 (1 19)
2-3 yea rs hence
3.21 (373)
3. 17(331)
2.94 (83)
3.04 (8 1) 2 .8 1 (187) 3.23 (53) 3.00 ( 153)
1-2 ycars hcncc
3.20 (427)
3. 16 (372)
2.85 (82)
W ithin the next ycar
3.2 1 (5 19)
Males
Fel1lOles
Une mployment
Males
Females
Males
First Job Entry
Females
5-Day Wo rking
Ma/es
F elIwies
Males
3.00(2 11 ) 3. 19(253)
3.08 ( 150)
3.3 1 (96)
3.4 1 (345)
2.98 (285)
3.22(3 18)
3. 13 (208)
2.93 (85) 2.84 (237) 3.02 (60) 2.97 (20 1) 2.99 (400)
3. 19 (462) 2.84 ( 103) 2.97 (99) 2.78 (255) 3.08 (62) 2.88 (283) 2.89 (589)
3.19 (374)
3.13 (249)
3.38 ( 11 2) 3.45 (399)
3.32 (157) 3.46 (493)
3.20 (461)
3. 12 (364)
3.36 ( 159)
3.49 (503)
3.27 (356)
3.44 (167)
3.54 (497)
Leads
Lags
0-1 years
3.56(3 16) 3.53 (34 1)
2.67 (72)
2.65 (66) 2.90 (236) 2.89 (62) 2.82 (285) 2.72 (60 1) 3.26 (457)
1-2 years
3.52 (396) 3.43 (383)
2.8 1 (70)
2.86 (73) 3.03 (230) 3.07 (46)
3.3 1 (238)
3.36 ( 166)
2-3 years
3.40 (385) 3.44 (337)
2.88 (66)
2.72 (69) 3.05 ( 193) 2.93 (45)
3.37( 138)
3.36 (122)
3-4 years
3.37 (342)
3.48 (297)
2.90 (58)
2.9 1 (54) 3.07 ( 163) 3. 14(36)
3.40( 12 1)
3.40 (97)
3.5 1 ( 164) 3.5 1 (502)
3.60 ( 111 ) 3.62 (342)
3.62 (76) 3.67 (240)
4-5 years
3.36 (3 10)
3.48 (275)
2.94 (48)
2.73 (40) 3.08 ( 130) 3. 11 (35)
3.45 (77)
3.43 (80)
3.57 (37)
3.65 (135)
5 or more yca rs
3.40 (790)
3.50 (6 13)
3.0 1 (73)
2.76 (68) 3.10 (262) 3.22 (55)
3.44( 108)
3.5 1 ( 105)
3.50 (38)
3.67 (85)
I or morc ycars
2 or morc
~'ears
2.69 (65)
3.00( 12)
2.87 (15)
Notcs: NUlllbcrs of ObSCfV:ltions arc calc ul atcd for indiv iduals agcd 16-60 (CXCCpl for widowhood 16-80) and di spla)'cd in bmckcts. Data: KLIPS 1998-2008.
T a ble 4: Effecl of Lire a nct Labor Markel Evenls o n Life Satisfac li o n
Widowhood
Divorcc
Marriagc
Uncmploymcnl
Firsl Job Enlry
5-D ay W orki ng
Fenw ies
Males
Fema les
Males
Females
Ma/es
FeJ1/Clles
Males
Fema les
Ma/es
Fell /a les
Males
( I)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
( 10)
( I I)
( 12)
3-4 ycars hcnc.:c
.054
-.077*
.111
-.046
.004
.05 1
-.002
.045
-.034
-.063
-.067
.020
2-3 ycars henec
.023
.0 19
.102
.0 16
.053
.159*
.007
-.005
-.0 10
-.015
-.04 1
.034
1-2 ycars hcncc
.007
.008
-.020
-.085
.058
-.067
-.027
-.0 11
-.038
-.028
-. 11 8**
.0 19
-.0 12
.028
-.057
-.068
-.064
-.022
-.096**
-.058**
-.024
-.006
-.092
.03 1
0- 1 years
.264***
.332***
-.286***
-.507***
-.003
-.264***
-. 190***
-.278***
-.050
.032
-.05 1
.060*
1-2 ycars
. 138***
.121 ***
-. 128
-.228***
.098*
-.136
-.249***
-.04 1
.05 1
-.024
.01 2
2-3 ycars
.025
.120***
-.082
-.375***
.11 3*
-.274*'
-.033
-.0 11
.039
.067*
3-4 ycars
-.007
.168***
-.1 66*
-.254***
.106*
-.1 05
-.039
.066
-.009
.069
-.056
.044
-.067
.06 1
-. 186***
.122
-.05 1
.005
Leads
Withi n the nex t year
Lags
4-5 ycars
-.004
.162***
-.052
-.3 26***
.092
-.115
5 or morc ycars
-.00 1
.197***
.0 14
- .29 1***
.070
-.024
-.103
I or more ycars
-.3 14*
2 or morc ~Ica rs
Observalions
45 ,049
43,953
50,0 14
48,799
59,282
56,639
49,953
46 ,284
42,633
4 1,527
11 ,046
24,923
Indi vid uals
5,923
6 ,087
6 ,798
6,946
7 ,609
7,637
6 ,823
6,6 19
6,703
6 ,836
2,156
3,902
NOIcs: U ncar fi xcd-cffccls cslimalor . Qlller contral \'ariablcs indude S-ycar-agc-cohorts. dununi es for houschold hcad :l.Ild spousc, ycars of school ing. houschold sizt.", num bcr of )'ou ng (0- 14 ycars) :lIld aid ( 1530 ycars ) chi ldren in the hOllschold, Illl rnbcr of aid fc nmlcs and malcs in thc houschold, dununies for marital st:ltus, log pcr-capita houschold incornc, log per-clpi ta regiomil incOIllC. hOuse owncrshi p, 14 rc"ional
dUTnmies, 10 year dummics. dummics conlrollin g for CUfrent occupati on st:ltus. *·*1·*'* ind icatc a parameter estirnalc is significan t al Ih c 1%15%/ 10% Icvel . Data : KLi PS 1998-2008.
(>
Table 5: Intra-Marriage Gender Happiness Model
Dep. Va ri. ble: Life s.tisfacti on -
Difference
Log per-capi ta hh income
(husb-wife)
Sum
(husb+wife) W ife
Husba nd
( I)
(2)
(3)
(4)
-.006
.103* *'
.055***
.048***
(.006)
(.008)
(.005)
( .005)
-.2 18**
.222
.2 18**
.004
(. 103)
(. 156)
(.094)
(.093)
-.036*
.04 1
.039**
.002
(.02 1)
(.032)
(.0 19)
(.01 9)
-.005
.002
.003
-.00 1
(.008)
( .0 13)
(.008)
(.008)
-.267***
-.160***
-.107*
( .096)
(.058)
(.057)
.0 12
-.078*'
- .044**
-.033*
(.020)
( .0:10)
(.01 8)
(.01 8)
.0 13
7 .133 ***
3.732***
3.402***
(.384)
(.849)
(.509)
(.507)
Obscrvati ons
27 ,5 54
27 ,554
27 ,554
27,5 54
Indil'id uals
4,976
4,976
4 ,976
4 .976
l nlra-couple bargaillillg
W ife's share in carnin gs
Age differencc
Schooling dirfe rence
Socia l comparisolJ
Log per-capita regional income
W ifc's share in earni ngs (hh /rcgional av cra ge)
Constant
Nolcs: Li ncar fixcd-cfrccts cstimation. Othcr control vnri ablcs incl udc )'cars of schooli ng of wife. own house, 10 occup:lIion
dumm ics husband. 10 occupation dummics wi fc. 5-year-agc-cohorls. 14 rcgional du mmic5 and 10 )'car dummics. Standard errors
in parentheses . ...
indieatc a parameter cslimale is significant al lhe 1%/5%/ 10% le\'el. Data: KLIPS 1998-2008 .
,**,*
Divorce female
MalTiage female
O,S
o,S
0,4
0,3
0,2
0,1
0,4
0,3
0,2
0,1
Widowhood female
o,S
0,4
0,3
0,2
0,1
°
°
°
·0,1
·0,2
·0,3
·0,4
-0, 1
-0,2
-0,3
-0,4
-4
-2
°
2
Marriage male
·2
°
2
4
O,S
o,S
0,4
0,3
0,2
0,1
1
1
~.
-o,S .....----.--...l!t.--~-_+~
-2
2
4
° and afler the event
No. of yea rs before
T
-4
Fig. 1: Adaptation to Life and Labor Market Events
-2
°
2
4
o,S ~
-0,1
-0,2
-0,3
·0,4
-o,S
l/T
Widowhood male
° t,~~--r-+------------
°
-0,1
-0, 2
-0,3
-0,4
-4
-4
Divorce male
0,4
0,3
0,2
0,1
~
1
T
·O,S
-4
4
ol
-0,1
-0,2
-0,3
-0,4
-o,S
-O,S
IJ
-2
2
4
° and after the event
No. of years bcfore
0,4
0,3
0,2
0,1
0t-~~~-i~r-~t-~~
-0,1
·0,2
-0,3
·0,4
·O,S
-4
-2
2
4
°
No. of ycars bcforc aud aftcr the cvent
(x ... signifi cant al the I % I cvcl ~ 11 ... significant at the 5% level; 0 . .. significant at the 10% level)
-Firstjob entry female
Unemployment female
0,5
0,4
0,3
0,2
0, 1
0,4
0,3
0,2
0, 1
0,5
0,4
0,3
0,2
0,1
"1
°-0,1
-
-0,2
-0,3
-0,4
-0,5
~
°t
-0,1 f
T
l ~~
<,
-0,3
T
T
1
1
l
l
.1
1
['..
1
-0,4
-0,5
-4
-2
2
°
4
-4
-t--...T
1
-0,1
°
-0,2
-0,3
-0,4
-0,5
-4
l'
0,5
0,4
0,3
0,2
0,1
°-
T
-0, 1 -"'...l
...,
4
2
°
No. of ycars bcfOl'c and aftcl' .he e vent
-2
-2
°
2
4
-0,2 ~
-0,3
-0,4
-0,5 -4
-
...
°-0,1 "
-0,2
-0,3
-0,4
-0,5 -4
-2
FiJ'stjob enh'y male
Unemployment male
0,5
0,4
0,3
0,2
0,1
S-day-working female
-
2
°
4
S-day-working male
o,S
T
T
T
1 1
1
-2
./
2
4
°
No. of ycars bcforc :lI1d ancr the cvcnt
Fig. 1: Ada ptali on 10 Life and L1bor Market Even ts (contd.)
0,4
0,3
0,2
0, 1
~ ~
°
~
T J ti.. _T_
J.
J.
-0,1
-0,2
-0,3
-0,4
-0,5
-4
.+. T
-2
2
4
No. of years before :lI1d ancr thc evcnt
°
(x ... significant at the 1% level; 6 ... signi fi can t at th e 5% level; 0 ... significant allhe 10% level)
Appendix
T abl e A I : D escription of Variables
V ari abie
CharaClcrist ic
Itu/i viduol charne/eris/ies
Lire sati sfaction
Overall li fe sati sfacti on; ordinal scale from I (very di ssati sfied) to 5 (very sa ti slicd)
Job sati sfac ti on
Overall sati sfaclion lI'ilh mai n job; ordinal scale from I (very di ssali sfïcd) 10 5 (very sal isfied)
Hours sati sfaction
Satisfacti on with workin g hours in the main job; ordinal seale from I (vcry dissati sficd) 105 (vcry sal isfied)
5- ycar-agc-cohorts
Dummy vmiable equal lo " I" if an individual belongs 10 a parlieular age cohort (20-24, 25-29 ,30-34, CIC.)
Schooling
Ycars of schooling
Hcad
Dummy variabie; H ead~ I if indi vidual is head of Ihe houschold
Spouse
Dumm y varia bic; Spousc= 1 if individual is SpOlISC of hOllschold head
Log of own ca rnin gs
Real monlhly carnin gs (in KRW I 0 ,000 of 2005); nalural log applied
Marital SIallIs
Never married
Indiv idual has never married
Marricd LEADS
e.g. 3-4 ycars hence: individual is observed 3 to 4 years beforc gelling married
Married LAGS
e.g . 1-2 ycars hencc: individual is observcd I to 2 years after gcttin g marri ed
O lher married
Rcmarri ed, marri ed beforc rïrst survcy in 1998, or not intervicwcd bcforc marri age
Sepcraled
Individual is still married , but scperated
Divorced
Individual is divoreed and not rel11arri ed
Widoll'cd
Individual is widowcd anel nOl rcmarried
OcclIparioll (maill activily eftlring last week)
Working
"Workcd" or "Tcmporaril y away rrom work"
HOllscwi fc
"Looked after famil y or home/chil d caring or preparation for marri age"AND "did nOl have ajob last wec k"
Retired
"Retired" or "aid age"
IIlness
"M enlal/physical illncss"
In cduca ti on
Attcndcd school, privalc academie institutes, or preparcd for higher cducation
Uncmployed
"Searchcd ror ajob" AND "was ablc to work (i r therc was a suitablc position) last weck"
Ol her oecupation
Rcferencc ca tcgory in Tabl e A2
Working hours
Avcrage actllal working hours per weck
Occupa ti on dllrnrnics
I I categori es ( I-di git Korcan Standared Classificalion of Occupati ons)
Indllslry dumrnies
17 categorics ( l -di git Korean Standared Industry Classilication)
Six-day workin g
Rcfcrcnce ca tcgory in T ablc 4 (columns 11 and 12) and Tablc A3. Dummy is cq ual to "I u if individual works
more 5.5 days or more in an avcrage wcek.
Othcr ti ve-day
Contral in Tablc 4 (columns II and 12) " nd T able A3 for those who work more than 4.5 and less than 5.5 days
a weck but that did not cxpcriencc a tran sition from six-day to five-day worki ng during the observed pcri ocJ.
Belall' li ve-day
A n individual thaL works Icss lhan 4.5 days in an avcrage weck.
HOl/sehold characteristics
HH size
Numbcr of indi viduals li ving in thc household
No of chi ld 0- 14
Number of chilelren of rcspcctivc couple aged
No of chi ld 15-30
Numbcr of chil dren of respcctivc coupl e aged 15 to 30 , cconomicall y depcndcnt anel li vi ng in th e hOllschold
No of aid fenmies
Number of females 65 years and older in the household
°to 14 and li ving in the houschold
No of old males
NlImber of males 65 years anel older in th e hOllschold
Age difference
A gc IlUsband minus age wifc
Schooling differenee
Y ears of schooling of hllsband minus years of schooling of wife
Wi fc's share in carnings
Wifc's earnin gs/(Wifc's earnings + Husband's carnings): in mOlllhl y term s
Log per-capita hh income/ Log hh pci
Real yearl y houschold income (in KRW I 0,000 of 2005) divided by houschold size; nat ura l log applied
Own house
Houschold is li ving in own house
Ilegiollal variables
Prov incia l el u111 111 i es
Dummies of residence for all CUITent 16 provinccs e"cept for the island of l ej u (not sampled)
Log per-capita reg inco l11el Log reg pei
Wifc's share in earnin gs (hh/ regional
avcrage)
Av erage per-capi ta hh income (sec above) by provincc; natllrall og applicd
Ollzer J1ariables
province of residencc
Yc.:'lr dlll11mics
Dummies for year of i nterview ( 1998 to 2008)
Wife's share in carn ings (sec above) in lhc houschold dividcd by wivcs' avcrage share in carnin gs in (he
Table A2: Effect of First Marriage o n Life Satisfaction
Males
F CI1/Clles
OIogit
BUC
FF
FE-OLS
OIogit
BUC
FE-OLS
(6)
FF
(7)
( I)
(2)
(3)
(4)
(5)
3-4 years hence
.296**
.189
.125
.054
2-3 years henee
.194*
.09 1
-.054
.023
-.042
-.290'
-.159
-.077*
.195*
.045
.072
.019
1-2 years henee
.160
.049
-.092
.007
.172
.003
.079
.008
Within nex t year
.170*
-.074
-. 180
-.0 12
.195 **
.1 15
.059
.028
0-1 years
1.10***
1.03***
.989***
.264* **
1.55 ***
1.33'**
1.34***
.332***
1-2 years
.573* **
.509**
.377**
.138***
.690***
.093
.039
.025
.703 ** *
.450***
.414'*
.121 ***
.091
.529* **
,497***
2-3 years
.684***
.168**'
(8)
TrallSilioll to mar
Leads
Lags
.120***
3-4 years
-.011
-.076
-. 194
-.007
.791 ***
.691 ***
4-5 years
-.022
-.032
-. 183
-.004
.707***
.679***
.599**'
.1 62***
.088
-.045
-.117
-.001
.800***
.83 1'**
.724'**
.197***
.195**
-.688**
-.915 ***
-.186***
.503***
.125
.059
.037
- 1.52***
-1 .04**'
-.7 19**
-,403 ***
~ .644***
-.899**
-.722**
-.227 ***
-.232'* *
~ .5 1 3***
-.862***
-.749*'*
~.219 ***
-.174**
-.099
-.890**
-1 .30'**
-.234**
5 or more years
Otlz mar slat
Other married
Seperated
-.804***
Oi vo rced
- ,48 1***
Widowed
.006
-1 .3 9* **
-.880***
-.605 *
Ycars schooling
.122**'
.048 **
.046** *
.011 ***
.129***
.022
.032*
.004
Workin g
03 00***
.330***
.334***
.090***
.5 18** *
.502***
.454***
.132***
HOllsewifc
.541 ***
.352* **
.324***
.097***
-.048
.070
.056
-.00 1
.629***
-.122
.556** *
-.516***
.486***
-.019
.5 32***
-.259*'
A05 ***
.538 ** *
-.287 **
.13 \ ***
-.0 11
.140* **
-.093 ***
.167 ***
.163***
llldh'idual
Retired
IIlness
In educati on
Unemployed
HOl/ schold
Log hh pci
.349***
-.019
.240**
.038
-.004
.0 13
~.507 ***
~ .407* **
- .488***
~ . 1 28 ***
.600***
.29 1***
.258 ***
.082***
~.688 ***
-.489***
~ .482***
-.155 ***
.049***
.3 16***
.162 ***
.164***
.048** *
Log reg pci
-.783***
-.686**'
- .783 ***
-.202 ***
~ .55 2 ***
~.s24* **
~ .656***
-.159***
Qwn house
.030* **
.344***
.327***
.089***
.584***
.322***
.290***
.083***
.0001
-.009
.0004
.034**
.039
.033
.013 **
.139***
.099***
.4 14** *
.5 16***
.434***
.149**'
.428
.776
.201
.197*
-.006
-.0 14*
.006
-.032
-.008
-.002
-.029
-.013
-.031
-.010
-.088***
Head
.435 ***
.466***
-.014
,46 1***
Spouse
,439***
.37 1**
.378***
No or child 0- 14
-.017
-.0 15
.006
No or ehild 15-30
-.060**
-.047
-.047
HH Sizc
-.008
.04 1
.048
.040
.005
-.033
-.020
-.032
No or a id males
-. 143**
-.093
-.055
-.03 1
-.203 ** *
-.358***
~ .344***
Log likelihood
-42,463
-22 ,902
-41 ,128
-21,8 10
- 16,430
Obscrva lions
45,049
63,907
29 ,024
- 17.353
4 1,947
43,953
61 .'83
27 ,372
40,482
43 ,953
5,274
6,087
No or a id remales
Indi viduals
5 ,170
45,049
5,923
Clusters
5,923
5,23 1
6,087
5 ,324
Nolcs: All rcgrcssioll s includc 5-ycar-agc-çohorls . 14 rcgional and 10 ycar dumrnics. Pooled cross-secli on"l ordercd logi l s pccifi calion s in ( I) and (5 )
includc agc and agc2 in slead of age cohorl5. These spccificilt ions' Slandard errors were cOrTccted for clu sterin g of obscrvations. *U/**/* indientc a
parameter cstimate is significant al the 1%/5%/ 10% level. Reference never marricd and in e ther occupaliolls. Data: KLiPS 1998-2008 .
Table A3: Working Days Reduclion and Various Sat isFaction Measures
Males
Feil/ales
Life Smisfat io n
Haurs Sati sfacti on
Jo b Salisfac lion
Lire Salisfation
Hours Sati sfacti on
.lob Sati sfaCLion
( I)
(2)
(3)
(4)
(5)
(6)
Ot her fivc-day
-.04 1*
.134* **
.029
.0 14
.129***
.029*
Below live day
-.050
.063 *
-M I
- .065 ***
-.0 15
-.098* **
3-4 yems henee
-.067
-.090
.057
.020
-.005
.002
2-3 ycars hence
-MI
-.030
.082
.034
.032
-.030
I -2 ycars hCllcc
-. 118**
-.03 I
.0 12
.019
.112***
MI
Within {he ncxt year
-.092
-.01 I
.109*
.031
.130'* *
.067*
Lags
0- I ycars
-.05 1
-.032
-.006
.324***
.113* **
1-2 ycars
-.024
.04 1
.050
.060'
.012
.280***
.103** *
Leads
2-3 ycars
.039
.018
.10 1
.067*
.255 ***
.11 5***
3-4 yc.ars
-.009
.095
.092
.069
.272***
.1 6 1***
4-5 yca rs
-.067
.058
.081
.061
.195 ***
.2 10***
5 or morc ycars
-.05 I
-.083
-.104
.005
.245***
.235** *
Obscrvations
11 .046
I I ,054
9,725
24,923
24.975
2 1,760
Individuals
2,156
2, 156
2,133
3,902
3,902
3.865
Notcs: Uncar fi xcd-cffccts cslirnalor. Ollier contral variables inc1udc 5-ycar-agc-cohorts. dUIlUllics for hOllschold head .. mI spousc. ycars of sc hooling. hOllschold sizc, numbcr of yOll ng (0- 14
ycars) and aid ( 15-30 ycars) childrcn in thc houschold . du mmi cs for marital status. log per-cu pita houschold incomc . log of own carnin gs . log per-capita rcgi orml incolllc. house owncrship. 14
rcgional dummics. 10 )'cnr dummics. 10 occupati on :ll1d 16 induslry dummics ... */**/* indicalc a parameter cSlimatc is signilïcan t at Ihc 1%/5%/ 10% Icv el . Data: KLlI'S 1998-2008: job
sati sfaclion onI)' 2000-2008 .