Embed Size (px)
Citation preview
Participation and Wage equations for
Married Women in European countries
Catia NicodemoUniversity of Tor Vergata, Rome∗
January 2007Preliminary and Incomplete Please do not quote
Abstract
This paper estimates the participation and the wage equationsfor married women, using panel data from the European Commu-nity Household Panel (EHCP) corresponding to the wave 1994-2001,for thirteen European countries and explores the difference across-countries in a labor supply contest.
The paper first shows the labor force participation equation that de-pends on the personal and family characteristics, the income in house-hold and, in some countries, on the labor status of the husband. Thereis a considerable variation in the degree of labor market participationrate of women across countries.
Finally, I estimate the wage equation for females in a frameworkof unbalanced panel data models with sample selection. A test forselection bias and a correction procedure are used. The coefficientsthat affect the wage are very different across-countries.
JEL. classification: J2, J3, C2, C3
Keywords: Female participation, labor supply, family benefits, un-balanced panel data
∗I’m grateful to Sergi Jimenez-Martin and Robert Waldman for their encouragementand suggestions. Thanks also to Franco Peracchi, Nick Longford, Olaf Jourgens for usefulcomments.
1
Introduction
Married women’s increased participation in the paid labor force was one ofthe most important social changes in Europe in the twentieth century. Themost rapid rise in married women’s participation came after 1950 in everycountry, although the intensity of female labor participation varies acrossEuropean countries.
The Presidency Conclusions of the European Union Council in Lisbon(March 2000) set target for female employment rate was 57 per cent of thepopulation of working age for the year 2005 and 60 per cent for the year2010. The targets for total employment are 67 and 70 per cent respectively.Not every country is required to achieve these targets but the evaluationof their progress is usually done by comparing each country’s performancewith the average.
The literature proposes several explanations for higher female labor forceparticipation, which include an increase in wage rates and educational at-tainment for women, reduction fertility and the increase of divorce rates, butthis trend isn’t uniform for all countries. The employment rates for femaleinactivity are much more different across countries than the unemploymentrates (see figures 1 and 2).
Both employment and participation are influenced by supply and de-mand factors. Employment may be low because many women don’t want toenter the labor market, or participation may be low because too few jobs arebeing offered to attract women into the labor market. In the first case, lowparticipation rates are explained by women’s preferences and in the secondby employers’ preferences and discouragement on the part of the women. Itis very difficult to disentangle these two effects, and it is made more difficultby the effect of wages on the participation and the employment rates.
In the frictionless neoclassical economic model, the participation ratedrives the employment rate: employment rates differ across countries be-cause of women’s preferences, given technology and wage levels. The coun-tries in our sample have similar technological structures and standards ofliving, so the differences in the employment and participation rates are morelikely due to their institutional structures than to women’s preferences.
Table 1 describes the evolution of the employment rate of women from1960 to 2001 and the distance needed to achieve the Lisbon target. Weobserve that the female employment rate has grown more in the last fortyyears.
Northern countries tend to have a higher employment rate than Mediter-ranean countries, although this rule is not true for Belgium, Ireland andPortugal.
This may reflect the substantial difference among the countries’ institu-tions, such as a rigid labor market, a limited option for part-time work andpoor benefits for the family. As a consequence, women participate less in
2
the labor market and have fewer children.The paper analyzes the intertemporal labor force participation behav-
ior of married women, using eight longitudinal waves carried out by theEuropean Community Household Panel (ECHP).
The Panel analyzed only 13 E.U. countries1 for which the required in-formation is available. Sweden and Luxembourg are excluded because theirdata are incomplete and the samples very small.
The sample is reduced to married women born between 1941 – 1965in the survey 1994 – 2001. The sample comprises data on households andindividuals.
This work compares the cross-country labor supply for married womenin different countries. Firstly investigates the relationship between the par-ticipation decision and some variables such as: women’s non-labor income,their education level, the presence of children, experience, the role of thefamily financial conditions and family benefits, the labor market status ofthe husband, and explores how these factors explain the differences in thelevel of participation of married women.
Afterwards I calculate the wage equation in a structure of panel datamodels with sample selection. I consider different econometric methods toestimate this equation. First I consider there isn’t sample selection so Iuse an OLS estimate. Secondly I try to estimate the wage equation with aHeckman Model, in a framework of panel data sample selection, and finallyI consider a model of 2SLS where the equation of interest contains endoge-nous explanatory variables as well as unobserved heterogeneity. Due to theincreased availability of longitudinal data and recent theoretical advances,panel data have become usually used in applied work in economics.
The wage equations of females have the following potential sources ofbias: first, unobserved heterogeneity for unobserved worker characteristics(ability). Second, sample selection bias, that occurs if unobservable charac-teristics, which affect the work decision, are correlated with characteristicsthat affect the process determining the work. Third, experience is likelyto be non-strictly exogenous, even after controlling for heterogeneity andsample selection.
The paper is organized as follows. In section one I discuss the data setused in the analysis and I describe the intertemporal participation behav-ior and demographic characteristics. Section two describes the econometricmethod used to estimate the participation equation of married females. Thisequation includes the personal and family characteristics of married womenand the effect of husband’s labor market status on the women’s labor supply.In Section three I estimate the participation equation and the wage equa-tion for married women, testing for selection bias and using a correction
1Austria, Belgium, Denmark, Finland, France,Germany, Greece, Italy, Ireland, Hol-lands, Portugal, Spain, The U.K.
3
procedure for this bias, as proposed by Wooldridge-Semikina (2006).There is a significant literature on married women’s labor force par-
ticipation (see Killingsworth and Heckman (1986), Blundell and MaCardy(1999)) but few works have compared the thirteen European countries, in aframework of panel data for eight years. To estimate the participation equa-tion I follow the work of Jimenez and Borrego (1997), where they comparecountries’ data set carried out from the CHER (Consortium of HouseholdEuropean Data) and where they estimate the participation and wage equa-tion for only three years with a two-step model of Heckman.
There is a lot of literature that estimates the wage equation and prob-lems related to heterogeneity and relativity under the assumption of strictlyexogenous explanatory variables. Verbeek and Nijman (1992) proposed twotests for selection in panel data.
Wooldridge (1995) proposed a test to correct the selection bias thatoccurs when unobserved effects are correlated with explanatory variables.Kyriazidou (1997) proposed a semiparametric approach for correcting theselection bias. Both the selection term and the unobserved effect are removedby difference between two periods. Rochina and Dustmann (2000) take intoaccount that the non strict exogeneity of regressors can be violated.
The extension literature for endogenous explanatory variables in the pri-mary equation is minimum and relatively recent.
Vella and Verbeek (1999) estimate a panel data model with censored en-dogenous regressors and selection, but don’t allow a correlation between theunobserved effect and exogenous variables in the interest equation. Kyriazi-dou (2001) makes an estimation of dynamic panel data models with selectionand lags of the dependent variable appear in the interest equation and in theselection equation. Garcia, Jimenez and Labega (2006) present an empiricalexample on Spanish data to estimate a wage equation and strike decisionwith a switching model. Finally we have Wooldridge and Semikina (2005)where they estimate a panel data sample selection and propose a test forselection bias and a procedure for correcting this.
1 Sample characteristics
Data
The data analyzed in this work come from a survey by the European Com-munity Household Panel (ECHP), a multi-country annual longitudinal sur-vey collected data since 19942 in 15 European Union Member States underEurostat (Statistical Office of the European Communities) coordination.The data set covers about 130,000 individuals from 60,000 households in
2Belgium, Germany, Hollands, The U.K., Denmark, France, Greece, Ireland, Portugal,Italy and Spain started in 1994 (wave 1), Austria jointed in 1995 (wave 2), Finland joinedin 1996 (wave 3).
4
the fifteen countries which were EU members in 2000, reflecting populationchanges over time through a continuous evolution of the sample. The paneldata cover a wide range of subjects such as demographics, labor force be-havior, income, health, education and training, housing, poverty and socialexclusion, etc. The survey is structured in the form of annual interviews witha particular representative sample of household members in each country.Interviews are conducted following a standardized questionnaire, althougheach country can modify the questionnaire’s wording to some extent, to re-flect their own institutional arrangements.The sample is constructed as an unbalanced panel of all women betweenthe ages of 31 and 53 years, who are married with or without children, andthus are old enough to have finished their formal education and too youngto retire. The size of this sample varies across the countries.
The variables refer to the personal characteristic of individual (age , workexperience, education, etc) and household family ( family income, familybenefits, number of people in the house etc.). Income, family and disabilitybenefits are deflated local currency units.
Information on their husbands has also been extracted (including laborstatus, education and unemployment benefit if received). The definition ofeach variable used in estimates is reported in Table 15.
In Table 2 I present the estimate labor force participation rate of marriedwomen in 1994− 2000 calculated dividing employed (at work or searching awork) by the total population in working age. There is a large heterogene-ity from Mediterranean European countries and Scandinavian countries likeFinland and Denmark. The range varies from 50% for the first group to90% for the second group, this may be caused by substantial differences inthe organization of the welfare state. In the last two years the rate has beenreduced for most of the countries, while Europe was in a recession businesscycle. Ireland seems to be an exception because its employment rate is moresimilar to the Mediterranean countries, which can be interpreted as the re-sult of cultural differences. Another exception is Portugal where the femaleparticipation rate is very high.
Descriptive Statistics
Table 3 present descriptive statistics on a selection of variables of interest,by country, for full sample married women, while in table 4 I compiled thesummary statistics for a subsample (women in the labor market excludingself-employed) in the year 2001.
The data demonstrate large differences in female and male educationlevels among countries, mainly at second and tertiary level of school (from13% of graduates in Portugal to 45% in Finland for females). The numberof males with tertiary level of school is higher than the number of women
5
with the same level of school, while the number of male with first level ofschool is greater in Mediterranean countries.
Concerning the subsample of employed married women, some of the aver-ages are quite different (although these differences are not statistically signif-icant) than the whole sample. Particularly, we observe a higher percentageof active women with a university degree and a few dependent children. Thefamily benefits that include Child Benefit and Lone Parent Allowance (fam-ilybenef ) are very small for Greece, Spain and Italy, both in the subsampleand in the full sample, while the unemployed benefits for their husbands aresmall for Italy, Greece and U. K. but very high for Spain.
The same trend is followed by the total disability benefits in the house-hold, except for U.K.
All variables of income are measured in local currency units deflated bythe average exchange rate in the sample year, and I take their logarithm.3.
From Table 5 to table 11 I compare the sample characteristics of womenfor which I consider the work activity status during the sample period(1994 − 2001) for each country. Given the large number of possible par-ticipation sequences during the panel, I choose only four of them.
In each table we have the descriptive characteristics of women for fullsample, for women who worked in each year (column 1), for women whonever worked (column 2), for women who had only a single transition fromnonemployment to employment (11111110, 01111111,.....) (column 3), forwomen who had only a single transition from employment to nonemployment(00000001, 10000000,....) (column 4) and in the last column for women whohave more than one transition in participation to labor market.In all countries women who always worked, and so with more experience,tend to be better educated and older, except for Finland where the maximumgrade of education is higher for all other sequences, and for France where it isgreater for females who never worked. In general we observe that women whohave worked for eight years have more children under age twelve and fewerdependent children, and that their household income is lower than women’swho never worked, with the exception of the Mediterranean countries (Italy,Portugal, Greece and Spain).
Family benefits are higher for women who don’t have experience in thelabor market (Denmark and Greece are the exception), while the disabilitybenefits paid to the family are greater for women who work except in Italy,Portugal and the U.K.
Regarding the level of education of their husbands, we note it’s higher inthe sample of women that worked every year, except for France and Irelandwhere we have the inverse, while for Italy husband’s education is in generallower than wives education.
3For France and Austria the wage and the unemployed benefit are in gross amount, Iuse the net/gross ratio
6
Below each table, I’ve calculated the distribution of years of women spentin employment by country, and the average rate of participation (partic).This analysis illustrates a significant persistence in employment decisions ofmarried women (full sample) observed annually. For example: if we takethe individual employment of Germany in an independent context from abinomial distribution with fixed probability of 69,59% (the average partic-ipation during the period) we conclude that the 70% of the women in thesample would be expected to work each year and 30% would not. Comparedthis with the distribution of the work years, we find that the sample relativefrequencies of women are 28% and 6% respectively. There is a considerabledifference in the work propensity of women if we compare the frequencydistribution of work years and the participation sequence. This heterogene-ity is observable in the different levels of education, age, nonlabor income,number of children, and in the policy that each government implements forfemale labor supply.
2 The Model
The participation equation is a discrete choice model, where the probabilityto participate pi is different for each individual and depends on the individualand household characteristics.
π = Φ(Xit)
where the Φ(Xit) is the cumulative distribution of the standard normal.The participation equation can be written as:
q∗it = ai + βXit + vit (1)
qit = 1[q∗it ≥ 0] (2)
We need to calculate the marginal partial effect for a unit of change ina particulary explanatory variable Xit, because the coefficients estimates inequation 11 are not directly provided by this information.
The marginal effects are computed in the following expression:
∂i = dΦ(Xit)/d(Xit) = φ(Xit)β (3)
Where φ(·) is the density of standard normal distribution.To analyze the wage equation we need a model for selection bias. The
Heckman bivariate normal selection model represents the classic way fordealing with selection on unobservable variables. The selection on unob-servable occurs when the error term in the outcome equation is correlated
7
with the treatment, or with the selection into the sample being used forestimation.
Starting with the Heckman model, I apply it to panel data because thesample selectivity is an acute problem in panel data and in cross sectionwhen we want to study the labor supply.
The bivariate normal selection model was developed in a context for esti-mating a population wage equation when only wage information on workersis observed (Gronau 1974 and Heckman 1976).
The difference between workers and non-workers determines the sampleselection bias because some components of the work decision are relevant tothe wage determining the process and the unobservable characteristics thataffect the work decision and the wage.
The usual setup is as follows. We have a wage equation
W = β′Xi + εi (4)
where W is the hourly wage and is observed only for workers, X areobserved variables related to productivity and εi is the error term that in-cludes all unobserved determinants of wages; it does not matter whether Xis observed for just workers or for everyone, as this information will only beused for workers. A reduced form employment equation is given by
$∗i = Ziγ + µi (5)
where $∗ is a latent index that can be thought of as representing thedifference between the observed wage and the reservation wage, that is thelowest wage at which the individual is willing to accept employment. $i isonly observable and equal to one if $∗
i > 0, where $i is an indicator variablefor employment.
The Heckman model requires the following assumptions:
(a) (ε, µ) ∼ N(0, 0, σ2ε , σ
2µ, ρεµ);
(b) (ε, µ) is independent of X and Z;
(c) varµ = σ2µ = 1
The first assumption represents a very strong functional form assump-tion -namely joint normality of the distribution of the error terms in theparticipation and outcome equations. The second equation assumes thatboth error terms are independent of both sets of observable variables. Thefinal assumption is the standard normalization for the probit selection equa-tion, which is identified only up to scale. Now if we take the expectationsof the wage equation conditional on working we have
E(Wi|$i, Xi) = E(Wi|Zi, Xi, µi) = βXi + E(εi|Zi, Xi, µi) (6)
8
The first equality just recognizes the fact that the variables determiningemployment in this model are Z and µ. The second equality comes from thefact that the expected value of X given X is just X. The final term can besimplified by noting that selection into employment does not depend on X,only on Z and µ, so we have:
E(Wi|$i, Xi) = βXi + E(εi|$i = 1) = βXi + E(εi|µi > −Ziγ) (7)
Thus, if we estimate the model using only data on workers, we do not getthe population wage equation, but rather something else. As a result of thisterm, OLS estimation on a sample of workers generally provides inconsistentestimates of the parameters of the population wage (or outcome) equation.
The first method to solve the problem of sample selection was proposedby Heckman in 1974 by a maximum likelihood estimator. With the as-sumption that εi and µi are i.i.d., N(0,
∑), where
∑is a variance matrix
covariance for the errors, and (εi, µi) are independent of Zi we write themaximum likelihood as:
L =1N
N∑i
{Ei ∗ ln[∫ ∞
−Ziγφεµ(W −Xiβ, µ)dµ]
+(1−$i) ∗ [ln∫ ∞
−Ziγ
∫ ∞
−∞φεµ(ε, µ)dµ]
where φεµ represents the probability density function for a bivariatenormal distribution. The previous expression is more similar to the Tobitestimator of type II. If φεµ = 0 then the equation 4 is reduced to product oftwo marginal likelihoods.
The second method for estimating the bivariate normal selection modelis that due to Heckman in 1979, it is sometimes called the ”Heckman two-step”.
The first step of the two-step approach estimates a probit model of par-ticipation. The estimate of γ from this probit model is then used to constructconsistent estimates of the inverse Mills ratio term:
λi(Ziγ) =φ(Ziγ)Φ(Ziγ)
; (8)
where φ(·) and Φ(·), denote the probability density and cumulative dis-tribution functions of the standard normal distribution. In the second stage,the outcome equation is estimated by OLS, where the equation wage includesboth the original X and the constructed value of inverse Mills ratio.
W = β′Xi + εi + νλi (9)
9
The inverse Mills ratio is sometimes called a ”control function”, a func-tion that controls for selection bias (Heckman and Robb, 1985). With theinverse Mills ratio included, and under the assumptions noted above, thecoefficients on the X represent consistent estimates of the parameters of thepopulation wage equation. The coefficient on the inverse Mills ratio termestimates ρσε. Since σε > 0 by definition, the sign of this coefficient isthe same as the sign of ρ. The sign of ρ is often substantively useful in-formation, as it indicates the correlation between the unobservable in theselection and outcome equations. The standard t-test of the null hypothesisthat βλ = 0 is a test of the null that there is no selection bias, conditionalon the assumptions of the model.
The bivariate normal selection model is formally identified even if X =Z.
The identification comes from the non-linearity of the inverse Mills ratio.A model that simply included the predicted probability of participationfrom a linear probability model into the outcome equation would not beidentified. However, the X = Z case often results in substantial collinearitybetween the predicted inverse Mills ratio term and the remaining covariatesin the outcome equation. This will be especially strong when there is notmuch variation in the predicted participation probabilities, because thenthe non-linearities will not play a major role. This collinearity will, asalways, lead to large standard errors. More generally, a large Monte Carloliterature illustrates the poor performance of the bivariate normal modelwith no exclusion restriction in finite samples. The exclusion restrictionhere is a variable that belongs in the participation equation but not in theoutcome equation. In other words, it is an instrument.
Panel data sample selection
Sample selection is more frequently used in studies for cross-section and lesscommon to estimate with panel data. Maddala (1993) defines panel data asdata sets on the same individual for different period of time.
The observation in panel data has two dimensions: a cross-section di-mension indicate by i and a time series time dimension indicate by t.Panel data have some benefits as: to control for individual heterogeneity, lesscollinearity among variables, more variability, large numbers of available in-struments, study for dynamics of adjustment etc. Limitations of panel dataare a problem for nonresponse, attrition, measurement of errors, design anddata collection problem etc. More panels are incomplete, especially whenthe panel concerns the household, because some of them move outside thepanel for different cause. In this case the panel is called unbalanced. Moreforms of selection bias and heterogeneity present in the panel data are elim-inated by the fixed effects estimator under the assumption of strictly exoge-nous explanatory variables(see Verbeek and Nijman 1992). Recent papers
10
have introduced some endogenous regressor as explanatory variables withselection bias and source of heterogeneity in equation of interest.
Consider the following model:
wit = xitβ + γi + µit (10)
$∗it = Zitγ + αi + εit (11)
$it = 1 if $∗it > 0 (12)
where i, (i=1,......,N) denotes the individual and t,(1,....,T) denotes thepanel. The dependent variable in the primary equation is only observedif $∗
it > 0, so selection bias is introduced. The errors are decomposed inindividual effects (αi, γi), and idiosyncratic errors (εit, µit), while xit is a 1xKvector that contains both exogenous and endogenous explanatory variablesand β is a Kx1 vector of unknown parameters. We allowed a correlationbetween the unobserved effects and the regressor, and some of the elementsof xit are correlated with the idiosyncratic errors εit.
Given the distributional assumptions, it’s possible to estimate the un-balanced panel data about a Heckman model for panel data and a two-stageleast squares regression model (2SLS) if we have an endogenous regressor.These are methods of extended regression to cover models which violateordinary least squares (OLS) regression’s assumption of recursivity, specif-ically models where the researcher must assume that the disturbance termof the dependent variable is correlated with the cause of the independentvariable.
If there exists a correlation between the regressor and the idiosyncraticerrors we assume a set of instruments denoted as zit that are strictly ex-ogenous on γi and not correlated with µit. We use these instruments inthe first stage of 2SLS to create the new variables (called instrumental vari-ables) which replace the problematic causal variables. The instruments arethe exogenous variables with direct or indirect causal paths to the problem-atic causal variable but which have no direct causal path to the endogenousvariable whose disturbance term is correlated with that of the problematiccausal variable. I use this estimation when I accept that the potential ex-perience is endogenous.
To test the sample selection we start with the model of Mundlak (1978),where if there is a correlation in selection equation between the individualαi and Zi, we need a set of individual exogenous instrument ξzi, so αi canbe written as:
αi = η + ξizi + fi (13)
11
where zi is a vector of individual exogenous variables averaged acrossperiod time t. The selection indicator $i is rewritten as:
$it = 1[Zitγ + ξizi + υit > 0] (14)
where υit = fii + εit has zero means normal distribution.If E(υit|µit) is linear, then we have:
E(µit|zi, γi, $i) = ρE(υit|zi, $it) (15)
and the wage equation becomes:
wit = xitβ + γi + ρE(υit|zi, $it) + eit (16)
where eit is an idiosyncratic error term uncorrelated with the regres-sor, the unobserved effect and the selection indicator. If $it is equal toone, using a probit estimation at each period t we obtain the estimationof E(υit|zi, $it = 1) that is equal to: λ(η + ξizi + Zitγ), where λ(·) is theinverse Mills ratio. We put the estimation of λit, in the wage equation andestimate this with a simple regression model or with 2SLS model if wehave endogenous regressor. We use t-statistic for testing the null hypothesisH0 : ρ = 0.
To add more flexibility to the model is possible calculate the interactionterms the λ which time dummies and test the selection with a Wald test.This procedure for correcting the bias and inverse Mills ratio is a consistentestimator of the parameters.
If the null hypothesis is true , so there isn’t selection, then OLS an consis-tent estimator for the primary equation if don’t have endogenous variables.
I applied this procedure for estimate the wage equation for marriedwomen in ECHP.
3 Empirical estimates
Female participation equation
Given the panel structure of data set, I start analyzing the determinants ofthe equation participation of married women. I estimate this equation witha probit random effects model for panel data, because if we try to estimateprobit with fixed effects we have serious problems due to large number of in-cidental parameters that make the estimator inconsistent, but a large T cansolve this problem (see Arellano-Hanhn, 2006). The participation equationwas write before as:
q∗it = ai + βXit + τi + vit (17)
where qit(women1), is a dummy variable which takes 1 if the woman par-ticipates in labor market (working or seeking work) and 0 if she doesn’t.
12
I excluded self-employed married women. The equation participation, q∗it,is positive only if the dummy variable equals one. The decision to partici-pate depends on a vector of explanatory variables Xit. This vector includesthe personal and family characteristics of the woman: age, education, chil-dren, family benefit and income house without her wage income, and thecharacteristic of her husband (status of work, education and if he receivesunemployed benefits); β is a vector of unknown parameters and τi and vit
are respectively time invariant effect specific to individual and individualtime-varying error.
Tables 12 and 13 display the estimations of probit participation equationand the marginal effects for probit analysis for each country.
The explanatory variables are divided into four blocks. The first set ofvariables contains personal characteristics such as age (age), the square ofage (agetwo), the woman education and the husband’s education.
The sign of age is positive and significant in all countries, while agetwoin each country is negative and significant, which means that the relationbetween age and female labor force participation decreases with the age, someans an inverted-U shape.
Probably the participation increases until women assume more familyresponsibilities.
Concerning education, I find a propensity to participate that increaseswith the education level. On the average the propensity to participate isabout 30% from secondary degree and university degree and the effect isvery large for several countries.
Education for females at the second and thirdly levels are significant foralmost all countries, while the education of husbands is significative in fewcountries.
The second block of variables describe the fertility of women. I constructthe variables for children children age 0-3, children age 3-6 and children age6-12, which are dummy variables that take 1 if women have children withage under three, six and twelve years. The results show that the coefficientsare significative and its sign in most countries is negative so women who havedependent children reduce their participation to work. Finally the effect ofolder children is not significative for most countries and generally positive.
The public policy on childcare could be very important to explain theparticipation of women across the countries.
The third set of variables represent the husband’s labor status (workstatus) and their unemployment benefit (longhunmp). In several countriesthe labor status of husbands is significant, not for Mediterranean countries,where having an unemployed husband doesn’t encourage women to partici-pate in the labor market, but the sign of coefficients is positive so it meansa positive influence on the participation of women, while the unemploymentbenefits are statistically insignificant for almost all countries but the sign isnegative.
13
The potential experiences most important for labor force participationdecision. A measure of potential experience assigns of market experienceto women of the same age and educational status even though their la-bor market histories might differ. The coefficient of potential experienceis significant than reflecting skill enhancement or seniority benefits, whilethe negative effect of potential experience captures reduce productivity overtime, fewer market options available to older individuals.
Finally we have different sources of income in the household: (logincome)the total non-labor income in the house where the employment wife’s incomehas been excluded, the disability benefits and family benefits. All incomevariables are deflated with CPI (Consumer price index) so a comparisonamong years is possible, but at the same time I use the PPP (Purchasingprice power) that allows a comparison among countries.
When we check the data we observe there is a negative income effect onmarried women’s labor participation. All the income variables are significantfor all European countries. The effect is larger for the income household thatfor the benefits in the family and disability benefits, so we may interpret thatother sources of income have more influence in the participation, for examplethe wage of husbands and financial family conditions.
Women’s wage equation
Now I estimate the wage equation with panel data sample selection, withthree different models, OLS, Heckman two step for panel data and 2SLS. Inthe last model I consider the potential experience as endogenous.
When we want to estimate the wage equation for married women we facedifferent problems: selection bias, endogeneity and eventually heterogeneity.
We have selection bias because the dependent variable of wage equationcan be measured only if the individual participates to the labor market.The literature offers estimators for correcting this problem (Heckman 1979,Powell 1994).
Heterogeneity is associated with the unobserved ability and motivationof an individual (ex. education), and if this unobserved individual effectsare correlated with the regressor of the model, the simple estimations withOLS are inconsistent, while panel data solve this problem.
After testing the selection equation described before, I take wage equa-tion 14 and I put inside the inverse Mill’s ratio estimated by probit model foreach period. I estimate this equation by OLS, Heckman and 2SLS modelsso the wage equation is:
wit = xitβ + γi + λit + eit (18)
The wage equation is analyzed on a sample that is limited for mar-ried women born between 1941–1965, excluding those self-employed. I’ve
14
dropped observations that are inconsistent, and excluded women when theyears of experience exceeds the age, when experience was missing, and whenthey reports a positive number of hours of work and zero wage.
The dependent variable wit is the logarithm of the real monthly wage ofmarried women. The vector of explanatory variables xit includes educationof women, time dummies, experience and experience square, education ofwomen in two different levels. The experience has been calculated as thedifference between the present age and the age when starting the first job.4
More papers use the actual experience but for the ECHP data set wecan’t observe any variables to calculate this, we observe only potential ex-perience.
I use both participants and no-participants married women to estimatethe selection equation, while to estimate the wage equation I use only themarried women that participate to labor market for at least two waves.
Furthermore, in few countries some variables where dropped for estima-tion because they where perfect predictors of one of the two alternatives.
In table 14 and 15 I represent the estimation of the wage equation usingthe OLS regression model, where there isn’t correction for sample selectionand endogeneity, and the Heckman model for panel data selection, where Iput in regression equation the inverse Mill’s ratio, calculated each year witha probit model (λit), to correct the sample selection and all regressors areconsiderate exogenous. In table 16 I tried to estimate a 2SLS model, consid-ering the experience variable as endogenous. The set of instruments zit usedto control the endogenous experience contain : age and it’s square, variablesfor children in three different ages, education of women and education oftheir husbands in two different levels, sources of income in household, andan indicator of the labor status of their husband. All this regressors are con-sidered exogenous. The vector xit in this case is a subset of zit that includeseducation of women, time dummies, experience and experience square.
The result confirms that the Heckman model for panel data is an appro-priate model to estimate the wage equation. First, not all variables whichaffect participation equation are also determinants to estimate the wageequation, and the magnitude of the effects are different. In table 15 we findthat the selectivity correction term λit is significant for most countries anda negative sign is present for countries with a higher rate of employment.
Concerning the experience variable, the effect is different between theOLS and the Heckman models. While in the OLS model the experiencedecreases the wage for most countries, in the Heckman model, experiencehave a positive and significant effect for all European countries. The squareof experience is negative and significant for both models in each country,which means that the relation between experience and wage is the same
4Similar regressions were also computed using experience calculated by age minus yearschooling minus 6. The results do not vary with the measure of experience.
15
after a high number of years of experience.Regarding the effect of education we find that a higher education level
increases the wage in several countries and the sign is statistically significantin panel data which are corrected for selection bias. We find the same resultsin the OLS regression, while a low level of education has a negative effecton the wage for both estimations.
The magnitude of the effect of each variable is different across-countriesin the Heckman model. In Northern countries the experience has a smallereffect than in Mediterranean countries. The effect of tertiary education isvery low in Italy, while in the rest of countries the magnitude is larger.Having a primary education influences negatively the wage while the effectis reduced for countries with a high female employment rate.
In table 16 I estimate the wage equation with a model of 2SLS, and findthat the experience has a positive effect on the wage, same the Heckmanmodel, but coefficients are lower than in the Heckman model and they aremore significant for Mediterranean countries. The education has the sameeffect in the Heckman model, a higher level of education incises negativelyon the wage equation.
Conclusion
In this paper I first analyze the participation in labor market for marriedwomen in thirteen countries in Europe for a panel of eight years. I foundthat the level of education has a positive influence on the participation, andwomen who have a university degree participate more in the labor market.Another factor that has a positive effect on participation is the age and thelabor status of husbands.
Variables that influence negatively the labor market of females are: thehousehold income, children, in particular dependent children, family anddisability benefits.
These results mean that the economic policy plays an important role forincreasing the participation equation, because for example an increase inthe policy of childcare and nursery development provides women with moretime to dedicate to the labor market. The targets imposed in the PresidencyConclusions of the European Union Council in Lisbon, in March 2000, arefar away for most countries.
Concerning the wage equation, both the experience and a higher level ofeducation have a positive effect on it, while a low level of education has anegative effect in most countries.
The magnitude of the effect of each variable is different across-countries.
16
References
[1] Askildsen, J.E., B.H. Baltagi and T.H. Holmas, 2003,Wage policy in the health care sector: a panel data anal-ysis of nurses labour supply, Health Economics 12, 705-719.
[8] Chamberlain, G., 1980, Analysis with qualitative data,Review of Economic Studies 47, 225-238.
[3] Dustmann C., and M.E. Rochina-Barrachina, 2000, Se-lection correction in panel data models: an applicationto labor supply and wages, IZA Discussion Paper 162.
[5] Kyriazidou, E., 1997, Estimation of a panel data sampleselection model, Econometrica 65, 1335-1364.
[5] Kyriazidou,E., 2001, Estimation of dynamic panel datasample selection models, Review of Economic Studies 68,543-572.
[6] Borrego-Jimenez, 1997, Married women labor supply : acomparation analysis, Cesp
[7] Mundlak, Y., 1978, On the pooling of time series andcross section data, Econometrica 46, 69-85.
[8] C. Nicoletti and F. Peracchi 2002: A cross-country com-parison of survey participation in the ECHP, Workingpapers of the. Institute for Social and Economic Research
[9] Semykina-Wooldridge, 2005: Estimating panel datamodels in the presence od endogeneity an selection: the-ory and application, working paper
[11] Vella, F. and M. Verbeek, 1999, Two-step estimation ofpanel data models with censored endogenous variablesand selection bias, Journal of Econometrics 90, 239-263.
[11] Verbeek, M. and T. Nijman, 1992, Testing for selectiv-ity bias in panel data models, International EconomicReview 33, 681-703.
[13] Wooldridge, J.M., 1995, Selection corrections for paneldata models under conditional mean independence as-sumptions, Journal of Econometrics 68, 115-132.
[13] Wooldridge, J.M., 2002, Econometric analysis of crosssection and panel data. MIT Press: Cambridge, MA. 33
17
Figure 1: Female inactivity and unemployment as per cent of pop-ulation of working age, 2000
Source: Oecd
18
Figure 2: North and Mediterranean Countries Female inactivityrate, 2000
19
Table 1: Female employment rate 1960 – 2000 Persons aged 15 to64 yearsCountry 1960 1980 2000 Men 2000 Lisbon Distance (a)Denmark 42,70 66,20 71,20 80,40 11,20Finland 54,90 65,00 64,30 69,70 4,30Norway 26,10 58,40 73,40 88,10 13,40Sweden 38,10 67,60 72,10 76,20 12,10U.K. 43,10 54,50 65,20 79,30 5,20
Belgium 29,60 35,00 51,10 69,80 -8,90Germany 35,00 34,80 58,10 73,50 -1,90Ireland - 32,20 52,20 74,00 -7,80
Hollands - 35,70 62,10 81,10 2,10Austria - 53,40 59,30 78,10 -0,70Greece - 30,70 40,40 70,20 -19,60Italy 28,10 33,20 39,70 68,50 -20,30Spain 21,00 28,40 40,30 70,30 -19,70
Portugal - 47,10 60,10 75,90 0,10France 42,90 50,00 53,10 69,80 -8,90
(a) Lisbon distance is the percentage differ-
ence between the female employment rate in
2000 and 60% (b) source OECD 2000
Table 2: Female employment rate
Country 1994 1995 1996 1997 1998 1999 2000 2001Germany 68,01 66,52 65,88 65,43 66,02 64,79 64,02 63,80Denmark 89,61 89,90 86,35 87,10 86,33 86,25 85,01 84,69Holland 56,23 64,93 58,53 58,19 59,46 57,21 56,38 57,47Belgium 70,41 70,63 70,64 69,19 67,71 66,55 64,50 65,61
France 64,31 66,52 65,81 63,77 63,91 64,12 62,79 61,69U.K. 64,27 63,97 65,29 65,55 64,31 63,15 63,40 60,87
Ireland 32,99 34,47 36,31 39,64 41,46 43,33 45,87 43,62Italy 40,79 40,49 39,31 39,91 39,16 39,25 37,16 37,47
Greece 38,39 38,40 36,16 35,84 34,99 30,89 32,41 31,26Spain 34,51 36,56 37,68 37,28 37,28 35,87 36,62 38,23
Portugal 59,95 58,49 58,52 57,87 57,62 57,88 55,80 55,26Austria - 51,61 52,20 50,33 50,50 49,63 49,88 49,51Finland - - 86,71 86,39 84,18 84,17 83,95 80,70
20
Tab
le3:
Sum
mar
yst
atis
tics
offu
llsa
mple
Coun-
try
n.o
bs
age
fe-
male
sec-
ondary
edu-
cati
on
fe-
male
ter-
tiary
edu-
cati
on
chil-
dre
nage
6-1
2
chil-
dre
nage
3-6
chil-
dre
nage
0-3
adult
houlo
gin
-com
e
dis
-abili-
ty-
benef
fam
-ily-
benef
work
statu
s
hus-
band
sec-
ondary
edu-
cati
on
hus-
band
ter-
tiary
edu-
cati
on
un-
em
-plo
y-
benef
Ger-
many
1203
46,5
50,5
70,1
80,1
70,0
40,0
12,5
610,6
70,3
04,7
40,8
50,5
60,3
30,6
9
6,9
70,5
00,3
80,3
80,1
90,1
10,7
50,7
81,6
23,7
10,3
50,5
00,4
72,2
7D
en-
mark
551
46,6
40,5
20,1
50,1
20,1
10,0
72,2
612,2
00,4
83,7
90,9
30,5
30,3
60,2
6
7,0
40,5
00,3
50,3
20,3
20,2
50,5
20,8
41,8
83,7
50,2
50,5
00,4
81,4
1H
ol-
land
1049
45,4
50,0
00,9
90,1
90,1
50,0
72,4
810,8
00,7
44,1
50,9
10,0
10,0
00,1
6
6,4
80,0
70,0
80,3
90,3
60,2
60,7
50,8
32,3
93,5
10,2
80,1
00,0
71,1
0B
el-
giu
m436
43,9
50,3
40,1
90,1
00,1
30,0
62,5
713,8
70,8
35,9
80,9
60,3
60,4
00,3
4
6,1
00,4
70,3
90,3
00,3
30,2
30,8
20,4
82,4
43,3
80,2
00,4
80,4
91,5
8
Fra
nce
850
45,8
50,1
20,6
10,2
20,1
00,0
62,6
111,9
20,6
52,9
90,9
00,1
00,2
50,4
7
6,2
70,3
20,4
90,4
10,3
00,2
30,7
80,7
62,1
13,6
80,3
00,3
00,4
31,9
5U
.K.
881
46,3
00,1
60,4
00,1
30,1
10,0
72,4
29,7
50,6
93,8
70,8
90,1
80,5
00,0
87,2
80,3
70,4
90,3
40,3
10,2
50,6
90,7
12,2
43,6
50,3
10,3
90,5
00,7
2Ir
e-
land
297
46,3
90,4
30,3
30,1
70,1
60,0
82,9
710,0
10,4
75,0
90,9
20,3
90,2
10,8
0
6,7
20,5
00,4
70,3
80,3
70,2
71,0
40,5
01,9
33,0
10,2
70,4
90,4
12,4
1It
aly
712
45,5
90,5
20,3
00,1
50,1
30,0
72,7
910,3
00,2
70,4
00,8
50,4
60,1
70,0
36,5
50,5
00,4
60,3
60,3
30,2
60,8
71,0
71,4
51,5
70,3
60,5
00,3
80,2
2
Gre
ece
341
44,8
70,3
30,3
50,1
40,1
10,0
72,8
915,1
00,2
20,3
80,8
90,2
70,3
50,2
8
6,2
90,4
70,4
80,3
50,3
10,2
60,9
51,5
31,3
11,6
00,3
20,4
40,4
81,4
1Spain
490
45,0
90,2
20,4
10,1
60,1
30,0
82,9
514,7
30,7
30,1
90,9
00,1
80,3
20,3
96,4
60,4
10,4
90,3
70,3
40,2
71,0
21,0
52,3
91,2
00,2
90,3
90,4
71,7
0Por-
tugal
618
45,9
20,1
20,7
00,1
20,0
90,0
42,7
814,4
60,4
43,7
30,8
80,1
20,1
20,2
2
7,0
30,3
20,4
60,3
20,2
80,2
00,8
61,1
11,7
92,8
80,3
20,3
20,3
31,3
3A
us-
tria
404
44,6
40,6
80,1
90,1
40,0
90,0
32,7
612,7
90,2
75,1
10,8
70,8
00,1
10,2
7
6,3
00,4
70,3
90,3
50,2
80,1
80,8
80,5
11,4
93,7
90,3
40,4
00,3
11,4
5Fin
-la
nd
713
47,3
70,3
90,1
60,1
10,1
20,0
52,3
711,7
61,7
63,8
30,8
80,3
90,3
90,7
8
6,9
00,4
90,3
70,3
10,3
20,2
30,6
10,7
73,1
23,5
80,3
30,4
90,4
92,3
4
Sta
ndar
ddev
iati
ons
init
alic
s,Full
sam
ple
:al
lm
arri
edw
omen
was
bor
n19
41-1
965
21
Tab
le4:
Sum
mar
yst
atis
tics
ofsu
b-s
ample
Coun-
try
num
.O
bs
age
fe-
male
sec-
ondary
edu-
cati
on
fe-
male
ter-
tiary
edu-
cati
on
chil-
dre
nage
6-1
2
chil-
dre
nage
3-6
chil-
dre
nage
0-3
adult
houlo
gin
-com
e
dis
-abili-
ty-
benef
fam
-ily-
benef
work
statu
s
hus-
band
sec-
ondary
edu-
cati
on
hus-
band
ter-
tiary
edu-
cati
on
un-
em
-plo
y-
benef
Ger-
many
1058
46.6
00.5
60.1
80.1
70.0
30.0
12.5
510.6
40.2
64.6
20.8
50.5
60.3
30.7
0
6.8
90.5
00.3
80.3
70.1
80.1
10.7
50.7
41.5
23.7
30.3
50.5
00.4
72.2
7D
en-
mark
546
46.5
90.5
10.1
40.1
20.1
20.0
72.2
512.2
00.4
73.8
10.9
30.5
30.3
60.2
6
7.0
10.5
00.3
50.3
20.3
20.2
50.5
20.8
51.8
63.7
50.2
50.5
00.4
81.4
1H
ol-
land
815
45.3
20.0
00.9
90.2
00.1
40.0
72.4
610.7
50.7
84.0
00.9
00.0
00.0
00.1
8
6.3
20.0
00.0
90.4
00.3
50.2
60.7
30.9
12.4
43.5
20.2
90.0
00.0
71.1
8B
el-
giu
m428
43.9
00.3
40.1
90.1
10.1
30.0
62.5
713.8
70.8
45.9
90.9
60.3
60.4
00.3
5
6.0
40.4
70.3
90.3
10.3
30.2
30.8
20.4
82.4
63.3
80.2
00.4
80.4
91.6
0
Fra
nce
850
45.8
50.1
20.6
10.2
20.1
00.0
62.6
111.9
20.6
52.9
90.9
00.1
00.2
50.4
7
6.2
70.3
20.4
90.4
10.3
00.2
30.7
80.7
62.1
13.6
80.3
00.3
00.4
31.9
5U
.K.
783
46.2
20.1
60.3
80.1
40.1
10.0
72.4
39.7
50.6
33.8
30.9
00.1
90.5
10.0
77.1
30.3
70.4
90.3
50.3
10.2
50.6
90.7
12.1
33.6
40.3
00.3
90.5
00.6
8Ir
e-
land
272
46.1
80.4
40.3
10.1
80.1
60.0
82.9
710.0
40.4
55.1
60.9
30.4
00.2
10.6
9
6.5
40.5
00.4
60.3
80.3
70.2
81.0
40.4
71.8
82.9
70.2
50.4
90.4
12.2
3It
aly
707
45.6
20.5
20.3
00.1
50.1
30.0
72.8
010.3
00.2
70.4
00.8
50.4
60.1
80.0
36.5
60.5
00.4
60.3
60.3
30.2
60.8
71.0
71.4
51.5
80.3
60.5
00.3
80.2
2
Gre
ece
335
44.8
40.3
30.3
40.1
40.1
10.0
72.8
915.1
00.2
20.3
80.8
90.2
70.3
60.2
4
6.2
70.4
70.4
80.3
50.3
10.2
60.9
41.5
41.3
31.6
10.3
10.4
50.4
81.3
1Spain
483
45.1
10.2
20.4
10.1
60.1
30.0
82.9
614.7
30.7
50.2
00.9
00.1
90.3
30.3
86.4
30.4
20.4
90.3
70.3
30.2
81.0
31.0
62.4
11.2
10.2
90.3
90.4
71.6
8Por-
tugal
614
45.9
10.1
20.7
00.1
20.0
90.0
42.7
814.4
50.4
33.7
20.8
80.1
20.1
20.2
0
7.0
40.3
20.4
60.3
20.2
80.2
10.8
51.1
11.7
72.8
80.3
20.3
20.3
31.2
9A
us-
tria
386
44.5
80.6
80.1
80.1
40.0
80.0
32.7
712.7
90.2
45.0
90.8
80.8
00.1
10.2
9
6.2
30.4
70.3
80.3
40.2
70.1
70.8
90.5
11.4
53.8
00.3
30.4
00.3
21.4
9Fin
-la
nd
707
47.3
50.3
90.1
60.1
10.1
20.0
62.3
711.7
51.7
43.8
30.8
80.3
90.3
90.7
7
6.8
90.4
90.3
70.3
10.3
20.2
30.6
10.7
73.0
93.5
80.3
30.4
90.4
92.3
3
Sta
ndar
ddev
iati
ons
init
alic
s,su
b-s
ample
:al
lm
arri
edw
omen
was
bor
n19
41-1
965,
emplo
yed
orse
arch
ajo
b
22
Tab
le5:
Sam
ple
char
acte
rist
ican
dye
ars
spen
tin
tow
ork
Ger
many
Den
mark
Em
pl.
8 years
Em
pl.
0 Years
Sin
gle
trans.
to Work
Sin
gle
tras.
from
Work
Mult
i.tr
ans.
Em
pl.
8 years
Em
pl.
0 Years
Sin
gle
trans.
to Work
Sin
gle
tras.
from
Work
Mult
i.tr
ans.
age
44.9
140.1
042.3
835.6
038.5
8age
44.3
443.5
744.3
233.9
237.6
3le
ssfe
male
sec-
ondary
educati
on
0.3
40.1
10.2
30.1
80.2
3le
ssfe
male
sec-
ondary
educati
on
0.4
10.0
00.2
40.3
70.4
1
fem
ale
secondary
educati
on
0.5
20.6
50.5
90.6
40.6
2fe
male
secondary
educati
on
0.4
80.2
90.6
40.5
50.5
2
fem
ale
tert
iary
ed-
ucati
on
0.1
30.2
40.1
70.1
70.1
5fe
male
tert
iary
ed-
ucati
on
0.1
10.7
10.1
30.0
80.0
8
childre
nage
6-1
20.1
80.1
80.2
20.1
20.1
5ch
ildre
nage
6-1
20.1
40.0
00.1
30.0
80.0
7ch
ildre
nage
3-6
0.0
30.2
20.0
70.2
40.1
7ch
ildre
nage
3-6
0.1
40.1
40.1
90.2
70.3
6ch
ildre
nage
0-3
0.0
30.1
00.0
50.2
00.1
7ch
ildre
nage
0-3
0.0
70.1
40.1
00.3
80.3
4lo
gin
com
e10.6
710.9
810.7
010.7
210.7
2lo
gin
com
e12.2
412.2
012.1
512.2
312.3
0dis
abilitybenef
0.3
00.3
20.2
70.1
60.1
3dis
abilitybenef
0.2
08.1
00.7
40.7
40.5
6fa
milybenef
5.2
77.4
95.8
85.4
35.6
3fa
milybenef
4.9
82.2
44.2
04.8
66.0
1husb
and
sec-
ondary
educati
on
0.5
20.5
70.5
70.5
80.5
8husb
and
sec-
ondary
educati
on
0.5
30.4
30.5
60.5
20.5
1
husb
and
tert
iary
educati
on
0.3
90.2
80.3
30.2
80.3
1husb
and
tert
iary
educati
on
0.4
00.1
40.2
70.4
00.3
5
unem
plo
ybenef
0.5
70.8
10.6
90.7
60.6
1unem
plo
ybenef
0.2
10.0
00.2
50.7
00.3
9part
ic.
1.0
00.0
0-
--
part
ic.
1.0
00.0
0-
--
N.Y
ears
work
N.Y
ears
work
0-
100
0-
100
1-
-23.4
69.1
114.3
61
--
39.1
011.9
215.6
72
--
11.3
913.2
414.8
92
--
10.1
513.2
516.7
43
--
10.2
513.7
212.9
93
--
6.3
912.9
114.2
64
--
9.4
513.2
412.4
14
--
6.7
712.1
413.0
55
--
8.0
913.6
611.6
65
--
5.2
613.3
611.0
56
--
8.8
813.3
611.7
76
--
4.1
413.6
910.7
97
--
10.2
513.5
411.3
67
--
5.6
412.1
49.4
28
100
-18.2
210.1
310.5
68
100
-22.5
610.6
09.0
2
23
Tab
le6:
Sam
ple
char
acte
rist
ican
dye
ars
spen
tin
tow
ork
Holland
Bel
giu
m
Em
pl.
8 years
Em
pl.
0 Years
Sin
gle
trans.
to Work
Sin
gle
tras.
from
Work
Mult
i.tr
ans.
Em
pl.
8 years
Em
pl.
0 Years
Sin
gle
trans.
to Work
Sin
gle
tras.
from
Work
Mult
i.tr
ans.
age
44.4
140.8
542.0
838.8
438.8
0age
42.9
840.6
740.7
134.7
436.1
2le
ssfe
male
sec-
ondary
educati
on
0.5
00.1
10.4
80.4
10.0
1le
ssfe
male
sec-
ondary
educati
on
0.5
30.1
70.6
00.5
40.5
3
fem
ale
secondary
educati
on
0.3
70.2
40.3
20.2
80.0
0fe
male
secondary
educati
on
0.3
10.1
70.3
10.2
00.3
4
fem
ale
tert
iary
ed-
ucati
on
0.1
20.6
40.1
50.3
10.9
9fe
male
tert
iary
ed-
ucati
on
0.1
60.6
70.1
00.2
60.1
3
childre
nage
6-1
20.1
50.1
60.1
40.1
50.1
5ch
ildre
nage
6-1
20.1
00.0
00.0
80.0
40.0
9ch
ildre
nage
3-6
0.1
70.3
00.2
50.2
10.2
6ch
ildre
nage
3-6
0.1
20.3
30.2
80.1
80.3
5ch
ildre
nage
0-3
0.0
80.1
50.1
90.1
70.2
7ch
ildre
nage
0-3
0.0
30.0
00.1
80.3
20.3
3lo
gin
com
e10.7
610.9
210.8
110.6
510.6
7lo
gin
com
e13.8
714.0
713.8
813.8
513.7
6dis
abilitybenef
0.6
10.3
70.1
70.6
00.6
2dis
abilitybenef
0.7
00.0
01.1
31.2
40.6
3fa
milybenef
3.9
46.1
74.7
14.2
04.6
7fa
milybenef
6.6
46.8
66.4
15.2
26.2
7husb
and
sec-
ondary
educati
on
0.3
10.1
50.4
30.2
30.0
0husb
and
sec-
ondary
educati
on
0.3
50.1
70.4
20.2
40.3
6
husb
and
tert
iary
educati
on
0.3
80.3
20.3
80.4
80.0
1husb
and
tert
iary
educati
on
0.4
20.3
30.3
60.5
00.4
4
unem
plo
ybenef
0.1
30.0
40.1
90.1
50.1
0unem
plo
ybenef
0.2
80.9
20.3
10.7
20.4
4part
ic.
1.0
00.0
0-
--
part
ic.
1.0
00.0
0-
--
N.Y
ears
work
N.Y
ears
work
0-
100
0-
100
1-
-13.9
511.3
010.3
11
--
41.7
810.5
714.2
52
--
9.3
013.3
411.8
02
--
8.0
013.2
715.3
73
--
9.2
013.8
512.7
53
--
6.6
713.7
914.9
04
--
9.8
913.6
513.7
54
--
5.7
813.6
613.6
95
--
10.1
913.7
513.2
35
--
5.3
313.5
312.4
86
--
10.7
813.0
313.3
56
--
5.3
313.4
010.9
77
--
11.8
712.1
213.0
07
--
4.8
912.5
09.8
98
100
-24.8
38.9
611.8
08
100
-22.2
29.2
88.4
6
24
Tab
le7:
Sam
ple
char
acte
rist
ican
dye
ars
spen
tin
tow
ork
Fra
nce
U.K
.
Em
pl.
8 years
Em
pl.
0 Years
Sin
gle
trans.
to Work
Sin
gle
tras.
from
Work
Mult
i.tr
ans.
Em
pl.
8 years
Em
pl.
0 Years
Sin
gle
trans.
to Work
Sin
gle
tras.
from
Work
Mult
i.tr
ans.
age
44.8
133.5
043.5
333.3
237.3
7age
44.4
738.2
742.3
235.5
338.1
7le
ssfe
male
sec-
ondary
educati
on
0.2
90.5
00.3
60.4
40.3
5le
ssfe
male
sec-
ondary
educati
on
0.4
60.3
10.4
80.4
50.5
3
fem
ale
secondary
educati
on
0.1
40.1
00.1
00.1
40.1
1fe
male
secondary
educati
on
0.1
40.2
20.1
40.3
20.2
0
fem
ale
tert
iary
ed-
ucati
on
0.5
80.3
90.5
30.4
10.5
2fe
male
tert
iary
ed-
ucati
on
0.3
90.4
70.3
80.2
30.2
7
childre
nage
6-1
20.2
70.2
50.1
90.0
30.1
0ch
ildre
nage
6-1
20.1
50.1
60.1
20.0
80.1
2ch
ildre
nage
3-6
0.0
90.2
50.1
40.2
90.3
2ch
ildre
nage
3-6
0.1
30.4
70.1
70.2
30.1
9ch
ildre
nage
0-3
0.0
40.5
00.1
30.3
70.2
9ch
ildre
nage
0-3
0.0
60.1
60.1
10.2
10.2
2lo
gin
com
e11.9
812.1
211.9
311.6
411.8
1lo
gin
com
e9.7
79.8
09.7
99.7
09.7
5dis
abilitybenef
0.4
80.2
10.6
50.4
70.6
8dis
abilitybenef
0.6
61.5
10.1
30.5
10.3
2fa
milybenef
3.2
13.5
73.5
24.0
24.0
2fa
milybenef
4.2
76.6
45.0
34.5
74.7
9husb
and
sec-
ondary
educati
on
0.1
10.1
50.1
20.1
20.0
9husb
and
sec-
ondary
educati
on
0.1
70.2
00.1
50.2
70.2
3
husb
and
tert
iary
educati
on
0.2
60.4
30.3
00.3
30.2
7husb
and
tert
iary
educati
on
0.5
20.4
90.5
30.5
20.5
5
unem
plo
ybenef
0.3
91.9
70.4
60.7
90.5
0unem
plo
ybenef
0.0
30.3
40.0
00.2
20.1
1part
ic.
1.0
00.0
0-
--
part
ic.
1.0
00.0
0-
--
N.Y
ears
work
N.Y
ears
work
0-
100
0-
100
1-
-35.8
111.1
313.9
61
--
24.2
89.7
210.7
42
--
9.2
313.6
615.6
62
--
9.8
513.1
612.0
33
--
9.6
813.7
915.9
03
--
7.9
813.4
113.3
94
--
6.5
312.8
213.0
14
--
7.6
413.8
313.2
25
--
6.3
113.7
912.1
25
--
7.8
113.4
113.7
16
--
3.8
312.1
011.4
66
--
9.6
813.5
813.5
17
--
5.4
113.1
48.3
67
100
-10.1
913.8
312.1
18
100
-23.2
09.5
89.5
322.5
89.0
511.2
8
25
Tab
le8:
Sam
ple
char
acte
rist
ican
dye
ars
spen
tin
tow
ork
Italy
Gre
ece
Em
pl.
8 years
Em
pl.
0 Years
Sin
gle
trans.
to Work
Sin
gle
tras.
from
Work
Mult
i.tr
ans.
Em
pl.
8 years
Em
pl.
0 Years
Sin
gle
trans.
to Work
Sin
gle
tras.
from
Work
Mult
i.tr
ans.
age
44.5
639.4
042.9
533.5
637.2
4age
43.4
139.8
641.9
134.0
037.9
0le
ssfe
male
sec-
ondary
educati
on
0.1
80.2
00.2
10.1
00.1
8le
ssfe
male
sec-
ondary
educati
on
0.4
30.1
40.4
20.1
90.3
2
fem
ale
secondary
educati
on
0.5
80.2
00.4
90.6
10.5
3fe
male
secondary
educati
on
0.4
00.4
30.4
60.5
10.3
6
fem
ale
tert
iary
ed-
ucati
on
0.2
40.6
00.3
00.2
90.3
0fe
male
tert
iary
ed-
ucati
on
0.1
70.4
30.1
20.3
00.3
2
childre
nage
6-1
20.1
70.1
00.1
30.0
40.0
7ch
ildre
nage
6-1
20.1
60.0
00.1
10.0
90.1
0ch
ildre
nage
3-6
0.1
20.4
00.1
90.1
50.3
0ch
ildre
nage
3-6
0.1
20.1
40.1
90.3
40.2
6ch
ildre
nage
0-3
0.0
60.2
00.1
50.2
20.3
2ch
ildre
nage
0-3
0.0
60.1
40.0
90.3
80.2
9lo
gin
com
e10.3
710.1
610.3
39.7
710.1
0lo
gin
com
e15.1
915.0
915.0
714.8
015.0
3dis
abilitybenef
0.0
40.0
80.0
40.0
40.0
2dis
abilitybenef
0.1
30.0
50.4
40.0
90.1
2fa
milybenef
0.0
50.0
70.0
30.0
30.0
7fa
milybenef
0.3
50.0
70.4
00.8
60.5
6husb
and
sec-
ondary
educati
on
0.5
20.2
00.5
40.4
80.4
2husb
and
sec-
ondary
educati
on
0.2
80.4
30.3
20.5
10.3
6
husb
and
tert
iary
educati
on
0.1
80.1
00.1
70.1
00.1
9husb
and
tert
iary
educati
on
0.4
90.1
40.5
30.2
10.2
8
unem
plo
ybenef
0.0
30.0
00.0
20.0
40.0
4unem
plo
ybenef
0.2
33.2
20.1
40.1
50.4
2part
ic.
1.0
00.0
0-
--
part
ic.
1.0
00.0
0-
--
N.Y
ears
work
N.Y
ears
work
0-
100
0-
100
1-
-31.7
310.1
812.3
81
--
40.2
59.2
411.9
52
--
10.8
213.6
114.2
72
--
11.0
213.6
514.3
73
--
7.9
313.7
714.2
93
--
5.5
114.0
614.2
64
--
7.2
112.8
513.0
54
--
5.5
113.2
512.8
05
--
6.9
713.8
612.6
45
--
8.0
513.4
512.4
66
100
-7.2
113.7
712.5
16
--
4.6
611.8
511.2
77
9.1
313.7
711.1
17
--
5.0
813.0
511.7
28
18.9
98.1
89.7
68
100
-19.9
211.4
511.1
6
26
Tab
le9:
Sam
ple
char
acte
rist
ican
dye
ars
spen
tin
tow
ork
Spain
Port
ugal
Em
pl.
8 years
Em
pl.
0 Years
Sin
gle
trans.
to Work
Sin
gle
tras.
from
Work
Mult
i.tr
ans.
Em
pl.
8 years
Em
pl.
0 Years
Sin
gle
trans.
to Work
Sin
gle
tras.
from
Work
Mult
i.tr
ans.
age
44.1
335.8
341.7
835.4
837.3
6age
43.9
043.2
542.2
433.2
337.3
5le
ssfe
male
sec-
ondary
educati
on
0.5
30.0
80.4
30.4
30.4
2le
ssfe
male
sec-
ondary
educati
on
0.2
60.0
80.2
00.2
00.1
7
fem
ale
secondary
educati
on
0.2
60.1
70.2
00.1
90.2
3fe
male
secondary
educati
on
0.1
60.2
90.1
00.2
90.1
9
fem
ale
tert
iary
ed-
ucati
on
0.2
10.8
30.3
70.3
70.3
5fe
male
tert
iary
ed-
ucati
on
0.5
80.6
40.7
00.5
10.6
5
childre
nage
6-1
20.2
20.1
70.1
20.0
60.0
9ch
ildre
nage
6-1
20.1
60.2
50.1
00.0
30.0
6ch
ildre
nage
3-6
0.1
80.1
70.1
60.0
70.2
0ch
ildre
nage
3-6
0.0
90.2
50.1
10.1
90.3
1ch
ildre
nage
0-3
0.0
80.1
70.1
00.1
30.2
8ch
ildre
nage
0-3
0.0
30.0
40.0
30.3
90.2
7lo
gin
com
e14.8
714.3
814.7
114.3
614.6
3lo
gin
com
e14.6
113.5
814.5
314.2
714.3
6dis
abilitybenef
0.3
30.0
60.8
20.3
10.5
9dis
abilitybenef
0.2
52.0
70.4
50.1
10.3
0fa
milybenef
0.1
92.0
20.1
40.2
70.5
4fa
milybenef
4.6
26.1
24.8
23.7
34.6
4husb
and
sec-
ondary
educati
on
0.2
50.1
50.2
20.2
20.1
9husb
and
sec-
ondary
educati
on
0.1
50.1
20.1
10.1
00.1
5
husb
and
tert
iary
educati
on
0.4
60.0
70.4
10.3
00.3
8husb
and
tert
iary
educati
on
0.1
50.0
70.1
80.1
20.1
2
unem
plo
ybenef
0.2
60.0
40.1
60.5
20.4
7unem
plo
ybenef
0.0
60.0
60.3
80.2
20.1
2part
ic.
1.0
00.0
0-
--
part
ic.
1.0
00.0
0-
--
N.Y
ears
work
N.Y
ears
work
0-
100
0-
100
1-
-26.7
210.4
411.4
41
--
29.0
48.5
611.6
42
--
9.2
413.5
712.9
02
--
11.4
012.8
413.4
43
--
8.4
012.9
412.3
53
--
6.9
913.6
713.1
74
--
8.5
713.3
612.7
14
--
8.4
613.5
612.6
35
--
7.7
313.7
812.5
45
--
4.4
113.0
812.5
36
--
6.5
513.5
713.1
66
--
6.6
213.4
412.0
87
--
9.4
112.1
112.9
37
--
7.7
213.6
712.6
38
100
-23.3
610.2
311.9
68
100
-25.3
711.1
811.8
8
27
Tab
le10
:Sam
ple
char
acte
rist
ican
dye
ars
spen
tin
tow
ork
Aust
ria
Fin
land
Em
pl.
8 years
Em
pl.
0 Years
Sin
gle
trans.
to Work
Sin
gle
tras.
from
Work
Mult
i.tr
ans.
Em
pl.
8 years
Em
pl.
0 Years
Sin
gle
trans.
to Work
Sin
gle
tras.
from
Work
Mult
i.tr
ans.
age
44.3
540.7
142.8
833.7
037.0
6age
44.1
942.5
043.2
536.2
738.2
9le
ssfe
male
sec-
ondary
educati
on
0.1
70.0
20.1
40.0
20.1
7le
ssfe
male
sec-
ondary
educati
on
0.4
90.5
00.5
70.5
20.4
6
fem
ale
secondary
educati
on
0.6
80.5
70.6
90.7
20.7
0fe
male
secondary
educati
on
0.3
90.2
50.3
20.3
00.4
5
fem
ale
tert
iary
ed-
ucati
on
0.1
50.4
00.1
70.2
60.1
3fe
male
tert
iary
ed-
ucati
on
0.1
20.2
50.1
00.1
80.0
9
childre
nage
6-1
20.1
50.2
10.1
50.0
90.0
9ch
ildre
nage
6-1
20.1
20.2
50.1
20.0
90.0
6ch
ildre
nage
3-6
0.0
50.2
90.1
50.2
10.2
6ch
ildre
nage
3-6
0.1
70.5
00.2
40.3
00.3
0ch
ildre
nage
0-3
0.0
30.0
70.0
70.3
30.1
6ch
ildre
nage
0-3
0.0
80.2
50.1
50.2
30.2
6lo
gin
com
e12.8
312.7
912.7
512.6
912.7
3lo
gin
com
e11.8
312.1
011.8
911.8
111.7
7dis
abilitybenef
0.1
90.0
00.1
30.2
00.0
9dis
abilitybenef
1.6
30.0
01.2
51.1
81.0
4fa
milybenef
5.4
56.9
35.2
26.3
36.1
8fa
milybenef
4.8
35.8
55.1
34.6
95.2
7husb
and
sec-
ondary
educati
on
0.8
20.7
90.8
20.8
40.7
5husb
and
sec-
ondary
educati
on
0.3
70.2
50.4
40.5
00.5
0
husb
and
tert
iary
educati
on
0.1
10.0
70.1
10.0
90.1
7husb
and
tert
iary
educati
on
0.4
40.5
00.3
90.2
50.3
2
unem
plo
ybenef
0.2
41.0
50.3
30.8
80.3
7unem
plo
ybenef
0.4
72.0
30.6
51.5
90.8
6part
ic.
1.0
00.0
0-
--
part
ic.
1.0
00.0
0-
--
N.Y
ears
work
N.Y
ears
work
0-
100
0-
100
1-
-37.5
510.3
114.3
91
--
45.9
115.2
821.0
22
--
15.7
116.0
916.5
12
--
13.9
817.4
723.1
03
--
9.2
015.7
814.8
83
--
11.3
517.4
719.9
14
--
5.3
615.9
414.7
54
--
11.6
118.2
317.7
55
--
8.0
515.4
714.8
15
--
5.5
418.0
19.7
66
--
7.6
615.1
613.0
56
100
-11.6
113.5
48.4
67
100
-16.4
811.2
511.6
0
28
Table 11: Sample characteristic and years spent in to work
IrelandEmpl. 8years
Empl. 0Years
Singletrans. toWork
Singletras. fromWork
Multi.trans.
age 43.59 43.67 41.59 36.13 40.47less female sec-ondary education
0.38 0.17 0.21 0.09 0.21
female secondary ed-ucation
0.49 0.33 0.65 0.52 0.47
female tertiary edu-cation
0.12 0.50 0.15 0.39 0.33
children age 6-12 0.14 0.17 0.18 0.17 0.14children age 3-6 0.31 0.17 0.24 0.22 0.21children age 0-3 0.16 0.08 0.18 0.39 0.23logincome 10.00 10.10 10.00 9.88 9.97disabilitybenef 0.43 0.00 0.00 0.40 0.30familybenef 5.91 6.48 5.85 4.86 5.72husband secondaryeducation
0.43 0.17 0.50 0.61 0.42
husband tertiary ed-ucation
0.32 0.33 0.24 0.09 0.18
unemploybenef 0.15 2.89 0.18 1.19 0.80partic. 1.00 0.00 - - -N. Years work0 - 1001 - - 41.55 9.63 13.122 - - 14.53 13.64 14.663 - - 8.11 13.37 14.754 - - 8.11 13.90 14.625 - - 8.45 13.64 14.096 - - 5.74 12.83 11.847 - - 5.74 13.90 9.328 100 - 7.77 9.09 7.60
29
Tab
le12
:P
robit
esti
mat
esof
labor
forc
epar
tici
pat
ion,(d
epen
dent
vari
able
:A
ctiv
e)
Coun-
try
age
agetw
o
fe-
male
ter-
tiary
edu-
cati
on
fe-
male
sec-
ondary
edu-
cati
on
chil-
dre
nage
6-1
2
chil-
dre
nage
3-6
chil-
dre
nage
0-3
login
-com
e
fam
-ily-
benef
dis
-abili-
ty-
benef
expe-
rience
hus-
band
sec-
ondary
edu-
cati
on
hus-
band
ter-
tiary
edu-
cati
on
work
statu
s
un-
em
-plo
y-
benef
Ger-
many
0.5
34∗∗∗
-0.0
63∗∗∗
0.8
83∗∗∗
0.2
39∗∗
0.1
21∗∗
-0.7
65∗∗∗
-0.6
04∗∗∗
-0.5
84∗∗∗
-0.0
31∗∗∗
-0.1
01∗∗∗
-0.0
36∗∗∗
0.3
34∗∗∗
0.3
88∗∗∗
-0.1
96∗
-0.0
21∗∗
0.0
39
0.0
04
0.1
37
0.0
93
0.0
61
0.0
69
0.0
84
0.0
42
0.0
11
0.0
11
0.0
09
0.1
0.1
19
0.1
17
0.0
09
Den-
mark
0.7
40∗∗∗
-0.0
84∗∗∗
0.9
59∗∗∗
0.5
55∗∗∗
0.0
44
0.0
04
-0.4
64∗∗∗
-0.7
67∗∗∗
-0.0
55∗∗∗
-0.1
06∗∗∗
-0.0
56∗∗∗
0.4
91∗∗∗
0.7
20∗∗∗
-0.2
81
-0.0
26∗
0.0
68
0.0
07
0.1
42
0.1
21
0.1
32
0.1
11
0.1
29
0.0
93
0.0
18
0.0
14
0.0
17
0.1
31
0.1
49
0.2
30.0
16
Hol-
land
0.6
97∗∗∗
-0.0
72∗∗∗
0.3
95∗∗∗
0.2
04∗∗
0.1
39
-0.2
90∗∗∗
-0.3
18∗∗∗
-0.2
92∗∗∗
-0.1
05∗∗∗
-0.1
24∗∗∗
-0.1
62∗∗∗
0.1
22
0.1
7-
0.4
94∗
-0.0
07
0.0
61
0.0
07
0.1
28
0.0
95
0.0
88
0.0
89
0.1
04
0.0
51
0.0
13
0.0
12
0.0
04
0.1
08
0.1
24
0.2
65
0.0
16
Bel-
giu
m0.7
17∗∗∗
-0.0
81∗∗∗
1.5
83∗∗∗
0.3
73∗∗∗
0.2
27∗
0.0
73
-0.0
47
-0.2
45∗∗∗
-0.0
08
-0.0
38∗∗∗
-0.1
37∗∗∗
0.2
29∗
0.0
91
-0.6
30∗∗
-0.0
19
0.0
84
0.0
09
0.1
73
0.1
19
0.1
27
0.1
13
0.1
33
0.0
70.0
18
0.0
15
0.0
11
0.1
20.1
59
0.2
80.0
2
Fra
nce
0.6
54∗∗∗
-0.0
75∗∗∗
0.8
80∗∗∗
0.2
72∗∗∗
0.2
47∗∗∗
-0.2
04∗∗∗
-0.5
83∗∗∗
-0.2
46∗∗∗
-0.0
62∗∗∗
-0.0
41∗∗∗
-0.0
63∗∗∗
0.2
97∗∗∗
-0.1
22
-0.0
27
0.0
21∗
0.0
51
0.0
06
0.1
43
0.0
79
0.0
94
0.0
73
0.0
88
0.0
44
0.0
10.0
12
0.0
04
0.0
78
0.1
35
0.1
40.0
12
U.K
.0.5
57∗∗∗
-0.0
62∗∗∗
0.5
29∗∗∗
0.0
84
0.1
68
-0.4
17∗∗∗
-0.9
55∗∗∗
-0.1
00∗∗
-0.0
80∗∗∗
-0.1
50∗∗∗
-0.0
75∗∗∗
0.2
10∗
0.0
47
-0.0
40∗
0.0
54
0.0
06
0.1
02
0.1
01
0.1
03
0.0
79
0.0
95
0.0
49
0.0
14
0.0
12
0.0
08
0.1
12
0.1
03
0.0
21
Ire-
land
0.2
76∗∗∗
-0.0
29∗∗∗
1.0
19∗∗∗
0.3
73∗∗∗
0.2
14∗∗
-0.5
51∗∗∗
-0.2
14∗∗
-0.3
85∗∗∗
-0.1
08∗∗∗
-0.1
08∗∗∗
-0.1
37∗∗∗
0.1
66∗
0.4
31∗∗∗
-0.3
56∗∗
-0.0
30∗∗
0.0
72
0.0
08
0.1
69
0.1
05
0.0
84
0.0
88
0.1
01
0.0
68
0.0
19
0.0
16
0.0
19
0.0
93
0.1
54
0.1
44
0.0
13
Italy
0.7
45∗∗∗
-0.0
78∗∗∗
1.5
98∗∗∗
1.1
94∗∗∗
0.0
68
-0.1
59∗∗
-0.2
67∗∗∗
-0.1
61∗∗∗
-0.0
34∗∗
-0.0
04
-0.1
49∗∗∗
0.1
96∗∗
0.8
16∗∗∗
-0.2
88∗∗
-0.0
42
0.0
52
0.0
06
0.2
02
0.1
0.0
78
0.0
81
0.0
95
0.0
27
0.0
15
0.0
17
0.0
05
0.0
97
0.1
95
0.1
45
0.0
84
Gre
ece
0.6
31∗∗∗
-0.0
67∗∗∗
1.8
20∗∗∗
0.5
38∗∗∗
0.0
09
0.0
05
-0.3
70∗∗∗
-0.0
76∗∗∗
-0.0
09
-0.0
70∗∗
-0.1
37∗∗∗
0.1
67
0.4
72∗∗∗
-0.1
78
-0.0
09
0.0
69
0.0
08
0.1
70.1
30.0
94
0.1
19
0.1
42
0.0
22
0.0
23
0.0
30.0
06
0.1
17
0.1
57
0.1
75
0.0
29
Spain
0.4
43∗∗∗
-0.0
47∗∗∗
1.7
59∗∗∗
0.7
51∗∗∗
0.1
11∗
-0.1
57∗∗
-0.3
51∗∗∗
-0.1
43∗∗∗
-0.0
05
-0.0
1-
0.0
96∗∗∗
0.1
10.3
53∗∗∗
0.0
19
0.0
02
0.0
49
0.0
05
0.1
26
0.0
95
0.0
67
0.0
72
0.0
87
0.0
25
0.0
17
0.0
12
0.0
05
0.0
92
0.1
10.0
89
0.0
1Por-
tugal
0.4
69∗∗∗
-0.0
53∗∗∗
1.0
50∗∗∗
0.3
04∗
-0.0
57
-0.2
49∗∗
-0.3
25∗∗∗
-0.1
60∗∗∗
0.0
29∗
-0.0
41∗∗∗
-0.1
22∗∗∗
0.2
02
0.6
25∗∗
0.0
57
-0.0
36∗
0.0
59
0.0
06
0.2
26
0.1
73
0.0
89
0.1
0.1
26
0.0
33
0.0
15
0.0
14
0.0
06
0.1
72
0.2
59
0.1
78
0.0
19
Aus-
tria
0.9
35∗∗∗
-0.1
16∗∗∗
1.4
83∗∗∗
0.5
18∗∗∗
0.1
26
-0.9
23∗∗∗
-0.2
16
-0.3
58∗∗∗
-0.0
42∗∗
-0.0
75∗∗∗
-0.0
50∗∗∗
0.0
13
0.1
05
-0.0
32
-0.0
40∗∗
0.0
83
0.0
09
0.3
55
0.1
68
0.1
14
0.1
22
0.1
50.0
82
0.0
17
0.0
19
0.0
16
0.2
04
0.3
60.2
85
0.0
19
Fin
-la
nd
1.0
71∗∗∗
-0.1
17∗∗∗
0.7
84∗∗∗
0.1
49
-0.0
66
-0.0
27
-0.9
30∗∗∗
-0.8
03∗∗∗
-0.0
52∗∗∗
-0.0
70∗∗∗
-0.0
69∗∗∗
0.3
44∗∗∗
0.5
71∗∗∗
-0.2
42
-0.0
31∗∗
0.0
81
0.0
09
0.1
53
0.1
17
0.1
14
0.1
12
0.1
41
0.0
97
0.0
18
0.0
12
0.0
17
0.1
11
0.1
44
0.1
87
0.0
14
not
epva
lue:
.01
-**
*;.0
5-**
;.1
-*;
Stan
dard
devi
atio
nin
ital
ics,
Yea
rdu
mm
yva
riab
les
are
in-
clud
edbu
tno
tre
port
ed
30
Tab
le13
:M
argi
nal
effec
tsof
labor
forc
epar
tici
pat
ion
Coun-
try
age
agetw
o
fe-
male
ter-
tiary
edu-
cati
on
fe-
male
sec-
ondary
edu-
cati
on
chil-
dre
nage
6-1
2
chil-
dre
nage
3-6
chil-
dre
nage
0-3
login
-com
e
fam
-ily-
benef
dis
-abili-
ty-
benef
expe-
rience
hus-
band
sec-
ondary
edu-
cati
on
hus-
band
ter-
tiary
edu-
cati
on
un-
em
-plo
y-
benef
work
statu
s
Ger-
many
0.0
86∗∗∗
-0.0
10∗∗∗
0.1
68∗∗∗
0.0
54∗∗∗
0.0
14
-0.1
98∗∗∗
-0.1
25∗∗∗
-0.1
60∗∗∗
-0.0
11∗∗∗
-0.0
34∗∗∗
-0.0
06∗∗∗
0.0
65∗∗∗
0.0
75∗∗∗
-0.0
11∗∗∗
-0.0
68∗∗
0.0
06
0.0
01
0.0
11
0.0
10.0
13
0.0
15
0.0
18
0.0
07
0.0
01
0.0
02
0.0
01
0.0
12
0.0
13
0.0
02
0.0
27
Den-
mark
0.0
85∗∗∗
-0.0
10∗∗∗
0.1
41∗∗∗
0.0
87∗∗∗
0.0
14
0.0
07
-0.0
66∗∗∗
-0.1
76∗∗∗
-0.0
03∗
-0.0
25∗∗∗
-0.0
04∗∗∗
0.0
48∗∗∗
0.0
72∗∗∗
-0.0
13∗∗∗
-0.0
67∗
0.0
06
0.0
01
0.0
11
0.0
10.0
18
0.0
14
0.0
19
0.0
11
0.0
02
0.0
02
0.0
01
0.0
11
0.0
12
0.0
02
0.0
4H
ol-
land
0.1
17∗∗∗
-0.0
11∗∗∗
0.1
68∗∗∗
0.0
89∗∗∗
0.0
62∗∗∗
-0.0
74∗∗∗
-0.0
26
-0.0
78∗∗∗
-0.0
23∗∗∗
-0.0
23∗∗∗
-0.0
30∗∗∗
0.0
26
0.0
17
-0.0
07∗∗∗
-0.0
6
0.0
08
0.0
01
0.0
16
0.0
14
0.0
15
0.0
16
0.0
19
0.0
09
0.0
02
0.0
02
0.0
04
0.0
17
0.0
20.0
03
0.0
52
Bel-
giu
m0.1
10∗∗∗
-0.0
12∗∗∗
0.2
90∗∗∗
0.0
81∗∗∗
0.0
75∗∗∗
-0.0
73∗∗∗
-0.0
94∗∗∗
-0.1
38∗∗∗
-0.0
02
-0.0
16∗∗∗
-0.0
23∗∗∗
0.0
20.0
05
-0.0
20∗∗∗
-0.2
07∗∗∗
0.0
10.0
01
0.0
15
0.0
15
0.0
22
0.0
20.0
23
0.0
12
0.0
02
0.0
02
0.0
01
0.0
16
0.0
17
0.0
03
0.0
55
Fra
nce
0.1
00∗∗∗
-0.0
11∗∗∗
0.1
66∗∗∗
0.0
88∗∗∗
0.0
45∗∗∗
-0.0
37∗∗
-0.1
00∗∗∗
-0.1
12∗∗∗
-0.0
19∗∗∗
-0.0
19∗∗∗
-0.0
07∗∗∗
0.0
56∗∗∗
-0.0
20
-0.0
83∗∗∗
0.0
08
0.0
01
0.0
15
0.0
12
0.0
15
0.0
15
0.0
17
0.0
08
0.0
02
0.0
02
0.0
02
0.0
12
0.0
15
0.0
02
0.0
29
U.K
.0.0
47∗∗∗
-0.0
05∗∗∗
0.1
07∗∗∗
0.0
54∗∗∗
0.0
68∗∗∗
-0.1
15∗∗∗
-0.1
73∗∗∗
-0.0
58∗∗∗
-0.0
17∗∗∗
-0.0
38∗∗∗
0.0
10∗∗∗
0.0
33∗∗
-0.0
27∗∗
-0.0
19∗∗∗
0.0
08
0.0
01
0.0
10.0
13
0.0
16
0.0
16
0.0
18
0.0
07
0.0
02
0.0
01
0.0
02
0.0
14
0.0
10.0
04
Ire-
land
0.0
71∗∗∗
-0.0
08∗∗∗
0.3
20∗∗∗
0.0
72∗∗∗
0.0
54∗∗∗
-0.1
14∗∗∗
-0.0
18
-0.1
63∗∗∗
-0.0
23∗∗∗
-0.0
22∗∗∗
-0.0
11∗∗
0.0
79∗∗∗
0.0
73∗∗∗
-0.0
11∗∗∗
-0.0
94∗∗∗
0.0
12
0.0
01
0.0
20.0
14
0.0
19
0.0
16
0.0
19
0.0
11
0.0
03
0.0
03
0.0
05
0.0
14
0.0
19
0.0
02
0.0
28
Italy
0.0
54∗∗∗
-0.0
05∗∗∗
0.3
27∗∗∗
0.1
91∗∗∗
-0.0
03
-0.0
41∗∗∗
-0.0
25∗∗
-0.0
53∗∗∗
-0.0
12∗∗∗
-0.0
01
0.0
14∗∗∗
0.0
29∗∗∗
0.0
56∗∗∗
0.0
14
-0.0
96∗∗∗
0.0
05
0.0
01
0.0
20.0
09
0.0
10.0
09
0.0
11
0.0
04
0.0
02
0.0
02
0.0
01
0.0
08
0.0
15
0.0
12
0.0
15
Gre
ece
0.0
60∗∗∗
-0.0
06∗∗∗
0.2
90∗∗∗
0.0
66∗∗∗
0.0
13
-0.0
02
-0.0
28∗∗
-0.0
40∗∗∗
0.0
04∗∗
-0.0
03
0.0
01
-0.0
01
0.0
79∗∗∗
-0.0
07∗
-0.0
13
0.0
06
0.0
01
0.0
19
0.0
11
0.0
11
0.0
12
0.0
12
0.0
04
0.0
02
0.0
03
0.0
01
0.0
10.0
14
0.0
03
0.0
21
Spain
0.0
43∗∗∗
-0.0
04∗∗∗
0.3
87∗∗∗
0.1
57∗∗∗
-0.0
05
-0.0
47∗∗∗
-0.0
45∗∗∗
-0.0
40∗∗∗
-0.0
05
0.0
01
0.0
02
0.0
15
0.0
49∗∗∗
00.0
25
0.0
06
0.0
01
0.0
14
0.0
13
0.0
11
0.0
10.0
12
0.0
04
0.0
03
0.0
02
0.0
01
0.0
11
0.0
11
0.0
02
0.0
17
Por-
tugal
0.0
11
00.3
80∗∗∗
0.1
16∗∗∗
-0.0
31∗
-0.0
71∗∗∗
-0.0
86∗∗∗
-0.0
62∗∗∗
0.0
05∗∗
-0.0
15∗∗∗
0.0
23∗∗∗
0.0
76∗∗∗
0.0
16
-0.0
02
0.0
51
0.0
09
0.0
01
0.0
21
0.0
22
0.0
18
0.0
19
0.0
22
0.0
06
0.0
02
0.0
02
0.0
02
0.0
22
0.0
31
0.0
04
0.0
43
Aus-
tria
0.1
09∗∗∗
-0.0
14∗∗∗
0.2
89∗∗∗
0.1
05∗∗∗
-0.0
02
-0.2
27∗∗∗
-0.0
19
-0.1
04∗∗∗
-0.0
14∗∗∗
-0.0
20∗∗∗
0.0
02
0.0
05
-0.0
43
-0.0
15∗∗∗
-0.0
31
0.0
14
0.0
02
0.0
25
0.0
16
0.0
24
0.0
22
0.0
30.0
13
0.0
02
0.0
03
0.0
05
0.0
21
0.0
34
0.0
03
0.0
65
Fin
-la
nd
0.1
05∗∗∗
-0.0
11∗∗∗
0.1
19∗∗∗
0.0
26∗∗
0.0
02
0.0
19
-0.1
55∗∗∗
-0.1
68∗∗∗
-0.0
03
-0.0
16∗∗∗
0.0
09∗∗
0.0
47∗∗∗
0.0
70∗∗∗
-0.0
12∗∗∗
-0.1
33∗∗∗
0.0
10.0
01
0.0
14
0.0
12
0.0
17
0.0
16
0.0
24
0.0
11
0.0
02
0.0
01
0.0
04
0.0
11
0.0
13
0.0
02
0.0
39
not
epva
lue:
.01
-**
*;.0
5-**
;.1
-*;
Stan
dard
devi
atio
nin
ital
ics,
Yea
rdu
mm
yva
riab
les
are
in-
clud
edbu
tno
tre
port
ed
31
Tab
le14
:Est
imat
esfo
rth
em
onth
lyw
age
equat
ion,O
LS
model
Coun-
try
expe-
rience
exp2
fe-
male
ter-
tiary
edu-
cati
on
less
fe-
male
sec-
ondary
edu-
cati
on
unm
-plo
y-
ment-
benef
login
-com
e
dis
-abili-
ty-
benef
fam
-ily-
benef
chil-
dre
nage6-
12
chil-
dre
nage
3-6
chil-
dre
nage
0-3
Ger-
many
0.1
89∗∗∗
-0.4
64∗∗∗
0.7
60∗∗∗
-0.3
63∗∗∗
-0.0
20∗∗
-0.5
02∗∗∗
-0.1
04∗∗∗
-0.0
26∗∗
0.1
53∗∗∗
-0.8
57∗∗∗
-0.8
26∗∗∗
0.0
23
0.0
42
0.1
05
0.0
99
0.0
09
0.0
44
0.0
14
0.0
10.0
58
0.0
89
0.1
03
Den-
mark
0.1
78∗∗∗
-0.4
40∗∗∗
0.3
19∗∗∗
-0.7
62∗∗∗
-0.0
34∗∗
-0.3
65∗∗∗
-0.1
19∗∗∗
-0.0
43∗∗∗
0.1
93∗
-0.0
62
-0.4
62∗∗∗
0.0
41
0.0
70.0
86
0.1
31
0.0
17
0.0
81
0.0
17
0.0
13
0.1
02
0.0
99
0.1
24
Fra
nce
-0.0
31∗∗∗
-0.0
76∗∗∗
0.8
12∗∗∗
-0.1
14
0.0
18
-0.1
45∗∗∗
-0.0
42∗∗∗
-0.0
44∗∗∗
0.3
74∗∗∗
-0.3
62∗∗∗
-0.7
85∗∗∗
0.0
11
0.0
21
0.1
62
0.0
81
0.0
12
0.0
41
0.0
12
0.0
09
0.0
84
0.0
77
0.0
95
U.K
.0.0
44∗∗
-0.2
34∗∗∗
0.4
92∗∗∗
-0.1
44∗
-0.0
31
-0.0
01
-0.1
44∗∗∗
-0.0
72∗∗∗
0.2
25∗∗∗
-0.6
03∗∗∗
-1.0
32∗∗∗
0.0
20.0
37
0.0
80.0
78
0.0
20.0
45
0.0
15
0.0
13
0.0
65
0.0
87
0.1
1Ir
e-
land
-0.0
09
-0.0
80∗∗
0.9
67∗∗∗
-0.5
39∗∗∗
-0.0
52∗∗∗
-0.0
13
-0.0
80∗∗∗
-0.0
65∗∗∗
0.3
25∗∗∗
-0.5
72∗∗∗
-0.3
24∗∗∗
0.0
26
0.0
39
0.1
67
0.0
96
0.0
13
0.0
52
0.0
17
0.0
16
0.0
89
0.0
94
0.1
08
Italy
-0.0
99∗∗∗
0.0
29∗∗
0.7
38∗∗∗
-0.5
63∗∗∗
-0.0
69
-0.0
69∗∗∗
-0.0
01
-0.0
06
0.0
66
-0.1
34∗∗∗
-0.2
73∗∗∗
0.0
10.0
14
0.2
23
0.0
73
0.0
65
0.0
15
0.0
07
0.0
08
0.0
44
0.0
49
0.0
6
Gre
ece
-0.0
85∗∗∗
0.0
39∗∗
0.7
04∗∗∗
-0.4
25∗∗∗
-0.0
24
-0.0
2-
0.0
22∗∗
-0.0
16
0.0
64
-0.1
24∗
-0.3
36∗∗∗
0.0
13
0.0
18
0.1
21
0.0
78
0.0
18
0.0
21
0.0
11
0.0
15
0.0
64
0.0
71
0.0
83
Spain
-0.0
69∗∗∗
0.0
21
1.0
53∗∗∗
-0.7
51∗∗∗
-0.0
12
-0.0
40∗∗
-0.0
11
-0.0
16
0.1
41∗∗
-0.2
78∗∗∗
-0.4
40∗∗∗
0.0
11
0.0
16
0.1
38
0.0
91
0.0
08
0.0
17
0.0
09
0.0
15
0.0
56
0.0
61
0.0
76
Por-
tugal
-0.0
57∗∗∗
-0.0
35∗
0.6
01∗∗∗
-0.7
85∗∗∗
-0.0
26
-0.0
87∗∗∗
-0.0
35∗∗∗
0.0
39∗∗∗
0.0
17
-0.2
29∗∗∗
-0.3
03∗∗∗
0.0
12
0.0
20.1
77
0.1
50.0
16
0.0
23
0.0
11
0.0
12
0.0
62
0.0
79
0.1
06
Aus-
tria
0.1
97∗∗∗
-0.5
20∗∗∗
1.1
47∗∗∗
-0.4
41∗∗∗
-0.0
37∗∗
-0.1
88∗∗∗
-0.0
57∗∗∗
-0.0
10.0
67
-0.9
47∗∗∗
-0.5
65∗∗∗
0.0
37
0.0
61
0.2
39
0.1
46
0.0
17
0.0
54
0.0
19
0.0
14
0.0
92
0.1
31
0.1
79
Fin
-la
nd
0.2
94∗∗∗
-0.6
66∗∗∗
0.6
25∗∗∗
-0.2
02∗
-0.0
30∗∗
-0.3
57∗∗∗
-0.0
61∗∗∗
-0.0
63∗∗∗
-0.0
38
-0.0
55
-0.8
04∗∗∗
0.0
28
0.0
51
0.0
96
0.1
10.0
14
0.0
66
0.0
12
0.0
15
0.0
76
0.1
23
0.1
27
not
epva
lue:
.01
-**
*;.0
5-
**;
.1-
*;Yea
rdum
my
variab
les
are
incl
uded
but
not
re-
por
ted.
Sta
ndar
der
ror
inital
ics
32
Table 15: Estimates for the monthly wage equation, Heckmanmodel
Countryexperi-ence
exp2femaletertiary
education
lessfemale
secondaryeducation
lambda
Germany 0.022∗∗∗ -0.033∗∗ 0.429∗∗∗ -0.067∗ -0.465∗∗∗
0.008 0.015 0.033 0.038 0.089Denmark 0.005 -0.022∗ 0.130∗∗∗ -0.093∗∗∗ -0.152∗∗∗
0.006 0.012 0.019 0.032 0.051France 0.015∗∗∗ -0.022∗∗ 0.452∗∗∗ -0.268∗∗∗ -0.002
0.005 0.01 0.035 0.031 0.054U.K. 0.001 -0.011 0.293∗∗∗ -0.167∗∗∗ -0.246∗∗∗
0.006 0.012 0.039 0.038 0.077Ireland 0.004 -0.026 0.550∗∗∗ -0.429∗∗∗ -0.159
0.017 0.031 0.046 0.052 0.124Italy 0.027∗∗∗ -0.049∗∗∗ 0.171∗∗∗ -0.347∗∗∗ 0.014
0.004 0.009 0.021 0.021 0.047Greece 0.038∗∗∗ -0.083∗∗∗ 0.241∗∗∗ -0.354∗∗∗ 0.104∗∗∗
0.006 0.016 0.03 0.038 0.035Spain 0.050∗∗∗ -0.096∗∗∗ 0.370∗∗∗ -0.472∗∗∗ 0.057
0.009 0.02 0.038 0.044 0.039Portugal 0.046∗∗∗ -0.099∗∗∗ 0.483∗∗∗ -0.652∗∗∗ 0.116∗∗∗
0.006 0.014 0.035 0.033 0.029Austria 0.031∗∗ -0.055∗∗ 0.396∗∗∗ -0.205∗∗∗ -0.106
0.013 0.026 0.058 0.05 0.083Finland 0.006 -0.011 0.236∗∗∗ -0.036 -0.145∗∗
0.006 0.011 0.021 0.024 0.062note pvalue: .01 - ***; .05 -**; .1 - *;Year dummy vari-ables are included but notreported. Standard error initalics
33
Table 16: Estimates for the hourly wage equation, 2SLS model
Coun-try
expe-rience
exp2lambda
lessfe-
malesec-
ondaryedu-
cation
fe-malesec-
ondaryedu-
cation
fe-maleter-tiaryedu-
cation
Ger-many
0.016∗ -0.021-
0.227∗∗∗ 0.474∗∗∗0.128 0.069
0.009 0.018 0.067 0.154 0.153 0.155Den-mark
0.001 -0.007-
0.109∗∗∗-
0.132∗∗∗-
0.230∗∗∗
0.006 0.012 0.030 0.013 0.016 0.025The
Nether-land
0.014 -0.024-
0.233∗∗∗-0.136
-0.436∗∗∗
-0.467∗∗∗
0.010 0.021 0.059 0.124 0.125 0.129Bel-gium 0.025∗∗
-0.036 -0.0380.158∗∗
-0.112-
0.326∗∗∗
0.010 0.023 0.024 0.067 0.070 0.074
France 0.064∗∗∗-
0.107∗∗∗0.073
0.624∗∗∗0.112 -0.145
0.010 0.020 0.050 0.125 0.124 0.124
U.K. 0.004 -0.008-
0.206∗∗∗ 0.543∗∗∗ 0.297∗∗0.118
0.010 0.018 0.061 0.126 0.127 0.125Ire-land
-0.036∗∗
0.058∗ -0.038-
0.201∗∗∗-
0.736∗∗∗-
1.054∗∗∗
0.016 0.031 0.115 0.040 0.036 0.050
Italy0.064∗∗∗
-0.119∗∗∗ 0.020** 0.723∗∗∗ 0.529∗∗∗
0.190∗
0.009 0.0200.047**
0.101 0.099 0.099
Greece 0.106∗∗∗-
0.238∗∗∗ 0.147∗∗∗ 0.528∗∗∗ 0.290∗∗∗
0.013 0.0340.041**
0.043 0.042
Spain0.111∗∗∗
-0.200∗∗∗ 0.138∗∗∗ 0.819∗∗∗ 0.393∗∗∗
0.013 0.026 0.042 0.038 0.042Por-tugal 0.058∗∗∗
-0.109∗∗∗ 0.108∗∗∗ 0.130∗∗∗
-0.443∗∗∗
-0.984∗∗∗
0.009 0.019 0.027 0.042 0.049 0.043Aus-tria 0.048∗∗
-0.071-
0.023**-
0.429∗∗∗-
0.638∗∗∗
0.023 0.0440.068** 0.060***
0.065
Fin-land 0.016∗∗
-0.024 -0.046-
0.249∗∗∗-
0.304∗∗∗
0.008**0.016 0.041 0.017 0.022
note pvalue: .01 - ***; .05 -**; .1 - *;Year dummy vari-ables are included but notreported. Standard error initalics
34
Tab
le17
:V
aria
ble
defi
nit
ions
Vari
able
Definit
ion
Wom
enSam
ple
ofw
om
enm
arr
ied
age
30-5
4
Wom
en1
=1
ifth
ew
om
an
work
or
seek
for
ajo
b,0
oth
erw
ise
Act
ive
=1
ifw
om
enpart
icip
ati
on
tola
bor
mark
et,0
oth
erw
ise
Dum
∗D
um
my
for
tim
e
Age
Age
ofin
div
iduals
Aget
wo
Square
ofage
/10
Fem
ale
educa
tion
less
than
seco
ndary
=1
ifth
ew
om
an
has
thir
dle
vel
educa
tion,0
oth
erw
ise
Fem
ale
seco
ndary
educa
tion
=1
ifth
ew
om
an
has
seco
nd
level
ofed
uca
tion,0
oth
erw
ise
Fem
ale
tert
iary
educa
tion
=1
ifth
ew
om
an
has
firs
tle
vel
ofed
uca
tion,0
oth
erw
ise
Childre
nage
6-1
2=
1if
the
wom
enhas
childre
nle
ssth
an
12yea
rsold
,0
oth
erw
ise
Childre
nage
3-6
childre
nw
ith
age
3–6
yea
rs
Childre
nage
0-3
childre
nw
ith
age
0–3
yea
rs
Ow
ner
house
Dum
my
=1
ifth
ehouse
isow
ner
Adult
hou
Num
ber
ofadult
sin
the
house
hold
Fam
ilyben
efLog
fam
ily
ben
efits
inth
ehouse
hold
Login
com
eLog
ofnola
bor
inco
me
inth
ehouse
hold
excl
usi
on
ofin
com
ew
ife
Dis
abilityben
efLog
realdis
ability
ben
efits
inth
ehouse
hold
Logw
age
Natu
rallo
gofth
ew
om
en’s
hourl
yre
alw
age
Exper
ience
Pote
nti
alex
per
ience
(pre
sent
age
-age
when
start
edw
ork
)
Exp2
Square
ofpote
nti
alex
per
ience
Ten
ure
Yea
rsofex
per
ience
incu
rren
tjo
b
Ten
2Square
ofte
nure
Husb
and
less
than
seco
ndary
school
=1
ifth
ehusb
and
has
firs
tle
vel
ofed
uca
tion,0
oth
erw
ise
Husb
and
seco
ndary
educa
tion
=1
ifth
ehusb
and
has
seco
nd
level
ofed
uca
tion,0
oth
erw
ise
Husb
and
tert
iary
educa
tion
=1
ifth
ehusb
and
has
seco
nd
level
ofed
uca
tion,0
oth
erw
ise
Unem
plo
yben
efLog
ofunem
plo
ym
ent
ben
efit
ofhusb
and
Work
statu
sSta
tus
ofw
ork
ofhusb
and
Lam
bda
Inver
seM
ills
rati
o
Lam
∗Lam
bda
inte
ract
ion
wit
hdum
my
tim
e
35
