Kids or courses? Gender differences in the effects of active labor market policies

Abstract

We investigate active labor market programs in Austria. We find only small effects, if any, for most of the programs. However, the programs may have unintended consequences for women. In particular for younger women, a key effect of the programs and one reason for the male–female effect differential that is observed in the literature is to reduce or postpone pregnancies and to increase their attachment to the labor force. Furthermore, the variables capturing pregnancies and times of parental leave play a key role in removing selection bias.

1 Introduction

There seems to be a consensus emerging that women benefit more from labor market programs than men, e.g., see the surveys by Bergemann and van den Berg (2006) for Europe and Heckman et al. (1999) for the USA. Our comprehensive evaluation of the Austrian active labor market programs shows that the effect differentials disappear once information on pregnancies and parental leave is incorporated or once the analysis is focused on unemployed with age above 40 years where pregnancies and parental leaves are hardly an issue anymore.

Many recent European studies have emphasized the role of effect heterogeneity on the program level.¹ In terms of participant heterogeneity Puhani (1999) and Kluve et al. (1999, 2008) find sex-specific program effects for Poland. Friedlander et al. (1997) and Heckman et al. (1999) feature sex differences for the US and other western economies. Lechner et al. (2009) look at employment effects for certain subgroups of participants in West Germany. They find effect heterogeneity with respect to residence, previous occupation, and sex. For East Germany, Lechner et al. (2007) find that for some training programs, the employment effects for women were much larger than for men. They attributed this heterogeneity to specifics of the selection process that resulted in a higher probability of men being trained with skills for the construction sector which then collapsed. However, such a precise identification of the reasons for gender differences is not always possible, and the puzzle remains in many other studies. Bergemann and van den Berg (2006) survey 15 studies on effect differentials for men and women in Europe. Thirteen of those studies report effect premia for women. Heckman et al. (1999) survey 16 studies for the USA and also provide broad evidence for effect premia for women in terms of earnings. The key explanations that are put forward are gender differences with respect to labor supply elasticity, eagerness to learn, responsiveness to wage changes, and with respect to the larger choice set for women, i.e., including times of parental leave in addition to work and leisure. Their overall conclusion is that labor market programs seem to work better for women in countries where the female labor force participation rate is relatively small, which is also the case in Austria.

Due to a unique and informative database of the Austrian labor force, we show that those estimated differentials between men and women consist of two components. The first component is a selection bias due to the lack of controlling for the occurrence of pregnancies before or at the start of the programs, leading to more pregnancies in the group of non-participants than in the group of participants. Thus, estimated effects that ignore this information show biases in favor of the programs. Second, the remaining differential in the employment effect appears because program participation postpones or reduces fertility, which in turn implies that programs have an adverse effect on other policies that are designed to foster birthrates. Once those two components are accounted for, the effect heterogeneity between men and women can be explained. In that context, it is interesting to note that information on pregnancy status has not been available in many studies surveyed by Bergemann and van den Berg (2006). In addition, one third of those studies even lacked information on dependent children. Thus, linking our findings to the latter study, we demonstrate that in countries like Austria with a low female labor force participation rate,² it is even more important to have information about the outside opportunities of women, in particular times of parental leave.

The underlying data are made available by the Federation of Austrian Social Insurance Institutions and the Austrian Public Employment Service. We possess a rich set of information on the employment history, times of unemployment, the counseling process, personal characteristics, parental leaves, and times of program participation as well as regional characteristics. Assuming conditional independence of the selection mechanism and potential outcomes, we employ an advanced version of a semi-parametric matching estimator that is very popular in the policy evaluation literature and was used previously, for instance, by Lechner et al. (2009).

The paper is organized as follows. Section 2 briefly summarizes the institutional background of the Austrian labor market policy. Section 3 introduces the underlying data and identification strategy as well as a first description of the population of interest. The estimation method and first results of the program allocation analysis can be found in Section 4. Section 5 shows estimation results and omitted variable checks and Section 6 concludes. Details concerning the data, the estimation method, and results are provided in an Internet appendix that can be downloaded from www.sew/lechner/at.

2 Labor market policies in Austria

The Public Employment Service Act constitutes the legal foundation of the Austrian labor market policy. It determines the objectives of the Public Employment Service by defining the following six principles. (1) The Public Employment Service has to match job seekers and vacancies efficiently, (2) remove any barrier that prevents this matching, (3) increase the flow of information about potential matches, (4) mitigate quantitative and qualitative differences between labor demand and supply, (5) secure sustainable employment, (6) and provide funds for the unemployed in case of a job loss. As many other countries, Austria uses active and passive labor market policies to implement those principles.

2.1 Passive labor market policy

Passive labor market policy in Austria is designed to cover earning losses caused by various types of non-employment. To receive unemployment benefit payments, the unemployed have to be registered at the Public Employment Service, be eligible and willing to work, and have a predefined record of employment with unemployment insurance (UI) contributions. The pre-unemployment employment requirement is a cumulated UI contribution period of 52 weeks within the last 24 months for the first draw on benefits. Subsequent benefits require 28 weeks within the previous 12 months. Exceptions regarding age exist.³ The standard replacement ratio is 55% of the former net income and the minimum entitlement period is 20 weeks. Extra payments depending on family status and the number of children may be added. After unemployment benefits expire, the unemployed are entitled to unemployment assistance if they are still available for work. Unemployment assistance payments are means-tested, but are not subject to a time limit.

2.2 Parental leave subsidies

There are three different types of subsidies for women in parental leave. Eight weeks before and after the scheduled confinement, women receive so-called confinement benefits, which are granted up to the average net wage of the previous 3 months.⁴ After the expiration of those benefits (and before January 2002), women had to apply for parental leave benefits. This benefit was granted subject to the same UI contribution requirements as unemployment benefits, which excluded women who failed to prove the required previous contribution times.⁵ After January 2002, women may apply for childcare benefit which is no longer linked to previous contribution times and granted to everyone with an amount of 14.53 euro a day for a maximum period of 30 months.⁶ Eighteen months of this entitlement period are counted as regular contribution times to the pension schemes. All periods in which such benefits have been paid out are recorded in the social security records. Therefore, this information is available in the current study. We use this information to construct the theoretical start date of a pregnancy by employing the usual 9-month pregnancy period as illustrated in Fig. 1.

Fig. 1

Construction of the pregnancy start date

2.3 Active labor market policy

Apart from counseling and placement services, there are a variety of active labor market programs offered by the Public Employment Service in order to overcome specific reintegration obstacles of the unemployed.

The first group of programs promotes vocational flexibility. Those programs can be classified into orientation measures, active job search, job coaching, and qualification measures. Orientation measures assess the individual situation and aptitude of the unemployed person and serve as an upfront decision process for subsequent (re-)integration activities. Active job search aims at improving job acquisition skills, like writing an application or interview training. Job coaching deals with the long-term unemployed and groups with specific placement handicaps, like disabled persons, by means of a combination of counseling, qualification, and on-the-job training. Qualification measures comprise further education and various forms of vocational training. The range of program stretches from courses requiring only basic skill levels, like catering courses, to high-level software courses and up to formal educational and vocational degrees. Participants are either allocated by the Public Employment Service or find a program on their own and then apply for course subsidies with the Public Employment Office.

Another group of programs consists of the so-called job creation schemes. Socioeconomic enterprises and non-profit sector projects are designed to capture the long-term unemployed individuals and other problematic cases, like, for instance, individuals with psychological diseases, etc. Such programs offer a quasi-realistic work environment. In principle, those jobs are restricted to 1 year. They are sometimes augmented by socio-pedagogical treatment to gradually reintegrate the participants into the regular labor market. With integration subsidies, the Public Employment Service supports special groups of unemployed, like the disabled, long-term unemployed, or older people, by means of a wage subsidy for the first 150 days of a new employment. Encouraging individuals to become self-employed, the Public Employment Service offers the so-called business start-up program, which supports young entrepreneurs starting with a business idea until the actual foundation of their own firm. Furthermore, the Public Employment Service supports young individuals who have been searching unsuccessfully for an apprenticeship by providing different courses aimed at endowing the participants with human capital that is similar to the level of the first year of a regular apprenticeship of 3 years. A hybrid form of labor market programs is the so-called beneficence for labor, which is organized in collaboration with local firms in order to compensate sudden local excess demand or supply of workers caused by, e.g., business foundation or sudden firm closures.⁷ Finally, there are also qualification programs for employees to enhance sustainable employment for workers threatened by unemployment.

To get an impression of the magnitude of the programs, Table 1 reports the overall expenditures and number of participants per program type. It can be seen that active job search and qualification measures are the most important programs with respect to the number of participants. Over time, we observe that the number of active job search programs increases, whereas participation in qualification measures drops to 70% in 2002 compared to 2000. Integration subsidies feature far less participants but a considerable amount of expenditures because the respective subsidies can amount up to 100% of the wage bill of the new employment. The same holds for socioeconomic enterprises and non-profit sector projects which are also characterized by high average costs per participation of, for instance, over 11,000 euros for non-profit sector projects in 2000. In contrast, course subsidies appear to be on average a rather low-cost measure of around 600 euros per participation in 2000.⁸ For the year 2002, we also calculated expenditures per day. Again, we can see that socioeconomic enterprises, non-profit sector projects, and integration subsidies are the most expensive measures per day. The former two are even more costly since the respective expenditures do not only cover the wage of the participants but also the coverage of potential losses of the job-creating firm. Beneficences for labor have very low costs per day since most of the costs are carried by the cooperating firms.

Table 1 Expenditures and number of participants by program type

3 Data and identification strategy

3.1 Data

The three data sources that are used for the program evaluation comprise administrative registers from the Federation of Austrian Social Insurance Institutions and the Austrian Public Employment Service, including information from the program register data. We make use of all the Austrian population instead of a random sample as is usually the case. Using the population increases computation time considerably, but maximizes the precision of our evaluation results. For example, due to the resulting large number of observations, it will be possible to non-parametrically estimate program effects fairly precisely even for smaller subgroups of participants and programs.

We use the Social Insurance data to obtain information about times in employment (employment states: employed, self-employed, or civil servants; with earnings and employer information), retirement, and other periods relevant for social insurance contributions from 1985 to 2005. Since all financial support during times of parental leave are granted relative to the scheduled confinement date, we identify not only times of parental leave but also the pregnancy status for women, which will be a key control variable in the analysis. Information about the counseling process of the Public Employment Service, i.e., beginning and end of an unemployment period, regional identifiers, personal characteristics like sex, marital status, nationality, current profession and desired profession, education, disability status, number of job offers received, or times of previous labor market program participation, is available from the Public Employment Service data from 1990 until 2005. Finally, the Public Employment Service data give us detailed information about the type of labor market program from 2000 to 2005.

Most of the data are available on a daily basis, but to condense the information into a manageable form, we chose to aggregate the daily information into 2-week intervals (which is more precise than the usual grids used by evaluation studies that are commonly based on monthly, quarterly, or even yearly information).

However, although this dataset is well suited for an evaluation exercise, the nature of the data nevertheless imposes some restrictions with respect to the definition of the participation window and the follow-up period, which will be discussed in detail later on. Furthermore, we have to rely on a quite broad definition of the type of qualification measures.⁹

3.2 Identification strategy

In the current analysis, we concentrate on the average program effects compared to non-participation. The identification problem in non-experimental program evaluations is that participants in one program differ, sometimes substantially, from potential comparison observations in the non-participation state with respect to characteristics that may influence the outcome variables under inspection as well. Since our data are very informative but contains no obvious instrumental variable, i.e., a variable that influences the outcome only by influencing the participation decision, we chose the so-called conditional independence assumption (CIA) to overcome the resulting identification problem. It states that if we are able to observe all factors that jointly influence the participation decision and the outcomes, then, conditional on those factors, we can learn the (potential) non-participation outcomes of the participants from the observable non-participation outcomes of the non-participants with the same distribution of characteristics, which identifies our parameter of interest. This identification strategy goes back to Rubin (1974) for the case of comparing participants to non-participants. Imbens (2000) and Lechner (2001) generalize this idea to the case of multiple treatments and provide similar identification conditions. However, the CIA strongly hinges on the availability of a comprehensive set of covariates. To justify its applicability in the present framework, we now discuss three important issues of program allocation: the allocation decision of the caseworker of the Public Employment Service, the willingness and collaboration of the unemployed individual, as well as relevant eligibility criteria in general.

The standard allocation procedure is initially based on a face-to-face interview between the unemployed person and the caseworker. Several aspects, like education, family affairs, past behavior on the labor market, features of the last employment, and individual program history, are discussed. As a result of this interview and in light of the local characteristics of the labor market, the caseworker decides whether or not the unemployed person should be sent into a specific labor market program. Multiple refusals of program offers can lead to temporary suspensions of the benefit payments. However, such punishments are rarely observed in the data. The data contain a large set of covariates that are suitable to map most of those aspects. In addition to variables like age, sex, foreigner status, family status, education, information on the job, and the previous sector of employment, we construct a rich set of variables that summarizes the entire labor market history of the unemployed person. This history covers up to 15 years before the actual entry into unemployment under inspection on a fine 2-week scale. We construct variables covering previous times of (un-)employment, program participation, times of childcare,¹⁰ military service, times of non-registration, which we call out-of-labor-force times (OLF) from now on. By means of this, we are also confident to possess suitable proxy variables for unobservable variables like motivation or the general attitude towards employment. In addition, we use characteristics of the local labor markets relevant for each specific individual.¹¹

From the perspective of the unemployed, all points mentioned above certainly play a role for the participation decision. Another component of the individual consideration might be the question whether the currently unemployed person was satisfied with the kind of his/her former job. Since we observe data on the current and the desired profession, we are able to identify or at least approximate this feature. Furthermore, since unemployment insurance contributions are paid during the time of program participation, the individual decision might take into account the remaining time of the unemployment benefits. Thus, we also compute the remaining unemployment benefits claim at the time of (hypothetical) program entry. Another important determinant will certainly be the existence of dependent children, which is available in this data. Thus, we control for all previous times in parental leave (and thus for the complete childbearing history), but also for the month of pregnancy for women during the unemployment spell under consideration, i.e., the pregnancy status right before the hypothetical program entry, which has not been available in previous studies on effect differentials for men and women. Furthermore, we assess all previous control variables to be also relevant for the outcome variable “parental leave” which is used later on, especially, age, marital status, and education. The desire for vocational change is also a key determinant, since unemployed who desire a vocational change are presumably less focused on family planning.

From an institutional point of view, a key eligibility requirement for program participation is being unemployed (whether the individual receives unemployment benefits or unemployment assistance is irrelevant). We will resolve this issue by choosing an adequate inflow of eligibles into unemployment. Finally, there are number of guidelines that define specific types of unemployed to receive preferential treatment in certain programs. We explicitly account for this by tailoring the selection models to the comparisons of the specific programs under consideration, i.e., using flexible specifications that include key determinants for the selection into the respective programs. Overall, we plausibly pin down most important factors that drive the allocation decision and the potential outcomes. Thus, assuming CIA appears to be a credible identification strategy.

3.3 Definition of the population and the programs of interest

To be included in our evaluation, programs and the respective participants have to meet five requirements. First, the identification strategy strongly hinges on the existence of a long labor market history before the entry into unemployment. Second, the follow-up period after program attendance should not be influenced by perturbing events like the possibility of (early) retirement. As a result of those arguments, we concentrate on the age groups of the labor force between 25 and 50 years. Third, the data must provide all relevant information about the selection into the different labor market programs. Fourth, since we employ non-parametric estimation techniques, the number of observations in the different programs has to be sufficiently large. Finally, we require the program content to be more substantial than the usual counseling process. Under those restrictions, we end up with six program types that can be credibly evaluated: socioeconomic enterprises, non-profit sector projects, job coaching, active job search, qualification measures, and course subsidies.

The nature of the data drives the definition of the population used in the estimation. Information on program participation is only available from the year 2000 onwards. The follow-up period is restricted by the end of the observation period in 2005. Hence, we consider the first inflow of individuals¹² from employment into unemployment or one of the six labor market programs between 2000 and 2002.¹³ By means of this, we observe enough participants in each program and have a follow-up period of at least 3 years, which enables us to identify effects that are less affected by initial lock-in effects. Doing so, we end up with a population of 797,034 persons.

The next step is to divide all persons who passed this criterion into participants and non-participants. In that population, we define a participant to be a person who took part in a program before the end of 2002 (without an employment spell between inflow and participation). Thus, non-participants are persons who moved from employment into unemployment and have not been allocated to a program between 2000 and 2002 or took up an employment before being allocated to a program. The resulting numbers of observations are shown in the first row of Table 2.

Table 2 Selection of the population used in the estimation

However, we impose a number of further restrictions. Some control variables, like the remaining unemployment insurance benefit claim or the duration in unemployment before the entry into a program, require a reference data (artificial program start date) for the non-participants. To obtain such a reference date, we employ an approach suggested by Lechner (1999). We simulate start dates for the non-participants by drawing start dates from the distribution of the participants. If the non-participant is not eligible at the simulated reference date, then this non-participant is not considered in the evaluation. The fairly drastic reduction in the number of observations is not particularly important though, since (1) participants are in abundant supply and (2) they serve only as comparison observations for participants and are not interesting per se. In Austria, temporary layoffs are widely spread. Especially workers in the tourism or construction sector are laid off with a more or less binding reemployment guarantee. Since such reemployment guarantees may differ substantially with respect to how binding they are and since we do not observe such differences that most likely influence participation and labor market outcomes, we require that all persons are laid off permanently. The age restriction, for reasons described above, is applied as well. Furthermore, we require the duration of the last employment before the inflow into our sample to be longer than 2 months. By means of this, we make sure that prior participants in subsidized employment are not employed further for a couple of days after the end of the program, which would cause a short employment spell before becoming unemployed again. We observe that especially the age restriction reduces the number of participants and non-participants considerably. The resulting number of observations, however, still allows reliable results from non-parametric estimation.

3.4 A descriptive analysis of the selection into the programs

As a first description of the selection process, Table 3 shows mean characteristics by participation status for selected variables.¹⁴ In general, the numbers exhibit many aspects of the institutional environment in Austria as well as the general allocation policy of the Public Employment Service. Except for socioeconomic enterprises and active job search, the fraction of female participants is above 50%. Qualification measures even have a female participation rate of 62%, which underlines gender mainstreaming requirements anchored in the Guiding Principles of the Federal Ministry of Economics and Labor. Consequently, those participants feature higher average mean durations in times of parental leave before the defining unemployment period. We computed the pregnancy state for women right before the hypothetical program start. It can be seen that pregnancy is hardly an issue for the program groups, except for non-profit sector projects. Furthermore, we observe that the fraction of pregnant female non-participants is higher compared to female participants of all programs so that pregnancy status is indeed an important variable to control for in the remainder of the analysis.

Table 3 Mean characteristics of selected variables (mean or share in %)

For programs which are specifically designed for unemployed with certain reintegration obstacles, like socioeconomic enterprises, non-profit sector projects, and job coaching, we observe a fraction of disabled participants of almost 22% which is nearly three times higher than for active job search and more than two times higher than for qualification measures and course subsidies. Participants of socioeconomic enterprises are also on average the oldest and the ones with the shortest mean duration in childcare. This distinction between programs for unemployed with stronger reintegration problems on the one hand and programs for unemployed with “usual” reintegration problems on the other hand can be observed in many dimensions. For the former group, we observe predominantly participants with compulsory schooling (9 years) as the highest education level, jobs in the production and construction sector, higher average times in unemployment, shorter durations of the last employment spell, and a lower overall fraction in employment over the entire observation period in the data. Furthermore, those participants have much lower remaining benefit claims at the time of program entry and lower past earnings.

A peculiarity that is observed for all program groups is that 15–23% of the participants attended a program of the same kind in a previous unemployment spell. Active job search measures, primarily used to endow participants with special job application and interview skills, are also used as a screening instrument for long-term unemployed in order to renew and tighten the contact to the local Public Employment Service office. This is also reflected in Table 3 since participants in active job search live in regions with a higher average fraction of long-term unemployment. It can be observed that non-profit sector projects and job coaching are rarely used in the state of Vienna compared to the rest of Austria. In terms of employment in the period after the program, we find that participants in socioeconomic enterprises, non-profit sector projects, and job coaching have considerably lower employment rates 1 year after the program start, which is not surprising given program lengths of up to 1 year. Participants in shorter programs exhibit higher employment rates. Overall, participants catch up after 2 or 3 years.

Figure 2 provides a more complete picture of pre- and post-program employment rates. The abscissa shows the months before and after the (hypothetical) start of the program. The ordinate measures the employment rate for different program groups. The left picture shows that participants in socioeconomic enterprises, non-profit sector projects, and job coaching differ quite substantially from non-participants with respect to their employment history 3 years before the program. Hence, interpreting post-program employment rate differences as program effects is not appropriate. The same holds for active job search, qualification measures, and course subsidies. Here, the pre-program differences are also visible, but not as large as for the first group of programs.

Fig. 2

Employment 3 years before and after program entry. NON non-participation, SEE socioeconomic enterprises, NSP non-profit sector projects, JC job coaching, AJS active job search, QM qualification measure, CS course subsidy

4 Econometric methodology

As discussed before, the identification of the program effects hinges on the existence of the variables that jointly influence program participation and potential outcomes. For every comparison of different program states (including non-participation), the estimation strategy is to form comparison groups that do not differ from the respective program groups with respect to the distribution of those conditioning variables. We employ an advanced version of a semi-parametric two-stage propensity score matching approach. This class of estimators is popular in the program evaluation literature because it allows for individual effect heterogeneity while not requiring a parametric specification for the relation of the outcome variable and the variables controlling for the selection bias correction. Rosenbaum and Rubin (1983) show that if the CIA holds, given all relevant covariates, then it also holds for a particular scalar function of those covariates, i.e., the participation probability conditional on the control variables (propensity scores). Hence, a first-step procedure estimates those conditional program participation probabilities. The advantage is that the construction of control groups can be done on the basis of the propensity score. Those points are discussed in Heckman et al. (1999) and Imbens (2004) for the binary treatment and in Imbens (2000) and Lechner (2001) for the multiple treatment case.

We model the propensity score by means of binary probit models for each program type and for men and women separately. The specifications differ sometimes considerably by program and gender as can be seen in the Internet appendix. The results give further insights into the program allocation of the caseworkers. Table 4 reports a selection of variables that appear frequently in all specifications.¹⁵ Despite the existence of considerable heterogeneity, some general determinants of program participation versus non-participation appear. For both sexes, we find a positive relation of participation to disability, desiring a vocational change, longer durations of the defining unemployment spell, and having higher average durations in past employment. Jobs in the law and administration and trade sector tend to increase the probability of being promoted in active job search, qualification measures, and course subsidies. Being a foreigner, having a university (entrance) degree as well as a previous occupation in the service sector decreases (if at all) the participation probability. For women, we find that the months of pregnancy reduce the participation probability for all programs, except socioeconomic enterprises. The overall previous time spent in parental leave prior to the defining unemployment spell under consideration reduces the participation probability for non-profit sector projects, job coaching, and active job search, but increases the one for course subsidies. For men, we find that having no vocational degree increases the probability of participating in socioeconomic enterprises, non-profit sector projects, and active job search, but decreases the one for participating in course subsidies. The remaining picture is less clear as can be seen in Table 4.

Table 4 Results of the propensity score estimation

To obtain the final estimates of the program effects, we use the extended propensity score matching procedure as proposed by Lechner et al. (2009) and modified by Wunsch and Lechner (2007). First, they allow for more than one good match, if available, by incorporating the idea of caliper matching as in Dehejia and Wahba (2002). Second, they incorporate a bias correction procedure to account for small mismatches of the matching step by exploiting the double robustness property as discussed in Rubin (1979) and Joffe et al. (2004). The Appendix contains a brief description of the way the estimator is implemented.

5 Results

5.1 Program effects by gender

The following figures illustrate program effects for participants in one program (listed at the top of each figure) compared to non-participation. The follow-up period relevant for outcome measurements starts at the day of program entry and ends 3 years later. Effects are estimated monthly as differences of percentage points for all outcome variables.¹⁶ If symbols appear on the different lines (denoting the program effects), it means that the respective effects are statistically different from 0 at the 5% level. Recall that the matching step is done on the basis of propensity scores that are estimated for men and women separately.¹⁷

The two graphs in the first row of Fig. 3 show employment effects of participating in socioeconomic enterprises, non-profit sector projects, and job coaching. Common to all graphs of Fig. 3 are negative employment effects for all programs right after the start, which is commonly labeled as lock-in effect (see Van Ours 2004, among others). The intuition is that participants reduce their search intensity while being in a program and therefore reenter less frequently into regular employment than non-participants. There are differences in the progression of the curves for men and women. For women, we observe that socioeconomic enterprises seem to increase the employment probability of the participants by 9% after 3 years. For qualification measures and course subsidies, presented in the second panel of Fig. 3, we find small positive effects at the very end of the follow-up period of about 2.5% for women.¹⁸ For male participants, we do not find positive effects for any program. Qualification measures even seem to harm the respective participants 3 years after program start. Interestingly, even after controlling for the pregnancy status which distinguishes this study from previous ones, we still find some positive effects after 3 years for women. Thus, for certain labor market programs, this study points to an effect premia for women as well.¹⁹

Fig. 3

Effects of program participation versus non-participation: Employment in % points. Results based on matching estimation. Abscissa, Months after program entry. Ordinate, Difference of employment rates. Symbols indicate that the effect is statistically different from zero at the 95% level. SEE socioeconomic enterprise, NSP non-profit sector project, JC job coaching, AJS active job search, QM qualification measure, CS course subsidy. Participants (male/female): SEE (343/340), NSP (300/347), JC (206/243), AJS (9,641/9,638), QM (6,869/11,330), CS (4,549/5,587)

5.2 Where do the positive effects for women come from?

Previous studies, like Lechner et al. (2009), showed that usually, the positive employment effects are not achieved by reducing the rate of registered unemployed participants but by increasing their labor force attachment, i.e., by reducing the rate of participants leaving the labor force. Therefore, Fig. 4 shows the program effects on times out of the labor force (OLF), defined as not being employed and not being registered as unemployed in the current study.

Fig. 4

Effects of program participation versus non-participation: OLF (not employed and not registered as unemployed) in % points. Results based on matching estimation. Abscissa, Months after program entry. Ordinate, Difference of out-of-the-labor-force rates. Symbols indicate that the effect is statistically different from 0 at the 95% level. SEE socioeconomic enterprise, NSP non-profit sector project, JC job coaching, AJS active job search, QM qualification measure, CS course subsidy. Participants (male/female): SEE (343/340), NSP (300/347), JC (206/243), AJS (9,641/9,638), QM (6,869/11,330), CS (4,549/5,587)

All programs reduce times in OLF. Comparing both sexes, especially in the lower panel of Fig. 4, we find the reduction of OLF to be higher for women than for men. Using a unique feature of our data, we disaggregate this effect further. Figure 5 shows the program effects on times of parental leave for men and on times of parental leave plus pregnancy (PP) for women.²⁰ For women, we find significant negative effects on PP for qualification measures and course subsidies.

Fig. 5

Effects of program participation versus non-participation: Parental leave and pregnancy. Results based on matching estimation. Abscissa, Months after program entry. Ordinate, Difference of shares in parental leave or pregnancy. Symbols indicate that the effect is statistically different from 0 at the 95% level. SEE socioeconomic enterprise, NSP non-profit sector project, JC job coaching, AJS active job search, QM qualification measure, CS course subsidy. Participants (male/female): SEE (343/340), NSP (300/347), JC (206/243), AJS (9,641/9,638), QM (6,869/11,330), CS (4,549/5,587)

For socioeconomic enterprises, job coaching, and active job search, we also find negative but insignificant effects. Only non-profit sector projects seem to have small positive effects on PP, though being insignificant. There are no significant effects on parental leave for men. To summarize, women who are not allocated to a labor market program, though being eligible, switch more frequently into PP. It seems as if some of those women are faced implicitly with the decision of being trained or using the time to realize family plans that would have been postponed otherwise, i.e., in case of a program allocation. Hence, we find that apart from effects on employment (positive or negative), programs may contradict other policies that are designed to increase birthrates. As a final check, we use an outcome variable which takes the value 1 for times in employment and PP and 0 otherwise. The results are presented in Fig. 6.

Fig. 6

Effects of program participation versus non-participation: Employment and pregnancy and parental leave. Results based on matching estimation. Abscissa, Months after program entry. Ordinate, Difference of employment rates. Symbols indicate that the effect is statistically different from 0 at the 95% level. SEE socioeconomic enterprise, NSP non-profit sector project, JC job coaching, AJS active job search, QM qualification measure, CS course subsidy. Participants (male/female): SEE (343/340), NSP (300/347), JC (206/243), AJS (9,641/9,638), QM (6,869/11,330), CS (4,549/5,587)

The result is rather striking. Three years after program start, we do not find any significant effect for any program type, neither for men nor for women. Moreover, we observe that the relative dominance of the women melted down towards zero. It appears that the only remaining difference appears for qualification measures with significant negative effects for men and insignificant effects for women. Hence, we do not find substantial effect premia for women as soon as we incorporate times of PP as an outcome variable. The female premia in Fig. 3 appeared because female non-participants take an additional outside opportunity, i.e., becoming a mother, which leads to comparably low employment rates for this pool of women (see also Fig. 7). Men are much less affected by such issues, and we therefore observe only the program effect, which is usually non-positive.

Fig. 7

Percentage of pregnant women by program status and age group. The pregnancy status is computed right before the (hypothetical) program start. For this illustration, we pooled all program categories due to the small fraction of pregnant women per program type

From the perspective of the policy maker, the message of the results should be ambiguous. If additional kids are considered as desirable as employment, then the programs are ineffective. If not, then the increase in the employment effect for women at the cost of reducing or postponing fertility may be desirable and considered as a “positive” outcome of the active labor market policies.

Due to the large size of the population in this study, it is possible to stratify female participants further, i.e., per age group, to get a clearer picture of the underlying heterogeneity relating to this effect. Thus, we divide all women into two age groups, below and above 40 years, to separate two groups that differ with respect to individual family plans. Doing so, 97.6% of all pregnancies just before the hypothetical program start are in the lower age group. Due to population size restrictions, we consider the three larger programs only, active job search, qualification measures, and course subsidies.

Figure 8 shows that splitting the female population according to age, we observe that the positive effect for course subsidies can be attached to women younger than 40 years. For this group, we observe that all programs have a negative effect on PP. For the older segment in turn, we do not observe such effects. Overall, for both age groups, we fail to detect positive effects once we take employment plus PP as the outcome variable. Additionally, but not surprisingly, the effects on employment for men above the age of 40 look exactly like the effects for women in that age group since pregnancies and parental leaves play hardly any role here.²¹ Both arguments confirm our result that once we correct for the selection bias (pregnancy status for women), the remaining (small) positive effects, here for course subsidies for younger women, only appear because eligible non-participants subsequently emphasize family planning, which leads to lower employment rates compared to those participants they have been matched to.

Fig. 8

Effects of program participation versus non-participation: Women per age group. Results based on matching estimation. Abscissa, Months after program entry. Ordinate, Difference of employment rates. Symbols indicate that the effect is statistically different from 0 at the 95% level. AJS active job search, QM qualification measure, CS course subsidy. Participants (≤40/>40): AJS (6,163/3,493), QM (7,569/3,762), CS (3,672/1,917)

5.3 Pregnancy bias—a sensitivity check for omitted variables

As became clear in the previous section, one important feature of this study is that we use information on parental leave and pregnancies as an outcome variable as well as for correcting for potential selection bias, as pregnant women are rarely observed in labor market programs. Thus, if this variable is not controlled for, it is likely that a larger share of pregnant women appears in the group of non-participants, which will bias the employment effects upwards.

Now, we analyze the size of this bias by comparing our results to results that would have been obtained without that information. First, we do not delete persons who are in parental leave right before the (hypothetical) program entry, and second, we leave out the month of pregnancy from the selection model, i.e., we allow for the selection bias that we suspect to be one driving factor for the positive effect differential for women. Figure 9 summarizes the results.

Fig. 9

Effects of program participation versus non-participation: Employment in % points. Results based on matching estimation. Abscissa, Months after program entry. Ordinate, Difference of employment rates. Symbols indicate that the effect is statistically different from 0 at the 95% level. SEE socioeconomic enterprise, NSP non-profit sector project, JC job coaching, AJS active job search, QM qualification measure, CS course subsidy. Participants (male/female): SEE (343/340), NSP (300/347), JC (206/243), AJS (9,642/9,640), QM (6,869/11,332), CS (4,552/5,594)

Obviously, the results for men are not affected by this change since parental leave is a minor issue here. For women, all effects increase by approximately two to three percentage points compared to Fig. 3. For socioeconomic enterprises, this results in a wider range of significant positive effects, especially at the end of the follow-up period. For active job search, qualification measures, and course subsidies, we now observe significant positive effects that are stable from the middle of the follow-up period onwards. According to these estimation results, we would conclude that we find clear evidence of positive effects for women for four out of six labor market programs, which is highly misleading as shown by the results in the previous sections.

To wrap up, we find two important impacts of the observability of times of parental leave. First, by constructing the pregnancy status for women, it removes the remaining omitted variable bias in the first-step selection model. Second, it can be used to better understand program effects in the follow-up period.

6 Conclusion

This study provides an econometric evaluation of several important active labor market programs in Austria. Large and informative administrative data are used to control for potential selection problems. As a particular advantage of the data, we identify times of pregnancy and parental leave. For women, this information turns out to be very important for reducing selection bias as well as for understanding the effects of the programs.

For men, we find the programs to be generally ineffective in increasing unsubsidized employment. However, without controlling for pregnancy status, most programs appear to be effective in increasing employment prospects for women. Those effects become smaller once the pregnancy information is taken into account, but they are still there. A closer investigation shows that the programs increase female employment by reducing the share of women leaving the labor force. The underlying mechanism is that the programs reduce the pregnancy rate of the participants, i.e., programs also have an adverse (and often unintended) effect on other policies designed to foster birthrates. Once that effect is subtracted from the employment effects, almost all gender differences disappear.

Our findings about the gender differences may explain results appearing in the survey by Bergemann and van den Berg (2006). They find that women’s effect premia predominantly occur in countries with a low female work force participation, indicating that times for taking care of infants and labor market participation are less compatible or exclusive. Moreover, none of those studies incorporated information on times of pregnancies. We demonstrate for the case of Austria that it is important to have information about the outside opportunities of women, like times of parental leave. The puzzle of women’s effect premia might be partially explained by the fact that important confounders (and outcome measures), like the ones discussed above, have not been available in other studies.²²

The question whether our results for women, namely, a positive employment effect and a zero effect on the fertility plus employment outcome, indicate that the definition of a program success depends on the value judgment of the policy makers. If additional (or earlier) kids are considered as desirable as employment, then the programs are ineffective. If employment is considered more important, then the increase in the employment effect for women at the cost of reducing or postponing fertility may be desirable and considered as a “positive” outcome of the Austrian active labor market policies. This conclusion is most likely true not only for Austria but for many other European countries as well.

Appendix: Matching protocol

The results presented contain the binary comparison of each particular program to the state of non-participation for the participants in that particular program. Table 5 shows the matching estimator that is used for each such comparison.

Table 5 A matching protocol for the estimation of ATET