German Active Labor Market Policies: The Use of Job Creation and ...

4 downloads 0 Views 49KB Size Report
started by Hans-Jürgen Krupp in 1984 (c.f. Krupp/Hanefeld 1987). In 1990,. 2,179 East German households containing 4,453 persons were surveyed (Wag-.
German Active Labor Market Policies: The Use of Job Creation and Training Programs Following Unification Lisa M. Amoroso and James C. Witte∗

Introduction The unification of Germany in 1989 brought serious economic implications for both the former East and West German Republics. The monetary cost to the existing West German states has been considerable in terms of subsidies and investments in the East - financed in large part by West German citizens through additional taxes, euphemistically called „solidarity contributions.” However, the financial burden placed on the West pales in comparison to the economic impact of Unification on the East as it makes the transition from a command to a marketoriented economy. No longer protected by trade barriers, subsidies and price supports, East German industries, firms and workers are now exposed to a competitive world market. The consequences for individual workers’ careers, the fortunes of firms, and the structure of entire industries have been profound. This paper reviews policy efforts on the part of the German government to manage and mitigate these consequences. Our emphasis is on mechanisms - active labor market policies (ALMPs) - that directly intervene in the market, rather than efforts to compensate individuals and firms for market outcomes (OECD 1996). In particular, we consider the use of training and employment programs by the German government after 1991 as the labor market moved from „socialist full employment” to a market system and unemployment became common. Our discussion focuses on the relative strengths of training and employment programs as different approaches to labor market policy issues. As such we add to a literature that has tended to focus on either the effects of ABM program participation (Steiner/Kraus 1996; Günther 1991) or ALMP job training (Pannenberg 1996;

∗ We are grateful to Markus Pannenberg for critique and most valuable suggestions.

200

Lechner 1996; Fitzenberger/Prey 1996), but only rarely compares (Hübler 1996) different types of ALMP programs.1

1

German Active Labor market policies: An Overview

Active labor market policy programs are intended to assist unemployed individuals in finding work and to make the labor market as a whole function more effectively (Steiner/Kraus 1996; OECD 1996). Active, labor market policies emphasize efforts to train and employ the long-term unemployed, rather than simply remedy their immediate hardships. Recent ALMP initiatives in Germany - ranging from employment training, job-search assistance and employment subsidies to direct job creation - find their legislative mandate in the Arbeitsförderungsgesetz, commonly referred to as AFG. The most active element of the AFG are Arbeitsbeschaffungsmaβnahmen (ABMs), measures to directly create employment that serves the public interest while providing the long-term unemployed with training and labor market experience. ALMPs, especially ABM programs, were widely used in the early 1980s in West Germany to address the issue of long-term unemployment, particularly that which resulted from regional, structural economic change (e.g., in the coal-mining region of the Ruhr Valley). During this time, the proportion of long-term unemployed persons among all unemployed persons increased from 12.6% to 30% (Hardes 1988). Far from an unambiguous success, this experience in the West suggested that a simple expansion of existing ALMPs was not a certain remedy for the structural changes and crises occurring in the five new eastern states (Siegers 1991). Other observers argue that as a rule ALMP programs decrease the effectiveness of labor market mechanisms and distort the incentives of the public employment service (PES), employers, and workers (Boeri/Burda 1996; Heckman 1994; Calmfors 1994).2 1 Hübler (1996) is an important exception, which uses longitudinal data from the Labor Market Monitor Study to study a variety of employment policy measures between 1990 and 1994. 2 Specifically, critics argue: (1) there are „dead-weight” effects because the PES is serving persons who would have eventually found jobs or training on their own; (2) there are „substitution” and „displacement” effects, meaning that employers have substituted subsidized jobs for regular, market jobs as employers replace non-ALMP workers and trainees with ALMP participants and thereby reduce overall labor costs; (3) a negative

201

In spite of these concerns, the institutional and political support for such programs throughout Europe, including Germany, is quite high. Empirical and theoretical arguments aside, ALMPs are widely perceived to effectively buffer workers from external market shocks and internal economic adjustment periods. Thus, the West German ALMP approach was rapidly extended to the eastern states as part of the „fast track” Unification strategy (OECD 1990; Steiner/Kraus 1996). The severity of the economic situation in the eastern states could not be ignored: output in eastern states had plummeted by 50% during 1990; over 340,000 skilled workers from the eastern states migrated to the west during 1990 (OECD 1990); full-employment in 1990 was replaced by 40% unemployment in some regions within one year (Günther 1991); while the level of regular employment in eastern Germany fell by 37% between 1989 and 1993. In these circumstances the immediate labor market priorities were to employ those who had lost jobs during the transition to a market economy and to retrain all workers to be productive and desirable in the new economic order (OECD 1996). Between November 1989 and November 1994 approximately three-fourths of all East Germans of working age, participated in one or more AFG programs - with more than half participating in some form of training or certification program and 11% in an ABM program (Hübler 1996). The central research question with respect to these programs is: „Do the programs decrease unemployment or simply provide another vehicle for delivering unemployment benefits?“ The analyses presented below directly consider the effects of ALMP participation on individuals' subsequent employment. To begin with, however, we consider the individual and structural characteristics related to the probability that individuals entered ALMPs following Unification.

2

Data and methods

The data used come from the German Socio-Economic Panel (GSOEP), an ongoing longitudinal study of households and individuals in East and West Germany

effect on employment after participation may result because of a stigma perceived by employers associated with prior program participation; (4) the macroeconomic effects of employment and training guarantees are to set higher reservation wages and training qualifications for workers.

202

started by Hans-Jürgen Krupp in 1984 (c.f. Krupp/Hanefeld 1987). In 1990, 2,179 East German households containing 4,453 persons were surveyed (Wagner et al. 1993). Individual data on age, gender, education, employment status (1991 through 1995), job characteristics, wages, and indicators of participation in federally-funded job creation and training programs from June 1990 through December 1994 are the basis for our analyses. During this observation period 710 East German respondents were either enrolled in a job training program3 or were employed in an ABM position.4 Our primary analytical interests are twofold. First, we assess the factors - individual (gender, employment status, age, and educational attainment) and structural (firm size and industry) - that affect the likelihood that individuals participate in an ALMP program. These models are standard logistic regression models. Second, we consider the extent to which participation in ALMP programs, along with individual and structural characteristics, influence employment status after program participation. To determine the manner in which the work force participant trajectories for ABM participants differ from structurally equivalent non-participants, we conduct a logistic regression model for East Germans where the dependent variable is the natural log odds of being in regular employment at the end of the observation period. Our primary interest is in the magnitude of these effects - the multiplicative change in the odds of being unemployed at time T3 due to participation in labor market programs in T1 and T2 .

3 GSOEP respondents were asked if they were enrolled in any type of vocational education or training at the time of each interview. Unfortunately, respondents were not explicitly asked if the training they were receiving was through an ALMP program. To avoid confounding ALMP participation with training received as a normal part of the education system, individuals were said to be receiving ALMP training only if they were in the labor force prior to receiving the training. 4 Of the 710 East Germans ALMP participants, 265 took part specifically in ABM programs. Typically German ABM positions are funded for one year, however people may re-enroll or stay in the program for a longer period. Of the 265 participants, 182 reported ABM participation at the time of one of the four yearly interviews, 64 respondents reported ABM participation on two occassions, 17 persons on three, and 2 individuals were in ABM for all four years.

203

3

Participation in Active Labor Market Policy Programs

The degree of ALMP program participation among East German GSOEP respondents is considerable. As Table 1 indicates the number of participants in such programs doubled between 1991 and 1992 - increasing from 8% to 16% of the labor force - and the participation rate remains above 15% through 1995. Approximately half of this change results from increased participation in job training programs, and the other half through participation in ABM employment.5 Table 1: Breakdown by Year of Participation (% and N) in ALMPs for East German Labor Market Participants

All Active Labor Market Programs ABM Vocational Training At Survey Date1 Registered Unemployed Total in Labor Market (N)

1991

1992

1993

1994

1995

8.09 253 n/a

16.50 485 4.25 125 12.69 373 18.40 541

17.22 452 3.20 90 13.13 369 19.17 539

15.69 390 2.32 65 11.61 326 19.49 547

15.02 388 3.24 89 11.02 303 16.48 453

8.09 253* 12.14 380 3,129

2,940

2,811

2,807

2,749

1)These numbers include those respondents who are in their initial vocation training and/or apprenticeships and so include some respondents who are excluded from the logistic regression analysis since they were not previously in the labor market and therefore not at risk of entering an ALMP. *)To determine this cell, we used calendar activity level data which asked about job training generally because East Germans were not asked about the same variety of job training options on the 1991 survey as they were in the remaining years. We compared the results of using the calendar level data with the annual survey questions and found no significant difference. We are using the annual indicators because the 1995 calendar data is not yet available (it is released with the 1996 data).

Source: GSOEP, own calculations.

In Table 2, we see that according to gender, age, and education level, individuals in the labor force in 1991 who later participated in an ABM program are demographically different than the labor force as a whole. Women make up a greater 5 It should be noted that the unemployment rate estimated using the GSOEP data is slightly greater than the official unemployment rates because older people who feel unemployed are not counted as unemployed in official statistics. Otherwise, the general trend apparent in Table 1 - a sharp increase, followed by continued high unemployment with a slight decrease again after 1994 - is consistent with the official numbers.

204

proportion of ALMP participants than in the labor force as a whole. More notably, young people and those with no vocational education represent a significantly greater proportion of ALMP participants than they do of the entire labor force. The shifting age and gender composition of ABM participants over time suggests that by 1994 the less marginalized people are participating less frequently in ABM than in the years immediately following unification. This could be interpreted as an economic recovery in that the economy is now functioning at a level where those who are traditionally able to secure jobs on their own are doing so. Beyond the effects of individual characteristics, Table 2 indicates that structural characteristics influence ALMP participation as well. About half of all East Germans work in firms with 200 or more employees and 19 percent are employed in firms with 2,000 or more employees. Rates of participation in ALMP programs roughly mirror the overall distribution of workers according to firm size. However, unification brought considerable change to the industrial makeup and there is noticeable variation in participation in ALMP programs by industry. Not surprisingly, since the legislative mandate for these programs stipulates that subsidized training and employment are to serve the public interest, ALMP participation is concentrated in the non-profit and local government sectors. Table 3 then presents results from multivariate logistic regression models considering the relationship between ALMP participation and these individual and structural characteristics. In the first column, Model 1 regresses the natural log of the odds of participation in any ALMP program on these char- acteristics, while Models 2 and 3 separately examine participation in ABM and job training programs. Beginning with Model 1, the significant negative coefficients attached to the years 1992, 1993 and 1994 indicate that individuals were less likely to enter ALMP programs after 1991. Regardless of the year involved, unemployed individuals were three and one-half times more likely to enter ALMP programs than employed persons.6 Though it may seem trivial to note that unemployed persons are more likely to enter ALMP programs, it is nonetheless important to establish that participants in these programs are not individuals who would otherwise be employed.

6 The strength of the relationship, as a percentage change in the conditional hazard of entering unemployment, is calculated as follows: 100(exp(b)-1).

205

Table 2: Individual and Structural Characteristics of the German Labor Market

Gender Female Male Age 16 - 19 20 - 29 30 - 39 40 - 49 50 + Education 3 No Vocational Ed Apprenticeship School-based Vocational Ed Higher Ed Firm Size (# of employees) 4 2000 Self Employed Industry 4 Agricultural and Extractive Manufacturing Construction Retail and Wholesale Sales Postal, Telephone, and Rail Service Sector Health and Education Real Estate and Legal Services Non-Profit and Local Government Other N

East German Labor Force1 1991

1991

1992

1993

1994

48.7 51.3

53.8 46.2

51.1 48.9

52.0 48.0

56.2 43.8

4.0 22.7 30.6 22.7 20.0

33.2 27.1 26.0 11.1 2.7

19.4 27.4 26.4 16.1 10.7

21.2 28.8 26.8 14.8 8.4

23.8 27.4 23.8 13.8 11.0

7.5 59.6 23.6 9.3

33.3 39.5 15.5 11.6

26.2 46.0 18.2 9.6

29.9 43.2 17.9 9.0

38.5 33.5 18.2 9.8

ALMP Participants2

6.8 11.5 32.4 28.5 18.2 2.0

3.5 10.0 23.9 35.3 26.9 0.5

4.5 16.7 34.7 23.8 19.6 0.8

7.2 20.0 31.9 22.2 17.8 0.8

7.2 19.8 32.4 20.8 18.4 1.4

10.9 25.5 8.0 11.1 7.1 4.0 14.9 1.5 8.9 8.1 3,129

16.1 23.7 11.4 5.7 4.7 2.8 16.6 0.5 8.5 10.0 262

5.8 23.8 11.6 9.0 4.8 6.3 18.5 1.1 14.8 4.2 485

1.7 16.6 18.3 6.9 5.1 10.9 17.7 4.6 16.6 1.7 452

2.6 17.9 13.9 10.6 4.0 9.3 17.2 4.6 15.2 4.6 390

1) Respondents were considered in the labor market if they had any level of employment including marginal part-time work. According to this definition about 78% of the East German adult population is considered in the labor market in 1991.- 2) This group consists of all respondents who participated in an ALMP program during the year indicated. 3) Less than 3% of the education data is missing for each year.- 4) Approximately 10% of the data for these categories is missing.

206

Source: GSOEP, own calculations.

Table 3: Logistic Regression Coefficients for Models Predicting the Odds of Entering an ALMP in T2 VARIABLES

Model 1

Model 2

Model 3

ALMP

ABM

JOB TRAINING1

Period (reference category = 1991) 1992 -0.364** 1993 -0.612** 1994 -0.448** Female 0.040 Unemployed at T1 1.511** Age (reference category = 20 -29) 16 - 19 1.517** 30 - 39 -0.104 40 - 49 -0.258* 50 + -0.879** Education at T1 (reference category = apprenticeship) No vocational ed. -0.243 School-based vocational ed. 0.257** Higher ed. 0.452** Firm size greater than 2000 em0.192 ployees at T1 Industry at T1 (reference category = manufacturing) Agricultural and extractive 0.348+ Construction -0.380* Retail and wholesale sales -0.435* Postal, telephone, and rail -0.641** Service sector 0.189 Health and education 0.347* Real estate and legal services 0.354 Non-profit and local government 0.589** Other 0.072 Constant -2.424** N2 9294 Pseudo R2 7.2%

-1.014** -1.094** -0.744** -0.034 1.744**

-0.062 -0.306* -0.250+ 0.094 1.802**

-0.491 0.184 0.226 0.124

1.812** -0.241* -0.409** -1.337**

-0.286 -0.257 -0.031 -0.315

-0.244 0.459** 0.604** 0.336*

0.858** -1.180* -0.680+ -1.176+ -1.182+ -0.073 -4.058 0.151 -0.163 -3.514** 10480 12.0%

-0.171 -0.245 -0.402+ -0.565+ 0.391 0.402* 0.653+ 0.553** 0.148 -2.890** 9549 6.2%

** α < .01; * α < .05; + α < .10 1) This analysis uses the monthly calendar indicators of job training participation but does exclude those respondents participating in initial job training and/or apprenticeships.- 2) These models have different Ns because in each case the risk set is defined as those not participating in an ALMP, ABM or Job Training at the initial time point. Since more individuals were in job training programs than ABM programs Model 3 has (as expected) a smaller N than Model 2. Additionally Model 1 has the smallest N since this is a measure of being in either ABM or job training at time 1.

207

Source: GSOEP, own calculations.

Model 1 also clearly shows that ALMP program participation is strongly correlated with age: young adults are significantly more likely to enter ALMP programs, while those aged forty and older are significantly less likely to participate.7 A significant positive relationship is also found between educational attainment and ALMP participation. There is essentially no difference between individuals with no post-secondary education and those with apprenticeship training in the probability of ALMP participation, while individuals with school-based vocational education and/or a university degree are significantly more likely to participate. The effects of the structural characteristics (firm size and industry) on ALMP participation are relatively small in Model 1. Individuals employed in the agricultural sector, the health care industry, and local government employees or more likely than those employed in manufacturing to participate in an ALMP program; by contrast, persons employed in construction, sales and and the postal service are significantly less likely to participate. In the former German Democratic Republic, agriculture was notoriously inefficient, so that after unification workers in this sector were prime candidates for ALMP programs. Meanwhile, creating positions and training for employees in healthcare and local government provided few administrative hurdles, as these areas are clearly consistent with the legislative mandate for ALMP programs. Construction and sales, on the other hand, were relatively strong sectors in the post-unification East German economy, while there was little need to use ABM programs for postal workers (including telecommunications) because these workers were already all federal employees. In Model 2, which looks exclusively at ABM program participation, most of the individual characteristics - gender, age and educational attainment - which were strongly significant in Model 1, are no longer statistically significant. Indeed, unemployment status remains the only individual characteristic that is significantly related to ABM program participation. At the structural level, firm size is no longer relevant, though nearly all of the same industry indicators remain significant. Based on these results, at the individual and structural level, ABM program participation is tightly coupled to immediate employment concerns. Regardless of age, gender or education level, individuals who need jobs participate in ABM;

7 Lower rates of participation among older workers are in part due to early retirement options that were among the passive labor market policies introduced immediately after unification.

208

moreover, these individuals are most likely to come from the weaker sectors of the economy. Model 3, which specifically looks at entry into ALMP job training programs, presents a different picture. Unemployment remains a statistically significant predictor of program participation. However, while unemployed persons are nearly five times more likely than employed persons to enter ABM programs (Model 2), the unemployed are only twice as likely to enter job training programs (Model 3) than those who are already working. Strong age effects are also associated with participation in job training programs: the youngest workers are more likely to enter such programs, while workers older than thirty are significantly less likely to participate. Strong education effects are also apparent, but not necessarily in the direction that one would expect. Those with the least education are significantly less likely to get training, and those with the most education are significantly more likely. These findings, combined with the insignificance of gender in Model 3 suggest that at the individual level, participation in ALMP job training is primarily motivated by human capital concerns: 1) young workers realize a return to investment in their human capital over a longer time period than older workers; 2) those who have already proven themselves capable of educational success are most likely to profit from additional investment. At the structural level, Model 3 indicates that the probability of participating in ALMP job training is higher among employees working in large firms. Not only are large firms able to benefit from economies of scale in providing training, but also are more likely to benefit from facilitating or providing training for their employees as turnover is usually lower than at small firms. Significant inter-industry variation in the probability of participation in job training is observed in Model 3 as well. As was the case with ABM program participation the probability of entering ALMP job training is significantly lower among postal employees and those in sales, but higher among local government employees and persons employed in the health care sector. It is also noteworthy that job training is also significantly higher among persons employed in the small, but growing real estate sector, which changed dramatically as East Germany moved from a command to a market economy.

209

4 Employment Effects of Active Labor Market Policy Programs The findings reported above indicate that ALMP program participation in East Germany since unification is by no means random; ABM employment and participation in job training programs is associated with individual and structural characteristics of workers and the jobs they have held in the past. ABM employment appears to be most closely tied to firm size and industrial sector, while ALMP job training is more closely linked to age and prior post-secondary educational attainment. A separate issue, however, is the impact that ABM and job training have on the careers of individuals once they have left the programs. To support the claim that ALMP programs are more than an alternative form of welfare, it is necessary to show that participation in job training and employment programs bring additional benefits to participants once they complete the programs and enter the competitive labor market. Results reported in Table 4 address this issue. These models regress the log of the odds of unemployment at T3 based on characteristics of the individual and the position occupied at T1 and employment status (including ALMP program participation) at T2 .8 In both models the risk set includes all persons either employed or enrolled in an ALMP program at T1 . The two models in Table 4 then consider the factors that affect the odds that an individual will be unemployed rather than employed or out of the labor force at T3 .9 Model 4 in Table 4, uses a single coefficient to indicate the import of ALMP program participation, while Model 5 substitutes two terms to distinguish between the effects of ABM program participation and job training.

8 More specifically, the models are based on a pooled data set that combines observations for all persons employed or participating in ALMP programs in 1991, 1992 and 1993 (T1 ) and considers their employment status in 1993, 1994 and 1995 ( T3 ) respectively, dependent on individual and structural chachteristics measured at T1 and employment status and ALMP program participation at T2 , i.e., 1992, 1993, 1994. 9 Combining employed persons and those out of the labor force in the contrast category for our dependent variable is done to emphasize the involuntary nature of unemployment.

210

Table 4: Logistic Regression Coefficients for Models Predicting the Odds of Being Unemployed at T3 VARIABLES

Model 4

Model 5

Employment Status Indicators at T1 (reference category = employed) ALMP -0.971** n/a ABM n/a 0.280 Vocational training n/a -8.175+ Employment Status Indicators at T2 (reference category = unemployed) ALMP 0.750** n/a ABM n/a 1.877** Vocational training n/a -0.846** Employed -2.572** -2.803** Period (reference category = 1991) 1992 -0.344** -0.392** 1993 -0.544** -0.544** Female 0.428** 0.489** Age (reference category = 20 -29) 16 - 19 -1.067* -0.302 30 - 39 -0.053 -0.170 40 - 49 0.260+ 0.138 50 + 1.301** 1.143** Education at T1 (reference category = apprenticeship) No vocational ed. 0.620** 0.834** School-based vocational ed. -0.404** -0.365** Higher ed. -0.355* -0.308+ Firm size greater than 2000 -0.260* -0.178 Industry (reference category = manufacturing) Agricultural and extractive 0.034 -0.088 Construction -0.370+ -0.329 Retail and wholesale sales -0.083 -0.060 Postal, telephone, and rail -0.124 0.128 Service sector -0.048 -0.020 Health and education -0.557** -0.558** Real estate and legal services -0.043 0.122 Non-profit and local government -0.038 -0.424* Other -0.118 -0.092 Constant 0.212 0.449** N 5864 5864 Pseudo R-Square 30.1% 26.0% + ** α < .01; * α < .05; α < .10. Risk set includes persons who are either employed or ALMP participants at T 1. Source: GSOEP, own calculations.

211

Turning first to Model 4, we see a negative and significant coefficient representing the effect of ALMP participation at T1 on unemployment at T3 . In other words, keeping in mind our definition of the risk set, those who were enrolled in an ALMP program at T1 were significantly less likely to be unemployed two years later than those who were simply employed at T1 . Yet, if we consider ALMP participation at T2 , the relationship is reversed. The coefficient attached to ALMP participation at T2 is positive and significant, while the coefficient associated with employment is negative and significant - where those who entered unemployment between T1 and T2 serve as the reference category. Initially this suggests that the relationship between ALMP participation and future employment varies with time: an issue to be explored more fully when we turn to Model 5. Model 5 (Table 4), which distinguishes between ABM and job training as the principle types of ALMP program participation, reveals an important difference between these types of ALMPs. ABM program participation at T1 is not significantly related to unemployment at T3 , while ABM program participation at T2 actually increases the probability that an individual will be unemployed at T3 . Moreover, the negative impact of ABM participation at T2 is by no means trivial, as these individuals are 5 times more likely to be unemployed at T3 than those who were already unemployed at T2 . Job training at T1 or T2 , on the other hand, significantly decreases the likelihood that an individual will become unemployed at T3 . Here, too, the differences are substantively as well as statistically significant: those engaged in job training at T1 are 100% less likely to be unemployed at T3 than those who were employed at T1 ; those who were in job training at T2 are nearly 60% less likely to be unemployed at T3 than those who were unemployed at T2 . It is not simply ALMP participation that matters, but rather different types of ALMP programs have different effects on participants. It is interesting to note that once the distinction between ABM and job training is included in the model, the relationship between employment in a large firm and the reduced risk of unemployment is no longer significant. Combined with the results in Table 3, which show that large firm size increases the probability that an individual will participate in job training, but not ABM, this suggests that firm size has little direct bearing on entry into unemployment. Rather, the risk of entering unemployment is indirectly reduced among persons employed in large firms because these individuals are more likely to benefit from ALMP job training programs.

212

5

Assessment and Policy Implications

A final assessment of the efficacy of ALMP programs requires that an important methodological point be considered. The analyses presented in this paper do not correct for selection bias. East Germans were not randomly assigned to ALMP programs, but rather, as our own analyses indicate, workers who were particularly vulnerable were more likely to participate in ALMP programs. The apparent failure of ABM programs noted in the results presented in Table 4 may be a function of the characteristics of those who participate in such programs and not a shortcoming of ABM programs per se. Hübler (1996) in analyses covering a slightly shorter time period (through 1994), however, does employ a variety of techniques to correct for selection effects and comes to essentially the same substantive conclusions as those presented here. Critics have argued that ALMP programs, especially ABM-type job creation programs, can not be used on a large scale and effect real change. Our findings, however, suggest that Germany has been relatively successful in its efforts to target ALMP policies to the individuals and industry sectors most vulnerable to the economic restructuring that has accompanied unification. Moreover, there appears to be a relatively effective use of job creation and employment training programs as distinct policy tools to address different goals. Our findings suggest that job creation programs have effectively dealt with structural issues, while jobtraining programs have been tailored to the needs of individual workers. On the other hand, if one considers outcomes beyond program participation, the results presented here raise considerable questions about the efficacy of all ALMP programs in the former East Germany. It is important to distinguish between job creation programs and job training programs. The latter appear to have a positive impact on individuals’ future employment status, but the former has little effect beyond providing temporary employment.

References Boeri, Tito and Michael C. Burda (1996), Active labor market policies, job matching and the Czech miracle, in: European Economic Review, pp. 805-817. Camfors, L. (1994), Active labor market policy and unemployment: a framework for the analysis of crucial design features, in: OECD Economic Studies 2, pp. 7-47.

213

Fitzenberger, Bernd and Hedwig Prey (1996), Training in East Germany: An Evaluation of the Effects on Employment and Wages. Günther, Horst (1991), Arbeitsbeschaffungsprogramme für Ostdeutschland, in: Wirtschaftsdienst, 71, pp. 111-19. Hardes, Hans-Dieter (1988), Langzeit-Arbeitslosigkeit und Arbeitsbeschaffungsmaßnahmen. Wirtschaftsdienst, vol. 78. Heckman, James (1994), Is Job Training Oversold?, in: Public Interest. 115 (Spring 1994), pp. 91-115. Hübler, Olaf (1997), Evaluation beschäftigungspolitischer Maβnahmen in Ostdeutschland, in: Jahrbücher für Nationalökonomie und Statistik , 216(1), pp. 21-44. Krupp, Hans-Jürgen and Ute Hanefeld (1987), Lebenslagen im Wandel: Analysen 1987. Frankfurt/New York. Lechner, Michael (1996), The Effects of Enterprise-related Continuous Vocational Training in East Germany on Individual Employment and Earnings. Institut für Volkswirtschaftslehre und Statistik, Discussion Paper 542-96. Mannheim. Organization for Economic Cooperation and Development (OECD) (1990), OECD Economic Surveys:Germany. Organization for Economic Cooperation and Development (OECD) (1996), The OECD jobs strategy: enhancing the effectiveness of active labour market policies. Pannenberg, Markus (1996), Zur Evaluation staatlicher Qualifizierungsma βnahmen in Ostdeutschland: Das Instrument Fortbildung und Umschulung (FuU), Institut für Wirtschaftsforschung Halle, Diskussionspapiere 38, Halle (Saale). Proposals for improving public employment agency services and for coordinating unemployment benefit measures with job training and placement efforts; based on the experience of 12 member countries. Australia, Denmark, Finland, Germany, Italy, Japan, Netherlands, Norway, Spain, Sweden, Switzerland, and Great Britain. Steiner, Viktor and Florian Krause (1995), Haben Teilnehmer an Arbeitsbeschaffungmaßnahmen in Ostdeutschland bessere Wiederbeschäftigungschancen als Arbeitslose? in: Steiner, V. and L. Bellmann (Eds.): Mikroökonomik des Arbeitsmarktes, Nürnberg. Wagner, Gert, Richard Burkhauser and Friederike Behringer (1993), The English Language Public Use File of the German Socio-Economic Panel, in: Journal of Human Resources, 28(2), pp. 429-33.

214