Before and After the Hartz Reforms: The Performance...

Before and After the Hartz Reforms:The Performance of Active Labour Market Policyin Germany*

Lena Jacobi and Jochen Kluve

Having faced high unemployment rates for more than a decade, the German govern-ment implemented a comprehensive set of labour market reforms during the period2003Ð2005. This paper describes the economic and institutional context of the Germanlabour market before and after these so-called Hartz reforms. Focussing on activepolicy measures, we delineate the rationale for reform and its main principles. Aspreliminary results of programme evaluation studies post-reform have become avail-able just now, we give a first assessment of the effectiveness of key elements of Germanactive labour market policy before and after the Hartz reforms. The evidence indicatesthat the re-organisation of public employment services was mainly successful, with theexception of the outsourcing of services. Re-designing training programmes seems tohave improved their effectiveness, while job creation schemes continue to be detrimen-tal for participants’ employment prospects. Wage subsidies and start-up subsidies showsignificantly positive effects. On balance, therefore, the reform seems to be moving theGerman labour market in the right direction.

Contents

1 Introduction

2 Economic situation and labour market institutions before Hartz

3 Core elements of the Hartz reforms

3.1 Increasing the effectiveness and efficiency of services and measures

3.2 Activation of the unemployed

3.3 Labour market deregulation

4 Labour market policy effectiveness before and after Hartz

4.1 Increasing the effectiveness and efficiency of labour market services andpolicy measures

4.2 Activating the unemployed


4.4 Macro-level evaluation

4.5 Summary of policy effectiveness before and after Hartz

5 Conclusion

References

* This paper was released for publication by the responsible editor in October 2006.

ZAF 1/2007, S. 45Ð64 45

Before and After the Hartz Reforms Lena Jacobi and Jochen Kluve

1 Introduction

In times of high unemployment, the necessity forlabour market reforms is discussed all over Europe.Many countries have either started or already fin-ished far-reaching reforms. Also in Germany theneed for reforms had become apparent over the pastdecades, when unemployment was rising constantlyand public budgets tightened. Finally in 2002, thegovernment took advantage of a scandal involvingthe Federal Employment Services1 to overcome theso-called reform bottleneck (“Reformstau”) andstart a series of rather radical Ð given the prior re-luctance Ð changes of active and passive labourmarket policies.

The resulting Hartz reforms Ð named after thechairman heading the independent expert commis-sion that worked out the blueprint for the reformpackage2 Ð are considered the most far-reaching re-form endeavour in the history of the German wel-fare state, and consist of four laws, the so-calledHartz IÐIV laws, that were implemented step bystep on Jan 1st 2003 (Hartz I and II), Jan 1st 2004(Hartz III) and Jan 1st 2005 (Hartz IV). While HartzIV constitutes a comprehensive modification of theunemployment benefit and social assistanceschemes, Hartz IÐIII modernised the organisationalstructure of the public employment services, modi-fied many of the already existing measures of ActiveLabour Market Policy (ALMP) and introduced a setof new ones. The laws contain a set of specific policymeasures that merge to a three-part reform strategy:(a) improving employment services and policymeasures, (b) activating the unemployed, and (c)stimulating employment demand by deregulatingthe labour market.

In this context, it is also the first time in the historyof the German welfare state that a policy reform isaccompanied by a comprehensive scientific evalua-tion. The government explicitly tied the implemen-tation of the Hartz laws to an evaluation mandate.The Hartz laws IÐIII were evaluated during theyears 2004 to 2006, and Hartz IV will be evaluatedfrom 2006 onwards. Given the scope of the reformendeavour, the evaluation of Hartz IÐIII was com-missioned by the government as a set of work pack-ages and modules, aiming at an evaluation of boththe Hartz reforms in their entirety and each particu-lar element on its own. In practice, the Hartz IÐIII evaluation has therefore involved more than 20

1 The Federal Employment Services were accused of massivefraud in the reporting of successful job placements.2 “Commission for Modern Labour Market Services” (Kommis-sion für Moderne Dienstleistungen am Arbeitsmarkt).

46 ZAF 1/2007

economic and sociological research institutes which,using methods based on qualitative case study ap-proaches as well as rigorous econometric analyses ofadministrative and survey data, face the challengeof disentangling the impacts of specific measuresin a setting characterised (a) by the simultaneousalteration of measures and institutional context, and(b) by the fact that many of the measures affect ev-ery worker, i. e. no comparison group exists. First re-sults of these evaluation studies have become availa-ble just now.

In this paper, we describe the economic and institu-tional context of the German labour market beforeand after the reforms. Focussing on active policymeasures, we delineate the rationale for reform andits main principles. We use the most recent empiricalevidence to discuss the effectiveness of key elementsof German ALMP before and after the Hartz re-forms. It has to be noted, however, that at presentthe evidence regarding policy effectiveness afterHartz is based on rather short observation periods,and therefore the findings are still preliminary at thepresent time.

The paper is organised as follows: section one givesan overview of the German economic situation sinceunification in 1990, briefly characterising the prob-lematic features of the eastern and western Germanlabour markets. We also describe the institutionalframework before Hartz, focussing on active labourmarket policies and the organisational structure ofpublic employment services, and discuss the mainweaknesses of the institutional setting that moti-vated the reform. Section three delineates core ele-ments of the Hartz reforms and the reform strategyin some detail. In section four we review both thehitherto existing and the most recent evidence fromevaluation research to assess the effectiveness of ac-tive labour market policy before and after the re-form. Section five concludes.

2 Economic situation and labourmarket institutions before Hartz

Since the 1990s, Germany has proved to be unableto benefit from favourable conditions in the globaleconomy. From 1991 until 2003 GDP grew by only18%, which is half the growth of the United King-dom (35%) or the Netherlands (34%) during thatperiod. At the same time, employment even de-creased slightly (by 0.4%) and unemployment ratesare higher than ever, ranging between 9.6% in west-ern and 18.6% in eastern Germany in 2005.

Certainly the unification in 1990 and its repercus-sions have contributed to Germany’s poor perform-

Lena Jacobi and Jochen Kluve Before and After the Hartz Reforms

ance. Unification suddenly increased the labourforce by roughly one third, a large share of whichwas inadequately trained for immediate employ-ment in an open market economy. Despite the needto first retrain the labour force and reshape the for-merly centrally planned economy, however, it wasa core political objective to adjust eastern Germanwages to the comparatively high western Germanlevels as quickly as possible. In contrast to otherCentral and Eastern European transition countrieshaving competitive wages levels to create sustain-able growth, the eastern German economy experi-enced rising unemployment and continuing depen-dence on federal subsidies and transfer paymentsfrom West to East. In addition to the high fiscalcosts of unification, the Maastricht criteria furtherreduced the government’s scope for expansivegrowth policies.

Only a small share of overall German unemploy-ment is thought to be attributable to business cyclefactors. Some studies argue that the GermanNAIRU has increased over the past decades (e.g.Franz 2001), indicating that structural factors playan important role in the German unemploymentproblem.

Figure 1 shows the trend in employment subject tosocial security contributions (sozialversicherungs-pflichtige Beschäftigung) and the trend in employ-ment with low pay for which exemption from socialsecurity contributions applies (geringfügig entlohnteBeschäftigung). In western Germany, we observe adecrease in the number of employees covered bysocial security since 2002. Almost at the same time,low-wage employment has steadily increased. Ineastern Germany, the decrease in regular employ-ment and the increase in low-wage employment canalready be found at the beginning of the observationperiod in 2000.

In Figure 2 we see the number of unemployed indi-viduals and the number of participants in trainingand job creation programmes for western and east-ern Germany during the time period 1991 to 2005.We observe a general increase in the number of un-employed individuals in both parts of the country.Whereas in western Germany the number of pro-gramme participants was relatively stable over thisperiod, decreasing only slightly, the figures for east-ern Germany document the massive use of ALMPmeasures during the early 1990s and a substantialdecrease over recent years.

Labour market institutions

Compulsory unemployment insurance was intro-duced in Germany as early as in 1927, complement-

ZAF 1/2007 47

ing the then already existing insurances for health,accident and old age. After World War II a generousbenefit system emerged, financed by contributionsand taxes. Active labour market measures were in-troduced when unemployment started to rise in the1970s. Both active and passive policy measures areadministered by the Federal Employment Services.In earlier years, when unemployment was still low,measures were designed to prevent rather than com-bat unemployment, adapting the workforce to struc-tural changes in labour demand. After unification in1990, active labour market measures played a cen-tral role in alleviating the social consequences of thebreakdown of the economy in eastern Germany. In1992, for instance, the number of people participat-ing in either a job creation or training programmeexceeded the number of unemployed (Figure 2).

The set-up of active and passive labour market pol-icy in Germany during the 1990s can be character-ised as follows: from the very beginning, unemploy-ment benefits were meant to maintain the worker’ssocial status during unemployment rather than pro-viding a safety net as a last resort. All paymentsmade to the individual over the entire period of un-employment were linked to his or her previous earn-ings. Unemployment benefits, which were paid forthe first 6 to 32 months of unemployment (depend-ing on previous employment duration and age),amounted to 67% of the last net income (60% with-out children), with a maximum level of 4250 EURper month. Unemployment assistance, which waspaid after entitlement to unemployment benefit ex-pired and without a time limit, still reached 57%(53%) of the last net income.

The unlimited duration of unemployment benefitpayments was an extraordinary feature of the Ger-man unemployment benefit system, leading to re-placement rates for long-term unemployed whichwere higher than in any other OECD country(OECD 2004). The replacement rates of short-termunemployed, in contrast, were Ð and still are Ð com-parable to many other OECD countries. Unemploy-ment benefits were financed by unemployment in-surance contributions shared by employers and em-ployees, while unemployment assistance was fi-nanced by taxes. In principle, unemployment assist-ance was means-tested on a yearly basis. It waspossible to supplement unemployment benefits withtax-financed social assistance. Generally, everyhousehold whose income fell below a certain incomethreshold qualified for social assistance.

The German benefit system combined generousbenefit levels with high benefit reduction rates thattaxed away most of a benefit recipient’s additional


earned income. Thus, incentives to take up a jobwere very low, especially for low-skilled workers.Engels (2001) calculates that a typical family withthree children receiving social assistance in westernGermany in the year 2000 received an income that

48 ZAF 1/2007

was only 15.3% below that of a comparable familywith a single earner receiving an average unskilledworker’s wage. In eastern Germany the respectivedifference was only 11.5%.


Compared with other countries, German active la-bour market policy in the 1990s was characterisedby high expenditure levels and long durations ofprogrammes. Training and public job creation meas-ures were the most important programmes in terms

ZAF 1/2007 49

of expenditure and number of participants. Meas-ures supporting the direct integration into regularemployment (e.g. wage subsidies and start-up subsi-dies) only played a minor role. Generally, job searchassistance and monitoring by the public employment


services was given rather low priority. Sanctions forlow engagement in job search activities were rarelyimplemented. For most programmes, the law nar-rowly defined the type of person who was eligiblefor participation in a programme. Assignment toprogrammes was not based on a systematic individ-ual profiling of each client, but rather on casework-ers’ discretion.

3 Core elements of the Hartz reforms

As unemployment continued to increase in the1990s, the social security system ran the risk of fi-nancial collapse and the need for a comprehensivereform of the institutional setting of labour marketpolicies became urgent. In both the political andacademic debates the benefit system was criticisedfor creating work disincentives and increasing long-term unemployment, deteriorating skills and thusworsening the mismatch on the labour market. Thepublic employment services were accused of operat-ing inefficiently, being customer-unfriendly and fail-ing to push job seekers sufficiently to search for ajob. The mix of active measures, focussing on train-ing measures and public job creation schemes withlong durations, was criticised for keeping partici-pants out of the open labour market instead of inte-grating them. Such criticism was based on evalua-tion studies of active measures that indicated severelocking-in effects and zero or even negative post-participation treatment effects of many programmes(e.g. Lechner 2000; Caliendo et al. 2003).

The so-called Hartz reforms, which were subse-quently implemented during the period 2003Ð2005,coalesce to a tripartite reform strategy (cf. Table 1).They aimed at (a) improving labour market servicesand policy measures in terms of effectiveness andefficiency, (b) activating the unemployed by enforc-

50 ZAF 1/2007

ing the so-called principle of “rights and duties”(Fördern und Fordern), and (c) stimulating employ-ment demand by deregulating the labour market.3

To this end, the reform modified many of the al-ready existing measures of active labour market pol-icy and introduced a set of new measures. It funda-mentally changed the general framework in whichthese measures operate and involved greater co-or-dination of institutional arrangements, especially be-tween active and passive policy measures. Deregu-lating measures concentrated on the temporarywork sector, while the biggest changes entailed bythe reform took place in the realm of job placementservices and the benefit system.

3.1 Increasing the effectiveness andefficiency of services and measures

The reform aimed to improve the performance ofplacement services and policy programmes mainlyby introducing market mechanisms to the realm ofplacement services and by streamlining public em-ployment services. Furthermore, cost-effectivenessin the specific context of each regional labour mar-ket is targeted to be the key criterion when choosingprogramme contents and participants.

First, regarding their organisational structure, thepublic employment services were modernised alongthe lines of New Public Management. The reformestablished results-based accountability and control-ling of local employment offices. Now, each employ-ment office has to fulfil quantitative goals which areindividually set for each type of agency, while at the

3 The Federal Employment Service translates “Fördern und For-dern” as “Challenge and Promotion”.


same time having a wider scope of discretion on thechoice of policy mix. The formerly hierarchically or-ganised employment offices are to be converted intocustomer-orientated one-stop-centres. The range ofservices provided has been extended, ranging fromadvising and counselling services to social servicesand the administration of benefit payments. Thecaseload of caseworkers is targeted to be reducedand every job seeker is assigned to a particular case-worker.

A second aspect regards quasi markets: the intro-duction of market forces is expected to improve thequality of services and to break up the informal andoften inefficient insider relationships between publicemployment management and private providers.The reform introduced, for instance, voucher sys-tems for placement services (Vermittlungsgutschein)and training measures (Bildungsgutschein). Each in-dividual whom the public employment service hasbeen unable to place after six weeks of unemploy-ment can choose an alternative private placementservice. The private service receives a lump sumpayment after having placed the job seeker success-fully. Providers of training measures, too, can bechosen freely by the client and paid for with avoucher.

The public employment service can choose to out-source services fully or partly, most importantlyplacement services. Public tendering became com-pulsory for external contracts. One example is place-ment via temporary work: since 2003, every localemployment office sets up a “Staff Service Agency”(Personal Service Agentur (PSA)) that acts as a tem-porary work agency for the unemployed. To thisend, the local employment office may either con-tract out to a private temporary work agency or, ifno provider is available, may run a PSA itself. Thelocal employment office may delegate hard-to-placeclients to the PSA, which in return receives amonthly lump sum fee for employing these workers.The PSA may hire out the worker temporarily toother firms or provide a permanent placement. ThePSA receives a bonus for each placement into em-ployment lasting at least three months and anotherbonus if the placed person is still employed after sixmonths. During periods of inactivity, the PSAshould provide training measures to the worker.Therefore, PSAs encompass aspects of both trainingmeasures and placement services.

Third, the reform aims at improving the targeting ofactive measures and the allocation of measures andresources. To this end, the statutory regulation of eli-gibility conditions is reduced, now leaving a widerscope for individually matching clients to measures.

ZAF 1/2007 51

The assignment of clients to measures is now basedon a profiling process, which is highly standardised.The caseworker will assess the client’s abilities,problems and potential labour market chances in aninterview and then assign the client to one of fourtypes: “Market clients” (Marktkunden) are consid-ered to have the highest chances of finding employ-ment, “clients for counselling and activation” (Bera-tungskunden aktivieren) range second and mainlyneed to be activated in their job search. “Clients forcounselling and support” (Beratungskunden för-dern) need more attention and are likely to be as-signed to a programme, while “clients in need of su-pervision” (Betreuungskunden) need special atten-tion since they face the lowest chances of re-employ-ment. Each type is linked to an action programme,defining the available measures for that type of jobseeker. Active labour market policy measures areavailable mainly for the types II “counselling andactivating” and III “counselling and support”. Thetype I “market client” is expected to re-integratewithout special assistance, while the fourth type,“supervision”, is deemed unlikely to benefit fromany measure and excluded from participation.

Many active policy measures were re-designed interms of their target population. For example, sincethe reform, selection into training measures deliber-ately uses cream skimming in order to choose thoseclients who will benefit most from training. Only cli-ents who are conjectured to have a 70% probabilityof finding a job after the measure will be admitted.Accordingly, training providers have to produce a70% success rate of respective participants in orderto be commissioned by the employment agency. Incontrast, job creation measures have been re-de-signed to target only the very hard-to-place unem-ployed. This means that public employment willconstitute market replacement and thereby preserveemployability for those who are not expected to finda way back into regular employment in the near fu-ture. Incentives for unemployed workers to take uppublic employment rather than regular employmentwere reduced as participants no longer regain eli-gibility for unemployment benefits after completingthe measure. The restrictive targeting of trainingand job creation schemes as well as the reduction ofprogramme durations induced a further reduction ofparticipants in and spending on these measures. Onaverage, participation in training will take place atan earlier stage in the unemployment spell.

Last but not least, the need for rigorous scientificevaluation of programme effectiveness, in order to beable to optimise existing programmes continuouslyon the basis of conclusive empirical evidence, wasrecognized by policy makers, and a corresponding


evaluation mandate was implemented with theHartz reforms. Hence, the Hartz reforms constitutethe first major reform in the history of the Germanwelfare state that is accompanied by a comprehen-sive scientific evaluation on behalf of the govern-ment. The process started with two competing pilotstudies developing a conceptual framework for theevaluation (Fertig et al. 2004; Hagen and Spermann2004) and was subsequently put out to tender. Cur-rently more than 20 economic and sociological re-search institutes with about 100 researchers are in-volved in the evaluation (cf. Bundesregierung 2006for details of the set-up).

3.2 Activation of the unemployed

The principle of “rights and duties” is the core ele-ment of the Hartz reforms. The activation strategyis implemented in virtually every element of the la-bour market policy framework. The new policy mixis more strongly orientated towards measures forjob seekers who actively seek to improve their situa-tion in a self-responsible way. There is more empha-sis on measures that promote the direct integrationinto the labour market as opposed to training meas-ures and public job creation schemes that keep par-ticipants out of the market for the duration of theprogramme. The active measures that best representthe new activation strategy are the start-up subsi-dies. These are a relatively new component of Ger-man active labour market policy. Though a so-called“bridging allowance” (Überbrückungsgeld) was in-troduced as early as 1986, the number of partici-pants has increased only recently. The benefit is paidfor 6 months and is equal to the unemployment ben-efit that the recipient had previously received orcould have received plus a flat-rate social securitycontribution. In order to receive the subsidy, thechamber of commerce has to approve the businessplan. The reform introduced an alternative subsidy,the so-called “Me, Inc.” (Ich-AG) subsidy, which isindependent of prior social security contributions.4

It is paid for a maximum period of three years aslong as the claimant’s income does not exceed25,000 EUR per year. It amounts to 600 EUR permonth in the first year, 360 EUR per month in thesecond and 240 EUR per month in the third year.

Furthermore, integration into paid employment maybe supported by several forms of wage subsidieswhich are paid to employers when hiring a certaintype of hard-to-place worker. The idea is to compen-

4 The Economist (Feb 2006) translates the subsidy as “Me-com-pany”.

52 ZAF 1/2007

sate the firm for the presumably lower productivityof this type of worker. The Hartz reforms simplifiedthe eligibility conditions of so-called integration sub-sidies in order to facilitate access to wage subsidies,giving priority to older and disabled workers. Gen-erally, the maximum duration varies between 6 and24 months, depending on the target group and, as arule, the maximum rate of subsidy should not ex-ceed 50% of the calculable remuneration. In orderto avoid substitution effects and free riding, pay-ments are not available if the employer has appar-ently dismissed a worker in order to receive the sub-sidy or if he had already employed the respectiveworker within the previous four years. Furthermore,the Hartz reform introduced social security subsi-dies for employers who recruit an older worker. Afirm which hires a worker aged 55 or older is exemptfrom paying contributions to the unemployment se-curity system for this worker. However, the contri-bution amounts to only 3.25% of the gross wage.

Finally, various reform elements intend to makework pay, aiming at increasing work incentives forthe unemployed. The reform did not reduce the veryhigh marginal taxes on people who move from un-employment to employment. Instead, new forms ofwage subsidies were introduced and existing oneswere modified or extended. The start-up subsidiesmentioned above are one example. Furthermore,the reform introduced incentives for workers aged50 and older to take up employment even if it paysless than previous employment. In these cases, olderworkers may receive a wage subsidy, the so-calledwage protection, when they accept a job offer thatpays less than their previous job. This wage subsidyamounts to 50% of the difference between the pre-vious wage and the current wage. It is paid for thesame duration as the unemployment benefit wouldhave been paid if the person had remained unem-ployed.

Before the reform, incomes of up to 325 Euros permonth were exempt from social security contribu-tions. The reform raised this threshold to 400 Euros,and substantially simplified the administrative pro-cedures for these now so-called “minijobs”. In addi-tion, the reform introduced “midijobs”, a type ofemployment with reduced social security contribu-tions for the income range of 400,01 to 800 Eurosper month. The amount of social security contribu-tions depends on the gross income and increasesfrom zero contributions at 400 Euros to full contri-butions at 800 Euros. Both minijobs and midijobsare generally applicable, i. e. they are not confinedto the unemployed.

The reform also introduced sanction elements, i. e.additional ways to effectively monitor the job seek-


er’s job search activities and personal efforts to re-integrate into the labour market. The job seeker isobliged to accept any offer of suitable work. Thedefinition of suitable work was broadened, e.g. in-cluding the obligation to move to a different areaunder certain circumstances. Benefit receipt isstrictly conditional on the availability for work andthe availability for programme participation. The in-dividual action plan that results from the profilingprocess is set out in a binding integration agreement(Eingliederungsvereinbarung). This written agree-ment states both the services that will be provided tothe job seeker as well as the job seeker’s obligationregarding job search activities and programme par-ticipation, where required. An unemployed individ-ual will be threatened with sanctions in the form oftemporary benefit reductions if he or she deviatesfrom the integration agreement or does not cooper-ate appropriately. A person’s availability can addi-tionally be tested by training or by workfare meas-ures in the public sector (the so-called 1-EURO-Jobs) or assignment to PSAs.

Although in this paper we focus on active labourmarket policy (Hartz IÐIII), for the sake of com-pleteness we will also briefly comment on the strat-egy for reforming the benefit system (Hartz IV).Whereas before the reform previous social securitycontributions were the key criterion for benefit ac-cess, now access to benefits and participation in ac-tive labour market policy programmes is strictlyconditional on a person’s ability to work. This is de-fined as being able to work at least 15 hours a week.Those capable of working are assigned to the em-ployment agencies and will be subject to activationpolicies on the basis of the principle of “rights andduties”. An unemployed person receives so-calledbenefit type I (Arbeitslosengeld I, or ALG I) for thefirst 6 to 12 months of unemployment. Thereafter,the person receives a flat-rate means-tested benefittype II (Arbeitslosengeld II, or ALG II). Individualswho have never paid social security contributionsbut who are deemed capable of working will receivebenefit type II right from the beginning. Benefittype II is not earnings-based, as the unemploymentassistance was in the previous system, and is lessgenerous than social assistance. Only those who arenot capable of working due to sickness, disability orcare responsibilities receive means-tested social as-sistance from the local authorities and are exemptfrom “duties”. Whereas benefit type I is adminis-tered by the local employment office, benefit typeII is administered at municipal level, either by a co-operation of the employment office and the munici-pality (Arbeitsgemeinschaft, ARGE) or by the mu-nicipality alone (Optierende Kommune).

ZAF 1/2007 53


Deregulation of labour market institutions tookplace regarding temporary work, dismissal protec-tion and the regulation of fixed-term contracts. Thewage setting process, on the other hand, remainshighly centralised.

Regarding temporary work regulations, the reform in-tended to facilitate the expansion of the alreadybooming sector of temporary work significantly. Tem-porary work agencies were legalised in 1967 and for-mally regulated in 1972. Since then, temporary workhad been regulated rather restrictively for manyyears, and had been forbidden completely in the con-struction industry. Since the late 1990s, the law hasgradually been liberalised. The Hartz reform finallyabolished restrictions on synchronisation, re-assign-ment, fixed-term contracts and the maximum dura-tion of temporary employment. A new rule was intro-duced requiring that a temporary work agency musteither guarantee equal pay and equal treatment oftemporary workers and regular workers or join a col-lective bargaining agreement between trade unionsand employers. Temporary work is now also allowedin the construction industry, provided that a collectivebargaining agreement applies. So far, however, nosuch contract has been agreed upon. As already men-tioned, the law introduced publicly sponsored StaffService Agencies as a new form of placement-orien-tated temporary work for hard-to-place individuals.

With respect to dismissal regulations and fixed-termcontract regulations, the reform did not deregulatestandard employment relations in a general way. In-stead, it simplified and widened the number of casesfor which exemptions from the generally rather re-strictive regulations apply. Before the reform, for in-stance, exemptions from restrictions on fixed-termcontracts applied for employees aged 58 and over. Forthese employees, fixed-term contracts could be re-newed repeatedly without justification. The reformreduced the minimum age for which this regulationapplies to 52 years.5 Furthermore, exemptions fromdismissal protection, which before the reform wereconceded to small firms with 5 employees or less, willnow apply to firms with up to 10 employees.

4 Labour market policy effectivenessbefore and after Hartz

For a long time the evaluation of German ALMPsuffered from a lack of suitable data. Only very re-

5 The reduction of the age threshold for exemptions regardingfixed-term contracts has meanwhile been declared null and voidby a ruling of the EU Court of Justice.


cently has good quality data become relativelywidely available to researchers. Early studies onALMP effectiveness were usually based on theGSOEP (German SocioEconomic Panel) or, foreastern Germany, the Labour Market Monitor East.The main drawback of these data is that, due torather small overall sample sizes and panel mortal-ity, they contain only few observations on partici-pants of active labour market policy measures. Re-searchers often had to group together heteroge-neous measures and some programmes could not beevaluated at all because participation was not docu-mented in the data.

It was only at the end of the 1990s that the govern-ment started to acknowledge the need for a thor-ough evaluation of active labour market policies,and, in the following years, considerable effort wasmade to derive large data sets from administrativedata at local employment office level (Bender et al.2005). These data provide a large number of obser-vations and cover rather long time periods. Theytherefore make it possible to detect short-term aswell as long-term effects and provide enough infor-mation to better distinguish different types of treat-ment and to analyse the optimal timing of events.These merged administrative data have recentlybeen made available to researchers and seem to beable to provide robust results. Most of the evalua-tion studies of the Hartz reforms make use of thistype of data.

Early evaluation studies mainly concentrated ontraining and job creation schemes, which for a longtime were the most important measures in terms ofexpenditure and number of participants (recall Fig-ure 2). Fitzenberger and Speckesser (2000) providea survey on early evaluation studies in Germany,most of which are based on the above-mentionedrather poor data. Caliendo and Steiner (2005), whoupdate that review, and Wunsch (2005), who dis-cusses the development of the German labour mar-ket since unification, include recent pre-Hartz stud-ies based on the new and better data. The post-Hartz studies we discuss in the following sectionsare first results of the evaluations of the Hartz lawsIÐIII. The evaluation of Hartz IV, which basicallycomprises the reform of the benefit system and theintroduction of benefit type II combining unemploy-ment and social assistance, will not begin until theautumn of 2006.

4.1 Increasing the effectiveness andefficiency of labour market servicesand policy measures

Major policy changes aiming at improving the ef-fectiveness and efficiency of existing measures took

54 ZAF 1/2007

place in the realm of placement services, trainingand job creation schemes. In the following, we willdiscuss findings from evaluation studies of thesemeasures before and after the reform. Other activepolicy measures that were re-designed under the re-form are discussed in the next sections.

4.1.1 Placement services

There are only two studies on the effectiveness ofplacement services before the reform. The DEA-based benchmarking study by Mosley et al. (2003)compares the relative efficiency in terms of activity(number of placements and programme entrants inrelation to staff inputs) and effectiveness (transi-tions into regular employment) of western Germanlocal employment offices. The analysis is supple-mented with case studies of eight employment offi-ces. The results indicate strong differences in rela-tive efficiency. The authors find that exogenous andendogenous factors each explain roughly 50% ofvariation in efficiency. They suggest that the averageefficiency of labour market offices could be in-creased by 19%. Hujer et al. (2005) study the effectsof two pilot measures of job search assistance (Stel-lenmarktoffensive) in 2001Ð2002 in the western Ger-man federal state of Hesse. The regional employ-ment office published a magazine for employerscontaining employment-wanted advertisementsamong other things. The study finds positive effectsof this measure, especially for women. Furthermore,the employment offices offered courses on jobsearch activities that provide advice on writing ap-plication letters, CVs, participating in job interviewsetc. Although the effects vary across groups, in gen-eral these measures do not produce positive results.

The Hartz reform converted the former employ-ment offices into customer-orientated service cen-tres (Kundenzentrum). These are assessed in thestudy by WZB and infas (2005) using a conditionaldifference-in-differences-analysis which exploits thefact that the customer service centres were intro-duced at different points in time. Ten employmentservice offices that have already been transformedinto customer service centres are matched with tenoffices that have not. Data is used from the inflowinto unemployment of the respective offices. The re-sults indicate positive effects of customer servicecentres on the integration into regular employment,though the effects are not significant. This might bedue to the fact that the number of offices used inthe analysis is small and the observation period atthe present time is a maximum of nine months. Theeffects are more positive in eastern Germany, wherelabour market conditions are worse, and seem towork better for men than for women.


Furthermore, the report studies the outsourcing ofplacement services by comparing clients who havemade use of external placement services with clientswho have remained under the public employmentservice. External placement services can be utilisedeither by the client himself, using a placementvoucher (Vermittlungsgutschein), or by the publicemployment service who can assign clients to exter-nal providers (Beauftragung Dritter). Selection iscontrolled for by performing a combination of exactmatching with propensity score matching. The ana-lysis is based on administrative data which is supple-mented by survey data from telephone interviewsfor a sub-sample. For placement vouchers, the re-sults fail to find any significant effect on theprospects of entering employment. It appears thatmany clients who had received placement vouchersdid not actually use them. In eastern Germany,those who found a job using a placement voucherremained in employment for a significantly shorterperiod than those who did not use placement vouch-ers. The assignment of clients to private placementproviders by the public employment service doesnot show significant effects either. Finally, the as-signment to PSAs produces significant locking-in ef-fects that delay the integration of workers into regu-lar, non-PSA employment.

4.1.2 Training measures

Training measures have been evaluated by numer-ous studies. Since studies mostly focus on eithereastern or western Germany, we present results foreach region separately. Early evaluation studies ontraining in eastern Germany include Lechner (1998,1999), Hujer and Wellner (2000) and Lechner(2000), which are based on the GSOEP. Studiesbased on the Labour Market Monitor East includeHübler (1997) and Fitzenberger and Prey (1998).Bergemann et al. (2000) use the Labour MarketMonitor East of the federal state of Saxony-Anhalt,as do Bergemann et al. (2004). Reinowski et al.(2003, 2004) use the Saxony microcensus. Recentstudies based on merged administrative data areHujer, Thomsen and Zeiss (2004), Fitzenberger andSpeckesser (2005) and Lechner et al. (2005).

Many of the early studies either find positive effectsor are unable to find any significant effects of train-ing programmes in eastern Germany. One exceptionis Hübler (1997), who finds negative effects forwomen. This result contrasts with the finding ob-tained by Bergemann et al. (2000) of significantlypositive effects of second treatments for womenonly. Bergemann et al. (2004) find positive effects inthe early 1990s and negative effects in later years. Ingeneral, studies published from 2000 onwards tend

ZAF 1/2007 55

to be more pessimistic. Besides Bergemann et al.(2004), also Lechner (2000) and Reinowski et al.(2003) find negative effects of training participation.By and large, the results are mixed and it is ratherunclear what lesson can be drawn from these studieson the programmes’ effectiveness.

Recent studies based on better administrative dataseem to derive more consistent results. Hujer,Thomsen and Zeiss (2004) use data from the period1999Ð2002 and Fitzenberger and Speckesser (2005)cover the period 1993Ð1997. The most comprehen-sive study is the one by Lechner et al. (2005), whouse data covering the period of 1993 to 2002. Dueto the richness of the data, various types of trainingcan be distinguished. Fitzenberger and Speckesser(2005) concentrate on the “provision of specific pro-fessional skills” which is a special type of furthervocational training programmes. Lechner et al.(2005) distinguish between short training (up to 6months), long training (over 6 months), retraining,and training in practice firms.

All of the studies based on the new data find signifi-cant evidence of locking-in effects for virtually alltypes of training, i. e. the labour market performanceis worse for participants compared with non-partici-pants during and shortly after participation. Thecentral question is whether there are positive effectsin the medium and long run that are big enough tocompensate for these negative short-run effects. Theanswer seems to depend on the outcome variable.For unemployment duration, Hujer, Thomsen andZeiss (2004) do not find significant long-term effectsof short and medium training programmes but findnegative effects of long programmes, which meansthey increase unemployment duration (here, a par-ticipating person is considered unemployed). Thiscontrasts with the other studies, Fitzenberger andSpeckesser (2005) and Lechner et al. (2005), whotake the employment rate as the outcome measureand find positive effects in the long run for pro-grammes that provide specific professional skills(Fitzenberger and Speckesser 2005) as well as forshort training and retraining programmes (Lechneret al. 2005). Lechner et al. (2005) also use monthlyearnings as an outcome variable and again find posi-tive effects in the long run.

Early studies on training in western Germany in-clude Pannenberg (1995), Hujer et al. (1998) andHujer and Wellner (2000). These studies use modelsof unemployment duration and are based onGSOEP data covering the second half of the 1980sand the early 1990s. Again, the results are mixed:Pannenberg (1995) and Hujer et al. (1998) do notfind significant positive effects, while Hujer and


Wellner (2000) do find positive effects, but for short-term programmes only. More recent studies basedon administrative data are Klose and Bender (2000),Lechner et al. (2004) and Fitzenberger and Speck-esser (2005). Klose and Bender (2000) use a prelimi-nary version of the data. Fitzenberger and Speck-esser (2005) use the final data covering the period1993Ð1997, while Lechner et al. (2004) base theirstudy on data covering the larger period of 1993 to2002.

Klose and Bender (2000) do not find any positiveeffects, which might be due to the preliminary char-acter of their data. In contrast, Lechner et al. (2004)as well as Fitzenberger and Speckesser (2005) cometo quite optimistic results. Fitzenberger and Speck-esser (2005) find negative locking-in effects on theemployment rate in the short run and significantlypositive effects in the long run for training pro-grammes providing specific professional skills. Thefindings of Lechner et al. (2004) suggest that shortand long training have positive effects on employ-ment rates in the short run. In the long run shorttraining and retraining show positive results. Fur-thermore, they find significantly positive effects onmonthly earnings for short and long training.

As delineated in section 3, the reform changed theusage and set up of training in various aspects. First,positive effects might be expected from the reduc-tion of participants and deliberate cream skimmingas part of the selection process. Second, the durationof programmes has been reduced and participationtakes place, on average, at an earlier stage in theunemployment spell. Third, course quality is ex-pected to improve due to increased competition be-tween providers and the priority of efficiency crite-ria.

The study evaluating training measures post-Hartzwas conducted by IZA, DIW and infas (2005) anduses administrative data and survey data to comparethe effect of training measures before the reform inthe period 2000 to 2003 with effects after the reformin the period 2003 to 2005. The results confirm theprevious results of severe locking-in effects. Theysuggest that the positive results in the medium andlong run are based on the positive employment ef-fect on people who otherwise would have driftedinto non-participation. Furthermore, effects of pre-reform measures seem to be less positive when tak-ing employment stability into account. As expected,the results indicate that the reform succeeded in sig-nificantly reducing locking-in effects, though evi-dence on long-term effects of the modified trainingmeasures are not yet available at the present time.The cost-effectiveness of measures before the re-

56 ZAF 1/2007

form was negative. By shortening course durationsand better targeting, the reform was able to reducethe gap between costs and benefits, though the bal-ance is still negative.

4.1.3 Job creation schemes

For a long time, job creation schemes could beevaluated only for eastern Germany because datasources that provide information on participation injob creation schemes were limited to eastern Ger-many only. These are the Labour Market MonitorEast which is used by Hübler (1997), the LabourMarket Monitor of the federal state of Saxony-An-halt, which is used by Bergemann et al. (2000), Eich-ler and Lechner (2002) and Bergemann (2005), andthe Saxony microcensus used by Reinowski et al.(2003).

None of the studies finds positive effects on the em-ployment rate, apart from Eichler and Lechner(2002) who find positive employment effects, al-though for men only. Reinowski et al. (2003) usethe hazard rate of transition from unemployment toemployment as a dependent variable, where unem-ployment spells include periods of participation.They do not find positive effects of programme par-ticipation. Bergemann (2005) finds that for womenparticipation significantly increases the re-employ-ment probability. Furthermore, she reports signifi-cantly positive effects on men’s and women’s prob-ability of remaining employed. Caliendo et al.(2003) use the recently derived administrative datafor the years 2000Ð2002, which provides informa-tion on programme effects in western Germany forthe first time. Their results are pessimistic, revealingnegative mean employment effects. Positive employ-ment effects are limited to a few socio-demographicgroups, namely women over 50, the long-term un-employed and hard-to-place women in western Ger-many as well as female long-term unemployed ineastern Germany. However, since the observationperiod is rather short, the negative effects might rep-resent locking-in effects similar to those found fortraining programmes.

The evaluation study by SÖSTRA, Compass, IMU,PIW (2005) provides new evidence for the period2000Ð2004. At the present time, the results on jobcreation schemes after Hartz are still preliminary.The authors use administrative data of people whoentered job creation measures in April of the years2000 to 2004. Control groups are constructed usingmatching methods. The programme effect on theprobability of leaving unemployment is assessed bycomparing the survival functions of treated and non-treated groups, where unemployment spells include


the time spent in unemployment before participa-tion started. The study confirms the generally nega-tive effects of participation in job creation measuresfor time periods before Hartz, though effects seemto be positive in the long run in western Germany.The results suggest that the detrimental effects ofjob creation measures regarding unemployment du-ration are entirely caused by the locking-in of parti-cipants, while stigma effects do not seem to play arole. Regarding the effects of the Hartz reforms, thestudy finds negative treatment effects for the post-reform period, too, although the magnitude of theeffects seems to have decreased. Since 2004 it hasbeen a statutory objective of job creation schemesto generate or preserve the “employability” of par-ticipants, rather than providing actual employmentonly.

4.2 Activating the unemployed

The Hartz reform shifts priority towards measuresthat require active behaviour of the unemployedand promote their direct integration into regularemployment (cf. section 3.2). To this end, the reformre-designed integration subsidies, introduced newforms of wage subsidies, start-up subsidies and jobswith reduced social security contributions. In the fol-lowing section we review evaluation studies of suchmeasures before the reform, if applicable, and sub-sequently focus on the novel evidence post-reform.

4.2.1 Wage subsidies to employers

The challenge of controlling adequately for self se-lection is especially difficult in the context of inte-gration subsidies. Here, selection into treatmentdoes not only depend on the characteristics of theparticipant, but also on characteristics of the poten-tial employer who will receive the subsidy. The caseworker, moreover, has a wide scope of discretion fordetermining whether the characteristics of employerand employee jointly satisfy the criteria required forthe subsidy. There are only two studies on the effectsof integration subsidies (Eingliederungszuschüsse)before Hartz. Jaenichen (2002) collects administra-tive data and survey data for a cohort of unem-ployed workers that started subsidised employmentbetween January and April 1999 and a control groupfrom selected Federal Employment Service districtsthroughout Germany. She finds that participating inintegration subsidy programmes significantly redu-ces the probability of a worker being registered asunemployed. However, this might be partly due tothe fact that, upon entering the programme, partici-pants are no longer counted as unemployed and

ZAF 1/2007 57

many participants are still in the programme at theend of the observation period. In the long run, whenthe subsidy has usually expired, the effect is stillpositive, although significantly so only in easternGermany. Hujer, Caliendo and Radi (2004) use firmdata to examine whether employing subsidizedworkers affects the employment development offirms. Based on the IAB establishment panel datacovering the years 1995Ð1999 they cannot find anysignificant effects.

The Hartz reform provided the opportunity to cir-cumvent the selection problems that were outlinedabove, by changing the eligibility conditions for inte-gration subsidies. The study by ZEW, IAB and IAT(2005) uses administrative data of the years 2000 to2003 and exploits the fact that since the beginningof 2002 older unemployed workers no longer needto satisfy the condition of being long-term unem-ployed in order to be eligible for integration subsi-dies. The effect of wage subsidies on the employ-ment prospects of older workers is assessed using adifference-in-differences estimator. The results indi-cate that integration subsidies increase the probabil-ity of being employed 6 months after entering un-employment by 2 percentage points. These effectsseem to stem from significant positive effects in east-ern Germany only, while in western Germany theeffects are insignificant. Within the target group thenumber of windfall beneficiaries seems to be low.The longer-term effects of integration subsidies onthe probability of being employed after the subsidyends is assessed using propensity score matching.The results suggest that, depending on the type,length and target group of the subsidy, this probabil-ity is 20 to 50 percentage points higher for the treat-ment group than for the comparison group. How-ever, there are some indications of windfall gains.

4.2.2 Start-up subsidies

Only little empirical evidence exists on the effective-ness of the “bridging allowance” start-up subsidy(Überbrückungsgeld) for the unemployed. The studyby Pfeiffer and Reize (2000) compares firm survivaland employment growth of start-ups by unemployedpersons receiving the bridging allowance and other,regular start-ups, based on firm data from 15 regionsin eastern and western Germany. The results indi-cate that the survival rate and employment effects ofsubsidised start-ups do not differ from unsubsidisedstart-ups. The Hartz evaluation study by IAB, DIW,Sinus, GfA, and infas (2005) also provides results onthe effectiveness of the bridging allowance beforethe reform. The analysis is based on administrativedata of cohorts entering unemployment in 2000which are observed until 2002. The effectiveness of


start-up subsidies is assessed using matching meth-ods. The authors suspect, however, that some selec-tion bias might remain due to unobservable charac-teristics of unemployed people who decide to startup a business. The results show that 6 months afterthe bridging allowance expires participants have alower probability of being unemployed. The effectis significant at a high level, especially for women.For the period after the reform, the data is supple-mented by survey data for the treatment and controlgroups. The study confirms the positive results ofthe bridging allowance after the reform.

For the second, new type of start-up subsidy, the so-called “Me, Inc.” (Ich-AG) subsidy, most partici-pants are still receiving the subsidy at the end ofthe observation period. Information on effects aftereligibility expires is therefore not available yet.However, the subsidy, which decreases gradually, isgenerally very low at the end of the observation pe-riod. Therefore, the significantly positive effects thatcan be observed at that stage might be expected topersist further as well. At the same time the studyreports indications of windfall gains as a substantialnumber of start-up subsidy recipients report thatthey would have started a business even without thesubsidy. These figures amount to around 25Ð60%for bridging allowance recipients, and to 60Ð70% ofindividuals in the “Me, Inc.” scheme. Even for thesebusinesses, however, the subsidy might still have ex-erted a positive effect during the first months inbusiness, rather than on business creation per se.

4.2.3 Wage protection for older workers

Wage protection (Entgeltsicherung) is a wage sub-sidy for workers aged 50 and older who take up em-ployment in a job that pays less than their previousjob. The effect of the subsidy on the employmentprospects of older workers is studied by ZEW, IABand IAT (2005). The analysis employs a difference-in-differences estimator using administrative datafrom the years 2002 and 2003, where workers aged50 or 51 are the treatment group and workers aged48 or 49 are the control group. The results indicatea positive though insignificant effect of the subsidyon employment probabilities. The authors suggestthat the insignificance might be due to low take-upof wage protection.

4.2.4 Employment with reduced socialsecurity contributions (minjob/midijob)

Various reform elements, including the introductionof minijobs and midijobs, are evaluated in terms oftheir impact on the general structure of employ-ment. These studies are part of the report by RWI

58 ZAF 1/2007

et al. (2005). The introduction of jobs with reducedsocial security contributions for the income rangebetween 400 and 800 EUR (midijobs), and the re-form of marginal employment waiving social secu-rity contributions for incomes below 400 EUR(minijobs) is expected to increase the number ofpeople working in these earnings segments. Mini-jobs and midijobs constitute a “universal treat-ment”, i. e. everybody in the labour market is af-fected, and no comparison group without treatmentexists. Its effects are examined by first estimatingthe individual probability of each worker being em-ployed in the respective earnings segment before thereform, using a fixed-effects linear probabilitymodel. Then, assuming that structural parameterswould have remained identical without the reform,the counterfactual probability of employment inthese segments is estimated on the basis of post-re-form data utilising the estimated pre-reform param-eters. The reform effect is then given by the differ-ence between the estimated post-reform probabili-ties using pre-reform coefficients and the actualpost-reform probabilities.

The results show that the introduction of midijobscaused a significant increase of about 125,000 in thenumber of employees in this income range, whilethe minijob reform caused a huge expansion of em-ployment in this earnings segment (+1.8 millionminijobs due to the reform). However, this increaseincludes approximately 250,000 persons employed inthe income range of 326 to 400 Euros prior to thereform, who by definition became “minijobbers”.Also included are around 850,000 persons who holdtheir minijob in addition to regular employment, astatus not previously captured in the data.

Moreover, the incidence of intra-enterprise dis-placement of regular jobs cannot be ruled out, andemployees who benefit from the regulations wererarely previously unemployed, in spite of hopes as-sociated with this policy that especially unemployedindividuals would increasingly take on minijobs andmidijobs. This might be due to the fact that the mar-ginal tax on people who move from unemploymentto employment is still very high.


As is the case for the evaluation of the minijobs andmidijobs reforms, deregulation policies regardingtemporary work and fixed-term contracts are evalu-ated with respect to the general employment struc-ture. The exemption of small firms from dismissalprotection regulations, however, was not subjectedto an evaluation.


4.3.1 Deregulation of temporary work

The reform of temporary work regulations is evalu-ated in the report by RWI et al. (2005) applying thesame methodology for universal treatments as forthe minijob/midijob evaluation outlined above. Theresults indicate that the reform significantly in-creased the number of employees in the temporarywork sector in the two quarters after the reform. Nodata is available yet for longer periods.

4.3.2 Fixed-term contracts for older workers

The effect of older workers being exempt from re-strictions on fixed-term contracts (erleichterte Befris-tung älterer Arbeitnehmer/innen) is assessed by RWIet al. (2005) using a difference-in-differences estima-tor. Workers just above the minimum age for ex-emption (52 to 53 years old) constitute the treat-ment group, while workers just below this age (50Ð51 years old) are the control group. The exemptionof workers over 52 years of age from restrictionson fixed-term contracts is expected to increase thenumber of older workers holding fixed-term con-tracts. The analysis does not reveal any significanttreatment effects. This might be due, however, to theshort observation period, since currently data areavailable for only three months after the exemptionwas introduced.

4.4 Macro-level evaluation

In the evidence presented above, we have concen-trated on micro-econometric studies dealing withsingle instruments of ALMP and their isolated ef-fects. These kinds of studies have their virtues, butthe evidence they provide is also limited. They areunable to uncover substitution and displacementprocesses and other effects that affect labour marketperformance at aggregate level. This could be thecase in regions like eastern Germany, where labourmarket programmes are implemented on a verylarge scale. A considerable number of studies try tograsp such indirect effects in Germany, focusing onthe macroeconomic impact of active labour marketpolicies. Recent studies include Schmid et al. (2001),Fertig et al. (2006), Hagen (2003), Blien et al.(2003), Vollkommer (2004) and Hujer et al. (2005).All of these studies use administrative data from re-gional employment offices. Most studies focus onsome indicators of matching efficiency. Exceptionsare Pannenberg and Schwarze (1996), who examinewage effects, and Blien et al. (2003), who use re-gional employment growth as a dependent variable.The study by Pannenberg and Schwarze (1996) indi-cates a small negative impact of training measures

ZAF 1/2007 59

on monthly wages in eastern Germany. Blien et al.(2003) provide evidence of positive effects of train-ing and job creation programmes on employmentgrowth in eastern Germany. There seems to be evi-dence of displacement effects by job creationschemes, as Hagen (2003) suggests. Further resultsregarding the effect on matching efficiency and un-employment rates are very mixed and rather incon-clusive.

The report by RWI et al. (2005) assesses the macro-economic effects of ALMP after the Hartz reformsusing administrative data of labour market out-comes and the respective “policy mix” at regionallevel. Regions are constructed as local labour mar-kets on the basis of commuter flows in order to con-trol for spill-over effects of regional policies. The re-sults suggest that spending a larger amount on start-up subsidies significantly increases the number ofnet exits from unemployment. The introduction of anew form of start-up subsidy by the Hartz reformreduced its magnitude, but the effect is still positive.In contrast, spending on job creation measures hasa significantly negative effect on net exits from un-employment in eastern Germany and no effect inwestern Germany. Also integration subsidies haveno effects on net exits. Generally, the entire reform,regardless of the specific policy mix, seems to havehad positive effects on the net exit rate of the short-term unemployed, while no effects could be foundfor the long-term unemployed. The positive resultsare mainly driven by positive effects on men.

4.5 Summary of policy effectivenessbefore and after Hartz

Table 2 presents an overview of the effectiveness ofvarious policy measures before and after Hartz. Thetable follows the previous structure of section 4(also used in section 3) in distinguishing betweenreform elements that aim at a) increasing the effec-tiveness and efficiency of labour market services andpolicy measures, b) activating the unemployed, andc) deregulating the labour market. The second col-umn summarises the estimated effects of the respec-tive policy arising from the available set of evalua-tion studies conducted before Hartz. Obviously, forseveral policies, such as the newly introduced place-ment vouchers, such an assessment is not applicable.The third column summarises the results obtainedfor each measure in the comprehensive evaluationof the effectiveness of the Hartz reforms (cf. alsoBundesregierung 2006 and Kaltenborn et al. 2006).Column 4 is then intended to assess the “before-after difference”, which Ð with explicit caution Ð


could be interpreted as a “reform effect” of themeasure. The caution results from the fact that,whereas the before-after difference may indeed becaused by the redesign of the measure, it is also con-ceivable that the overall change in the institutional

60 ZAF 1/2007

framework of the labour and placement markets isresponsible for differentially effective policies pre-Hartz and post-Hartz. Also, some evaluation studiesbefore Hartz are quite distinct in their specific focusfrom those conducted after Hartz Ð e.g. the studies


concerning general placement services Ð and henceare not directly comparable. Another caveat wheninterpreting the results is the fact that the post-re-form observation period is short-term only.

The table summarizes the results discussed in moredetail in the previous sections. Training and publicjob creation schemes, both of which were alreadyused and evaluated extensively before the reform,appear to have been improved. The “improvement”of public job creation schemes, however, unfortu-nately merely implies that since the reform they ap-pear less detrimental for participants’ employmentprospects than before. For another re-designedmeasure, wage subsidies, the reform provided thefirst opportunity to consistently evaluate its effectsand thus to reveal its apparently high effectiveness.

Several new measures Ð placement vouchers, as-signment to private placement providers, fixed-termcontracts for older workers Ð do not display signifi-cant effects, which may be due to de facto ineffectiveor small-scale policies, or perhaps due to the factthat the post-reform observation period is not yetlong enough. While placement via temporary work(PSA) shows negative treatment effects, the newstart-up subsidy significantly reduces the risk of un-employment. Both the deregulation of the tempo-rary work sector and the introduction and reform,respectively, of jobs with reduced social securitycontributions (midijobs and minijobs) appear tohave created additional employment opportunitiesin the respective labour market segments. However,intra-enterprise displacement effects cannot beruled out. The redesign of the customer service ofpublic employment offices appears promising.

On balance, we therefore find that the Hartz re-forms in their entirety seem to have contributed toa better functioning of the German labour marketand the effectiveness of specific active labour mar-ket policies. This positive assessment, however, hasto be qualified somewhat in the light of the fact thatthe starting situation upon which the reforms in-tended to improve was quite dismal.

5 Conclusion

In this paper we have described the features of Ger-man labour market policy and delineated the ration-ale for the Hartz reforms implemented in the years2003Ð2005. We have described the main underlyingprinciples and the corresponding policy changes.Following this we have surveyed the existing evi-dence on the effects of active labour market policybefore and after the reforms.

ZAF 1/2007 61

German ALMP before Hartz was dominated bytraining and public job creation measures. Thesemeasures were characterised by a long durationcompared with other countries. Especially in easternGermany the extensive use of job creation measurescreated a sheltered labour market of substantialmagnitude. In contrast, measures directly supportingintegration into regular employment (e.g. wage sub-sidies and start-up subsidies) were introduced rela-tively recently and played a minor role before. As-signment to programmes was not based on a system-atic profiling of clients. Generally, job search assist-ance and monitoring by the public employmentagency was given rather low priority. It was arguedthat the main weaknesses of the former labour mar-ket policy in Germany were, firstly, public employ-ment services operating inefficiently, and secondly,the fact that the interplay of active policy measureswith the generous benefit system created work disin-centives that retained the unemployed in passivityrather than stimulating them to integrate into theregular labour market.

The Hartz reforms aimed at improving employmentservices and policy measures, and activating the un-employed. First, public employment services weremodernised along the lines of New Public Manage-ment. This includes results-based accountability oflocal employment offices, the outsourcing of manyservices and open competition between private ser-vice providers. The former employment offices wereconverted into customer-orientated one-stop-cen-tres, offering individual profiling, job search assist-ance, social services and administration of benefitpayments. Second, various policy changes imple-ment an activation strategy according to the princi-ples of “rights and duties”. The entire benefit systemwas re-designed. Unemployment benefit levels anddurations were reduced. Eligibility for subsistenceallowances now differs according to a person’s abil-ity to work rather than according to previous contri-bution payments, as was the case before. Benefit re-cipients may also be subject to sanctions, mainlybenefit reductions, if duties are not complied with.Priority is given to measures that support unem-ployed workers who are pro-actively seeking inte-gration into regular employment, most importantlywage subsidies and start-up subsidies. Jobs with re-duced social security contributions were introduced(midijobs), and the regulation for jobs exempt fromany social security contributions was reformed(minijobs), both with the intention of providinggreater incentives for individuals to take up employ-ment in the low-wage sector. Third, the temporarywork sector was deregulated and exemptions fromrestrictions on fixed-term contracts and dismissalprotection were introduced.


Results of evaluation studies of German active la-bour market policies have been rather inconsistentfor many years. This was probably due to a lack ofappropriate data, or to the sensitivity of results withrespect to different identification strategies. How-ever, due to better data, advances in methodologyand a higher consensus on identification strategies,recent evaluation studies seem to be able to providemore robust and consistent results. For the pre-Hartz period, it can be concluded for example thatmost training measures seem to show considerabledynamics in programme effects, having negative(locking-in) effects in the short run and a tendencytowards positive employment effects in the long run.Based on such results, future cost-benefit analysesmight be able to trade costs of negative short-runeffects against benefits of positive long-run effects.Moreover, there is evidence that job creationschemes perform badly on average in the short run,and actually result in impaired employmentprospects for participants. The (limited) evidence onwage subsidies and start-up subsidies pre-Hartz indi-cates modestly positive effects of such measures.

The comprehensive evaluation of the Hartz reforms,which involves more than 20 research institutionsand a total of about 100 researchers, has producedrich evidence giving the general impression that theeffectiveness of measures has improved modestly.The results, which are still preliminary at the presenttime, indicate that the re-organisation of public em-ployment services was mainly successful, with theexception of the outsourcing of services. Re-design-ing training programmes seems to have improvedtheir effectiveness, while job creation schemes con-tinue to be detrimental to the probability of re-em-ployment. Policy measures such as the redesignedwage subsidies and start-up subsidies show signifi-cantly positive effects. Thus, the new strategy, withmore emphasis on wage subsidies and start-up subsi-dies and less emphasis on training and public jobcreation schemes, seems to be a promising mix ofactive labour market policies, moving the Germanlabour market in the right direction. It seems clear,however, considering the continuing crisis of theGerman economy, that further steps have to betaken.

References

Bender, S./A. Bergemann/B. Fitzenberger/M. Lechner/R.Miquel/S. Speckesser/C. Wunsch (2005): Über die Wirk-samkeit von FuU-Maßnahmen, Ein Evaluationsversuchmit prozessproduzierten Daten aus dem IAB, Nürnberg,Beiträge zur Arbeitsmarkt- und Berufsforschung.

62 ZAF 1/2007

Bergemann, A. (2005): Do Job Creation Schemes InitiatePositive Dynamic Employment Effects?, mimeo, De-partment of Economics, Free University Amsterdam.

Bergemann, A./B. Fitzenberger/B. Schultz/S. Speckesser(2000): Multiple Active Labour Market Policy Participa-tion in East Germany: An Assessment of Outcomes,Konjunkturpolitik, Beiheft Nr. 1, S. 195Ð243.

Bergemann, A./B. Fitzenberger/S. Speckesser (2004): Evalu-ating the Dynamic Employment Effects of Training Pro-grammes in East Germany Using Conditional Differ-ence-in-Differences, ZEW DP No. 04-41.

Blien, U./E. Maierhofer/D. Vollkommer/K. Wolf (2003):Einflussfaktoren der Entwicklung ostdeutscher Regio-nen Ð Theorie, Daten, Deskriptionen und quantitativeAnalysen, in: U. Blien (ed.), Die Entwicklung der ost-deutschen Regionen, Beiträge zur Arbeitsmarkt- undBerufsforschung 267, 67Ð254.

Bundesregierung (2006): Die Wirksamkeit modernerDienstleistungen am Arbeitsmarkt Ð Bericht 2005 derBundesregierung zur Wirkung der Umsetzung der Vor-schläge der Kommission Moderne Dienstleistungen amArbeitsmarkt (Ohne Grundsicherung für Arbeitssuchen-de), http://www.bmas.bund.de/BMAS/Redaktion/Pdf/hartz-evaluation-volltext,property=pdf,bereich=bmas,sprache=de,rwb=true.pdf (Feb 1st, 2006).

Caliendo, M./R. Hujer/S. L. Thomsen (2003): The Employ-ment Effects of Job Creation Schemes in Germany: AMicroeconometric Evaluation, IZA DP No. 1512.

Caliendo, M./V. Steiner (2005): Aktive Arbeitsmarktpolitikin Deutschland: Bestandsaufnahme und Bewertung dermikroökonomischen Evaluationsergebnisse, DIW DP515.

Eichler, M./M. Lechner (2002): An Evaluation of PublicEmployment Programs in the East German State ofSachsen-Anhalt, Labour Economics 9, 143Ð186.

Engels, D. (2001): Abstand zwischen Sozialhilfe und an-derem Arbeitseinkommen: Neue Ergebnisse zu einer al-ten Kontroverse, Sozialer Fortschritt 3/2001.

Fertig, M./J. Kluve/C. M. Schmidt/H. Apel/W. Friedrich/H.Hägele (2004): Die Hartz-Gesetze zur Arbeitsmarktpoli-tik Ð Ein umfassendes Evaluationskonzept, RWI-Schrif-ten 74, Duncker & Humblot: Berlin.

Fertig, M./C. M. Schmidt/H. Schneider (2006): Active La-bour Market Policy in Germany Ð Is There a SuccessfulStrategy?, Regional Science and Urban Economics,forthcoming.

Fitzenberger, B./H. Prey (1998): Beschäftigungs- und Ver-dienstmöglichkeiten von Weiterbildungsteilnehmern imostdeutschen Transformationsprozeß: Eine Methoden-kritik, in Pfeiffer, F. and W. Pohlmeier (eds): Qualifika-tion, Weiterbildung und Arbeitsmarkterfolg, ZEW Wirt-schaftsanalysen 31, Baden-Baden: Nomos.

Fitzenberger, B./S. Speckesser (2000): Zur wissenschaftli-chen Evaluation der Aktiven Arbeitsmarktpolitik in


Deutschland: Ein Überblick, Mitteilungen aus der Ar-beitsmarkt- und Berufsforschung 33, 357Ð370.

Fitzenberger B./S. Speckesser (2005): Employment Effectsof the Provision of Specific Professional Skills and Tech-niques in Germany, mimeo, Department of Economics,Goethe-University, Frankfurt am Main.

Franz, W. (2001): Neues von der NAIRU?, Jahrbücher fürNationalökonomie und Statistik, 221, 256Ð284.

Hagen, T. (2003): Three Approaches to the Evaluation ofActive Labour Market Policy in East Germany UsingRegional Data, ZEW DP No. 03-27.

Hagen, T./A. Spermann (2004): Hartz-Gesetze: Metho-dische Ansätze zu einer Evaluierung, ZEW Wirtschafts-analysen 74, Nomos: Baden-Baden.

Hübler, O. (1997): Evaluation beschäftigungspolitischerMaßnahmen in Ostdeutschland, Jahrbücher für Natio-nalökonomie und Statistik 216, 21Ð44.

Hujer, R./Blien, U./Caliendo, M./Zeiss, C. (2005): Macro-econometric Evaluation of Active Labour Market Poli-cies in Germany Ð A Dynamic Panel Approach UsingRegional Data, forthcoming in: Carloeo, F. E., Deste-fanis, S. (eds.): Regions, Europe and the Labour Market.Recent Problems and Developments, Physica Verlag.

Hujer, R./M. Caliendo/D. Radi (2004): Estimating the ef-fects of wage subsidies on the labour demand in WestGermany using the IAB establishment panel, MIKAS ÐMikroanalysen und amtliche Statistik, Wiesbaden: Sta-tistisches Bundesamt, 249Ð283.

Hujer, R./M. Caliendo/S. Thomsen (2005): Mikroökonome-trische Evaluation des Integrationserfolges, in Schaade,P. (ed): Evaluation des hessischen Modells der Stellen-marktoffensive, Beiträge zur Arbeitsmarkt- und Berufs-forschung 291, Nürnberg.

Hujer, R./K. Maurer/M. Wellner (1998): Kurz- und langfris-tige Effekte von Weiterbildungsmaßnahmen auf die Ar-beitslosigkeitsdauer in Westdeutschland, in Pfeiffer, F.and W. Pohlmeier (eds): Qualifikation, Weiterbildungund Arbeitsmarkterfolg, ZEW Wirtschaftsanalysen 31,Baden-Baden: Nomos.

Hujer, R./S. Thomsen/C. Zeiss (2004): The Effects of Voca-tional Training Programmes on the Duration of Unem-ployment in Germany, IZA DP No. 1117.

Hujer, R./M. Wellner (2000): Berufliche Weiterbildung undindividuelle Arbeitslosigkeitsdauer in West- und Ost-deutschland: Eine mikroökonometrische Analyse, Mit-teilungen aus der Arbeitsmarkt- und Berufsforschung33, 405Ð420.

IAB, DIW, sinus, GfA, infas (2005): Evaluation der Maß-nahmen zur Umsetzung der Vorschläge der Hartz-Kom-mission Ð Modul 1e: Existenzgründungen, June. BMASForschungsbericht (Research report of the Federal Min-istry for Labour and Social Affairs).

IZA, DIW, infas (2005): Evaluation der Maßnahmen zurUmsetzung der Vorschläge der Hartz-Kommission Ð

ZAF 1/2007 63

Modul 1b: Förderung beruflicher Weiterbildung undTransferleistungen, June. BMAS Forschungsbericht (Re-search report of the Federal Ministry for Labour andSocial Affairs).

Jaenichen, U. (2002): Lohnkostenzuschüsse und individu-elle Arbeitslosigkeit. Analysen auf der Grundlage kom-binierter Erhebungs- und Prozessdaten unter Anwen-dung von Propensity Score Matching, Mitteilungen ausder Arbeitsmarkt- und Berufsforschung 35, 327Ð351.

Kaltenborn, B./P. Knerr/J. Schiwarov (2006): Hartz: Bilanzder Arbeitsmarkt- und Beschäftigungspolitik, Blick-punkt Arbeit und Wirtschaft 3/2006, Berlin.

Klose, C./S. Bender (2000): Berufliche Weiterbildung fürArbeitslose Ð Ein Weg zurück in Beschäftigung? Ana-lyse einer Abgängerkohorte des Jahres 1986 aus Maß-nahmen zur Fortbildung und Umschulung mit einerergänzten IAB-Beschäftigungsstichprobe 1975Ð1990,Mitteilungen aus der Arbeitsmarkt- und Berufsfor-schung 24, 421Ð444.

Lechner, M. (1998): Training the East German LabourForce Ð Microeconometric Evaluations of ContinuousVocational Training after Unification. Heidelberg: Phy-sica-Verlag.

Lechner, M. (1999): Earnings and Employment Effects ofContinuous Off-the-Job Training in East Germany afterUnification, Journal of Business and Economic Statistics17, 74Ð90.

Lechner, M. (2000): An Evaluation of Public Sector Spon-sored Continuous Vocational Training Programmes inEast Germany, Journal of Human Resources 35, 347Ð375.

Lechner, M./R. Miquel/C. Wunsch (2004): Long-Run Ef-fects of Public Sector Sponsored Training in West Ger-many, IZA DP No. 1443.

Lechner, M./R. Miquel/C. Wunsch (2005): The Curse andBlessing of Training the Unemployed in a ChangingEconomy Ð The Case of East Germany After Unifica-tion, IAB Discussion Paper, 14/2005.

Mosley, H./H. Schütz/ G. Schmid (2003): Effizienz der Ar-beitsämter: Leistungsvergleich und Reformpraxis, edi-tion sigma: Berlin.

OECD (2004): Benefits and Wages: OECD Indicators,Paris.

Pannenberg, M. (1995): Weiterbildungsaktivitäten und Er-werbsbiographie, Eine empirische Analyse für Deutsch-land, Frankfurt am Main, New York: Campus Verlag.

Pannenberg, M./J. Schwarze (1996): Unemployment, LaborMarket Training Programmes and Regional Wages: AnExtended Wage Curve Approach. DIW DP 139, Berlin.

Pfeiffer, F./F. Reize (2000): Business start-ups by the unem-ployed Ð an econometric analysis based on firm data,Labour Economics 7, 629Ð663.


Reinowski, E./B. Schultz/J. Wiemers (2003): Evaluation vonMaßnahmen der aktiven Arbeitsmarktpolitik mit Hilfeeines iterativen Matching-Algorithmus Ð Eine Fallstudieüber langzeitarbeitslose Maßnahmeteilnehmer in Sach-sen, IWH DP Nr. 173.

Reinowski, E./B. Schultz/J. Wiemers (2004): Evaluation ofFurther Training Programmes with an Optimal MatchingAlgorithm, mimeo, IWH.

RWI, ISG, GISA, IWH, M. Burda und M. Kvasnicka(2005): Evaluation der Maßnahmen zur Umsetzung derVorschläge der Hartz-Kommission Ð Modul 1 f: Verbes-serung der beschäftigungspolitischen Rahmenbedingun-gen und Makrowirkungen der aktiven Arbeitsmarktpoli-tik, June. BMAS Forschungsbericht (Research report ofthe Federal Ministry for Labour and Social Affairs).

Schmid G./S. Speckesser/C. Hilbert (2001): Does Active La-bour Market Policy Matter? An Aggregate Analysis forGermany, in de Koning, J. and H. Mosley (eds.): LabourMarket Policy and Unemployment Ð Impact and Proc-ess Evaluations in Selected European Countries, Chel-tenham. E. Elgar.

64 ZAF 1/2007

SÖSTRA, Compass, IMU, PIW (2005): Evaluation derMaßnahmen zur Umsetzung der Vorschläge der Hartz-Kommission Ð Modul 1c: Arbeitsbeschaffungsmaßnah-men, June. BMAS Forschungsbericht (Research reportof the Federal Ministry for Labour and Social Affairs).

Vollkommer, D. (2004): Regionalisierung der aktiven Ar-beitsmarktpolitik, Nürnberg, Beiträge zur Arbeitsmarkt-und Berufsforschung 287.

Wunsch, C. (2005): Labour Market Policy in Germany: In-stitutions, Instruments and Reforms since Unification,Discussion paper 2005Ð06, Universität St. Gallen.

WZB, infas (2005): Evaluation der Maßnahmen zur Um-setzung der Vorschläge der Hartz-Kommission Ð Modul1a: Neuausrichtung der Arbeitsvermittlung, June.BMAS Forschungsbericht (Research report of the Fed-eral Ministry for Labour and Social Affairs).

ZEW, IAB, IAT (2005): Evaluation der Maßnahmen zurUmsetzung der Vorschläge der Hartz-Kommission ÐModul 1d: Eingliederungszuschüsse und Entgeltsiche-rung, June. BMAS Forschungsbericht (Research reportof the Federal Ministry for Labour and Social Affairs).

Date post:	18-Sep-2018
Category:	Documents
Upload:	tranphuc
View:	214 times
Download:	0 times

Before and After the Hartz Reforms: The Performance...

Documents