The Hartz IV reforms are consistently regarded as a successful German model for the reduction of structural unemployment. Is it therefore appropriate to advise other member states of the European monetary union to carry out similar reforms? This paper strongly disagrees, arguing that the reduction in unemployment in Germany since 2005 is mainly due to cyclical factors and to the petering out of the negative employment effects caused by the economic transformation after German unification.
The reforms of the Agenda 2010, which were announced by German Chancellor Gerhard Schröder in March 2003 and for the most part implemented in January 2005, are widely seen as the main cause of the impressive performance of the German economy today. This is particularly true of the marked improvement in the condition of the labour market. Accordingly, similar reforms are now recommended for the Economic and Monetary Union (EMU) member states suffering from serious unemployment problems. However, this success story has very seldom been challenged analytically. Is the impressive economic competitiveness of German firms really the result of more flexible labour markets and, above all, the reduction in support for the long-term unemployed? Or is it due to the pronounced wage moderation in Germany that had started years before the Hartz reforms were implemented? For the functioning of the EMU, these are decisive questions. While more flexibility could be regarded as a positive sum game, wage moderation is a definitive zero-sum game which is associated with a deflationary bias for the euro area.
The “sick man” of Europe: a flawed diagnosis
In the discourse of economists and politicians today, the diagnosis of Germany as the “sick man” of Europe in the first half of the 2000s plays a decisive role. It serves as the basis for the narrative that the economy was miraculously healed by Chancellor Schröder’s Agenda 2010. But this finding, which was analytically supported by Hans-Werner Sinn’s best-selling book Can Germany Be Saved, was simply wrong. This becomes clear when one considers the therapy that Sinn prescribed in 2005. He called for a wage reduction in Germany of 10 to 15%, and for low-skilled workers he even envisaged a reduction of one-third.1
I have already questioned the “sick man” diagnosis in my 2004 book Wir sind besser, als wir glauben (English translation: We Are Better Than We Believe We Are).2 In particular, I pointed out that many of the problems that existed at that time were caused by the challenging economic transformation taking place in East Germany. At the beginning of the 2000s, transformation-related unemployment in East Germany was still very high, so it is not surprising that the German social security system was strained.
Considering the huge economic challenges of reunification, the German economy exhibited impressive strength even in the early 2000s. West Germany, with its 61 million inhabitants in 1990, was from the very start able to provide the 16 million citizens of the former German Democratic Republic a high level of social benefits and the rapid modernisation of the mostly obsolete infrastructure. Moreover, from 1989 to 2006, 2.8 million immigrants from the former Soviet Union arrived in Germany and obtained German citizenship immediately. This gave them full access to the social security net. While the unemployment rate of this group was very high, it was never recorded separately in the unemployment statistics.3
Direct labour market effects of the Hartz IV reform
The core of the Agenda 2010 is the so-called Hartz reform, in particular the reduction of benefits for the long-term unemployed (Hartz IV). Before this reform, the long-term unemployed received benefits under the unemployment assistance scheme (Arbeitslosenhilfe). The benefits amounted to 53% of previous net income. Hartz IV combined unemployment assistance with social assistance that was granted to people without an entitlement to unemployment benefits. As a result, the long-term unemployed received monthly financial support that was no longer related to their previous income. Thus, Hartz IV implies a relatively strong reduction of the benefits for high-skilled workers, while for low-skilled workers, the fixed payment under Hartz IV can be equivalent to or even higher than the 53% replacement rate of the unemployment assistance scheme. This is especially the case for families with children.4
To most observers, the success of these reforms becomes obvious in the light of the trend break in the development of unemployment in Germany in 2005. Up until this point, unemployment had been increasing in West Germany since the 1970s and in reunified Germany since 1990. The peak was reached at a seasonally adjusted figure of over 5 million unemployed people in March 2005.5 Since then, the employment situation has steadily improved, with the number of unemployed individuals falling to only 2.5 million in autumn 2017. Between 2005 and 2016, the average number of unemployed persons in West Germany fell by 1.3 million and in East Germany by 900 000.
But this simple evidence requires a more detailed analysis that considers cyclical factors and the specific effects of reunification. In order to identify cyclical factors, it is necessary to compare years with similar output gaps.6 According to OECD estimates, the year 2005 exhibited a strong negative relative output gap of -1.7%. It was obviously a recession year. This is also reflected by the ratio of unemployed to vacancies. In 2005 there were 22 unemployed for every vacancy.
For a comparison with 2016 – with a positive output gap of 1.1% – the most suitable year is 2001, for which the OECD estimates an identical positive output gap of 1.1%. A comparison of the unemployment figures for these years in West Germany shows a reduction in the number of unemployed of just 350 000. Thus, there remains very little support for the “Hartz miracle”. In East Germany, the reduction in the number of unemployed by 820 000 remains impressive. However, this can primarily be attributed to the phasing out of the transformation-induced employment effects. As employment in East Germany has remained constant compared to 2001, the decline in unemployment must be mainly related to the shift of workers into the pension system.
The development of long-term unemployment figures presents a similar picture. It is often argued that Hartz IV induced especially strong incentive effects in this regard. If one once again compares the years 2001 and 2016, a very small decline in long-term unemployment is recorded for West Germany, from 817 000 to 727 000 persons. Again, the picture is different for East Germany, with a reduction of 270 000 persons.
Earnings replacement (including housing benefits) for a long-term unemployed single person
Note: Figures assume a previous income of 67% of the average wage.
Source: OECD: Benefits and Wages.
These findings are supported by more detailed studies analysing the duration of long-term unemployment. Fehr and Vobruba show that since the Hartz reforms, the episodes of unemployment for recipients of Hartz IV and the predecessor system of Arbeitslosenhilfe have not become shorter.7 Indeed, when taking socio-demographic effects and the labour market situation into account, recipients of Hartz IV remained unemployed for even longer than recipients of Arbeitslosenhilfe.
Launov and Wälde provide a simple explanation for these findings. In line with the argumentation of this paper, their study comes to the conclusion that the effects of Hartz IV are very limited. They estimate that it has led to a decline in the unemployment rate of less than 0.1 percentage points.8 The authors attribute this to the above-mentioned fact that for low-skilled workers, the difference between the pre-Hartz IV (Arbeitslosenhilfe) and the post-Hartz IV (Unemployment Benefit II) benefits is rather low. For high-skilled workers, Hartz IV resulted in a clear financial deterioration. But for this group, the employment situation was always relatively beneficial. In 2005 the unemployment rate for people with a professional education was 8.6%, whereas for unskilled workers it was 26%. Launov and Wälde therefore assume that qualified workers are usually able to find a new job before they can become long-term unemployed under the regime of Hartz IV. It is not surprising that studies which do not differentiate between the effects of Hartz IV on high-skilled and low-skilled workers find significantly higher positive employment effects for Hartz IV.9
The critical assessment of the labour market effects of Hartz IV is also supported by the fact that even after the Hartz reforms, the benefits for the long-term unemployed in Germany are still quite high when compared internationally (see Figure 1). If very low benefits for the long-term unemployed were the solution to unemployment problems, the labour markets in Italy and Greece ought to be in a particularly good shape. In these two countries, the long-term unemployed do not receive any benefits at all.
Indirect effects of the Hartz IV reform
It is often argued that the Hartz reforms and the resulting pressure on the unemployed have been responsible for the pronounced overall wage moderation in Germany. The reforms are also regarded as a main cause of the development of a low-wage sector and the strong growth of part-time employment.
The so-called wage moderation has played an important role for the economic development of Germany since the 1990s.10 Dustmann et al. come to the conclusion that the positive effects of wage moderation were significantly higher than the effects of the Hartz reforms.11 The existence of wage moderation in the EMU can be inferred from the fact that the increase in unit labour costs falls short of the European Central Bank’s objective for price stability of below, but close to, two per cent. The process of wage moderation had started long before the implementation of the Hartz IV reforms. It can be traced back to the second half of the 1990s. In this period, the German trade unions made the strategic decision to secure jobs by renouncing real wage increases. As early as 1995, the chairman of the metal workers union put forth the proposal of a Bündnis für Arbeit (pact for work). He explicitly declared his willingness to accept a stagnation of real wages, i.e. nominal wage increases that compensate for inflation only, if the employers were willing to create new jobs.12 This led to the creation of the Bündnis für Arbeit, Ausbildung und Wettbewerbsfähigkeit (pact for work, education and competitiveness), which was established by Chancellor Schröder in 1998. On 9 January 2000 trade unions and employers’ associations explicitly declared that productivity increases should not be used as argumentation for increases in real wages, but instead to secure agreements that increase employment.13
Another important contribution of the trade unions to wage restraint is the so-called Pforzheim agreement, which was signed on 12 February 2004 in Pforzheim by the tariff partners in the metal and electrical industries. The agreement made it possible to deviate from the wages collectively agreed upon at the firm level. Thus, the Hartz IV reform came long after wage moderation in Germany had set in.
Considering the clear readiness of the German trade unions to secure jobs through wage moderation, an approach which is internationally unique, it is more than surprising that Hans-Werner Sinn criticised the unions in 2004 for not taking notice of “the laws of the economy”.14 He even blamed them for adopting an unscrupulous cartel policy.15
The impact of the Hartz reforms on the emergence of a low-wage sector in Germany is also overestimated. According to calculations by Kalina and Weinkopf, in terms of total employment, the share of employees in the low-wage sector increased from 18.7% in 1995 to 22.6% in 2004.16 After the introduction of the Hartz reforms, the trend slowed down. In 2013 the share was 24.4%. The same applies to part-time employment. The share of part-time employment increased from 17.9% in 1991 to 33.3% in 2004. Since then the increase has continued, but the speed has been reduced. In 2016 the share was 39%.
German reforms as a model for other countries?
The analysis carried out in this paper not only shows that the effects of the Hartz IV reform for the German economy are very limited, but it also highlights the fact that many of the so-called problem countries in Europe provide lower and more restrictive benefits for the long-term unemployed than Germany does. Germany is a questionable role model when it comes to the flexibility of the labour market. In its Annual Report of 2013/14, the German Council of Economic Experts noted that the German labour market is one of “the most heavily regulated in the world”.17 This finding is supported by the OECD indicators of employment protection legislation, according to which the employment protection for regular workers in Germany is especially high.
In essence, the improvement of the German labour market performance in the past 15 years can be explained primarily by the petering out of the labour market shock that followed German reunification. This understanding makes the traditional strength of German firms in the global markets even more apparent. From a longer-term perspective, the strong wage moderation also had positive effects. While it suffocated domestic demand in the short term, it boosted export demand in the longer term.
Recommending wage moderation as a strategy for other EMU member states is not without problems for Germany. The more consistently other countries follow this advice, the more the price competitiveness of German companies within the EMU will deteriorate. In fact, the same result could be achieved if wages in Germany were to increase over a period of time by more than the sum of the national productivity increase and the target inflation rate of the ECB. Curiously, in the German debate, the former approach of foreign wage moderation is widely accepted, while the approach of increasing wages in Germany is rejected outright.
Average annual GDP growth rates in EU member states and US states, 1999-2015
Sources: Bureau of Economic Analysis; and Ameco.
It is, of course, an open question how the two different approaches would impact the real exchange rate of the euro against the currencies of third countries. Assuming, in the medium term, the validity of the purchasing power theory of exchange rates, the third-country effects of both approaches for the euro area as a whole would be similar. For the deflationary approach, i.e. wage restraint in the other member states, a higher valuation of the euro would be expected. Thus, the positive effects of wage moderation on global price competitiveness would be fully cancelled out. In the case of higher wage increases in Germany, a devaluation of the euro versus third countries would be likely. This would compensate for the negative effects of higher wages on global price competitiveness.
In the end, what matters for the improvement of the labour market situation via wage moderation in the problem countries is Germany’s willingness to accept a certain deterioration of its price competitiveness. This leads to the insight that any attempt to improve a country’s competitiveness via wage moderation has the character of a zero-sum game. For Germany, this strategy worked as long as the other member states remained passive or even accepted excessively high wage increases. Due to the impact of high unemployment rates in several member states, this is no longer the case.
That is why it is problematic for German politicians to recommend reforms to other countries that would make their labour markets more flexible. A closer look at German labour market regulations shows that employment protection is relatively high and unemployment benefits are relatively generous, even after the Hartz reforms. Thus, greater flexibility is not a precondition for high employment. Furthermore, if the recommended reforms imply measures that would directly or indirectly reduce labour costs, Germany would have to be willing to accept the resulting deterioration of its price competitiveness, which is not very likely.
Does Michigan need structural reforms?
The critical assessment of Hartz reforms and labour market flexibility in general raises the question of how to enable a country like Italy to improve its growth. According to the prevailing narrative, Italy’s bad economic situation is primarily the result of policy failure, i.e. a lack of “reforms”. But could it not also be the case that in the process of globalisation there will necessarily be regions which are winners and regions which are losers? For example, for the United States, Autor et al. have identified the negative consequences of the “China shock” for local labour markets.18
Could Italy not also be the victim of a “China shock”? To answer this question, one may compare the average GDP growth rate of the member states of the euro area with that of the individual states in the US from 1999 to 2015 (see Figure 2). As expected, Italy performs poorly. Surprisingly, however, the US state of Michigan exhibits even less favourable development. But while the political and academic literature is inclined to regard Italy as a culprit who has not pushed reforms forward, Michigan is regarded as a victim of processes that are beyond its control. No one would ask Michigan to implement “structural reforms”, as the state has essentially the same institutional framework as those states in the US which are successful.
Concerning the attitude of Germany with regard to Italy, such a change in perspective would have far-reaching consequences. Germany could no longer just sit back and criticise the policy failure in Italy. As the obvious winners in the globalisation process, supported by the institutional framework of the monetary union, Germany would have to consider the extent to which it should compensate the loser. Otherwise, there is a danger that Italy may one day decide to abandon the euro, which would be the most expensive outcome for Germany.
- 1 H.-W. Sinn: Ist Deutschland noch zu retten?, Berlin 2005, Ullstein, p. 113.
- 2 P. Bofinger: Wir sind besser, als wir glauben, Munich 2004, Pearson Studium.
- 3 A. Brück-Klingberg, C. Burkert, H. Seibert, R. Wapler: Spätaussieder mit höherer Bildung sind öfter arbeitslos, IAB-Kurzbericht No. 8/2007, 2007.
- 4 This is confirmed by H. Goecke, J. Niehues: Verteilungswirkungen der Agenda 2010. Eine Mikrosimulationsanalyse der Hartz-IV-Reform, Cologne Institute for Economic Research, 11 December 2014. In this paper, distribution effects were calculated for 2011. Regarding the distribution effects until 2010, it has to be considered that a temporary premium was granted to individuals receiving Unemployment Benefit II. This premium was granted in the amount of a maximum of €160 for a single person and €320 for a married couple per month. For every child, €60 were added. The specific premium for a given person amounted to two-thirds of the difference between the unemployment benefit the person received most recently (in addition to an accommodation allowance) and the Unemployment Benefit II the person and the dependants are granted.
- 5 It must be considered that through the introduction of Hartz IV, unemployment increased by about 500 000 people for statistical reasons alone. Employable recipients of benefits, who thus far had not registered as unemployed, were incorporated into the unemployment statistics via Hartz IV. In contrast, many unemployed might have forgone reporting unemployment after the introduction of Hartz IV if they were not entitled to Unemployment Benefit II due to their domestic situation.
- 6 The lack of regard for the economic situation is not only characteristic of the popular debate, but of scientific research as well. For example, Krause and Uhlig assume that the unemployment quota in 2005 virtually corresponded with the equilibrium unemployment rate. They use a model for their analysis which does not allow for cyclical unemployment or for unemployment arising from transformation processes. See M.U. Krause, H. Uhlig: Transitions in the German labor market: Structure and crisis, in: Journal of Monetary Economics, Vol. 59, No. 1, 2012, pp. 64-79.
- 7 S. Fehr, G. Vobruba: Die Arbeitslosigkeitsfalle vor und nach der Hartz-IV-Reform, WSI Mitteilungen No. 5/2011, Institute of Economic and Social Research, 2011.
- 8 A. Launov, K. Wälde: Estimating Incentive and Welfare Effects of Non-Stationary Unemployment Benefits, in: International Economic Review, Vol. 54, No. 4, 2013, pp. 1159-1198.
- 9 T. Krebs, M. Scheffel: Macroeconomic Evaluation of Labor Market Reform in Germany, IMF Working Paper No. 13/42, 2013.
- 10 P. Bofinger: German wage moderation and the EZ Crisis, VoxEU, 30 November 2015.
- 11 C. Dustmann, B. Fitzenberger, U. Schönberg, A. Spitz-Oener: From Sick Man of Europe to Economic Superstar: Germany’s Resurgent Economy, in: Journal of Economic Perspectives, Vol. 28, No. 1, 2014, pp. 167-188.
- 12 M. Wolf: Von der „Konzertierten Aktion“ zum „Bündnis für Arbeit“, in: UTOPIE kreativ, No. 117, 2000, pp. 669-680.
- 13 Bündnis für Arbeit, Ausbildung und Wettbewerbsfähigkeit: Gemeinsame Erklärung zu den Ergebnissen des Spitzengesprächs, in: Blätter für deutsche und internationale Politik, No. 2/2000, pp. 250-252.
- 14 H.-W. Sinn, op. cit., p. 106.
- 15 Ibid., p. 523.
- 16 T. Kalina, C. Weinkopf: Niedriglohnbeschäftigung 2013: Stagnation auf hohem Niveau, IAQ Report No. 2015-03, Institute for Work, Skills and Training, 2015.
- 17 German Council of Economic Experts: Against a backward-looking economic policy, Annual Report 2013/14, 2013.
- 18 D.H. Autor, D. Dorn, G.H. Hanson: The China Shock: Learning from Labor-Market Adjustment to Large Changes in Trade, in: Annual Review of Economics, Vol. 8, No. 1, 2016, pp. 205-240.