This Selected Issues paper attempts to analyze the end-point problem and improve the quality of potential GDP estimates for Germany. It projects that Germany’s potential GDP growth will slow over the coming decade, mainly because of declining labor input. The paper focuses on a long-term fiscal scenario for Germany based on current policies. The paper also attempts to construct a preliminary public sector balance sheet for Germany, and analyzes the performance of its nonfinancial corporate sector.


This Selected Issues paper attempts to analyze the end-point problem and improve the quality of potential GDP estimates for Germany. It projects that Germany’s potential GDP growth will slow over the coming decade, mainly because of declining labor input. The paper focuses on a long-term fiscal scenario for Germany based on current policies. The paper also attempts to construct a preliminary public sector balance sheet for Germany, and analyzes the performance of its nonfinancial corporate sector.

VII. Does Excessive Regulation Impede Growth in Germany?46

A. Introduction

177.Excessive regulation is often cited as one reason behind the lackluster performance of the German economy since the 1990s. On average, German GDP grew by 1.3 percent in real terms, compared to 1.8 percent in Euro area countries and more than 2½ percent in the UK and the US (Table 1). Much of this disappointing performance was reflected in a lack of employment growth, where Germany fell on average one half percentage points behind its European peers: precisely the difference in GDP growth. The persistence of this performance gap has raised questions whether institutional differences can explain these discrepancies.

Table 1.

Germany: Stylized Facts, 1992-2004

(Average annualized growth rates)

article image


178. The role of institutions and, more specifically, barriers to competition in product and labor markets have recently received new attention among academics and policy makers. While sectoral regulations may not appear excessive when viewed in isolation, in combination and applied over long periods of time, they may have become a growth trap resulting in low potential growth and high unemployment. Regulation is a particularly topical issue for Germany in light of recent government initiatives to implement significant supply-side reforms (“Hartz”) and the intensive discussion of a new EU directive aimed at deregulating the service sector.

179. It is important to note that a certain degree of regulation is a prerequisite for growth and, more generally, economic welfare. Regulatory activities such as establishing property rights, safeguarding competition, and managing natural monopolies are a precondition for efficient economic activity rather than an obstacle to it. Thus, up to a certain level, regulation should bolster efficiency and real economic activity. In fact, efficient capital and labor market regulation were an integral part of the “German model” that helped sustain the high growth rates of the post-war period (Carlin 1996).

180. From a theoretical point of view it is often unclear when regulation becomes detrimental to economic activity. Consider the example of labor market regulation. While certain basic rules will be conducive for an efficiently operating labor market (e.g., health requirements, safety standards, or basic contract enforcement), further government intervention might hamper both efficiency and employment creation. Similar arguments can be made with regard to product market regulation or other fields of regulatory activity, which may create monopolistic rents in some areas while they support the functioning of natural monopolies in others (Jean and Nicoletti 2004). Ultimately, it is an empirical question whether regulation has efficiency costs.

181. The tenor of the empirical research so far is that current levels of regulation are at least partly responsible for the divergences in economic performance among industrial countries. A host of new data on regulatory activity has allowed a fresh look at regulatory activity across sectors, countries, and over time. One of the facts emerging from this analysis is that excessive regulation has a measurable negative effect on economic activity, and the intensity of regulation appears highly correlated across different markets. This suggests that any empirical analysis examining the impact of regulation on economic activity must also take a broad view, incorporating, for instance, both labor and product market regulation (Berger and Danninger 2005). Another result worth highlighting is that the impact of regulation often hinges on the implementation of legal norms, including, for instance, the role of labor courts.

182. So does excessive regulation impede growth in Germany? The paper tries to answer this question in two ways. Section B summarizes evidence on barriers to competition by comparing regulatory activity in Germany with its peers. The basic insight is that at an aggregate level product market regulations in Germany do not appear excessive. However there is some indication that administrative restrictions in the service sector are high. Labor market regulations mirror these findings. Procedural complexities are significant, and easy access to labor courts in the case of employment conflicts have raised the cost of employment. Section C takes up this latter theme and assesses whether the implementation of labor market regulations through labor courts have contributed to Germany’s rise in unemployment. The answer is affirmative.

B. Stylized Facts on German Regulation

183.Measurement of market regulation in OECD countries has improved markedly in recent years. Spearheaded by the OECD and the EU commission, several new data sets quantify the regulatory environment. These new indicators rely both on “objective” information, such as the legal framework, and on “subjective” evaluation derived from expert surveys or interviews with business people. Despite the methodological differences in data collection, Nicoletti and Pryor (2001) found that the results from different data sets are significantly correlated. An important feature of the indicators is the breadth of coverage.

The measures are often based on specific information at disaggregated levels, which is then aggregated along a variety of economically relevant dimensions. Based on this new data a number of negative linkages between macro-economic performance and market regulation have been established (Box 1).

184.This section describes Germany’s comparative position and dynamics of market regulation in three different areas. We first discuss the overall level of product market regulation, drawing on an OECD data set covering the years 1998 and 2003. The data include a comprehensive summary index distilled from 16 dimensions of product market regulation. Next we turn to market restrictiveness in the service sector. The data sources are two studies commissioned by the EU commission.47 The final part explores the degree of labor market regulation. The information draws on the OECD employment protection database (OECD 2004b). Data sources and transformations into comparable indices are described in Box 2

General Findings

185. At an aggregated level Germany’s regulation of product markets appears average. Figure 1 compares Germany’s regulation indices for three different areas with the EU 15 average, which has been normalized to zero. Each index is broken up into subindices with weights depending on the data sources. The first observation is that overall product market regulation (top observation) is on par with the EU15 average. This relative ranking is based on 2003 data, but it also applies to the 1998 observation in the OECD data set. Other countries with a similar “middle-of the-road” score are the Nordic countries and a number of small continental European countries. France and Italy have a somewhat higher score, while the UK has a less regulated environment (see further below).

Figure 1.
Figure 1.

Market Regulation: Germany Relative to the EU-15 1/

Citation: IMF Staff Country Reports 2006, 017; 10.5089/9781451810509.002.A007

Sources: OECD, Copenhagen economics (2005), Institute of Advanced Studies (2003).1/EU-15 average normalized to zero.

The Economic Effects of Regulation in Literature

article image

Sources and Construction of Regulation Indices

The empirical section relies on four data sources: (i) economy wide indices of product market regulation for 1998 and 2003 (OECD 2005); (ii) measures of market restrictiveness in the service sector in 2002 (Copenhagen Economics 2005); (iii) a regulation index of liberal professions in 2002 (Institute for Advanced Studies, 2003); and (iv) employment protection indices for the years 1988, 1998, and 2002 (OECD 2004b). Figure 1 depicts Germany’s comparative score of the most recent regulation data on a standardized scale (i.e. adjusted for the EU15 average and divided by the standard deviation).

Data sources and transformations prior to standardization are as follows:

  • (i) The index for product market regulation was designed to measure economy wide restrictions to competition and private governance. Data were derived from detailed questionnaires submitted by governments. Data items were transformed onto a common (0-6) scale increasing in the degree of restriction and then grouped into 16 regulatory dimensions. Aggregate indices (i.e. the overall level of regulation and administrative and economic regulation) were constructed based on a robust weighting scheme. For a detailed discussion see Conway, Janod and Nicoletti (2005). All data used in this study prior to standardization reflect published values.

  • (ii) and (iii)Service sector indices were derived from detailed questionnaires sent to governments and relevant public bodies. The restrictiveness measures (Copenhagen Economics 2005) covers four service sector industries with separate assessments for domestic and foreign firms. The index of regulation for liberal professions (Institute for Advanced Studies 2003) covers six professions and distinguishes between market entry and market conduct restrictions. All data used in the middle panel of Figure 1 reflect published values with two exceptions. The indices of regulation for regulated professions and other sectors show the average value for domestic and foreign firms. The overall index for the service sector is computed as the arithmetic average of the standardized sectoral scores from Copenhagen Economics (2005) and Institute for Advanced Studies (2003).

  • (iv)The employment protection data reflect legislative restrictions on temporary and regular employment arrangements as assessed in OECD (2004b). Indicators for restrictiveness on temporary and regular employment are averages of sub-indicators measuring various aspect of employment such as layoff restrictions, notification requirement, procedural complexity and other aspects. All data reported prior to standardization reflect published values.

186. The overall product market index masks some important differences. The index represents the weighted average of economic and administrative regulatory burdens (depicted below the overall index), where the respective weights are based on evaluations by OECD-experts.48 For the economic index—which refers to aspects of government control (public ownership) and direct regulation (price or quantity ceilings) and the tariff or trade barriers—Germany’s score is somewhat below average. But for the administrative index a less favorable image emerges. The administrative index synthesizes (again using weighted-average techniques) transparency aspects and the complexity of administrative processes which can be especially costly for small enterprises. Germany’s higher score in this category is due to the complexity of its administrative system which puts a comparatively high burden on new entries.

187. Germany’s high administrative regulation is also visible in the high regulation scores for the service sector. Measures of the restrictiveness of competition in the service sector show a comparative weakness in some sub-areas. While sectors such as distributive trade (wholesale and retail) and IT services are on par with EU 15 average, Germany’s regulatory system for regulated professions is more restrictive. Remnants of the guild system with its extensive licensing procedures and qualification requirements may be one factor explaining this difference. The general finding for the service sector is confirmed by another study which focuses exclusively on regulations in the liberal professions. Again, this study shows that Germany’s index for product market regulation is more restrictive than the EU average (discussed further below).

188. Turning to the labor market, the measure of employment protection legislation (EPL) is close to the EU average. This is perhaps surprising, since it is commonly believed that Germany’s labor market regulation is comparatively strict. One explanation is that the average score is the results of two offsetting measures: loose regulations in temporary arrangements (where employment is increasing) and stringent rules for regular employment (where employment is declining). Since the bulk of employment falls under regular employment contracts, the perceived strictness is consistent with the data.

189. However, measuring the tightness of labor market regulation is not straightforward. Actual EPL can differ from what has been written into the regulations along two dimensions. The first issue is completeness. The regular employment indicators depicted in the graph pertain to procedural complexity and restrictions on dismissal practices. The index on temporary employment measures the relative ease and scope to establish contracts, and the permitted duration and renewal of temporary contracts. What is missing, however, are other important institutional restrictions such as the collective wage bargaining system. Another factor not captured by the EPL indices is enforcement. As discussed further below, much of German EPL is implemented through German labor courts, which have significant leeway in interpreting the underlying legal norms. As a consequence, a tighter application of EPL by the courts may have also led to a tighter system than picked-up by the OECD indicators.

Product markets: Areas of High Regulation and Scope for Improvement

190. The level of product market regulation (PMR) in Germany is average and has decreased in recent years.49 The scatter plot shows the level of overall product market regulation in 1998 and 2003 on a scale between zero and six, with six implying most restrictive. The vertical and horizontal lines indicate the OECD average in the respective year. Values below (above) the diagonal indicate a shift towards less (more) regulation between 1998 and 2003. As already highlighted above, Germany lies in the middle of the distribution. Moreover, like all other countries in the sample, PMR levels have decreased since the late 1990s.

191.A more nuanced picture of product market regulation emerges from the various sub-indicators. The overall PMR index is made up of sixteen subcomponents. The first eleven sub indices relate to economic dimensions such as the degree of state control, legal barriers to entrepreneurship, and the regulation of trade and investment. The remainders describe administrative aspects of regulation.50


Overall Product Market Index

Citation: IMF Staff Country Reports 2006, 017; 10.5089/9781451810509.002.A007


Barriers to Foreign Ownership

Citation: IMF Staff Country Reports 2006, 017; 10.5089/9781451810509.002.A007


Complexity of Licensing System

Citation: IMF Staff Country Reports 2006, 017; 10.5089/9781451810509.002.A007

Source: Copenhagen economics.

192. In many legal aspects of economic regulation, Germany has a relatively liberal environment. Variables pertaining to the level of state control (e.g. control over price or enterprises) are at or below the OECD average. Similarly, legal barriers to entrepreneurship, or rules pertaining to antitrust legislation, are comparatively less restrictive. There is also little evidence of excessive protective measures creating disincentives for inward-FDI: discriminatory procedures and barriers to foreign ownership appear lower than in most OECD countries (see figure) and thus should not create disincentives or cost disadvantages.

193. Germany measures up less favorably on administrative features of regulation with potentially negative repercussions for entrepreneurship. The main areas where Germany has higher scores compared to its peers are complexity of administrative procedures and transparency of information. License and permit systems are comparatively complex. These results imply an opaque system of regulations and procedures, which makes entering markets difficult and costly. With a large share of economic activity conducted by small and medium-sized enterprises (“Mittelstand”), such barriers to entry can have significant costs in terms of constrained economic dynamism and limited output and employment growth. In the area of crafts, entry requirements still apply to 90 percent of all businesses, despite some easing of licensing requirements through a selective waiver of “master” certification (OECD 2004a). This sector employs about 5½ million people and generates more than 10 percent of value added (Siebert 2005).

194. Using the top-three country average as a benchmark, Germany shows the largest scope for improvement in licensing procedures and permit systems (Table 2). While Germany has increased competition—measured by a decline of the PMR index—since 1998, it did so in parallel with other countries leaving its relative ranking unchanged. Underlying this effort was considerable progress in removing administrative constraints. Simplifications of rules, improvements in communication, and the lowering of administrative burdens for small enterprises have progressed faster than in peer countries. The overall gap to lead countries has however not narrowed. The largest discrepancies and, thus, the greatest scope for improvement still remains in the area of administrative regulations and within this category licensing and permit systems. Compared to these areas, discrepancies in the area of public sector size and scope, while significant, would seem somewhat less stark. Here Germany’s weaker relative position is influenced, to some extent, by privatization programs in some of the new member EU countries, which have reduced public ownership to a minimal amount.

Table 2.

Germany: Scope for Improvement in Product Market Regulation

article image
Source: OECD, Copenhagen Economics, Institute for Advanced Studies.

195. Focusing on the service sector, regulation of liberal professions appears particularly restrictive.51 A 2003 study initiated by the EU provides indicators of services regulation in a number of sub-sectors in EU countries (Copenhagen Economics 2005). The regulatory indicators show that the overall level of restrictiveness in the distributive trades and the IT sector are low across EU countries, while regulatory constraints are somewhat higher in the accounting sector—with Germany being no exception.52 The index for Germany lies in the middle range for the less constrained areas and above average for the accounting profession.

196. Going a step further than the OECD studies discussed so far, Copenhagen Economics also provides information of the economic relevance of these indicators. Thestudy estimates the regulation-induced cost or price (rent) mark-up in enterprise profitability equations from a large enterprise data set. Results were produced for three service sector industries: business services, distributive trades, and regulated professions, and distinguished between domestic and foreign firms.53Estimates of the implied cost and price mark-up at the sectoral level are shown in Figure 2below.

Figure 2.
Figure 2.

EU: Estimated Cost and Rent Mark-Up Due to Market Restrictiveness 1/

(In percent, 2002)

Citation: IMF Staff Country Reports 2006, 017; 10.5089/9781451810509.002.A007

1/ The terms “Rent” (surplus accruing to the firm’s owners) and “Cost” (surplus to the inputs) mark-up refers to the estimated percentage price or cost increase in high relative to low regulated industries in a cross-country sample.

197. The estimated economic effects of regulation broadly match the findings above.The most competitive areas are business services with mark-ups of 1-2 percent and distributive trades with mark-ups around 1-4 percent. In contrast, costs of market restrictiveness in the regulated professions are quite significant. The average mark-up is around 10 percent on average. The estimates for Germany broadly follow this overall pattern. In other words, not all service sector areas appear to have competitive barriers and by implication higher costs. Germany, like most other European countries, faces the largest costs and economic losses in the area of regulated professions, which are typically important in the value chain of all businesses in an economy.

198.Higher regulation in service sector professions is confirmed by another recent study showing higher barriers to market entry and constraints on market conduct. A 2003 report on the economic impact of regulation in the liberal professions finds that in comparison to the EU15, Germany has a rather restrictive regulatory environment.54 The text figure depicts measures of restrictiveness for two different aspects of regulation: market entry (e.g., licensing and qualification systems) and market conduct (e.g. restrictions on pricing, location, diversification, and advertisement). The vertical and horizontal lines indicate average scores. Compulsory membership in professional bodies and binding pricing prescription appear to be some factors responsible for the comparatively high score in Germany.

Labor Market Regulation

199.Indicators of labor market regulation aim to capture the legal environment of employment relations. They are thus only one element of a broader set of labor market institutions such as, wage setting institutions, unionization, and the unemployment benefit system. The standard measures of labor market regulation—used in this study—focus on different aspects of employment protection. They include measures of the permissiveness of temporary employment contracts (e.g. duration and renewal), constraints and procedures for layoffs, and special requirements for collective dismissals. Data for these dimensions are available since the late 1980s and thus allow some longer-term comparison.

International comparison

200.As reported above, cross-country data do not identify Germany as a particular outlier when it comes to labor market regulation. The overall employment protection index shows Germany at the EU15 average in 2002 (text table). Germany’s present position is the result of a gradual deregulation process since the late 1980s. While all EU countries have loosened their labor market regulations over time, Germany has done so at a faster pace. More recently, as part of the Agenda 2010, the government amended the existing employment protection legislation in January 2004 with the aim to loosen existing rules. The revisions eased restrictions on temporary employment and also clarified the criteria for legal dismissals.55 The impact of the new regulations is currently unclear, and a mandatory review of the effects of the revised EPL is due in 2007.

Index of Employment Protection

article image
Source: OECD and staff estimates

201.This overall achievement masks deficits in certain areas. Progress was limited to the area of temporary employment, which has become less regulated than in most EU countries. Germany removed time limits on the duration of temporary work contracts, eased regulations on fixed term contracts, and loosened rules for temporary work agencies (Table 3). On the other hand, levels of employment protection in regular work arrangements have slightly increased since the 1980s, and regulations for collective dismissals have tightened since the 1990s (by increasing the mandatory waiting period). To the extent that temporary work is a substitute for regular work arrangements, deregulation of temporary work has the potential to increase overall labor market efficiency. However, substitutability might be limited, and so the lack of reform concerning regular contracts remains a reason for concern.

Table 3.

Germany: Scope for Improvement in Labor Market Regulation

article image
Source: OECD and staff estimates.

Actual improvement refers to the period 1998 to 2003.

202. Assessing the scope of improvement shows several areas for potential catch-up to leaders. Using the top-three country average as a benchmark shows that even temporary employment is still regulated more tightly than in lead countries. The largest improvement score is found for procedural inconveniences of regular employment. This variable essentially measures permissible steps of appeal against dismissals. In 2003 workers with regular jobs have retained essentially the same rights of legal recourse to layoffs as in 1988. Labor courts can easily be invoked in a dispute creating a comparatively high procedural burden. Similarly, difficulty of dismissals of regular employment has remained high reflected by a relatively limited number of permissible causes for dismissal. Also, regulation of collective dismissal processes is quite stringent when compared to the top three liberal regimes.

203. Recent research hints at weaknesses in these indicators. Working with a panel of OECD countries, Nickell, Ochel, and Nunziata (2005) find that institutional factors explain about 55 percent of the rise in unemployment rate between 1960-2000. Of this part, the largest contributions come from changes in the benefits system, higher labor taxes, and shifts in union density. Only about 14 percent is attributable to changes in employment protection legislation—i.e., the regulation of labor markets in a narrow sense.56 This finding is somewhat puzzling as it contradicts the common belief that labor market regulations have contributed to the rise in unemployment.

Implementation of EPL

204. Differences in implementation can explain why differences in regulatory norms do not map into differences in economic performance. Many indicators focus on formal regulatory norms—not least because they are most easily measured—thereby ignoring differences, both across countries and time, in informal regulatory practice and implementation. This issue might be particularly pervasive regarding labor market regulation, were labor courts play a major role in implementing existing rules. The employment protection indicators discussed above are an important example. While these measures incorporate some aspects of judicial practices, the role of labor courts “is likely to be somewhat understated in the information presented...” (OECD 2004b, p. 66).

205. Labor courts are likely to play a crucial role in the translation of labor market rules into effective regulation—in particular when it comes to the regulation of firing costs.57 Courts interpret existing law, translating employment protection rules into monetary or non-monetary firing (e.g. by temporarily extending a work contract intended to end) costs for firms. Labor court activity in this respect is likely to influence firing costs both directly and indirectly. If employees had reason to believe that courts increased potential transfers in case of dismissal compared to what their severance package might include, they would be more inclined to involve labor courts when dismissed, leading to an increase in case load and, ultimately, higher actual firing cost. This direct action is likely to have an additional indirect influence by informing private negotiations between employees and employers, as both sides will internalize any change in court policy when discussing the conditions for a dismissal.

206. Recent studies support the notion that courts have important influence on the intensity of effective employment protection.58 The available cross-country evidence, while scarce and not always allowing generalization, suggests that the direct impact of labor courts on EPL is often amplified through the “threat” of ruling—that is, relatively few cases may actually reach courts (OECD 2004b, Table 2.1).59 In recent work on the US, Autor and others (2004a, 2004b) describe in some detail how changes in US labor courts decision-making influenced firing costs and, ultimately, employment across states. They stress that some of the impact of innovations in court policies precede the publication of a fully elaborated decision behind a new common-law doctrine, as employers may already have responded to the initial precedent-setting. All this suggests that any evaluation of the true—or implemented—extent of employment protection in a country needs to take into account its ongoing interpretation by labor courts.

207. There might be a temptation for labor courts to use their discretionary power to pursue employment protection policies of their own. There are, for instance, indications that labor courts’ activity might not be neutral with regard to labor market conditions. Ichino and others (2002) show that Italian labor courts do not base decisions on the characteristics of employee’s (alleged) misconduct alone: the same behavior might be considered sufficient for terminating an ongoing conduct in a booming labor market but insufficient otherwise. This implies that higher unemployment could endogenously lead to higher firing costs, with possible repercussions for higher unemployment. Bertola and others (1999) discuss evidence for other OECD countries pointing in the same direction.

208. German labor courts, much like their counterparts elsewhere, enjoy significant leeway in their interpretation of existing laws. Contrary to other legal fields, the basic German Protection Against Dismissal Law of 1951 (“Kundigungsschutzgesetz”) places only mild restrictions on court behavior. Dismissals are considered illegal, if they are “socially unjustified” and lack an “important” reason, but the specific meaning of either term is a matter of interpretation. As a consequence, it is mostly labor courts that determine the size of actual firing costs (Richardi and Wlotzke 1992). Courts rule on a case-by-case basis and ask, as a matter of principle, whether a dismissal was indeed the “ultima ratio” or could have been avoided—in other words, the presumption is that the work contract should be continued—with the burden of proof whether a dismissal was justified placed solely on the employers.60

209. Labor courts may have their strongest impact on actual firing costs—not least because of the uncertainty surrounding their interpretation of the underlying laws. The German Sachverstandigenrat (2003, p. 385) argues that the interpretative freedom given to labor courts in this respect renders the outcome of judicial action all but “unpredictable”, thereby severely reducing the transparency of employment protection even for the legal expert. This uncertainty is of quantitative relevance. According to the Sachverstandigenrat (2003, p. 386), in 2001 the ratio of decided labor court cases to the number dismissed employees was about 27 percent suggesting a relatively high probability of court involvement.61

C. Regulation and Labor Market Outcomes

210. This section examines the impact of labor market regulation on unemployment. It builds on the discussion of the importance of labor courts for the interpretation and implementation of labor market regulation, in particular EPL. In addition to shedding some light on the issues raised in the previous section, the focus on labor courts has the advantage of allowing the use of straightforward time-series techniques in the econometric analysis of the effects of regulation—a feat more difficult to achieve with more traditional regulation indicators.

A Look at the Role of Labor Courts

211. Labor court activity in Germany shows significant changes over time. Figure 3 reports labor court activity in percent of regular civil court activity over the post-war period. For reference, the unemployment rate and real GDP growth (all in percent) are also shown. Ideally, the role of German labor courts in employment protection would be described not only along the dimension of their actual involvement in labor market regulation (i.e., court activity), but also with regard to the quality of their decisions (i.e., changes in court interpretation of existing laws). Unfortunately, such data are not available. Bertola and others (1999, p.21) argue, however, that higher court activity may be indicative of more employee-friendly rulings. Based on expert surveys for a number of OECD countries they conclude that countries where courts are more often involved in disputes over the termination of work contracts tend also to be “those to have the highest percentage cases favorable to employees.”

212. The striking time-pattern of labor court activity seems to reflect both demand and supply factors.62 Labor court activity in the post-war period as described by Figure 3 was hardly constant: from a stationary holding-pattern in the 1950s and 1960s, court activity suddenly tripled in the early 1970s, receded somewhat in the 1980s, only to continue its increase during the 1990s. The visible jump in the early 1970s hints at demand factors, as it coincides with the—equally dramatic—jump in unemployment following the first oil shock. On the other hand, this does not explain the fact that numbers of court cases remained high afterwards, pointing at supply-side factors.

Figure 3.
Figure 3.

Germany: Labor Court Activity, Unemployment Rate, and Real GDP Growth Rate 1951-2002

Citation: IMF Staff Country Reports 2006, 017; 10.5089/9781451810509.002.A007

213. According to anecdotal evidence, an important reason for the permanent upward shift in labor court activity was the labor courts themselves. Franz (1994) reports that, sometime in the late 1960s, a new generation of judges seems to have drifted towards a more employee-friendly interpretation of existing labor laws, giving priority to safeguarding existing work contracts over the interest of job-seekers (“Bestandsschutz”)63 Since accessibility of labor courts in Germany is high for employees, firms found it decidedly more difficult to dismiss workers, which, in turn, led to a sharp increase in actual firing costs (Soltwedel (1984).64 Interestingly, the temporary reduction in court activity in the mid 1980s coincides with the enactment of the 1985 Employment Security Act (“Beschaftigungssicherungsgesetz”), an effort by the new conservative government to curtail the discretionary power of labor judges.65 In addition, the law liberalized temporary work contracts, which were highly restricted in their use both by legislative and labor court regulation, thereby allowing firms to hire without facing restrictions when discontinuing (or not renewing) the work contract. The renewed upward-movement during the 1990s is a little harder to interpret.66 In part, demand factors—that is, the further increase in unemployment—will have played a role. There is also evidence, however, that the frequency with which dismissed employees went to court increased during this period (Sachverstandigenrat 2003), which points toward supply-side factors as well.

214.To learn more about the effects of labor court activity on unemployment, it is worthwhile to take a closer look at the time-series behavior of the series depicted in Figure 3. A first observation is that there is obvious co-movement in unemployment and labor court activity levels. Moreover, as one would expect, the level of the unemployment rate and real GDP growth also seem related. In fact, a formal test points toward two co-integration relationships among the three variables. A second observation relates to the dynamics underlying the series. At first glance, changes in labor court activity seem to pre–date changes in the unemployment rate.

Table 4.

Germany: Test of Granger-Causality between Labor Court Activity and Changes in Unemployment Rate 1/

article image
Source: Own calculations.

Sample 1952-2002. The lag length has been set to 2. The F-statistic allows rejection of the hypothesis that changes in unemployment are independent of changes in labor court activity—but not the other way around—at the 5 percent-level.

215. Table 4 shows that court activity seems to play a more important role in the explanation of the dynamics of unemployment than vice versa. As with co-integration, this result need not necessarily imply economic causality, however. For instance, it could be argued that increases in the unemployment rate are pre-dated by higher labor market turnover, which, in turn, could increase labor court activity ahead of visible shifts in the unemployment rate.

216. To gauge the size of the implied effect of court activity a Vector Error Correction model was estimated and is discussed in the annex. The main message stemming from this exercise is that labor court decisions have a clear positive impact on the unemployment rate—even after controlling for the possible endogeneity of the latter with regard to real activity. The effect builds up over about 5 years and shows surprising rigidity suggesting an almost permanent effect. There is, in addition, a smaller negative impact of labor court activity on real GDP growth. Higher unemployment has only a short-lived positive impact on courts and there is almost no reaction of court activity on real GDP growth shocks.

217. It is important to note the limitations of the above result: lacking all but the simplest of economic structure, the model is ill-suited for going beyond the qualitative assessment discussed. Further analysis would be required to ensure that the aggregate relationship holds up against micro evidence, that is, whether the court’s use of their discretionary powers indeed have negative repercussions for the labor market at the individual level. But, at the very least, the evidence appears to support the idea that any attempt to influence effective employment protection in Germany must take into account— and perhaps limit—the autonomy of labor courts.67

D. Conclusions and Policy Implications

218. Market regulation is a candidate in explaining growth discrepancies between industrialized countries. Policy makers and academics alike are increasingly interested in pursuing this link between institutions and economic activity.

219. It is important to note, however, that the theoretical implications of regulation per se are ambiguous. A basic regulatory framework is a requirement for growth, and only empirical analysis can show whether regulation overall, or in a particular sector of economic activity, is excessive. As others have argued, a leaner, more growth-oriented regulatory framework helped Germany’s “Wirtschaftswunder” during the earlier post-war period (Carlin 1996).

220. Recently a host of new data on regulation has become available. While a number of problems remain—for instance, with regard to measurement of implementation or the possible endogeneity of regulatory indicators—the data allow comparison of regulatory regimes across countries and (to a lesser degree) across time.

221. Empirical studies find that excessive product and labor market regulation have a measurable negative impact on economic activity. Among the findings is the result that employment protection can magnify the effects of economic shocks and increase unemployment of disadvantaged groups on the labor market (such as the young, women, or unskilled workers). Product market regulation may reduce productivity and, thus, potential growth, and also tends to increase labor costs. Finally, there is evidence that regulatory activity in different sectors is cascading. For instance, the effect of employment protection in a particular sector depends in part, on product market regulation in this sector.

222. Germany’s overall level of market regulation appears average in OECD countries. However there is some indication of higher barriers to competition in the service sector, especially in the area of regulated professions. Administrative burdens tend to be relatively high. Complex permit and licensing requirements further dampen entrepreneurship and economic activity compared to the best practice in the EU15. The picture emerging from a comparative look at German labor market regulations is similar in many ways. While the overall level of German labor market regulation does not stand out as such, there is ample scope for improvement when it comes to specific areas. In particular, employment protection for regular employment is high due to procedural complexities which can be traced back, among other things, to ease of access to labor courts.

223. One area sometimes ignored—but of particular relevance in the German case—is the role of labor courts. In Germany and elsewhere, labor courts play a key role in implementing and interpreting employment protection. There is (so far mostly anecdotal) evidence pointing toward a surprisingly independent role of German courts, which may have added to employment and firing costs for firms. Our analysis provides some empirical support for the hypothesis that labor court activity had a detrimental effect on employment.

224. Summing up, there is reason to believe that targeted reforms could help Germany’s regulatory environment to become as conducive for growth as it once was. Tasks identified from comparisons with best practices within the EU15 include reducing administrative burdens and licensing restrictions in the services sector and the easing the protection of regular employment contracts. The independent role of German labor courts poses a particular challenge related to reducing employment and firing costs.

APPENDIX A Vector Error Correction Model of Labor Court Activity and Unemployment

A simple econometric exercise provides further insights, while controlling for the possible endogeneity of court activity. To better gauge the impact of changes in labor court activity on unemployment, while taking into account repercussions of developments in unemployment and real growth on court activity itself, we estimate a simple Vector Error Correction (VEC) model allowing for two co-integration relationships over the 1954-2002 period (49 observations with annual data).

Formally, we estimate


where the vector xt=(ut, yt, lct)’ contains the observations on the unemployment rate, the rate of real GDP growth, and labor court activity in period t, D is the difference operator, and the coefficients αulculculcγlcγlcγlc, and βi are to be estimated. The index z’=1,.., n captures the lag structure of the VAR, with n=2 in the estimated model.68εt is a vector of residuals assumed to follow standard assumptions. In short, the VEC-model combines the notion of co-movement in the levels of all three variables (thought as representing long-run equilibrium relations) with a standard VAR in first differences, allowing a relatively rich description of their short-run interactions.69

The figure below illustrates the results using impulse response functions. The various graphs show the reaction of the three variables included in the model to a “shock” in the variable itself (indicating the degree of inertia) as well as in the two other variables over a simulated ten-year period.

The main message stemming from the figure is that labor court decisions matter for unemployment. There is a clear positive impact of higher labor court activity on the unemployment rate—even after controlling for the possible endogeneity of the latter withregard to real activity. The effect builds up over about 5 years and shows surprising rigidity even beyond the period shown in the figure, suggesting an almost permanent effect. There is, in addition, a smaller negative impact of labor court activity on real GDP growth.70 At the same time, higher unemployment has only a short-lived positive impact on courts (even turning negative after about 2 years) and there almost no reaction of court activity on real GDP growth shocks.


Figure. Germany: Impulse Response Functions from a simple VEC Model

Citation: IMF Staff Country Reports 2006, 017; 10.5089/9781451810509.002.A007

Source: Own computation.Notes: The graphs show the reaction to a one-standard deviation shock in the respective residuals based on the Cholesky procedure using the order real GDP growth (Y), unemployment (U), labor court activity (LC) with sample degrees of freedom correction. The model assumes that real-growth shocks influence all other variables contemporaneously, while unemployment shocks only influence labor court activity within the period, and there is no contemporaneous impact of labor court activity on the other variables.


  • Alesina, Alberto, Silvia Ardagna, Guiseppe Nicoletti, and Fabio Schiantarelli, 2003, “Regulation and Investment,National Bureau of Economic Research, Working Paper No. 9560.

    • Search Google Scholar
    • Export Citation
  • Autor, David, John J. Donohue III, and Stewart J. Schwab, 2004, “The Costs of Wrongful Discharge Laws” (unpublished; Cambridge, Massachusetts: MIT).

    • Search Google Scholar
    • Export Citation
  • Autor, David, John J. Donohue III, and Stewart J. Schwab, 2004, “The Employment Consequences of Wrongful-Discharge Laws: Large, Small, or None at All?,American Economic Review, 94 (2) (May 2004): pp. 44046.

    • Search Google Scholar
    • Export Citation
  • Bauer Thomas, Stefan Bender, and Holger Bonon 2004Dismissal Protection and Worker Flows in Small Establishments”, IZA Discussion Paper Series No. 1105

    • Search Google Scholar
    • Export Citation
  • Berger Helge and Stephan Danninger 2005:Labor and Product Market Deregulation: Partial, Sequential, or Simultaneous Reform?”, IMF Working Paper, forthcoming.

    • Search Google Scholar
    • Export Citation
  • Berger, Helge, 1998, “Regulation in Germany: Some Stylized Facts about Its Time Path, Causes, and Consequences,Zeitschriftfur Wirtschafts- und Sozialwissenschaften, 118 (2), 185220.

    • Search Google Scholar
    • Export Citation
  • Bertola, Giuseppe, Tito Boeri, and Sandrine Cazes, 1999, “Employment Protection and Labor Market Adjustment in OECD Countries: Evolving Institutions and Variable Enforcement,Employment and Training Papers, No. 48 (Geneva: International Labor Office).

    • Search Google Scholar
    • Export Citation
  • Bertola G., F. Blau and L. Kahn, 2002, “Comparative Analysis of Labor Market Outcomes: Lessons for the US from International long-run evidence,” in A. Krueger and R. Solow (eds.), The Roaring Nineties: Can Full Employment Be Sustained? pp. 159218 (Russell Sage and Century Foundations)

    • Search Google Scholar
    • Export Citation
  • Blanchard, O. and J. Wolfers, 2000, “The Role of Shocks and Institutions in the Rise of European Unemployment: The Aggregate Evidence,Economic Journal, Vol. 110, No. 462, pp. C133.

    • Search Google Scholar
    • Export Citation
  • Blanchard, Olivier, and Francesco Giavazzi, 2001, “Macroeconomic Effects of Regulation and Deregulation in Goods and Labor Markets,National Bureau of Economic Research, Working Paper No. 8120.

    • Search Google Scholar
    • Export Citation
  • Carlin, Wendy (1996), “West German Growth and Institutions,” in: Crafts, Nicholas and Toniolo, Gianni (eds.), Economic Growth in Europe Since 1945, Cambridge, Cambridge University Press, 45597.

    • Search Google Scholar
    • Export Citation
  • Conway, Paul, Veronique Janod, Giuseppe Nicoletti, 2005, “Product Market Regulation in OECD Countries: 1998 to 2003,Working Papers No. 419 (Paris: OECD Economics Department).

    • Search Google Scholar
    • Export Citation
  • Copenhagen Economics, 2005, “Economic Assessment of the Barriers to the Internal Market for Services,Copenhagen Economics.

  • Ebell, Monique, and Christian Haefke, 2004, Product Market Deregulation and Labor Market Outcomes, IZA Discussion Paper No. 957.

  • Estevao M., 2005, “The Benefits of Wage Moderation,IMF Working Paper, forthcoming.

  • Franz, Wolfgang, 1994, “Chancen und Risiken einer Flexibilisierung des Arbeitsrechts aus okonomischer Sicht,Zeitschriftfur Arbeitsrecht, 25(3), 43962.

    • Search Google Scholar
    • Export Citation
  • Ichino, Andrea, Michele Polo, and Enrico Rettore, 2002, “Are Judges Biased by Labor Market Conditions?” (unpublished; Florence, Italy: European University Institute).

    • Search Google Scholar
    • Export Citation
  • Institute for Advanced Studies, 2003, “Economic Impact of Regulation in the Field of Liberal Professions in Different Member States, ” Research Report, HIS (Vienna).

    • Search Google Scholar
    • Export Citation
  • Jean, Sebastien, and Giuseppe Nicoletti (2004), “Regulation and Wage Premia,CEPII, No. 2004-12.

  • Kugler Adriana, Giovanni Pica, 2003, “Effects of Employment Protection and Product Market Regulations on the Italian Labor Market,Working Paper.

    • Search Google Scholar
    • Export Citation
  • Nicoletti, Giuseppe, and Stefano Scarpetta, 2003, “Regulation, Productivity, and Growth: OECD Evidence,World Bank Policy Research Working Paper (Economics Department) No. 2944.

    • Search Google Scholar
    • Export Citation
  • Nicoletti, Giuseppe, and Frederic L. Pryor, 2001, Subjective and Objective Measures of the Extent of Governmental Regulations.

  • Nickell, Stephan, Luca Nunziata, and Wolfgang Ochel, 2005, “Unemployment in the OECD Since the 1960s. What do We Know?The Economics Journal, 115, pp. 127.

    • Search Google Scholar
    • Export Citation
  • OECD, 2004a, Economic Survey 2004 Germany, OECD (Paris).

  • OECD 2004b, Employment Outlook, OECD (Paris).

  • OECD 2005 “Going for Growth,” Economic Policy Reforms (Paris).

  • Richardi, Reinhard and Ottfried Wlotzke (ed.), 1992, “Munchener Handbuch zum Arbeitsrecht, Band 1,” Munchen, Beck.

  • Ruthers, Bernd, 1995, “35 Jahre Arbeitsrecht in Deutschland,Rechtder Arbeit, 6, 32633.

  • “Sachverstandgenrat zur Begutachtung der gesamtwirtschaftlichen Entwicklung, Jahresgutachten 2003/04.

  • Siebert Horst, 2005, The German Economy: Beyond the Social Market: Princeton University Press, 416 pp.

  • Soltwedel, Rudiger, 1984, “Employment Problems in West Germany— The Role of Institutions, Labor Law, and Government Intervention,” Carnegie-Rochester Series on Public Policy, 28, 153220.

    • Search Google Scholar
    • Export Citation
  • Young, David, 2003, “Employment Protection Legislation: Its Economic Impact and the Case for Reform,European Economy, 186.


Prepared by Helge Berger and Stephan Danninger.


Copenhagen Economics (2005) and Institute for Advanced Studies (2003).


The components of the economic index have a greater weight within the overall index, since they account for eleven out of the sixteen sub-indices. Sensitivity tests show that relative country rankings are robust to marginal changes in the distribution of subindices along administrative and economic dimensions (Conway, Janod, and Nicoletti 2005).


Regulation indices here as well as in the following sections have not been transformed and thus are identical to published data.


A detailed discussion of the different variables is given in Conway, Janod, Nicoletti (2005).


Liberal professions are generally defined as occupations requiring special training in the arts or sciences, including lawyers, notaries, accountants, architects, engineers or pharmacists.


See Box 1 above for a discussion. The panel describes the results for the restrictiveness indicators. All findings distinguish between domestic and foreign firms. A higher value on the index means a more restrictive environment.


The sectoral groupings are derived from a data matching process between the sectoral data from the regulation index and the sectoral information contained in the enterprise level data


The surveyed fields include accounting, legal and pharmaceutical professions, engineering, and architecture.


The main changes were (i) an extended duration of temporary contracts in newly founded enterprises (up to four years), (ii) the use of severance pay in exchange for waiving appeal to labor courts, (iii) greater exemptions from EPL for small firms, (iv) and an exact definition of the social criteria that can be considered in a disputed dismissal case.


Studies measuring directly the effect of dismissal protection legislation on German micro data also find little evidence of a significant impact (Bauer et al 2004).


This section, in part, draws on Berger (1998).


See, for instance, Bertola and others (1999), and OECD (2004b).


See also the survey in Young (2003).


Since 1999 a change in leadership at the supreme labor court led to a gradual shift away from case-by case jurisdiction in favor of a more standardized interpretation of legal compliance.


Survey data indicate that during the years 1991–1998 roughly one third of dismissed employees received some form of transfer from their former employers. This figure might be a lower bound. Especially smaller firms, which probably are less willing to pay for the legal fees and other costs associated with going to labor court will be inclined to settle outside court (Sachverständigenrat 2003). Thus, with about one third of all dismissed going to court, the actual ratio of those receiving some sort of transfer to the number of dismissed could well be higher. Another relevant stylized fact emerging from surveys is that currently about one third of all exits from labor contracts is due to dismissal by firms.


Note that the denominator of the labor court activity series shown in Figure 3 excludes labor court cases. Moreover, the number of new dismissal cases arriving at the courts behaves similar to the number of finished cases. Therefore, the observed time pattern of labor court activity should mostly be due to changes in the nominator and be broadly independent of court capacity. See the discussion in Soltwedel (1984).


The Sachverständigenrat (2003) also provides a discussion of the development of the Bestandsschutz-idea by the courts.


Court fees are low, employees can call upon trade unions for legal advice before the courts, and the defeated party is not charged with the legal costs of the opposition.


More specifically, the law reduced the obligation of new firms to provide social plans for employees following larger-scale lay-offs. More generally, it constrained court decisions on the applicability of social plans: instead of being determined solely by the judges, applicability was made a predictable function of firm size and the number of employees affected.


This is not an artifact of unification. As the notes to Figure 3 point out, up to 1994, the labor court activity indicator refers to the Old Länder only. Thus, while unification might have influenced the second half of the 1990s, the steep increase before must have different reasons.


The Sachverständigenrat (2003, p. 389) demands more specific legal norms to curtail the courts’ discretionary leeway and to decrease uncertainty regarding the effective level of employment protection in Germany.


The lag length selection was based on standard criteria. Introducing a linear trend in the cointegration (i.e., error-correction) part of the model or altering the lag length in the VAR part does not alter the result much. All three series are non-stationary based on standard criteria.


Theoretically, the demand for labor court action should be positively correlated with changes of unemployment or deviations from its trend, i.e. the flow of people dismissed not the stock of people out of work. As a matter of fact, the annual number of dismissals decreased as German unemployment rates rose (Soltwedel et al. 1990).


Berger (1998), based on an endogenous growth model and using a somewhat shorter sample period, also reports a small negative impact of labor court activity on real GDP growth. A one standard-deviation of labor court activity (based on the change of the raw series in logs) leads to a growth loss of about .14 percentage points. The result is based on a two-stage IV-approach in which labor market regulation enters lagged one period instrumented by past regulatory activity, past growth, and past values of other likely determinants of regulatory activity.

Germany: Selected Issues
Author: International Monetary Fund