This Selected Issues paper examines Germany’s growth record in 1992–2001 and analyzes how future performance might be enhanced. The paper focuses on the longer-term strains on the public finances. It reviews Germany’s external competitiveness, which deteriorated substantially in the wake of unification, and concludes that, by the beginning of the current decade, competitiveness had been largely restored. The paper also examines the recent slowdown in credit, which has gone beyond what might be expected on cyclical grounds.


This Selected Issues paper examines Germany’s growth record in 1992–2001 and analyzes how future performance might be enhanced. The paper focuses on the longer-term strains on the public finances. It reviews Germany’s external competitiveness, which deteriorated substantially in the wake of unification, and concludes that, by the beginning of the current decade, competitiveness had been largely restored. The paper also examines the recent slowdown in credit, which has gone beyond what might be expected on cyclical grounds.

I. Growth and Adjustment in Germany: Perspectives and Prospects1

A. Introduction and Summary

1. This chapter advances various hypotheses to help understand Germany’s low real GDP growth during 1992-2001 with a view to informing the discussion on policies to improve the country’s economic performance. Per capita real GDP growth during this period was a modest 1 ⅓ percent a year. Much of the retrospective analysis relies on cross-country comparisons with France, Italy, the Netherlands, and the United Kingdom (UK), against which Germany’s growth rate lagged by ⅓-1 percentage point. These countries have similar per capita income levels to Germany and offer enough variance in terms of growth performances and policies to form some tentative conclusions.2 Developments in eastern Germany are also compared to those in the Czech Republic, Hungary, Poland, and Portugal to assess the structural transformation that has taken place since unification.

2. At the aggregate level, the main reasons for Germany’s poor economic performance relative to the comparator countries are:

  • Cyclical developments. Following a 15 percent expansion of the west German economy during 1989-91, Germany entered the 1990s with unemployment below its natural rate by a larger margin than in the other countries. The unwinding of the economic tensions from the unification-related boom slowed real GDP growth in the 1990s. Today, Germany’s economy is operating with more slack than the other countries reviewed.

  • Smaller efficiency gains. Even after adjusting for the business cycle, a per capita real GDP growth gap of about ½ percentage point per annum remains relative to the Netherlands and the UK. Germany was less successful than these countries in pushing out its production possibility frontier through a better use of labor resources. In both the Netherlands and the UK, unemployment rates declined for a given level of labor costs, indicating a structural increase in labor supply. This increase was supported by reforms to the welfare system. By contrast, in Germany generous benefits for the jobless and tax and contributions increases to fund the extension of the west’s welfare system to the east hindered a structural increase in labor supply.

3. Behind the aggregate developments were much larger sectoral and structural changes than in the other countries reviewed:

  • First, there were major shifts in fiscal policies. A sizable expansion of the fiscal deficit took place during the early years of unification to fund the extension of infrastructure and the welfare system to the new Länder. Subsequently, Germany—together with Italy—undertook the largest fiscal effort among the comparator countries to put its public finances on a sounder footing. In that process, government employment was cut by about 0.8 million and revenues were increased significantly. Consolidation likely dampened demand. But tax and contributions increases may also have had a significant impact on aggregate supply.

  • Second, a large appreciation of the real exchange rate hastened the downsizing of the manufacturing sector. Germany started the 1990s much more exposed to industry and thus less resilient to external shocks. Over the past decade, the share of industry in value added shrunk considerably—a process that was accelerated by the appreciation of the deutsche mark (DM) around unification—to reach a size that is broadly comparable to that in the other countries reviewed.

  • Third, the construction sector went from boom to bust. The construction sector was initially a major winner from unification. But the rebuilding boom eventually collapsed and the construction sector has been in a deep recession since 1997.

  • Fourth, the economic landscape of new Länder was transformed. While unification-related policy changes and sectoral adjustments held back growth in the west, the growth performance in the east was fairly impressive: in terms of per capita GDP, the new Länder now resemble Portugal. However, overly rapid convergence of wages toward western levels has left a legacy of high unemployment in the east.

4. Weathering these adjustments more successfully would have required more resilient economic institutions—and, in particular, better functioning labor markets. Rather than fostering the rapid reintegration of job losers into the labor market, the welfare system essentially absorbed them into long-term unemployment or de facto early retirement. In addition, employment protection legislation and regulated product markets inhibited the reallocation of jobs to expanding sectors.

5. The conclusion that the labor market must take part of the blame for Germany’s poor growth performance is well documented in other studies. For example, EU (2002a) attributes about one third of the difference in Germany’s real GDP growth differential with its EU partners between 1995-2001 to structural factors. It specifically singles out the labor market—insufficient wage differentiation across skills and regions, poverty traps, and a lack of general flexibility and mobility. Taking a longer view, Scarpetta and others (2000) find that failure to increase labor utilization in the euro area, and particularly in Germany, accounts for why per capita GDP in Europe has been falling further behind that of the United States since 1985.

6. For the next decade, unification strains will be having a significantly lower impact on the economy, but an improved growth performance cannot be assured without reforms. On present trends, per capita potential real GDP growth could in principle reach about 1¾ percent a year. Achieving or exceeding this would require:

  • Putting more people to work, particularly in the new Länder. Although Germany’s employment-to-population ratio is not low by EU standards, ample scope remains for raising employment. In the first instance, this might be achieved by raising the employment rate of 55-64 year olds, which stands at under 40 percent, and reducing the high unemployment rate in the east. In addition, the incentives of the low skilled—which, according to EU (2000a), account for over one third of the unemployed in the western Länder—to accept full-time jobs need to be strengthened. This could be accomplished by: (i) for older workers, reducing the length of unemployment benefit duration from 32 months to the standard length of 12 months, while subsidizing their on-the-job training with social security exemptions; (ii) eliminating poverty traps, by linking part of the receipt of benefits for the jobless to mandatory training or the performance of publicly useful work; and (iii) continuing with moderate wage settlements and creating more room for wage differentiation.

  • Raising productivity. Germany’s total factor productivity (TFP) growth does not compare unfavorably with the other countries reviewed. But European economies missed out on the surge of TFP growth seen in the United States in the 1990s. The reasons for the acceleration in the United States and its sustainability are not well understood. Nonetheless, ample scope remains for Germany to improve the quality of its education and deregulate its product markets, two drivers of TFP growth. In addition, the new Länder experienced a marked slowdown in TFP growth during the second half of the 1990s. More infrastructure and transfer of business skills to the east would help to speed closure of the remaining productivity gap with the west.3

7. The chapter is organized as follows. Section B provides a descriptive analysis of Germany’s growth performance. Section C investigates supply-side developments, focusing on the contributions of labor, capital, and TFP to real GDP growth. Section D reviews fiscal policy, sectoral changes in output, and the integration of eastern Germany in the 1990s. Section E concludes.

B. The Economic Performance

8. For the period 1992-2001, the economic cycle works against Germany in growth comparisons with other EU countries (Table I-1). In particular, this period does not include the unification-boom years. Unification caused a surge in investment and consumption and boosted real GDP in western Germany by some 15 percent during 1989-91 (Table I-2). This surge meant that Germany’s 10-year average growth rate remained above that in all the other countries reviewed, except for the Netherlands, through 1999.4 In 2001, the year when the 10-year average no longer includes any of the unification-boom years, Germany falls to the bottom of the league.5

Table I-1.

Germany’s Economic Performance, 1992-2001

(Geometric average in percent, unless otherwise indicated)

article image
Sources: OECD Economic Outlook database; and IMF WEO database

Based on working age population.

In percent of potential output.

Table I-2.

The Years Preceding German Unification. 1989-1991

(Geometric average, unless otherwise indicated)

article image
Sources: WEO database; and SM/94/213 for Germany in 1989-91.Note: Data for 1989-91 are for West Germany only. Data for 1989-2001 are for West Germany for 1989-91 and unified Germany for 1992-2001.

Arithmetic average.

9. Nonetheless, the economic cycle explains only part of the growth gap relative to the high growth countries. By focusing on potential real GDP growth in per capita terms, the effects of the cycle and of population growth are removed. On that measure, Germany still ranks low, although together with France and Italy. The growth gaps relative to the other countries vary somewhat depending on whether OECD or IMF staff estimates of potential output growth are used, but broadly suggest that Germany’s potential output growth lagged that of the Netherlands and the UK by about ½ percentage point.

10. Sharp reductions in the estimated natural rate of unemployment underlie the estimates of higher potential output growth for the Netherlands and the UK in the 1990s. Relative to Germany, the natural rate in these countries is estimated to have declined by roughly ¼ percentage point per annum since 1991 (Table III-3). As (hours adjusted) annual labor productivity growth has typically fluctuated between 1½ -2 percent for all three countries, this decline basically accounts for the full ½ percentage point growth differential in per capita potential output.

11. More broadly, Germany’s labor market performance slipped relative to the other countries in the past decade. At the beginning of the 1990s, Germany’s employment ratio (actual or hours adjusted) compared favorably with the other countries, coming second only to that in the UK (see Table I-3).6 But a relatively poor job-creation record allowed the other countries to make up ground—or, in the case of the UK and Netherlands, to surge ahead. And relative to France and Italy, Germany still has a lower structural unemployment rate, but the differential has narrowed. Structural unemployment is much lower in the UK and Netherlands.

Table I-3.

Germany’s Labor Market Performance, 1992-2001

(Arithmetic average in percent, unless otherwise indicated)

article image
Sources: OECD Economic Outlook and Analytical databases; EUROSTAT; and IMF, World Economic Outlook database.

Average hours worked by employees in the business sector; in percent of Germany.

Employment rate multiplied by index of hours worked.

C. Supply-Side Developments

12. A factor decomposition of business sector real GDP growth over 1992-2001 reveals that Germany lagged the Netherlands and the UK significantly with respect to the contribution from labor; it also lagged France over 1997-2001 (Table I-4).7 The conclusion is the same if cyclical differences between the countries are smoothed out with 10-year backward moving averages (Figure I-1). Structural labor supply was held back in Germany by income and payroll tax hikes as well as by generous welfare benefits that fostered long-term unemployment and early retirement in response to economic shocks. Germany’s TFP performance was relatively strong. But the contribution of capital to growth was surprisingly low considering that unification effectively increased the supply of labor.

Table I-4.

Germany: Growth Accounting, 1992-2001

(Geometric averages in percent, unless otherwise noted)

article image
Sources: OECD Economic Outlook database; and author’s calculations.

Adjustment made by subtracting from the labor contribution working age population growth multiplied by the average labor share in business sector GDP.

Figure I-1.
Figure I-1.

A Factor Decomposition of Business Sector Real GDP Growth, 1980-2001

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Sources: OECD Economic outlook database; and author’s calculations.

The contribution of labor

13. The lower contribution of labor to output growth accounts fully for the per capita potential output growth gap with the Netherlands and the UK. Over the 1990s, the labor contribution to output growth was some ¾ percentage point lower per annum in Germany than in these countries. Adjusting for differences in population growth cuts that gap to about ½ percentage point, or to the size of the difference between per capita potential output growth rates.

14. Employment growth was slowed by higher labor cost growth in Germany compared to the other countries reviewed, particularly during the first half of the decade (Table I-5). The appropriate measure for labor costs is real effective compensation, defined as real wage compensation deflated by the ratio of TFP and the labor share.8 Over the 1990s, real effective compensation fell in all the countries reviewed but least of all in Germany and the UK. In addition, as factor proportions adjust slowly to changes in factor costs, Germany’s much larger increase in labor costs during the first half of the decade would have had major relevance for its recent labor market performance. However, it is important to separate out the extent to which the rapid labor cost growth in the early 1990s may have reflected an economy that was operating beyond its “normal” production possibility frontier—and hence a tight labor market—as opposed to underlying factors.

Table I-5.

Germany: Labor Cost Developments

(In percent, unless otherwise noted)

article image
Sources: OECD Economic Outlook database; and author’s calculations.

For 1997-2001, reflects largely substitution of payroll taxes with a tax on value added (IRAP).

Change in percentage points.

15. Adjusting for temporary distortions in labor costs due to the economic cycle and fluctuations in productivity reveals that structural labor supply has increased in all the countries reviewed over the last 30 years, except for Germany (Figure I-2). A firmer understanding of labor market developments requires identifying structural changes in labor supply: that is changes in labor costs (unemployment) for a given level of unemployment (labor costs). Appendix I explains how labor costs and unemployment can be related in a typical labor supply schedule.9 Smoothing shifts of such a supply schedule with a 5-year backward moving average to remove temporary distortions to labor costs uncovers a structural decrease in German labor supply in the first half of the 1990s, unlike in any of the other countries. However, a turnaround is evident around 1996. The equivalent interpretation is that Germany established structural moderation of wages much later than in the other countries.

Figure I-2.
Figure I-2.

Labor Supply and Factor Proportions, 1980-2001

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Sources: OECD Economic outlook database; and author’s calculations.

16. The generosity, duration and administration of benefits for the jobless probably had a significant impact on labor supply developments. The benefit system fostered long-term unemployment and early retirement in response to labor market shocks, rather than downward adjustments in wages. Time series evidence suggests that, unlike the Netherlands and the UK, Germany did little to lower the gross replacement rates for the unemployed (Figure I-3). 10 Even more important was the “duration dimension” of benefits in Germany, which does not show up in the data on replacement rates (Figures I-3 and I-4 and Table I-6). A more detailed analysis of employment ratios across age categories reveals that employment gaps relative to the UK—the country with the highest employment rate—are concentrated on the young and older workers.11 For those older than 53 years, benefit duration in Germany was raised from the standard 12 months to a maximum of 32 months during 1985-87 to facilitate early retirement. When the benefit period expires, unemployed persons qualify for unemployment assistance--with standard replacement ratios of 56-58 percent, as opposed to 63-68 percent under unemployment benefits—right up to the statutory retirement age of 65 years.12 Furthermore, for older workers in the new Länder who lost their jobs during 1990-93 a special program (Altersübergangsgeld) provides unemployment benefits until they qualify for an old-age pension.

Table I-6.

Net Replacement Rates, 1999

(In percent, after tax and including family and housing benefits)

article image
Source: OECD (1999)
Figure I-3.
Figure I-3.

Unemployment and Gross Replacement Rates, 1971-2001

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Sources: OECD Economic Outlook database; and OECD Benefit and Work Incentives database.1 Replacement rates are an average for various individuals in various earning categories. For Italy, the replacement rate includes only unemployment insurance.
Figure I-4.
Figure I-4.

Taxes and Contributions, 1980-2001

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Source: OECD Tax and Benefits database.

Germany and the United Kingdom: Employment by Age Groups, 2001

(In percent of corresponding population)

article image
Source: Eurostat.

Additional jobs if Germany had the same employment ratios for each age group as the U.K.

17. While Germany made benefits for the jobless more generous in some crucial respects, the Netherlands and the UK did the opposite. The Netherlands reduced benefits during the decade up to 1995 and the UK did so mostly in the 1980s. Reductions in unemployment, disability, and sick leave benefits in the Netherlands are well documented; in addition, eligibility criteria for the various programs were tightened considerably.13 Similarly, benefits were lowered and eligibility criteria were tightened in the UK, the latter through various measures, including compulsory counseling in 1986; the disqualification from unemployment assistance of the 16-17 year olds in 1988; and the “Actively Seeking Work Rule” in 1989 (Nickell and Quintini, 2002). These examples illustrate that the administration of benefits, as well as their level and duration are crucial for labor supply.14

18. An increase in structural labor supply was also hampered by increases in taxation and employee-paid social security contributions (Figure 1-4). As gross replacement rates remained essentially unchanged, rising taxes and contributions acted to slow labor supply and boost wages. In comparison with the Netherlands and the UK, the average rate for total payments less cash benefits rose by 11-15 percentage points for a single filer; for a married filer with one child, the rate rose by 2-6 percentage points. Some estimates suggest that a 10 percentage point decrease in the tax wedge—an amount lying roughly in the middle of the range of changes observed for the Netherlands and the UK relative to Germany—would lead to a 25 percent reduction of the unemployment rate (see Box).

Taxation and Labor Supply

The extent to which labor taxes affect the relation between post-tax pay and reservation wages both in the short and long run is fundamentally an empirical issue. Layard and Nickell (1999) in a review of the literature argue that “[t]here is some evidence that overall labor tax rates have a short-run, and possibly long-run, impact on unemployment rates.” They put the reduction in unemployment that can be obtained with a 5 percentage point reduction in the tax wedge at 13 percent (e.g., from 8 percent to 7 percent of the labor force).

Rising labor taxation can pose more problems than high labor taxation. Assuming that capital is internationally mobile while labor is not, the incidence of labor taxes ultimately falls fully on labor. High labor tax countries thus need not have higher unemployment rates. However, following payroll and other tax hikes, unemployment may well rise until unions agree to lower wage floors or cease to seek offsetting wage increases.

To understand the role of labor taxation, recall that the relation between post-tax pay and the reservation wage matters for labor supply; the analysis here has focused on pre-tax pay. Assume, for simplicity, that reservation wages are based on alternative earnings, which are also taxed, and that taxation is proportional. Standard models suggest that taxation does not matter then: the real wage drops in line with the reservation wage. However, what if individuals’ utility depends on both the level of income and its change? There is supportive evidence for such behavior from work in the area of adaptation theory. Standard models would then suggest real wage resistance in the face of tax increases. Similarly, wage floors can prevent an adjustment in market wages in response to tax hikes; and progressive taxation can narrow differentials between market and reservation wages. Lastly, the reservation wage may be a function of untaxed nonmarket activities, such as tending to children or the elderly.

19. The more recent improvement in structural labor supply partly reflects moderate wage settlements. While wage moderation will secure existing jobs and boost employment prospects for labor market entrants, its sustainability is questionable absent welfare reform. In the interim, wage moderation might have a depressing effect on aggregate demand until factor proportions react. Some studies of investment find a lag of two to three years between the beginning and completion of an investment project.15 For projects that entail fundamental changes in factor proportions, the lags are likely to be longer. For example, Blanchard (1998) assumes a mean lag of four years in the adjustment of factor proportions. Notice that in all the other countries reviewed, labor market conditions improved only many years after wage moderation had set in: in the Netherlands, toward the end of the 1990s; in the UK in the mid-1990s; and in France and Italy only toward the end of the 1990s. In the Netherlands, wage moderation started with the Wasenaar agreement in November 1982. In the UK, more moderate wage settlements owed to the decline in union power, stemming partly from changes in trade union laws (Nickell and Quintini, 2002).

20. Germany’s comparatively subdued labor market performance does not appear to be related to structural declines in the demand for labor that may have been prompted by changes in the production structure. A satisfactory characterization of labor demand with good empirical backing is much harder to obtain than an aggregate labor supply schedule (Appendix I). With a simple Cobb-Douglas production function, a technology-driven decline in labor demand can be thought of as a drop in the labor share in output—although a drop in labor demand because of less union bargaining power is observationally equivalent. In this regard, Germany did not fare significantly worse than any of the other countries, except for the UK (Figure I-5).16 Allowing for costs in the adjustment of factor proportions does not materially change the picture. The relatively favorable structural labor demand in Germany might reflect more cooperative labor relations during the 1970s and part of the 1980s than in the other countries reviewed.

Figure I-5.
Figure I-5.

Labor Demand and Payroll Taxes, 1970-2001

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Sources: OECD Economic Outlook database; and author’s calculations.

21. A better performance of labor demand in Germany was, however, obstructed by payroll tax hikes (Figure I-5). Unless payroll tax hikes are compensated with more moderate wages, they boost unemployment. In the Netherlands and the UK, the effective payroll tax rate generally fell during the 1980s and 1990s, thereby boosting labor demand. By contrast, in Germany the rate generally increased, in particular over 1992-1998. Taking a longer view, none of the countries have managed to reduce the effective payroll tax rate, except for the UK. In the Netherlands, considerable progress was made during 1985-95, but this appears to have been undone recently. However, as the payroll tax rates were raised, they were compensated for with more moderate wage growth.

The contribution of capital

22. The contribution of capital to output in Germany in the 1990s was lower than in the other countries reviewed. Since the quality of German capital stock data was affected by unification, the differences between the countries—which are not large—should not be overemphasized. Nonetheless, from the perspective of neoclassical growth theory and supporting empirical evidence Germany’s performance was disappointing.

23. In fact, with its productivity level at the beginning of the 1990s below that in the other countries, the forces of income convergence should have boosted Germany’s relative productivity growth. Because of a higher return to capital, and thus a higher rate of per capita capital accumulation, Barro and Sala-î-Martin (1995) show that the catch-up of output per capita of poorer with richer entities (“β-convergence”) across European regions proceeded at a rate of 1 percent per annum over 1980-90.17 A 1 percent rate of β-convergence implies that a country with a 10 percent labor productivity gap relative to the average should post 0.1 percentage point faster per capita real GDP growth per year, other variables remaining unchanged. For capital accumulation, this translates into a higher per capita growth rate of about 0.3 percentage points, as the share of capital in output was roughly one third. At the beginning of the 1990s, because of the low level of productivity in the new Länder, productivity in Germany was significantly lower than in the other countries reviewed, except for the UK (Table I-7). Simple calculations suggest that this should have meant that capital accumulation exceeded the rates in France, Italy, and the Netherlands by roughly ⅓ - ¾ percentage point per annum. Instead, Germany’s actual performance fell short of that in those countries.

Table I-7.

Labor Productivity Gaps and Capital Stock Growth, 1992-2001

article image
Source: OECD 2000.Note: Productivity is GDP per man-hour.

24. The key element behind the subdued rate of capital accumulation in Germany is likely to have been the behavior of wages. While productivity dropped considerably with unification, average wages declined much less: in 1991, unit labor costs were some 50 percent higher in the new Länder than in the old Länder. Wage growth in the new Länder continued to be fueled in the following years by wage catch-up to the west. With wages not behaving as assumed in the neoclassical growth framework, investment failed to stage a stronger performance than in the other countries.18

The contribution of total factor productivity

25. Germany’s TFP growth in the 1990s was not significantly different from that in the other countries reviewed. In the short run, TFP is highly cyclical, mainly because it picks up changes in capacity utilization; including or excluding a few years in a growth comparison can make a large difference. A 10-year backward-looking geometric moving average of TFP growth reveals that Germany’s performance has generally been strong. The only significant lag relative to another country, namely the UK, emerges in 2001, when none of the unification-related boom years are included (see Figure I-1). 19 Removing cyclical effects using the Hodrick-Prescott (HP) filter series also suggests that Germany’s TFP growth has been comparatively strong (Table I-8).

Table I-8.

TFP Growth, 1981-2001

article image
Source: OECD Economic Outlook database and author’s calculations

26. That said, TFP growth slowed hi all the countries reviewed in the 1990s, in contrast to the experience of the US. The reasons for the TFP acceleration in the US are not well understood, nor is it clear how sustainable it is. It is partly related to information technology developments in the US but “how close [the] connection is and how exactly it works is not certain” (Baily, 2002). According to detailed case studies by McKinsey Global Institute (2001), the key factor behind the productivity acceleration in the US was competitive pressure that stimulated innovation. Also, a variety of sources contributed to boosting productivity, not just the information technology industry. Studies for EU countries have generally not been able to detect a similar contribution of information technology to growth as in the US. However, they have generally reached optimistic conclusions for the future, including for Germany (IMF, 2000).

27. If competition boosts innovation, then it could be a driving element of TFP growth. The literature case is ambiguous. On the one hand, more competition reduces firms’ funding for research and development (R&D) and their ability to appropriate profits from innovation. On the other hand, neck-and-neck competition induces innovation by incumbents, particularly if the survival of firm and management are at stake. Using R&D expenditure as a proxy for innovation, Bassanini and Ernst (2002) find a positive relation between less regulated product markets and innovation across OECD countries. Of direct relevance for the relation between TFP growth and competition are the results of Nickell (1996), who substantiates a positive link in a panel of 670 UK companies.

28. Germany compares favorably to most of the other countries reviewed with respect to indicators of product market regulation and spending on R&D. For various reasons, competition may affect innovation in ways that are not captured by R&D spending and thus the two sets of data should be considered complementary in assessing the TFP friendliness of the regulatory and institutional environment. According to Nicoletti and others (2000, 2001), Germany is ahead of the other countries reviewed—except for the UK—in both the current state and progress in improving product market regulation (Table 1-9). Key remaining weaknesses are barriers to entry and excessive administrative regulation. Germany also spends a higher share of GDP on R&D than all the other countries reviewed and only marginally less than the US (Table I-10).

Table I-9a.

Germany: Regulatory Reform in Product Markets, 1978-19981/

(Scale 0-6 from least to most restrictive)

article image
Source: Nicoletti et al(2001).

Simple Average of indicators for 7 industries; gas, electricity, post, telecoms, air transport, railways, freight.

2/ In percent

Table I-9b.

Germany: Detailed Indicators of Product Market Regulation, 1998 1/

(Scale 0-6 from least to most restrictive)

article image
Source: Nicoletti et al. (2000).

Sectors covered include retail distribution, road freight, air passenger and railway transport, and telecommunications.

Table I-10.

Germany: R&D Expenditure, 1980-2000 1/

(Averages, in percent of GDP)

article image
Source: OECD Science and Technology database.

Comparisons across time are affected by data breaks.

29. The link between labor market regulation, in which Germany fares no worse than most of the other countries, and innovation/TFP is unclear. The link depends on the interplay of policies and institutions. Strong employment protection legislation (EPL) may induce “inside” workers to appropriate some of the benefits of innovation by renegotiating wages. Thus it may have a depressing effect on innovation and TFP growth. However, if labor relations are of a cooperative nature, the appropriation of innovation rents under strong EPL is less of an issue. Also, firms would then adjust production in response to innovation by retraining and redeploying workers rather than hiring and firing. Germany’s employment protection legislation is rather restrictive if compared with the UK (Table 1-11), but not relative to the other countries reviewed. However, Germany is also considered a country with cooperative labor relations: on that basis, Calmfors and Driffell (1988) rank Germany ahead of the UK in terms of delivering low wage growth and unemployment.

Table I-11.

Germany: Employment Protection Legislation

(Scale 0-6 from least to most restrictive)

article image
Source: OECD, Labor Market Indicators database

30. Declining relative education standards could have dampened TFP growth. While the effect of education should, in principle, be captured by a quality-adjusted labor input in the growth decomposition, adequate data to do so are not available. Based on educational attainment and assumptions about how many years of education a particular level of achievement represents, Germany leads the other countries reviewed in terms of levels of education (Table 1-12). However, the change in educational achievement during the 1990s was the lowest. Moreover, the data are not adjusted for differences in education quality, for which the OECD’s PISA study (2002) revealed important weaknesses in Germany (Table 1-13).

Table I-12.

Germany: Education, 1970-98

(Average number of years of education in working-age population)

article image
Source: Bassanini et al. (2001).

In percent.

Table I-13.

Germany: Ranking Educational Achievement Based on 2000 PISA Study

(Highest and lowest possible rank among 32 industrial and developing countries)

article image
Source: Bassanini et al. (2001)

31. In sum, Germany’s TFP growth and its driving forces do not compare unfavorably with the other countries reviewed, but there is ample scope to improve the quality of education and foster product market competition though streamlining administrative regulations and removing barriers to entry. With fierce product market competition believed to be the driving force behind the acceleration in US TFP growth over the 1990s, deregulation and a vigilant anti-trust policy could play an important role in boosting German output growth in the future.

D. A Decade of Economic Adjustments

32. Changes in aggregate demand and the real exchange rate played a key role in unwinding unification-related tensions and set off adjustments in the structure of output and employment. First, following an expansion in the fiscal stance at the beginning of the 1990s, a major adjustment effort was undertaken to fund the integration of the new Länder without an undue expansion in government debt. Second, the fiscal expansion and unification-related construction boom raised the demand for nontradables, causing a sharp appreciation of the deutsche mark that accelerated the decline of Germany’s relatively large industrial sector. This section reviews the impact of these developments on Germany’s growth performance and takes stock of the progress in the economic integration of the new Länder.

The demand side: public sector adjustment and the interplay with monetary conditions

33. The two countries with the lowest real GDP growth in 1992-2001—Italy and Germany—also made the greatest efforts to adjust their public finances (Figure I-6). Germany’s adjustment was primarily achieved by raising revenue (Figure I-7), although the composition of spending changed as well. The size, composition, and length of the adjustment effort likely affected growth, with the revenue hikes limiting the potential non-Keynesian effects of fiscal adjustment. The consolidation effort followed a large loosening of policies in 1990 to fund unification.20 This loosening may have had an expansionary effect even in the subsequent years. Nevertheless, this is not considered in the following exercise, which explores the extent to which the policies and conditions prevailing during 1992-2001 can explain Germany’s economic growth performance during those years.

Figure I-6.
Figure I-6.

Fiscal Policy and Monetary Conditions, 1991-2001

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Sources: IMF, World Economic Outlook; OECD Economic Survey for Germany (1996); and author’s calculations.1 The “wide” primary structural balance for Germany includes the unification-related deficits of the major public enterprises and the Treuhand Anstalt.
Figure I-7.
Figure I-7.

Government Revenue and Primary Expenditure, 1991-2001

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Source: IMF, World Economic Outlook.

34. The direct contribution of the public sector to real GDP explains only some of the growth gap with France and virtually none of the gap with the other countries. Using business sector data, which exclude the public sector, the growth gaps relative to Italy, the Netherlands, and the UK are little changed from the overall GDP gaps (Table I-14). Only the gap relative to France is virtually eliminated. Nonetheless, Germany reduced the government payroll throughout the 1990s by some 0.8 million, an achievement comparable to that of the UK. This can be seen as a structural decline in labor demand in Germany. Avoiding a fall-out on unemployment would have required a more resilient labor market, notably a less generous welfare system.

Table I-14.

The Role of the Public Sector, 1992-2001

(Geometric average in percent, unless otherwise indicated)

article image
Sources: OECD Economic Outlook database; and IMF WEO database.

Based on working age population.

Employment level, in millions.

35. Similarly, simulations with the OEF macroeconomic model suggest that different fiscal policies and monetary conditions can explain the lower growth during 1992-2001 relative to France but not relative to the other countries reviewed (Appendix II). Such simulations allow for the interaction between the government and the private sector. The specific question addressed was how would real GDP growth in Germany have fared if the country had adopted the same fiscal policy and benefited from the same change in monetary conditions as the other countries. The simulations show that if Germany had had the fiscal policies of France and the Netherlands, annual real GDP growth in Germany might have been higher by about ½ percentage point (Table I-15). This largely eliminates the growth gap relative to France but not the Netherlands. If Germany had had the policies and monetary conditions of Italy and the UK, real GDP growth in Germany would have been below the rate recorded during 1992-2001.

Table I-15.

Explaining the Per Capita Real GDP Growth Differences, 1992-2001

(In percent)

article image
Sources: OECD Analytical Database; WEO; and author’s simulations with OEF model

36. The dynamics of the model and the simulation exercise may miss some of the effects of different fiscal policies. First, the model might not adequately capture the long-run effects of changes in taxation on economic performance. The debate on taxation and growth is not settled but the estimates that are available suggest that rising taxes and contributions in Germany would have reduced potential growth (see above, and Box). Second, it is hard to do justice with the model to the interaction of a large shock, such as unification, with a substantial change in the policy mix.

37. Even so, perhaps the conclusion to draw is that only some of Germany’s slow growth relative to the other countries in 1992-2001 can be explained by fiscal policy and monetary conditions. In particular, the performances of the Netherlands and the UK were fundamentally stronger, consistent with estimates of potential output growth rates.

The effects of changes in the sectoral composition of output

38. During the 1990s, the sectoral changes in Germany were substantially larger than those in the other countries reviewed. The contraction of the industrial sector accelerated sharply during the early 1990s, in tandem with the appreciation of the real exchange rate. In addition, the construction sector underwent a major boom-bust cycle that was related to unification. Such changes raise frictions that hold back structural labor supply and demand. Weathering them while keeping growth strong would have required a more resilient labor market.

39. Slow growth of the industry and energy sector accounts for an important part in Germany’s subdued growth performance, especially during 1992-96 (Tables I-16).21 To a much lesser extent, this sector also posted below-average growth in the UK and, during the second half of the 1990s, in the Netherlands. Excluding industry and energy from value added, puts Germany’s growth performance over 1992-2001 ahead of the of that of France and Italy by some ¾ percentage point but still leaves a considerable gap relative to the Netherlands and the UK (Table I-17).22 The same calculation for 1997-2001 suggests that Germany’s increase in value added remains modestly ahead of that of France and Italy. However, it falls further behind that of the Netherlands and the UK—and this despite the appreciation of the pound sterling.

Table I-16.

The Size of Industry, Energy and the Construction Sectors

(In percent of total value added)

article image
Sources: Eurostat; and author’s calculation
Table I-17.

Value Added Excluding Construction, Industry, and Energy

(Average annual percent change)

article image
Sources: Eurostat; and author’s calculations

40. The unification-related boom only temporarily slowed the secular decline of the share of industrial output, with the subsequent downsizing greatly exceeding that in the other countries reviewed. The 1993 recession, together with the continued real appreciation of the deutsche mark, led to a major contraction of industry (Figure 1-8). Simple calculations show that over 1973-91 the size of the sector in total value added contracted by over 0.1 percentage points per annum. Allowing for a once-and-for-all reduction in the share of industry in value added by 2 percentage points in 1991-93, the 1973-91 relation continued to hold through 2001. The downscaling of industry caused job losses totaling almost 2.8 million over 1992-2001 (Table 1-18). About 2.6 million of these jobs were lost during 1992-95 alone, of which more than 1 million were in the new Länder. By comparison, the losses in industry in France and Italy over 1992-2001 were about 0.5 million. In the UK they reached almost 0.9 million.

Figure I-8.
Figure I-8.

Value Added in Industry and Energy and the Real Exchange Rate, 1980-2001

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Sources: Eurostat; and IFS database.
Table I-18.

Job Growth in Manufacturing

(In percent per annum)

article image
Sources: OECD; and author’s calculations

41. The contribution of the construction sector does not make much direct difference to the cross-country growth comparison for the full period 1992-2001. Upon excluding it from value added, Germany’s growth is broadly on par with that of France or Italy, but relative to the Netherlands and the UK, the gaps do not differ much from those for total value added. The construction sector does make a difference in the more recent period 1997-2001, explaining some of the rising growth gap with France and accounting for almost half of the growth gap with the UK. However, as in the case of the downsizing of industry, the boom-bust in construction did lead to major shifts in employment that the labor market had to process.

42. Overall, Germany experienced significantly more structural change than the other countries in the 1990s. The standard deviation of sectoral output growth rates was highest in Germany, reaching 2.8 percent, about 17 percent higher than in the high-growth (Table 1-19). And the standard deviation of sectoral employment growth averaged 3.8 percent over 1992-2001, exceeding that of the UK by some 20 percent. Finally, the ratio of the standard deviations of sectoral employment and output growth rates was also highest in Germany.23

Table I-19.

Turbulence: Standard Deviation of Sectoral Growth Rates

(Average over 1992-2001, in percent)

article image
Sources: OECD; Eurostat and author’s calculations
Table I-20.

Germany: Public Sector Balances

(In percent of GDP)

article image
Source: OECD Economic Surveys, Germany 1996, Table 6, p. 26.Note: Treuhand Anstalt: Operations wound up in at end 1994; successor organization financed directly from government budget. Post and Telecoms were incorporated in 1995 while railways now receive subsidies directly both from state and federal governments; the debt of the railways was taken over by the government at end 1994. Excludes East German housing debt taken over by government in 1995.

For 2000, includes receipts from the sale of UMTS licenses equivalent to roughly 2.5 percent of GDP. These receipts are excluded from calculations of the structural balance.

Table I-21.

Primary Structural Fiscal Balances, 1991-2001

(In percent of potential GDP)

article image
Sources: WEO database; and author’s calculations.

Including the spending by public enterprises and the Treuhand Anstalt shown in Table 16.

43. In sum, the German labor market processed considerable change, suggesting that the wage bargaining system and employment protection legislation might, at this stage, not be the binding constraint on a better growth performance. Instead, potential labor resources have been siphoned off into long-term unemployment and early retirement by the welfare system, thereby dragging down Germany’s growth performance. However, the sectoral changes are slowing and this may set the scene for higher growth in the future.

Integrating the new Länder

44. Eastern Germany’s income catch-up with the west has been impressive, considering the findings in the literature on convergence and the performance of selected eastern European countries. However, it has also been disappointing in one dimension: the utilization of labor. The unemployment rate in the new Länder is more than double that in the western Länder. This is partly a legacy of overly rapid wage growth at the beginning of the decade—or equivalently the decline in structural labor supply that slowed the contribution of labor to growth.

45. For a broad sample of advanced and developing countries, Barro and Sala-i-Martin (1995) find that per capita incomes converge at a rate of 3 percent per annum, holding various “other variables” (e.g., human capital, political stability, and the size of government) constant in per capita terms.24 Following unification, the new Länder largely adopted West Germany’s institutions and received substantial support for infrastructure and purchasing power from the west. Thus the “other variables” should, if anything, have contributed to a higher than 3 percent rate of productivity convergence, considering further that the new political and legal environment should have fostered investment in the east. But in other respects, unification might also have put eastern Germany at a disadvantage relative to an emerging market country. In particular, the west made for an attractive emigration target and the east had to adopt a tax and benefit system as well as labor market regulation that have stymied incentives to work and accumulate capital.

46. The rate of income convergence of the new Länder has far outstripped the theoretical pace. In 1991, real GDP per employed in the east stood at 35 percent of the west, implying that productivity should have grown at least 2 percentage points faster annually than in the west, according to the Barro and Sala-î-Martin findings. In fact, productivity grew some 7.3 percentage points faster than in the west over 1992-2001 to reach 71 percent of the western level in 2001 (Figure 1-9). Only to a minor extent was this performance a reflection of employment losses: per capita GDP in the new Länder also started from 35 percent of the level in the old Länder to climb to 64 percent.25 As a result, per capita GDP in eastern Germany is now only about 6 percent lower than in Portugal. None of the Eastern European countries reviewed here—the Czech Republic, Hungary, and Poland—achieved such a rapid improvement in productivity and GDP per capita. In the Czech Republic, the most productive among the eastern European countries reviewed, GDP per capita it is still 13 percent lower than in the new Länder.

Figure I-9.
Figure I-9.

Eastern Germany; Economic Convergence Compared, 1991-2001

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Sources: OECD Economic Outlook database, Volkswirtschaftliche Gesamtrechnungen (for East Germany), and staff calculations.

47. The productivity and per capita GDP gains in the new Länder were concentrated in the early years of unification, raising the question of whether convergence might be stalling. Productivity in the east has only grown 1 percentage point faster than in the west over 1997-2001. However, a lower rate of convergence is partly natural: in terms of productivity and per capita GDP the new Länder now compare to the less-rich advanced European regions. Using Barro and Sala-î-Martin’s results for such regions, which may now be more applicable, productivity growth in the east should now be higher relative to the west by only 0.3 percentage points a year.

48. Moreover, adjusting for special developments in the construction sector, which were driven by the need to improve the housing stock and infrastructure in the new Länder, income convergence continues. A construction boom at the beginning of the 1990s partly explains the rapid convergence initially and the recent slowdown (Figure I-10). By contrast, the manufacturing sector—where productivity is easiest to measure—has continued to post an impressive performance: even over 1997-2001, productivity converged at a rate of almost 4.8 percentage points per annum while employment expanded. In short, the forces of convergence remain alive, even if the rapid gains during the early years following unification can no longer be expected.

Figure I-10.
Figure I-10.

Germany: East-West Convergence, 1991-2001

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Sources: Volkswirtschaftliche Gesamtrechnungen; Eurostat; and author’s calculations.

49. The new Länder have considerably reduced their dependence on transfers from the west. In 1999, per capita disposable income in the new Länder amounted to 82 percent of the level in the old Länder;26 for nominal GDP, the ratio was 64 percent. The same ratios for 1991 were 57 percent and 35 percent, respectively. According to measures conventionally used in the literature on interregional income redistribution, this puts the amount of redistribution at about 25 percent of nominal GDP in per capita terms in 1999, down from over 60 percent in 1991. The amount of redistribution is thus no longer extraordinary by international standards. A number of papers have examined redistribution across regions within Canada, France, Italy, the UK, and the US. The lowest amounts of redistribution are found for the US, with most estimates in the range of 10-15 percent of gross state product.27 The highest estimates of redistribution are found for France and reach almost 40 percent (Melitz and Zumer, 1998). What is unusual in the case of eastern Germany is that the poor regions are concentrated in one area and that the income gap is large.28

50. From one perspective the contribution of the new Länder to economic growth in Germany has been disappointing: unemployment rates of 15-20 percent in the east testify to a large pool of unused resources. This is a legacy of overly rapid convergence of eastern wages toward western levels. Unit labor costs stood almost 50 percent above the western level in 1991 (Figure I-11). Today eastern unit labor costs in the economy excluding the primary sector and in manufacturing still remain considerably higher, albeit by a much smaller margin than 50 percent.

Figure I-11.
Figure I-11.

Germany: East-West Unemployment and Labor Costs, 1991-2001

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Source: Volkswirtschaftliche Gesamtrechnungen; Eurostat; and author’s calculations.

51. The employment experience in the new Länder exemplifies the interaction of Germany’s welfare benefit system with shocks. With benefits linked to past wages, the system provided an incentive for unions to push for large wage increases when faced with the inevitable restructuring of east Germany’s uncompetitive industry. And an eastern entrepreneurship that had no ownership stake in its companies was only too ready to cave in to demands. The result exacerbated job destruction.

52. To cut unemployment, further wage moderation in the new Länder is warranted, although recent developments show that some equilibrating mechanisms are at work. Employers are increasingly leaving their associations, and hence the collective bargaining “table”, to strike better pay deals. Moreover, many of the remaining collective wage bargaining agreements in the east now contain opening clauses that allow lower pay or longer working hours to secure jobs.

Firms’ Membership in Employers’ Associations in the New Länder, 1993-2000

article image
Source: Burda and Hunt (2001), “From Reunification to Economic Integration.” Brookings Papers of Economic Activity

E. Economic Adjustment: The Road Ahead

53. Many points of the above analysis suggest that the low growth rate in 1992-2001 ought not to be representative of Germany’s medium-term outlook. Assuming per capita TFP were to grow by at least at 1 percent per annum, the contribution of the capital stock to growth was at least ½ percent per annum, and—thanks to the wage moderation of 1996-2000—the contribution of labor was small but positive, per capita potential output growth would be 1¾ percent. But absent policies to sustain wage moderation, labor’s contribution would drop close to zero, the projected growth rate for the working-age population, once unemployment closes in on the estimated natural rate of 7 percent.

54. However, the persistent, current slowdown dampens confidence in the view that Germany has put its decade of slow growth behind it. At this stage, the depth of the slowdown relative to most of the comparator countries chosen here appears to be more a matter of somewhat earlier timing as well as some specific factors:

  • Investment in machinery and equipment had peaked earlier than in the other countries, partly because tax changes prompted entrepreneurs to advance expenditure on machinery and equipment to a record level in 2000. At the same time, the post-unification slump in construction continued to exacerbate the fall in total investment.

  • Unlike elsewhere, residential real estate prices have not buffered the effect of the decline in the stock market on investment and consumption (Figure I-12).29

  • Structural changes in the German banking system are perhaps holding back the supply of funding.30

  • As in the past, Germany, along with the Netherlands, was faster and perhaps more severely affected by the 3½ percentage point slowdown in US GDP growth in 2001. Econometric evidence for the 1990s suggests that the GDP growth rate in Germany and the Netherlands falls by some 32-43 percent of this amount in two quarters and by 42-62 percent in four quarters, with the Netherlands at the upper end of the ranges (Figure 1-13). In France, the adjustment after four quarters is similar, but it is smaller after two quarters. In Italy and the UK the drop after both two and four quarters is considerably lower. Except for Italy, this is consistent with the countries’ shares in merchandise exports. Italy, however, stands out in that it is much less involved with the US economy via foreign direct investment.31 For the UK, the country’s smaller exposure to oil price changes might have mattered.

  • The structural developments in the labor market continued to play a role. Given the lags with which wage moderation affects labor supply, the preceding upswing was less job-rich in Germany than in the other countries reviewed (see Table 1-4 and IMF, 2001). With factor proportions still adjusting to a lesser extent to wage moderation than in the other countries reviewed, employment and employees’ real disposable incomes failed to accelerate, despite a sizable tax cut in 2001.

Figure I-12.
Figure I-12.

Real GDP Growth and Asset Prices

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Sources: WEO database; BIS; and Bloomberg.
Figure I-13a.
Figure I-13a.

Real GDP Growth Rate in Selected Contries: Response to Innovation in US Real GDP Growth Rate, 1992:1-2001:4 1

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Source: IMF, World Economic Outlook and staff calculations.1/ Obtained from running bivariate VARs. Size of innovation is one standard error. Error bands are two standard errors wide
Figure I-13b.
Figure I-13b.

Real GDP Deviation from HP-Filtered Real GDP in Selected Contries: Response to Innovation in Deviation of US Real GDP from US HP-Filtered Real GDP, 1992:1-2001:41/

Citation: IMF Staff Country Reports 2002, 240; 10.5089/9781451810431.002.A001

Source: IMF, World Economic Outlook and staff calculations.1/ Obtained from running bivariate VARs. Size of innovation is one standard error. Error bands are two standard errors wide

Fixed Investment, 1992-2001

(Growth rates, in percent)

article image
Source: Eurostat and staff calculations

Exports of Goods

(In percent of GDP)

article image

55. And absent reforms, the economy is likely to continue to lack resilience in the face of shocks. The 1990s marked a decade of major shocks and adjustments in the German economy, unlike in any of the other countries reviewed.32 A specific problem in Germany has been the interaction of the welfare benefit system with large economic shocks. Without change, there would be a risk that shocks would once again lead to an upward ratcheting of unemployment. Ljungqvist and Sargent (1998) model the interaction of shocks with benefit systems for the jobless. Shocks destroy employment and the jobless lose firm-specific and (potentially) industry-specific human capital.33 Having suffered human capital losses, the jobless would have to agree to a wage cut to find a new job. However, in Germany, the incentive to do so is limited by generous benefits that are a function of previous earnings (Grund (1999) and Burda and Mertens (2001)). As a result, shocks cause an increase in the duration of unemployment spells that slows job creation and real GDP growth.

56. Finally, although the effects of the unification-related shocks should be unwinding, reforms are needed to deal with the “stock problem” posed by the low employment rate of older workers and workers in the new Länder. While various measures should be contemplated—including looser employment protection legislation, more scope for temporary employment, and better job matching—a reform of benefits for the jobless is the perhaps the single most crucial step. Meanwhile, the effect of measures should be leveraged with reforms that foster productivity growth, by enhancing product market competition and improving the quality of education.

APPENDIX I Identifying Labor Supply and Demand

Labor supply

57. Labor supply can be characterized by a wage-setting relationship relating real consumption wages to unemployment rates.34 Define the real effective consumption wage as ωc, with ωc=log(ωc/α) then the wage-setting relation is given by ωc=-βU-Z where U stands for the unemployment rate. The relation assumes reservation (and thus actual) wages rise with productivity: history clearly suggests no relation between productivity and employment in the long run. However, if workers’ aspirations lag reality, reservation wages are unlikely to adjust instantaneously with TFP, giving rise to labor supply shocks in this model.

58. Is it sensible to build on such a wage relation? One perspective is offered by the job matching approach and wage bargaining.35 Workers cannot costlessly relocate to find a new job nor can they costlessly and instantaneously be replaced by their employers. In a depressed labor market, workers will settle for a wage close to the reservation wage because it is hard to find a job elsewhere. The opposite holds for a booming labor market. Alternatively, models focus on the firm-worker relationship, arguing that wages affect productivity. Firms may want to pay more to workers than their reservation wage to economize on turnover costs, motivate greater effort, or for social considerations, such as fairness.36 The following equation can thus be taken to characterize the evolution of labor supply over time:


59. With data for U, w, and α available, while a can be obtained from the growth accounting exercise, assessing labor supply developments requires an estimate for the parameter β. Blanchflower and Oswald (1995) investigate the relation between real wages and unemployment for three out of the five countries considered here (Italy, the Netherlands, and the UK). Using internationally broadly comparable microeconomic data, they estimate a cross-sectional earnings equation for each country, in which, together with the usual set of control variables,37 the regional unemployment rate is entered as an explanatory variable. They demonstrate that there appears to be an empirical regularity in international pay and unemployment data, whereby estimates of the unemployment elasticity of pay cluster around-0.1. This result is broadly consistent with β=1, considering the average unemployment rates in the countries, implying an increase in unemployment by 1 percentage point decreases effective wages by 1 percent. Other studies have found estimates for β close to 2 but qualitatively that difference is not important.

Labor demand

60. The specification of labor demand begins with the following CES-type production:


where n stand for labor, k for capital, and σ for the (constant) elasticity of substitution between effective labor (an) and capital. To find labor demand, set the marginal product of labor equal to the real effective product compensation ((1+tp)w/a) multiplied by a mark-up (1+μ)


with the mark-up resulting, for example, for “efficient bargaining.”38 Take natural logarithms and rearrange to find:


61. There are three perspectives to changes in labor demand. A technology-driven decline in labor demand can be thought of as a drop in α for unchanged μ. Alternatively, a drop in demand because of less union bargaining power can be thought of as an increase in μ. for given α. And a drop in the demand for labor because of a rise in payroll taxes can be thought of an increase in the payroll tax rate tP, with α and μ unchanged. Observationally these changes in labor demand Ld are equivalent and their change over time is given by:


62. Conveniently, for σ=1, the evolution of labor demand is given simply by the natural logarithm of the labor share in income. Moreover, the effect of payroll taxes on labor demand can be approximated by the percentage point change in the effective tax rate. Note that equations (3) and (4) are correct only in the absence of costs of adjusting factor proportions. If it is costly for firms to adjust factor proportions, an increase in the wage will be associated with little contemporaneous change in (an/k) and thus (an/y); this in turn will lead to a decrease in the measured wedge. To allow for adjustment costs in factor proportions Blanchard (1998) proposes to replace log(w/a) with log(wt*)=0.8 log wt-1*+0.2 log (wt/at) in computing Ld, implying a mean lag in the adjustment of factor proportions of four years.39

APPENDIX II Macroeconomic Model Simulations

63. The simulations were carried out with the help of the Oxford Economic Forecasting (OEF) model. The OEF model is a mainstream economic model, situated in the middle ground between purely statistical models of the economy (e.g., vector autoregressions (VAR)) and computable general equilibrium models (CGEMs). Like CGEMs, the model imposes certain parameter restrictions based on theoretical priors. However, like VARs, it also estimates other parameter coefficients with country data to ensure that the model does well in reproducing the short-run behavior of key variables. The model exhibits Keynesian features in the short run and neoclassical ones in the long run. Its key characteristics are: (i) countries have a natural long-run growth rate but aggregate demand need not match supply in the short run; (ii) monetary policy, which is characterized by a Taylor rule, cannot affect output in the long run; (iii) consumption is a function of real incomes, real financial wealth, the real interest rate, and inflation; (iv) investment equations are influenced by q-theories, in which investment is a function of its opportunity cost after taking taxes into account, as well as accelerator effects; and (v) the country is small with terms of trade determined and exports a function of world demand and external demand. The model is backward-looking, with equations featuring an error correction component. In the simulations, the response of monetary policy to fiscal policy is muted.

64. The specific question addressed was how would real GDP growth in Germany have fared if the country had adopted the same fiscal policy and benefited from the same change in monetary conditions as the other countries:

  • A “same fiscal policy” was defined as a policy that would have imparted the same fiscal impulse to the economy as in the other countries reviewed.40 Because the actual fiscal adjustment in Germany was largely achieved by revenue increases, the simulated change in the impulse to make it match that in the other countries entirely took the form of a higher or lower tax bill.

  • For Germany, the fiscal balance was adjusted for the unification-related expenditure by the major public enterprises—the post and telecommunications company and the railway company—that were outside the general government sector. Together with the operations by the Treuhand Anstalt, borrowing by the public enterprises increased the fiscal deficit by some 1½ percentage points of GDP during 1991-95 (see Tables I-20 and I-21). Much of this borrowing had been eliminated or assimilated into the general government by 1995.

  • The “same change in monetary conditions” was defined as the same change in the short-term real interest rate. The cumulative differences in real interest rate changes between countries were relatively small over 1992-2001. Both interest rate levels and changes matter for investment. However, to the extent that the stock of capital had adjusted to a level consistent with interest rates prevailing in 1990-91, interest rate changes may be more relevant. A more fundamental problem is that the simulations should really use ex ante rather than the observed ex post real interest rates: higher ex post real interest rates in Italy during the early 1990s partly reflected market views about the probability of disinflation failure.

  • The real exchange rate, the other determinant of monetary conditions, was ignored. In real effective terms, the deutsche mark appreciated by about 20 percent in the first half of the 1990s. By contrast, the French franc and the Dutch guilder appreciated between 0-5 percent, while the lira and pound depreciated to different extents. EU (2002a) argues that the effects of the appreciation during the first half of the 1990s should have been unwound by now. And pre-EMU versions of the OEF model suggest that real exchange rate movements have only a small, short-run effect on GDP in Germany. A key issue is the interaction of monetary policy and the exchange rate: in the OEF model, a 5 percent appreciation of the exchange rate triggers a 0.5 percentage point cut in interest rates after one year. Thus, GDP drops only by 0.2 percentage points initially but returns to the baseline after three years.41 However, the real exchange rate powerfully affects output in the industrial sector in the OEF model, an issue which is discussed further in the main text.42


  • Akerlof, George A. and Janet L. Yellen (1990), The Fair Wage-Effort Hypothesis and Unemployment,” Quarterly Journal of Economics, Vol. 105, pp. 22583.

    • Search Google Scholar
    • Export Citation
  • Argyle, M. (1987), The Psychology of Happiness, Methuen, London.

  • Barro, R. J. and X. Sala-î-Martin (1995), Economic Growth, McGraw-Hill, New York.

  • Baily, M. N. (2002), “The New Economy: Post Mortem or Second Wind?,Journal of Economic Perspectives, Vol. 16, pp. 322.

  • Bassanini, A. and E. Ernst (2002), “Labour Market Institutions, Product Market Regulation, and Innovation: Cross-Country Evidence,Economics Department Working Papers, No. 316, OECD.

    • Search Google Scholar
    • Export Citation
  • Bassanini, A., Scarpetta S,. and P. Hemmings (2001), “Economic Growth: The Role of Policies and Institutions,Economics Department Working Papers, No. 283, OECD.

    • Search Google Scholar
    • Export Citation
  • Blanchard and Katz (1997), “What We Know and Do Not Know About the Natural Rate of Unemployment,Journal of Economic Perspectives, Vol. 11 (1), pp. 5172.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Blanchard, O. (1998), “Revisiting European Unemployment, Capital Accumulation, and Factor Prices,mimeo, MIT, Cambridge (MA).

  • Blanchard, O. and J. Wolfers (1999), “The Role of Shocks and Institutions in the Rise of European Unemployment: the Aggregate Evidence,NBER Working Paper No. 7282.

    • Search Google Scholar
    • Export Citation
  • Blanchflower, D. and A. Oswald (1994), The Wage Curve, MIT Press, Cambridge (MA).

  • Blanchflower, D. and A. Oswald (1995), “International Wage Curves”, in Freeman and Katz (eds.), Differences and Changes in Wage Structures, University of Chicago Press, Chicago.

    • Search Google Scholar
    • Export Citation
  • Borga, M. and D. R. Yorgason (2002), Direct Investment Positions for 2001: Country and Industry Detail, Bureau of Economic Analysis.

  • Bundesbank (2000), Monthly Report, August.

  • Bundesbank (2001), Monthly Report, July.

  • Burda, M. C. and J. Hunt (2001), “From Reunification to Economic Integration: Productivity and the Labor Market in East Germany”, Brookings Papers on Economic Activity (2), pp. 192.

    • Search Google Scholar
    • Export Citation
  • Burda, Michael C. and Antje Mertens (2001), “Estimating Wage Losses of Displaced Workers in Germany”, Labour Economics Vol. 8, pp. 1541.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Calmfors, Lars and John Driffill (1988), “Bargaining Structure, Corporatism, and Macroeconomic Performance,Economic Policy, Vol. 6, pp. 1461.

    • Search Google Scholar
    • Export Citation
  • Edison, H. and T. Sløk (2001a), “Wealth Effects of the New Economy”, IMF Working Paper WP/01/77.

  • Edison, H. and T. Sløk (2001b), “New Economy Stock Valuations and Investment in the 1990”, IMF Working Paper WP/01/77.

  • European Union (2002a), “Germany’s Growth Performance in the 1990s,Economic Paper No. 170, Directorate General Economic and Financial Affairs