The United States as a Job Creation Machine: an Example for Germany?--Flemming Larsen

October 10, 1997

Address by Flemming Larsen
Deputy Director of Research, International Monetary Fund
at a conference by the Christian Social Union
Bavaria, Germany, October 10, 1997

Thank you, Mr. Chairman. It is an honor to address this forum of the Christian Social Union. I was very pleased to accept Finance Minister Waigel's kind invitation to discuss whether the extraordinary employment record of the United States holds useful lessons for Germany.

Let me say at the outset that I do not wish to be overly critical of the German model which after all has produced a Wirtschaftswunder envied by most other countries, with one of the highest living standards in the world. Nevertheless, I do believe that you are interested in a frank assessment.

The contrast is stark indeed. In Germany, unemployment has risen relentlessly since 1991 and has now reached 4 1/2 million or 11.7 percent of the labor force. In the United States, the unemployment rate has stabilized close to 5 percent--a 25-year low--with inflation remaining remarkably well contained. The contrasts are even sharper when one looks at employment trends. In the United States, total employment has increased by 9 percent since 1991. In Germany, it has fallen a staggering 6 percent.

It is true of, of course, that the United States and Germany are in different stages of the business cycle and that part of Germany's unemployment is cyclical in the sense that it can be expected to decline as the recovery that now seems to be underway gains momentum. However, even after full cyclical recovery most analysts believe that the unemployment rate will remain very high, perhaps close to 9 percent, which we consider to be the structural unemployment rate. In the United States, the current unemployment rate is probably slightly below the structural unemployment rate which we judge to be in the range of 5 to 5 1/2 percent--this suggests that the U.S. is now probably close to the peak in its business cycle and that growth will need to moderate to prevent a rise in inflationary pressures.

How does one then explain the difference in structural unemployment rates between the United States and Germany? And how does one explain that the structural unemployment rate has remained fairly stable in the United States since the early 1970s while it has been steadily increasing in Germany (and also in France and Italy)? These are complex questions but I believe that there is a growing consensus among economists regarding the key role of labor market institutions.

In the remainder of my presentation I will first compare the U.S. and German labor markets and address some of the commonly perceived adverse social consequences of the deregulated U.S. labor market. I will then discuss the drawbacks of the German (and European) model and point to some of the reforms that are likely to be needed to substantially reduce unemployment over the medium term. I will finally discuss the issue of labor and product market reform from the perspective of the planned Economic and Monetary Union.

Characteristics of the U.S. Labor Market compared with Germany
Comparing the German and U.S. labor markets will shed some light on the large differences in job creation, despite similar technology and external conditions. In the United States enterprises and workers respond directly to market incentives:

  • the government and union presence in the labor market is more limited
  • wage determination is less centralized and developments are guided by market forces; there is little emphasis on incomes policies
  • social welfare benefits are not as generous, they are available for relatively short periods, and they are designed to give incentives for recipients to search for new and better job opportunities
  • and the wedge between gross wages and take-home-pay, caused by income taxes and social security contributions, is much smaller.
In the tradeoff between allocative efficiency and distributional equality, the United States has emphasized allocative efficiency, while Germany has placed more stress on equity as a cornerstone of the Soziale Marktwirtschaft model. But the U.S. system provides both firms and individuals equal access to the market place and the economic opportunities that are available. The U.S. system stresses the incentives for individual initiative and responsibility for one's economic well-being--that is, rewards to innovation, risk taking, and investment in oneself (i.e., human capital). The social safety net attempts to limit the downside risks associated with this system but not to eliminate inequality of outcomes. From a European perspective, it is interesting that the vast majority of Americans do not question their economic system.

Let me now elaborate on some of the specific differences:

Regarding government intervention in the labor market, job protection legislation and labor standards are more limited in the United States than in Germany and tend not to interfere with decisions by individual enterprises. To illustrate this, the OECD has constructed an employment protection index based on the legal framework governing hiring and dismissal. It shows the United States as the least restrictive country (1 on a 20 point scale), while Germany is one of the most restrictive in the OECD (15).1Greater employment protection reduces labor turnover and job creation in part because it increases the costs of hiring and of dismissals and is positively correlated with higher unemployment rates.

Wage formation is decentralized (at the firm level) in the United States, while the wage bargaining process in Germany (like the Netherlands) is gauged by the OECD as intermediate--industry-wide but less centralized than the economy-wide negotiations more common in the Nordic countries. The U.S. workforce is less unionized (only 16 percent belong to unions compared with 29 percent in Germany). Union agreements cover only 18 percent of all U.S. wage agreements compared with 92 percent in Germany. And unlike in Germany, there is no coordinated or pattern bargaining.

Aggregate real wages and relative wages respond more quickly to market forces in the United States. Trade unions in Germany have tended to push up hourly wages to the benefit of the employed (insiders) but to the detriment of the unemployed, of workers who may become unemployed, and of those discouraged from entering the labor force (jointly known as outsiders). Economic research suggests that real wage developments in Germany are only half as responsive to unemployment as in the United States.

Recent Fund staff analysis indicates that labor costs in Germany exceed substantially (by about 10 percent in the old Länder) the level derived under the assumption of competitive markets and full employment. A labor force that is overpriced by international standards has induced firms to seek less expensive foreign sources for labor-intensive production and to invest directly abroad. The ratio of net foreign direct investment outflows to GDP in Germany has been, on average, five times higher during the 1990s than it was during the 1970s. High labor costs have also forced firms to invest primarily in labor-saving technology; this has artificially boosted labor productivity and reduced the job intensity of growth.

Wage dispersion has increased in the United States largely because of higher earnings paid to workers in the upper ten percent. But wage dispersion in Germany has narrowed mainly because of higher earnings paid to workers in the lower portions of the income distribution. This trend also reflects the faster growth of real wages for low-skilled workers in Germany compared with the United States. Over the period 1985-1995, real earnings of low paid workers increased nearly 20 percent in Germany but they decreased by 5 percent in the United States. Skill-biased technological change has probably been the main factor underlying the greater wage dispersion in the United States: the wage premium for the college educated doubled and workers who use a computer on the job now earn significantly more than those who do not. Changing technology has created substantial economy-wide employment losses for low-skilled workers in Germany (4 percent per annum during 1982-93), which have exceeded those in the United States (about 1/2 percent per annum during 1982-93). By contrast, employment has increased for skilled workers in both economies.

Perhaps even more important than the income dispersion among categories of workers is the income differential between the employed and unemployed--this, after all, is a major incentive to work. The provision of high social welfare benefits in Germany-- both unemployment and social assistance--is a key difference distinguishing the United States and Germany. While alleviating some of the hardships associated with unemployment, such benefits change labor market incentives by increasing the reservation wage, reducing job search (and thereby trapping people in unemployment), and by lowering the costs associated with excessive wage demands.2 Eligibility criteria for receiving unemployment benefits are also less stringent in Germany, particularly conditions (such as wage levels and travel time) that allow the unemployed to reject a job offer and still retain unemployment benefits. Effectively, the relatively generous benefits in Germany have allowed unions to press for higher wages with little concern for the unemployed. Once unemployed, workers in Germany confront a welfare/unemployment trap: high effective marginal tax rates (75-90 percent) when returning to work, which undermine the economic incentive to find employment. Not surprisingly, the share of long-term unemployment (12 months or longer) in total unemployment has been dramatically lower in the United States (9.5 percent in 1995) than in Germany (48.3 percent in 1995).

The tax and social contribution wedge also affects employment decisions. Taken together, the tax and social security systems impose a much smaller "wedge" between gross and net wages in the United States than in Germany; there is thus a smaller disincentive effect on both employers and employees. The effective average rate for personal income taxes and social security contributions was somewhat below 30 percent in the United States, while in Germany, the effective average rate was around 47 percent. Moreover, because of differences in wage bargaining, the cost of this wedge is shifted onto employers in Germany, while in the United States wages tend to adjust downward so that total labor costs remain unchanged.

Corporate taxation. The effective average corporate tax rate in the United States is roughly equal to the effective average rate for personal income. By contrast, the effective average corporate tax rate in Germany was only 10 percent owing to generous depreciation allowances and liberal accounting standards. There are two implications of these large differences in effective tax rates on personal and corporate incomes--or perhaps more precisely on labor and capital.

  • One, the tax system in conjunction with the wage bargaining process has helped to skew output growth away from relatively high cost (and taxed) labor and toward higher capital intensity.
  • Two, government revenues are more highly sensitive to changes in the distribution of income between labor and capital in Germany than in the United States. This sensitivity has been especially evident this year in Germany as the fiscal balances were affected by the lopsided recovery that favored capital relative to labor.
Unification also highlights the structural roots of German unemployment. The extension of the wage determination and social welfare systems to the new Länder contributed to an overly rapid convergence of wages compared with the convergence of productivity. As a result, unit labor costs in the east are one third above those in the west.3 Consequently, a high unemployment rate developed in the east (17 percent, or 25 percent if the underemployed are included), and has persisted, requiring massive and enduring fiscal transfers from west to east. The labor market situation in the new Länder contrasts with the labor market conditions in Poland, which began its transition at about the same time (but with less assistance). Poland now has an open unemployment rate more than 7 percentage points below the comparable rate in the new Länder. Rapid export-led growth in Poland fuels convergence with the EU, while exports play virtually no role in the new Länder and the convergence process with the old Länder has stalled.

There are some important areas where Germany and other countries in the European Union clearly outperform the United States. In particular, investment rates in the major European economies have averaged around 20 percent of GDP (and above in Germany) in the first half of this decade, compared with 17 percent in the United States. Consequently, Europe's capital stock and ratio of capital to labor have been rising faster than in the United States. European productivity growth has also been higher, and levels of productivity are converging to that in the United States. Overall savings rates are also higher in Europe, and critically, the pre-university education system in Europe is widely thought to work better than that in the United States. By comparison with the United States, Germany's and Europe's labor market problems do not seem to reflect shortages of either physical or human capital.

Perceptions of social consequences of U.S. Model
So far, I have alluded to some of the key elements of how the U.S. approach has been more effective in creating new jobs. But are these jobs any good? Does the U.S. system produce an army of working poor with no future? Does the U.S. system achieve its efficiency at too high a social cost?

The U.S. labor market model is often rejected in Europe because of a widespread perception that the U.S. economy is a "wild west" economy without rules and with unacceptable social implications. To be sure, there are few restrictions on layoffs for economic reasons, and according to recent polls, workers in larger firms are increasingly concerned about losing their jobs. Fear of job loss is heightened by the U.S. health system where an individual's medical care is tied to employment. In fact, this concern has been linked to muted wage increases despite low levels of unemployment that have been associated with rising inflation pressures in the past. While labor market "anxiety" is clearly not welcome from an individual point of view, mobility in the job market has advantages. Managers can take risks in hiring when they see expanding opportunities for their products, and they can restructure so that workers are matched to jobs according to productive skills. This degree of flexibility promotes a dynamic environment in which established businesses and entrepreneurs are able to develop new products, services, and production techniques more easily and quickly.

Criticisms of the U.S. labor market model often revolve around concerns over income distribution and equity. It is true that the measured distribution of income in the United States has become more unequal in the past two decades, particularly because real wages at the lowest end of the range have not kept up with those at the upper end. However, there is a great deal of upward income mobility in the United States, much of it linked to education and skill upgrading, and taking this into account substantially reduces--by upwards of 25 percent--the degree of income inequality in the United States.

It is true that the flexibility of wages and lack of wage floors in the United States means that full time work can provide low incomes for the low skilled. The tradeoff, however, is that such flexibility in compensation allows employers to offer jobs to the low skilled in the first place, reducing the risk of long-term unemployment and social exclusion. Rather than having wage disparity reflect skill levels, the distinction in Europe is more fundamentally between those with jobs and those without. Because of their lack of flexibility, European labor markets "clear" at lower levels of employment and total output.

Critics often contend that the jobs created in the United States over past decade have been low-pay, low-skill positions in the service sector--the so-called hamburger flippers at McDonalds. It is true that all the net new jobs created in the United States between 1989 and mid-1997 are found in the service sector. However, these service sector jobs are actually quite diverse and many jobs are in high-wage positions in financial services, health care, computers, and professional and business services (which in some cases have been outsourced from manufacturing firms). A recent US study found that almost 70 percent of the net growth in full-time employment occurred in jobs that paid above the median wage. Moreover, there are also opportunities for advancement in these occupations, promising higher wages in the future.

What is needed in Germany and elsewhere in Europe?
Compared with the United States, there is little doubt in my view that the root causes of Europe's poor labor market performance are the adverse (and unintended) consequences of elaborate job and income protection arrangements that raise the cost of labor (including through high social security contributions), discourage job creation and job search, and favor substitution of capital for labor. Product market distortions also appear to have contributed by reducing efficiency and impeding competition. These problems have been exacerbated by technological changes that demand increased flexibility and adaptability. In contrast, the forces of globalization and increased trade with low-cost countries appear to have played only a minor role (a topic we have analyzed extensively in the May 1997 World Economic Outlook).

In recognition of the need to allow market forces a greater role, a wide variety of reforms have been implemented across Europe in recent years, but many have been postponed and some measures may in fact have exacerbated structural problems in labor markets. A common problem has been the failure to implement sufficiently comprehensive labor market reforms because of strong opposition from insiders--those who have jobs and feel they benefit, rather than suffer, from existing labor (and product) market institutions and regulations. In fact, labor market policies have often sought to mask the underlying problems by promoting early retirement or work sharing. Such measures appear to be intended to reduce open unemployment not by increasing the demand for labor but by reducing labor supply. But with unreformed labor markets, such measures tend to improve the bargaining position of insiders and raise their real wages, with little benefit to outsiders who are likely to remain unemployed. Such measures also involve costs to producers or to government budgets, creating a vicious cycle of higher taxes, higher labor costs, further employment losses, and so on. Instead, what is needed are reforms that give outsiders a better chance to compete for jobs, thereby raising the effective supply of labor, and that also increase labor demand. In addition to reducing unemployment on a durable basis, such reforms would have important positive effects on growth and public finances. They would in fact allow most of the remaining fiscal imbalances in Europe to be worked off through reduced costs of income support for the unemployed as well as higher tax revenues associated with rising employment and higher levels of national income. Failure to implement reforms that reduce structural unemployment would require much tighter fiscal policies over the medium term, reduce the scope for tax reduction and productive public expenditure, and in all likelihood permit only modest growth of output and real incomes--that is, a continuation of recent experience.

The basic strategy was recognized by the German Government when it launched its 50 Point Program in early 1996. These reforms focussed on reducing rigidities in labor markets, lowering nonwage labor costs, strengthening job search incentives, and deregulation. The Government has acted to ease dismissal rules, promote part-time jobs, and extend the use of fixed-term contracts--all of which are also helpful to stimulate employment. Still, the measures adopted so far have been piecemeal and inadequate.

There are five major areas where reforms are needed to improve employment prospects in Germany:

First, labor costs need to be lowered in many sectors to gain competitiveness and stem the tide of job losses. There have been some encouraging signs of wage moderation and greater flexibility in the workplace from employers and unions. These developments must be encouraged, inter alia by sensible government wage settlements. The Government should also resist making wage agreements generally binding on all parties within the sector (Allgemeinverbindlichkeitserklärungen) as was done in the construction sector.4 To narrow the unit labor cost gap in the new Länder, as part of a joint initiative with the Government, employers and unions have agreed to stabilize labor costs in eastern Germany. But more action is necessary: greater realism in wage negotiations and more flexibility in the centralized collective bargaining process is needed so that wage levels correspond more clearly to worker productivity and the financial conditions of individual firms. In this connection, a shortening of the work week to 32 hours with no change in weekly pay, as recently proposed, would clearly be counterproductive.

Second, it is critical to reduce non-wage labor costs and, in particular, to limit the need for further increases in social security contributions as the German population ages. The Government's objective is to reduce social security contributions from currently more than 42 percent of gross wages (shared equally by employees and employers) to below 40 percent by 2001; this will be difficult to achieve. To ease the upward pressure on the already high contribution rates, reform of the pension system is essential. Reducing the net pension replacement rate and tightening of eligibility rules for early retirement should be the cornerstone of any future pension reform. Systematic reforms, such as the shifting of a significant portion of public pension provisions to the (currently) small privately funded occupational pension plans should be seriously considered.

Third, disincentives for the unemployed to seek jobs stem from the indefinite duration of social benefits for the unemployed and have fostered unrealistically high reservation wages. To stimulate job search, efforts should aim at lowering net benefit replacement ratios, to limit the duration of unemployment assistance, and to enforce strictly the recently tightened rules on refusing a job offer without loss of benefits. To reduce "unemployment traps", the Commission on Alternative Tax-Transfer Systems suggested in 1996 that the different benefit systems be aligned and that benefits be withdrawn only slowly as wage income rises. Studies in the United States indicate that the earned income tax credit unambiguously strengthens the incentive to take a job by increasing net income in work at all levels of earnings up to the end of the phase-out.

Fourth, income tax reform has to be a central element of the reform agenda for Germany. Tax rules are too complex and relatively opaque. The proliferation of preferential rules has distorted investment decisions, while the extensive use of tax shelters has created the perception of a lack of tax equity. Unfortunately, the opposition in the Bundesrat vetoed the Government's tax reform proposal. The tax reform would have commendably lowered and flattened the steep marginal tax rate schedule, reduced the tax wedge on labor income, broadened the narrow tax base by reducing tax allowances, and lessened tax distortions. It would have also implied a shift from direct to indirect taxation.

But even the Government's reform proposals left scope for a further broadening of the tax base, both on personal and corporate income.

Last but not least, there is a need to substantially curtail the many subsidies and rents that remain such a dominant feature of the economic landscape in Germany. And it does not matter whether such subsidies are paid over the budget or through higher prices to consumers or enterprises. Regulations, cartels, subsidies and other impediments to market forces in electricity generation/distribution, postal services, telecommunications, trucking, banking, coal mining, shipbuilding, and agriculture all contribute to excessive cost levels throughout the economy that hamper competitiveness and job creation.

In fact, rather than protecting employment such actual or implicit subsidies constitute a tax on the competitive sectors in the economy. It is encouraging that reforms are underway in some of these areas but Germany is clearly lagging behind the United States by many years in the process of deregulation and product market reform.

Diagnosis is easier than curing the patient. It is clear that many of these needed reforms will be politically difficult to implement, especially now as the electoral season gets into full swing here. A major task will be to galvanize political will and to educate the public about the causes and costs of Germany's most pressing problem. We in the Fund are looking to Germany to set an example for the rest of Europe.

Labor market flexibility and EMU
This brings me to the last issue: the importance of labor market reforms for the success of EMU. Just two weeks ago, the IMF published an extensive analysis of the implications of EMU and on the conditions we believe need to be fulfilled to make EMU a success and allow it to help strengthen Europe's economic performance.5

The IMF strongly supports the efforts that have been under way for over 40 years to promote European economic and monetary interpretation, including the adoption of a single common currency. We are particularly pleased with the strong commitment to price stability embodied in the charter for the European Central Bank. We are also impressed by the sensible rule for the conduct of fiscal policy embodied in the Stability and Growth Pact. The macroeconomic policy regime in the euro area clearly promises to allow the euro to become a robust international reserve currency which over time is likely to rival the role of the U.S. dollar in the international monetary system. Regarding the conditions for a successful EMU, it is our assessment that the macroeconomic convergence criteria set out in the Maastricht Treaty are now so close to being fulfilled that it seems safe to proceed with the third and final stage of EMU at the end of next year. Of course we have also underscored the need for further fiscal reforms to safeguard long-term fiscal stability.

This being said, we have also stressed the importance for the long-term success of EMU to step up efforts to enhance the flexibility of European labor and product markets. This is of course needed with or without EMU, but it becomes particularly important when countries can no longer resort to exchange rate changes when they are faced with exceptional shocks that require adjustment of real wages in order to limit the adverse consequences for activity and employment. A failure to implement sufficiently comprehensive reforms could well result in further increases in unemployment, lead to persistent budgetary difficulties, and ultimately undermine public support for the project.

* * *

In conclusion, while the proper balance between market efficiency and equity differs from country to country, I do believe that Germany and Europe have a lot to learn from the dynamism of the U.S. labor market. In fact, several countries in Europe, including the United Kingdom, the Netherlands and Denmark have already made good progress in reforming the labor market and their structural unemployment rates have begun to decline.

This does not mean that market forces can solve all problems. Clearly, there are market failures that government policies need to deal with in order to foster social cohesion. However, such policies will need to work to a greater extent through taxes and transfers and through the education system and to a lesser extent through regulations and other impediments to the effective functioning of market forces.

I would like to end on a positive note by putting this discussion into a cyclical perspective. After a long period of dismal growth, Germany and the rest of continental Europe now seems poised for a period of economic recovery. With low inflation; with much of the fiscal retrenchments needed to satisfy Maastricht now behind us; with long-term interest rates at record lows in most of Europe; with confidence improving as uncertainties about EMU are waning: all of these factors should help support relatively robust growth during the rest of the decade. As a result, we are likely to see some decline in cyclical unemployment and the budgetary situation will suddenly look a lot better. Under these circumstances, there will be a temptation, also because of adjustment fatigue, to sit back and relax the reform efforts. That would be a pity, because the underlying problems would not be going away. Instead, the next 3 to 4 years are likely to provide perhaps the best economic environment Europe has seen for a long time--and is likely to see for some time in the future, for implementing the fundamental reforms that are so badly needed to restore Germany's and Europe's long-term economic dynamism.

1Labor standards cover several dimensionsÑworking time, fixed-term contracts, employment protection, minimum wages, and employeesÕ representation rightsÑwere measured by an index ranging from the least (zero) to the most (ten) intervention. The United States was scored at zero and Germany at 6 with only three OECD countries above it (i.e., Italy, Spain and Sweden).
2Net replacement rates in the United States in the first month of unemployment range from 60-68 percent for a couple, while in Germany, comparable rates range from 60-78 percent. However, the net replacement ratio, including social assistance, in the 60th month of unemployment is only 0-17 percent in the United States for a couple with and without children compared with 37-71 percent in Germany. The duration of unemployment insurance in the United States is relatively short (26-39 weeks, depending on state unemployment levels); in Germany, benefit periods extend up to 32 months and for unemployment and social assistance have indefinite duration.
3Standard unit labor cost calculations for the new LŠnder could be biased by price distortions in the base period. Thus, for both the new and old LŠnder unit labor costs are defined as total labor costs divided by nominal GDP per employed person. However, this method tends to understate the differences in unit labor costs.
4The purpose of this action was mainly to protect German construction workers from competition from foreign construction workers.
5World Economic Outlook, October 1997.


Public Affairs    Media Relations
E-mail: E-mail:
Fax: 202-623-6278 Phone: 202-623-7100