Abstract
This Selected Issues paper for France provides an analytical framework to explain the consequences of the downward shift in the unemployment/wages relationship. This framework is also used to analyze possible changes in the equilibrium unemployment rate resulting from cuts in employers’ social security contributions and movements in the user cost of capital. The contribution of wage moderation to the reduction in the equilibrium unemployment is quantified. The paper also addresses the question of fiscal benefits of job-rich growth in France during 1997–2000.
I. Wage Moderation and Long-run Unemployment in France2
A. Introduction
3. The economic expansion of the second half of the 1990s was characterized by a sharp rise in employment and reduction in unemployment. Employment increased by about 1.8 million people, a record in such a short period of time for France, and the unemployment rate fell from a peak of 12.3 percent at the beginning of 1997 to 8.6 percent in mid-2001. This performance is all the more noteworthy in that output grew less during this period than in previous expansions. Finally, nominal wages were surprisingly sluggish during the expansion and, as a result, real wage growth inched up 1¾ percent from 1997 to 2000. The combination of high employment and relatively modest output growth led some analysts to characterize the upswing as “rich in employment.”3
4. Existing studies suggest that job-rich growth was caused in part by changes in the basic parameters of the wage setting mechanism resulting in a rightward shift in the labor–supply like relationship between real wages and the unemployment rate.4 This improved the trade-off between wages and employment (the wage-setting relationship) in the 1990s, implying a much lower equilibrium unemployment rate than prevailed at the beginning of the decade.
5. This chapter provides an analytical framework to explain the consequences of the downward shift in the unemployment/wages relationship. This framework is also used to analyze possible changes in the equilibrium unemployment rate resulting from cuts in employer’s social security contributions and movements in the user cost of capital. The model not only accounts for the large job growth associated with sluggish wages observed in the second half of the 1990s, but also predicts a pickup in investment rates as the economy converges to its long-run equilibrium. So, abstracting from business cycle effects, the previous job-rich growth phase could be followed by a “capital-rich” growth period. The contribution of wage moderation to the reduction in the equilibrium unemployment is quantified in the final part of the chapter. It is shown that if this moderation is not reversed in future years, the long-run unemployment rate will be less than half of its 1996 reading.
6. The rest of the chapter is organized as follows: section B discusses recent labor market performance in France; section C develops a model to analyze the job-rich nature of growth in the last part of the 1990s and presents an estimate of the decline in the long-term unemployment rate resulting from wage moderation; Section D concludes the chapter.
B. Some Figures on Job-Rich Growth
7. The nature of employment growth during the second half of the 1990s can be illustrated by a comparison with the previous expansionary period at the end of the 1980s (Table I.1).5 Annualized labor productivity growth—defined as changes in the output/employment ratio—in the business sector was 2 percentage points lower in the most recent episode, and employment growth was 1½ percentage points higher. If labor productivity growth at the end of the 1980s were used to back out employment growth given the rate of output change in the more recent period, employment would have risen only by about half of the rate posted at the end of the 1980s (fifth column, lines 1 and 2). The “extra” employment produced in the business sector during the most recent expansion (i.e., the number of jobs due to lower productivity growth for a given rate of output increase) was about 1.2 million jobs (the difference between the actual cumulated change in employment in column 3 and the counterfactual figure in column 5).
Job-Rich Growth
Adjusted employment for the latest expansion is equal to output growth minus productivity growth in the previous expansion.
Employment adjustment for the latest expansion obtained by applying the adjusted rate of growth to levels from 1996Q4 forward.
Job-Rich Growth
| Percent change during the period indicated at an annual rate | |||||||
|---|---|---|---|---|---|---|---|
| Labor productivity | Employment | Adjusted employment 1/ | |||||
| Business sector | Aggregate | Business sector | Aggregate | Business sector | Aggregate | ||
| Expansions | |||||||
| 1987Q2-1990Q3 | 3.15 | 2.64 | 1.09 | 1.05 | 1.09 | 1.05 | |
| 1997Q1-2001Q2 | 1.16 | 1.39 | 2.47 | 1.77 | 0.50 | 0.54 | |
| Accumulated changes in thousands 2/ | |||||||
| Expansions | |||||||
| 1987Q2-1990Q3 | 532 | 822 | 532 | 822 | |||
| 1997Q1-2001Q2 | 1540 | 1856 | 303 | 556 | |||
Adjusted employment for the latest expansion is equal to output growth minus productivity growth in the previous expansion.
Employment adjustment for the latest expansion obtained by applying the adjusted rate of growth to levels from 1996Q4 forward.
Job-Rich Growth
| Percent change during the period indicated at an annual rate | |||||||
|---|---|---|---|---|---|---|---|
| Labor productivity | Employment | Adjusted employment 1/ | |||||
| Business sector | Aggregate | Business sector | Aggregate | Business sector | Aggregate | ||
| Expansions | |||||||
| 1987Q2-1990Q3 | 3.15 | 2.64 | 1.09 | 1.05 | 1.09 | 1.05 | |
| 1997Q1-2001Q2 | 1.16 | 1.39 | 2.47 | 1.77 | 0.50 | 0.54 | |
| Accumulated changes in thousands 2/ | |||||||
| Expansions | |||||||
| 1987Q2-1990Q3 | 532 | 822 | 532 | 822 | |||
| 1997Q1-2001Q2 | 1540 | 1856 | 303 | 556 | |||
Adjusted employment for the latest expansion is equal to output growth minus productivity growth in the previous expansion.
Employment adjustment for the latest expansion obtained by applying the adjusted rate of growth to levels from 1996Q4 forward.
8. In considering these facts, it is worth noting that average hours of work have been declining steadily in France since the mid-1990s, thereby dampening the increase in total labor input used in the production process.6 That reflects the higher use of part-time work as well as the three laws introduced after 1996 to reduce the workweek.7 As shown in Table I.2, labor productivity growth—defined as changes in the ratio of output to hours of work—has in fact been declining steadily since the 1970s as a result of both slower capital deepening (i.e., lower growth in the capital/labor ratio, K/L) and slower TFP growth. The decline in capital deepening between the late 1980s and the late 1990s seems to have been significantly larger than the reduction in TFP growth, but such a comparison is tainted by possible errors in the way the capital stock is computed.8 The last column shows the evolution of total factor productivity weighted by the labor share, a measure of labor–augmenting technological growth (see Blanchard, 1997).
Decomposing Labor Productivity Growth in the Business Sector
(Percent change during the period indicated at an annual rate)
Decomposing Labor Productivity Growth in the Business Sector
(Percent change during the period indicated at an annual rate)
| Labor productivity | K/L | TFP | TFP/labor share | |
|---|---|---|---|---|
| 1970-1980 | 4.12 | 5.22 | 2.66 | 3.72 |
| 1980-1990 | 3.38 | 4.27 | 2.18 | 3.10 |
| 1990-2000 | 1.85 | 2.57 | 0.91 | 1.47 |
| 1987-1989 | 3.88 | 2.94 | 2.89 | 4.40 |
| 1997-2000 | 1.87 | 0.73 | 1.58 | 2.59 |
Decomposing Labor Productivity Growth in the Business Sector
(Percent change during the period indicated at an annual rate)
| Labor productivity | K/L | TFP | TFP/labor share | |
|---|---|---|---|---|
| 1970-1980 | 4.12 | 5.22 | 2.66 | 3.72 |
| 1980-1990 | 3.38 | 4.27 | 2.18 | 3.10 |
| 1990-2000 | 1.85 | 2.57 | 0.91 | 1.47 |
| 1987-1989 | 3.88 | 2.94 | 2.89 | 4.40 |
| 1997-2000 | 1.87 | 0.73 | 1.58 | 2.59 |
9. Another stylized fact concerning labor markets in France in the 1990s is the deceleration in wages. Using data from the Ênquete Emploi, Estevão and Nargis (2002) show that real hourly wages increased only by 3.2 percent during 1990–2000 in France (an average annual rate of 0.3 percent), well below total factor productivity growth adjusted for the labor share (about 1.5 percent per year)—our measure of labor–augmenting technical progress. Real wages grew only 1¾ percent between 1997 and 2000.9 The strong job creation associated with this tame real wage growth suggest either a large labor demand shock with important composition effects toward less-skilled workers or a positive labor supply shock or both.
10. The 1990s have actually been marked by many policy changes in France, most of them targeted to the hiring of less-skilled workers. Among the measures affecting labor demand directly were legislated cuts in firm’s social security contributions and in the size of the workweek. Among the measures to improve the quality of the labor pool were policies to enhance the employability of young workers through targeted job creation in the public sector (emplois jeunes) and to raise labor skills (contrats de formation et alternance). As will be further discussed below, there is some evidence that the first type of policy had a positive effect on employment but, in the absence of composition effects, they should have boosted wages as well. There is no convincing evidence of the effects of the second type of policies on employment in France although new results for a panel of OECD countries suggest a positive effect (see Estevão and De Coninck, 2002).
C. Wage Moderation and Variations in Structural Unemployment
11. Box I.1 presents an equilibrium labor market model to evaluate the impact of shifts in labor supply and labor demand on the short- and long-run unemployment rates and wages. Figure I.1 summarizes the important features of the model, which comprises: 1) a short-run labor demand equation that results from firm’s profit maximization process for a given level of capital (SLD curve); 2) a flat long-run labor demand curve that results from the hypothesis of a small open economy—interest rates, the user cost of capital, and thus the marginal product of capital are determined in the world market (LLD curve); 3) a contract curve relating wages to the unemployment rate (assumed, for the sake of simplicity, to be equal to 1-N, where N is the employment level) that results from bargaining between firms and unions (WS curve). Wages are defined as a ratio of the technology level (or, as it is commonly called, “in efficiency units”), implying that in equilibrium (for instance, point E in Figure I.1) wages grow at the rate of labor-augmenting technological growth.
Wage Determination and Structural Unemployment
Following Blanchard (2000) and Carnot and Quinet (2000), assume the economy grows along a balanced path determined by the rate of labor augmenting (Harrod-neutral) technological growth, ga. The production function combines labor and capital according to a constant returns to scale technology:
Y, N, K and A denote output, labor, capital, and the technology level. Assuming that capital is fixed in the short-run and firms maximize profits, the short-run labor demand curve is:
The link between wages in efficiency units, W/A, and labor is affected by the level of capital in efficiency units and the rate of employers’ social security contributions, te. This relationship is drawn as the curve SLD in Figure I.1, where the labor force is normalized to 1 and employment is N = 1-u (u is the unemployment rate). In the long run, capital is allowed to vary and, assuming interest rates are determined abroad, the user cost of capital, C, is exogenously given. In this case, labor cost in efficiency units is set to equalize the profit rate to the user cost of capital independently of the unemployment rate (LLD in Figure 1).
A “labor supply-like” relationship can be modeled according to the right-to-manage model (see Layard and others, 1991) in which firms and unions bargain over wages given the short-run labor demand. A version of such a model, developed in Estevão and Nargis (2002), generates:
where, B and τ stand for, respectively, the income a worker would receive if unemployed and the ratio between the fiscal wedge on unemployment income and the fiscal wedge on labor income; m is a structural parameter determining the position of the wage curve and its steepness, which is affected among other things by unions’ bargaining power and the relative importance of employment vis-à-vis wages in workers’ utility function. For a given rate of unemployment, wages in efficiency units depend on unemployment income (net of the relative tax wedge) and on the position of the wage curve, a function of m. Ceteris paribus, wage demands are higher the larger is unemployment income, as the outcome in case of disagreement is less unattractive. On the other hand, when the unemployment rate increases, the probability of not finding a job also augments and wage demands are more subdued. Whenever workers’ bargaining power becomes weaker, or whenever workers value employment more, the parameter m decreases and wages are lower for a given rate of unemployment.
Point E in Figure I.1 represents the long-run equilibrium in the labor market, where wages are such that the profit rate equals the worldwide user cost of capital. In this steady state, output, capital, and employment in efficiency units (AN) grow at ga percent.
Wage Moderation and Long-Run Adjustment
Citation: IMF Staff Country Reports 2002, 249; 10.5089/9781451813531.002.A001
12. Under the hypothesis of a significant downward shift in the contract curve between wages in efficiency units (W/A) and employment (N=l-u)—due for instance to a shift in workers preference toward employment and away from wages—wages will grow less than technological progress and the unemployment rate will decline as the economy moves along a negatively sloped short-run labor demand curve and reaches the short run-equilibrium point E1. However, in this situation firms will have a large incentive to invest in capital, as low wages raise profit rates to a level above the user cost of capital. The short-run labor demand will then shift outwards, moving along the labor supply relationship, until the profit rate and the unit cost of capital are equal, point E2. Structural unemployment is lower than in E but wages in efficiency units are unchanged.
13. In terms of this model, the movement in the WS curve between 1996 and 2000 was quite large. Using microeconomic data from the French labor force survey (Enquête Emplof,) Estevão and Nargis (2002) document that the WS curve moved to the right during the 1990s. The shift was stronger after 1996 and cannot be explained by reductions in the relative tax wedge on labor and unemployment income, by declines in either unemployment benefits or replacement rates, or by technological changes. Their analysis also accounts for compositional changes that may be driving the sluggish behavior in aggregate wages.
14. According to these estimates, the shift in the aggregate wage setting relationship and the accompanying outward shift in labor demand (movement from E to E2 in Figure I.1) would have more than halved the long-run equilibrium unemployment rate with respect to the mid-1990s. The full effect of the shift in WS takes place in the long-run though, as firms adjust their capital stock to align profitability to the user cost of capital. Point E1 in Figure I.1 corresponds to a “short-run” NAIRU of about 8½ percent.10
15. These calculations illustrate the importance of shifts in the wage–setting relationship for the sharp increase in employment during the period but the precise causes of the shift are difficult to determine. Estevão and Nargis (2002) attribute such a shift to a change of unions’ preferences toward more employment and away from wages. A variation of this interpretation is offered by Blanchard (2000). According to him, the unemployment rate rose between the mid-1970s and the mid-1980s because of workers’ failure to perceive the (exogenous) reduction in technological growth that took place in those years (Table I.2). As a result, wages grew faster than productivity. By the end of the 1980s, workers adjusted wage demands to the new reality, and the WS curve began to move back to its long-run equilibrium, with wages growing by less than productivity.
16. Wage moderation could also have been the result of bargaining among social partners or between social partners and the government, where unions limited their wage demand in exchange for other concessions (for example, the 35-hour workweek and training programs). Last, the employment policies implemented in the 1990s might have improved worker quality (mainly, among the long-term unemployed and younger individuals), increased labor availability, and, therefore, contributed to depress wages. Estevão and De Coninck (2002) show that larger expenditures in active labor market policies as a proportion of GDP during the 1990s have, in fact, been associated, on average, with higher employment rates in the business sector among OECD countries. The same was not true for the 1980s, which explains in part the negative results on the link between these policies and employment performance in studies written in the mid 1990s.
17. At the same time, however, the downward shift in the wage-setting relationship is likely not the only factor behind the recent positive labor market performance in France. First, as discussed in Blanchard (2000), changes in world interest rates have also moved long-run labor demand since the 1970s. While the increase in interest rates depressed investment and reduced the capital stock in the 1980s, causing an inward shift in labor demand, the decrease in these rates during the 1990s moved long-run labor demand outward and pushed structural unemployment further down.11 Second, the various cuts in firms’ social security contributions, targeted to the hiring of low-wage workers, allowed firms to offer higher wages at the same overall cost and likely caused an outward shift in labor demand and less structural unemployment. Crépon and Dezplatz (2001) calculate that about 450,000 jobs were either created or maintained between 1994 and 1997 due to reductions in employers’ social security contributions. Even though this is a very large number, there is no direct evidence of the isolated effect of this policy on employment growth between 1996 and 2000.12
D. Conclusions
18. The observed wage moderation and part of the employment increase in the second half of the 1990s can be viewed as the first phase of a labor market adjustment process where a structural change in wage bargaining in France dominates outward shifts in labor demand. After this stage, a “capital-intensive” phase is expected as investment rates increase to bring profit rates down until they match the user cost of capital. When this effect becomes sufficiently strong, wages should grow more rapidly than observed in the 1990s. In addition, these gains should not be taken for granted and labor cost increases (like the one coming from the unification of the minimum wage (SMIC) and monthly income guarantees by 2005) may easily reverse them. Finally, past cuts in employers’ social security contributions and reductions in world interest rates likely contributed significantly to a lower structural unemployment rate.
19. Wage moderation and the subsequent outward movement in labor demand is expected to cut the equilibrium unemployment rate in half in the long run. This amelioration comes, however, after a likely upward shift in the wage setting curve in the 1970s and 1980s, suggesting that the wage/unemployment locus is somewhat volatile in France.
20. Because of the importance of shifts in this curve to determine structural unemployment, wage behavior and, ultimately, inflation and economic growth, more research on the determinants of the trade-off between wages and unemployment in France is needed. Also, given our ignorance about the fundamental mechanisms behind wage moderation in France and the fact that world interest rates are not under the control of French authorities, a continuation of policies to lower labor costs could be crucial to ensure reduced unemployment rates.
References
Blanchard, Olivier, 2000, “The Economics of Unemployment: Shocks, Institutions, and Interactions,” Lionel Robbins Lecture, London School of Economics, October.
Blanchard, Olivier, 1997, “The Medium Run,” Brookings Papers on Economic Activity, 2, pp. 89–158.
Carnot, Nicolas and Alain Quinet, 2000, “L’Enrichissement du Contenu en Emploi de la Croissance: Une Tentative de Clarification,” Complément B in Jean Pisani-Ferry, Plein Emploi, Conseil d’Analyse Économique, Paris.
Crépon, Bruno, and Rozenne Dezplatz, 2001, “Une Nouvelle Évaluation des Effets des Allégements de Charges Sociales sur les Bas Salaries,” Economic et Statistique, no. 348, pp. 3–24, August.
Decressin, Jorg, Marcello Estevão, Phillip Gerson, and Christopher Klingen, 2001, “Job-Rich Growth in Europe,” Background paper to the French, German, Italian and Spanish 2001 Article IV consultation, IMF.
Estevão, Marcello, and Raphael De Coninck, 2002, “Active Labor Market Policies and Business Employment in OECD countries,” unpublished, IMF, August.
Estevão, Marcello, Nargis Nigar, 2002, “Wage Moderation in France,” IMF Working Paper, no. WP/02/151, Washington, D.C..
Hagneré, Cyrille, Alain Trannoy, 2001, “L’impact conjugué de trois ans de réforme sur les trappes inactivité,” Économie et Statistique, INSEE, 346/347:161–185.
Hamermesh, Daniel, 1993, Labor Demand, Princeton University Press, Princeton.
Kramarz, Francis, and Thomas Philippon, 2000, “The impact of differential payroll tax subsidies on minimum wage employment,” Discussion Paper, no. 00/488:1-30, University of Bristol, June, forthcoming in Journal of Public Economics.
Laroque, Guy, and Salanié, Bernard, 2000, “Une décomposition du non-emploi en France,” Economie et Statistique, INSEE, 331:47-66.
Laroque, Guy, and Salanié, Bernard, 1999, “Prélèvement et transferts sociaux: une analyse descriptive des incitations financières au travail,” Économie et Statistique, 328:3-20.
Layard, R., S., Nickell, and R. Jackman, 1991, Unemployment, Oxford, Oxford University Press.
Pisani-Ferry, J., 2000, Plein Emploi, Conseil d’Analyse Economique, Paris.
See IMF Country Report No 01/198: Labor Market Developments and Wage Moderation in France in the 1990s.
Prepared by Enrica Detragiache and Marcello Estevão.
See, for instance, Pisani-Ferry (2000) and Decressin and others (2001). As discussed in the latter paper, in the second half of the 1990s, the unemployment rate fell sharply and wage growth was weak also in Spain and, to a lesser extent, in Italy. The Netherlands and the United Kingdom posted very large declines in unemployment beginning in the 1980s.
Comparing labor productivity growth during expansions controls for business cycle effects. The end of the 1980s was the last fime France posted strong sustained quarterly output growth. See for instance, Pisani-Ferry (2000), page 29. The expansions were dated using two consecutive declines in the quarterly unemployment rate as their beginning and two consecutive increases as their end. Slight modifications of this criterion change the exact dating of these expansions but do not qualitatively alter the conclusions derived from Table I.1
Using microeconomic data from the French labor force survey (Ênquete Emploi) average weekly hours of work are estimated to have declined fiom 40.4 hours in 1991 to 38.5 hours in 2000, notwithstanding the economic expansion in the second half of the 1990s.
These laws provided financial incentives to firms adopting the 35-hour workweek voluntarily (laws Robien, June 1996 and Aubry I, June 1998) and established deadlines for the compulsory introduction of this shorter work time (laws Aubry I and Aubry II, January 2000.) Firms employing more than 20 employees were obliged to adopt the 35-hour workweek in February 2000 while smaller firms were obliged to do so in January 2002. The law Aubry II, which regulates the compulsory introduction of the shorter workweek, also reduces firms' social security contributions to partly offset the expected increase in labor costs.
According to a survey conducted by the Banque de France, capital operating time has been trending up in the second half of the 1990s. Because capital stock data does not take into account movements in factor utilization, the TFP growth series, which is obtained by residual, is biased upwards. Therefore, it is possible that there have been more capital deepening and less TFP growth than shown in Table I.2
As discussed in Estevão and Nargis (2002), part of the substantial real wage increase between 1990 and 1993 was actually a figment of composition effects during the recession years: less-skilled workers (and, thus, low-wage earners) were the first to be fired and that raised average wages.
The “short-run” NAIRU depends on the elasticity of the short-run labor demand with respect to variations in labor costs. Kramarz and Philippon (2000) estimated this elasticity at about -1.5 for individuals earning the minimum wages. As summarized in Hamermesh (1993), the short-run elasticity of demand for lower-skill (and, therefore, lower-paid) workers tends to be larger than for higher-skill individuals, leading to an average elasticity of demand for labor in France below -1.5. Using an elasticity of-1, assuming a given labor force size, and setting the NAIRU in the mid-1990s at around 12 percent generates the 8½ percent figure.
The logic is simple: given a lower (higher) user cost of capital, the profit rate needs to decline (increase) to keep the zero profit condition underlying the long-run relationship (4). This only happens if wages in efficiency units increase (decline).
Their results could be used to partially reconcile an outward shift in labor demand with observed wage moderation. They show that about half of the jobs created or maintained belonged to workers earning about the minimum wage, even though these low-paid workers accounted for only about 15 percent of the workforce in 1994 (our own calculations using the Ênquete Emploi). By 2000, the share of workers earning the minimum wage or less in France grew to 20 percent. Estevão and Nargis (2002) account for this composition effect and still find a large downward shift in the wage-setting relationship.
