Edward C. Prescott: Why Do Americans Work so Much More Than Europeans? Federal Reserve Bank of Minneapolis Quarterly Review Vol. 28, No.1, July 2004, pp. 2-13 Steinar Holden, August 2005 1
Output, Labor Supply, and Productivity In selected countries in 1993-96 and 1970-74 Relative to United States (U.S. = 100) Period Country Output per person Hours worked per person Output per hour worked 1993-1996 Germany 74 75 99 France 74 68 110 Italy 57 64 90 Canada 79 88 89 United Kingdom 67 88 76 Japan 78 104 74 United States 100 100 100 1970-74 Germany 75 105 72 France 77 105 74 Italy 53 82 65 Canada 86 94 91 United Kingdom 68 110 62 Japan 62 127 49 United States 100 100 100 2
Key stylized facts: Comparing G-7 countries: Output per person in Europe is approximately 70% of output per person in the US. In the early 1970s, the lower output was due to lower productivity. In the 1990s, the lower output was due to lower labor supply (fewer hours per person). How can this be explained? 3
Theoretical framework: Utility of representative household given by Cobb-Douglas specification over consumption and leisure Calibrated real business cycle model, where household sets current consumption and current labor supply to maximize utility Transfers to households treated as lump sum, public expenditure, except military, treated as perfect substitutes to private consumption All time outside market work is treated as leisure 4
Optimal labor supply is given by where hours = labor share c α labor share + y 1 τ c/y is the current consumption ratio, α a parameter measuring the value of leisure relative to consumption (set to 1,54), τ is the effective tax rate on labor relative to leisure labor share is given by technology, and set to 0.68 The effect of c/y represents intertemporal concerns; lower expected future taxes will lead to higher consumption and lower labor supply 5
Empirical estimation: Using OECD/SNA data, Prescott estimates the actual tax-rates on labor and savings. Marginal taxes are set to 1.6 times average taxes, based on US evidence. Using data for output and consumption, Prescott predicts the representative household s labor supply in each country / time period. The results can be summarized in the following table: 6
Actual and predicted labor supply In selected countries in 1993-96 and 1970-74 Labor supply Period Country Actual Predicted Differences (predicted less actual) Prediction factors Tax rate Cons./Output 1993-96 Germany 19,3 19,5 0,2 0,59 0,74 France 17,5 19,5 2,0 0,59 0,74 Italy 16,5 18,8 2,3 0,64 0,69 Canada 22,9 21,3-1,6 0,52 0,77 UK 22,8 22,8 0 0,44 0,83 Japan 27,0 29,0 2,0 0,37 0,68 US 25,9 24,6-1,3 0,40 0,81 1970-74 Germany 24,6 24,6 0 0,52 0,66 France 24,4 25,4 1,0 0,49 0,66 Italy 19,2 28,3 9,1 0,41 0,66 Canada 22,2 25,6 3,4 0,44 0,72 UK 25,9 24,0-1,9 0,45 0,77 Japan 29,8 35,8 6,0 0,25 0,60 US 23,5 26,4 2,9 0,40 0,74 Labor supply is measured in hours worked per person aged 15-64 per week 7
Labor supply and inverse tax rates 1993-1996, deviation from sample mean 20.0 15.0 10.0 5.0 0.0-5.0 Labor supply Tax rate -10.0-15.0 Germany France Italy Canada United Kingdom Japan United States Figure 1: Constructed from the table above. 8
Labor supply and inverse tax rates 1970-1974, deviation from sample mean 20.0 15.0 10.0 5.0 0.0-5.0-10.0 Labor supply Tax rate -15.0 Germany France Italy Canada United Kingdom Japan United States Figure 2: Constructed from the table above. 9
Key findings: Prescott concludes that virtually all large differences between the U.S. labor supply and those of Germany and France are due to differences in tax systems. The relation between taxes and labor supply seems stronger in the 1990s than in the early 1970s. U.S. labor supply increased from 1970s to 1990s, in contrast to model predictions. (where the high consumption ratio in the 1990s should reflect low expected future taxes, leading to low current labor supply) Prescott explains the increase with a change in the tax system, reducing taxes on two-income households, which is not reflected in the estimate for the marginal tax rate 10
Alberto Alesina, Edward Glaeser and Bruce Sacerdote: Work and leisure in the US and Europe: Why so different? NBER Working Paper series. Working paper 11278 11
Alesina et.al argue that European labor market regulations, advocated by unions in declining European industries who argued work less work all explain the bulk of the difference between the US and Europe. These policies do not seem to have increased employment, but they may have had a more society-wide influence on leisure patterns because of a social multiplier where the returns to leisure increase as more people are taking longer vacations. 12
Critique of Prescott Prescott uses Cobb-Douglas utility function which forces the elasticity of labor to about.77, i.e. one percent higher wages increase labor supply by.77 percent Is this reasonable? Higher taxes on labor income has substitution effect, which will make people work less income effect, which will make people work more By treating public expenditure as a perfect substitute to private consumption, the income effect of higher taxes is removed. o Not clear that this is appropriate, as public consumption is on other products 13
Micro-estimates of uncompensated labor supply elasticities (which includes both income and substitution effects) are often close to zero Micro-estimates of compensated labor supply elasticities for men and single women are also much lower than Prescott s Reasonable elasticity estimates can at best explain one-half of the difference in hours worked between the U.S. and Europe Micro-estimates of compensated labor supply elasticities for married women are closer to Prescott s estimate Tax rate differences can potentially explain the difference in hours worked for married women 14
Cross-country evidence on labor supply and taxes Cross-country evidence suggests much stronger negative link between labor supply and taxes than micro-evidence does. However, cross-country evidence not really credible because of o difference from micro evidence o poor times series evidence Why is cross-country evidence misleading? Omitted variables (unions and labor market institutions) o Strong positive correlation between unions and taxes, and negative correlation between unions and labor supply Social multipliers (i.e. the utility from leisure time is higher if others do the same) 15
Unionization, high taxes and low labor supply Strong evidence that unions have been a strong advocate for the reduction in hours in many European countries o Hunt (1998, 1999) documents how German and French unions pursued a policy of work sharing, with slogans like work less work all. Unions also correlated with leftist government and higher taxes By compressing wages, unions reduce the gain from working hard, thus possibly reducing labor supply. 16
Why did unions in Europe fight for lower work hours? Unions care about size, when faced with sectoral shocks unions that are interested in maintaining their membership will cut hours worked. o Size is important for the union s political power. Higher taxes may have led union members to demand lower working hours (argument more in line with Prescott, 2004). Unions helped coordinate a demand for lower hours due to an income effect (argument in line with Blanchard, 2004). The union simply responded to the increasing income level of their members who demanded more leisure. The 1970s and 1980s were a period with high inflation and high unemployment in Europe. This made demands for increased real wages politically difficult/unacceptable. Asking for higher hourly wages combined with reduced working hours to share the jobs were however a powerful rhetoric. 17
The social multiplier Plausible to assume that the utility of vacation and leisure is higher if other people (friends and family) also have vacation or leisure time Also complementarities of work, better to work when others also work If not, why do so many people have Saturday and Sunday as days off? o in spite of commuting problems, and idle capital in weekend Macro-elasticities might capture such social multipliers and thus be relevant However, Alesina et al still conclude that elasticities are too small to explain the differences in hours worked 18
Conclusions (Alesina et al) Europeans work less than American due to policies of unions in the 1970s 1990s and because of labor market regulations Should lower hours be viewed as distortion, or just as an effect of welfare improving coordination? Happiness is strongly correlated with length of vacation for individuals But this may reflect omitted variables and reverse causality Places with more mandated vacation do seem a bit happier But the gap between this finding and any sort of policy recommendation remains vast 19
Some reflections and speculations Rat race negative external effects in working time o Promotions depend on you doing a better job than your colleagues o Your firm s success depends on better performance than the competitors Firm level unions can avoid negative externalities at the work-place, while industry unions or national unions can mitigate negative externalities at more aggregate levels Seem to be stronger pattern bargaining effects for working hours than for wages, where hours reduction for some groups of workers lead to similar reductions for all o This does not work in the same way under individual wage setting Unions dominated by workers with lower education, which have stronger preference for leisure than workers with higher education? 20
Over time, productivity growth must lead to higher real income or fewer hours Firms will favor higher income, to attract workers who want to work o Biased relative to workers preferences?????????? Unions will be closer to workers true preferences??? 21