It's no secret that Europeans work less than Americans do. Every Labor Day the media tell us that Europeans have just enjoyed weeks of summer vacation while Americans have been toiling away. These stories often depict Americans as hard-working drones who revere material possessions above all else. Europeans meanwhile bask in the good life of long lunches and months at the beach.
There is some truth to the portrayal, at least in terms of hours worked. The International Labor Organization reports that the average American worked 1,815 hours in 2002, well above the comparable figures for France (1,545) and Germany (1,444), for example. (The average South Korean, on the other hand, worked over 2,400 hours.)
But if it's widely acknowledged that Americans work more hours than Europeans, it remains a puzzle quite why there's such a large difference. With similar economies and social structures—at least relative to the rest of the world—it would seem that labor patterns should also be alike. Social scientists have been hard-pressed to explain the disparity.
Most accounts focus on cultural explanations. The most popular is the notion that Europeans have a fuller appreciation of la dolce vita—the sweet life—the Italian version of the idea that life is to be enjoyed, not endured. Work is a means to an end, not an end in itself.
The idea of cultural and religious influences on economic activity isn't new. German sociologist Max Weber wrote The Protestant Ethic and the Spirit of Capitalism nearly a century ago, attributing the rise of capitalist economies to the "Protestant work ethic." It was an immensely persuasive theory in its time, and derivative explanations have held great sway ever since.
"Why do Europeans and Americans differ so much in their attitude toward work and leisure? I can think of two reasons," opines a recent Time magazine essay. "Broadly speaking, Americans value stuff—SUVs, 7,000-sq.-ft. houses—more than they value time, while for Europeans it's the opposite. Second, ... in the puritanical version of Christianity that has always appealed to Americans, religion comes packaged with the stern message that hard work is good for the soul. Modern Europe has avoided so melancholy a lesson."
"It all comes down to what people feel is important and how they feel about their lives," argues a September 2003 U.S. News & World Report editorial. "We value more money and more stuff; they value more leisure time. ... We are proud of being busy—it is a virtue; being idle is perceived as a vice."
Economic explanations
Economists have always been suspect of such cultural explanations. Standard economic theory assumes that people's preferences are, on average, homogeneous, and that choices depend largely on economic factors. Still, while economists agree that dollars and cents lie behind the work pattern differential, there is little harmony among them as to the right economic explanation.
Some economists say work regulations keep Europeans from working longer hours and point favorably to recent European reforms on vacation time. Others argue that greater inequality in the United States motivates workers to try harder to get ahead. Most of these explanations come from the perspective of labor economics and its core belief that social structures and institutions such as unions are the major determinants of labor patterns.
But in a recent series of papers and lectures, Edward C. Prescott, senior monetary adviser to the Minneapolis Fed and economist at Arizona State University, looks at the labor supply question through the prism of the growth model—a different perspective altogether—and provides a convincing and remarkably straightforward explanation for the dramatic differences in hours worked. It is an explanation that has far-reaching implications for policymakers—and for anyone else who's ever received a paycheck.
According to Prescott, the reason for these large differences in labor supply is not culture. "French, Japanese, and U.S. workers all have similar preferences," he writes. "The French are not better at enjoying leisure. The Japanese are not compulsive savers." The reason for the wide range in working hours is, in a word, taxes. Europeans supply less labor because there's a much larger wedge in most European countries between what a worker is paid and what that worker actually gets to keep after taxes are taken out. This tax wedge, argues Prescott, distorts the trade-off people make between consumption and leisure by making consumption more expensive. And since people work, ultimately, to earn money to pay for consumption goods, they'll supply less labor if consumption goods become relatively more expensive. The cheaper alternative: leisure. Hello, Riviera.
If the concept seems straightforward, its evolution was anything but. Like most ideas that seem obvious in retrospect, the awareness that taxes distort labor markets dramatically and account for major international differences in work patterns came about indirectly and as a revelation to those who happened upon it.
The discovery
Prescott's discovery about the role of taxes in labor supply variation began, simply enough, in his classroom at the University of Minnesota, where he taught from 1980 to 2003. "I was making up exercises for my students," he recalled in a recent interview. "I said, 'use this nice little growth model.'"
The "nice little model" he presented to his students is the workhorse of modern macroeconomics; it says, mathematically, that a nation's total output (or gross domestic product, GDP) is dependent on three sources: labor, capital and the efficiency (or productivity) with which it merges them to create economic value. The other key part of this standard theory is, in the jargon of economics, a utility function: a formula representing the notion that households try to maximize their happiness by finding the best possible combination of leisure and consumption, given their resources.
Prescott wanted his students to become familiar with this model by looking at how it performed in different nations over time, and how key variables—capital endowments, productivity, labor supply—could account for differences among nations in per capita GDP. "I wanted to try to get across the basic ideas and the importance of productivity," said Prescott. "And then I thought, let's put a few taxes in."
The intuition was far more significant than Prescott suspected, but that became clear only after looking at the relative contributions of capital, productivity and labor. The data, compiled by the United Nations and the Organization for Economic Cooperation and Development, showed that in the mid-1990s among developed countries—the United States, much of Europe and Japan—relative levels of capital differ little and explain just a small portion of the variation in per capita GDP (see adjacent table). "The capital factor is not an important factor in accounting for differences in incomes across the OECD countries," writes Prescott in his 2002 Richard T. Ely Lecture to the American Economic Association. "[It] contributes at most 8 percent to the differences in income between any of these countries."
Capital, Labor, Productivity
and GDP |
||||
Country |
Capital/ |
Hours |
Productivity: |
GDP |
Germany |
2.7 |
19.3 |
99 |
74 |
France |
2.2 |
17.5 |
110 |
74 |
Italy |
2.6 |
16.5 |
90 |
57 |
Canada |
na |
22.9 |
89 |
79 |
United Kingdom |
2.6 |
22.8 |
76 |
67 |
Japan |
2.5 |
27.0 |
74 |
78 |
United States |
2.3 |
25.9 |
100 |
100 |
Source: Federal Reserve Bank of Minneapolis Research Department Staff Report 321 and Working Paper 618. |
Productivity, on the other hand, is very important, at least for some national differences. Japan and the United States, for example, have similar levels of labor and capital, but per capita GDP in Japan is far below that in the United States because its productivity is less than three-quarters that of the United States.
But what of European countries like France, Italy and Germany? Why are their levels of per capita GDP so much lower? All these nations have capital endowments comparable to the United States. Their productivity levels also are similar to U.S. rates, or in the case of France, even higher. The data suggest that the differences in wealth are due almost exclusively to the markedly lower number of hours worked in these European countries. Germany, for instance, had a slightly higher capital endowment than the United States and an equal level of productivity, but just 74 percent of the U.S. per capita GDP. The evident reason: Its workers supplied just over 19 hours of labor per week compared to nearly 26 hours a week per American worker.
While many believe that cultural differences lead to fewer hours worked in Europe than in the United States, Prescott doubts it. After all, data from the early 1970s show that the French actually worked more hours per week than did Americans at that time. Has French culture changed radically over the last two decades? Probably not: They still like good wine, aged cheese and, inexplicably, Jerry Lewis. Prescott's hunch was that differences in marginal tax rates might explain the differences in labor supplied and thus account for differences in per capita GDP.
Enter the tax wedge
"What is important is the price of consumption relative to leisure," Prescott writes in the lecture he gave in April 2003 as he accepted Northwestern University's prestigious Erwin Plein Nemmers Prize in Economics. "And it is determined by the consumption tax rate and the labor income tax rate." (See the lecture, "Why Do Americans Work So Much More Than Europeans?")
By introducing these taxes into the growth model, and making standard microeconomic assumptions, Prescott derived what he calls "the key equilibrium relation."1 It's a mathematical formula for labor supply that says workers will supply labor dependent on, among other things, their preference for consumption now over consumption later (spend or save?), their preference for leisure relative to consumption (play or work?) and the effective tax rate. Holding the first two variables fixed and looking empirically at different national tax rates enables Prescott to see if tax differences can account, fully or partially, for variations in labor hours supplied.
Estimating the effective tax rates in these countries was, in itself, a major accounting exercise. Consumption taxes include value-added taxes, sales taxes, excise taxes and property taxes. Labor is subject to both income taxes and Social Security taxes. For each nation under consideration, Prescott and his students crunched the numbers, determined a tax rate, plugged it into the formula along with fixed estimates of the other variables, and derived predictions of labor hours supplied per week per worker.
How good were the predictions? Dead-on for Germany and the United Kingdom, a bit low for Canada and the United States, and a bit high for the other countries (see table below). Given measurement inaccuracies, the rough nature of the tax-rate estimates and the difficulty of international comparisons, writes Prescott, the model's predictions were "surprisingly close to the actual."
Tax Rates and Labor
Supply |
||||
Country |
Tax Rate |
Actual |
Predicted |
Difference |
Germany |
59 |
19.3 |
19.5 |
0.2 |
France |
59 |
17.5 |
19.5 |
2.0 |
Italy |
64 |
16.5 |
18.8 |
2.3 |
Canada |
52 |
22.9 |
21.3 |
-1.6 |
United Kingdom |
44 |
22.8 |
22.8 |
0.0 |
Japan |
37 |
27.0 |
29.0 |
2.0 |
United States |
40 |
25.9 |
24.6 |
-1.3 |
Source: “Why Do Americans Work So Much More Than Europeans?” Federal Reserve Bank of Minneapolis Research Department Staff Report 321. |
Here, notes Prescott, "the important observation is that the low labor supplies in Germany, France and Italy are due to high tax rates. In these countries if someone works more and produces 100 additional euros of output, that individual gets to consume only 40 euros of additional consumption and pays directly or indirectly 60 euros in taxes." Put in such stark terms, it seems obvious that many Europeans might opt to work less, while Americans and Japanese, taxed more lightly, would be keen to put in extra hours.
Confirmation and implications
Prescott found further confirmation for his hypothesis when he looked at tax rates and labor supply in the early 1970s (see table below). While his model's predictions of labor hours supplied diverge from the actual in several cases—Italy and Japan, in particular—Prescott observes that "when European and U.S. tax rates were comparable, European and U.S. labor supplies were roughly equal."
Tax Rates and Labor
Supply |
||||
Country |
Tax Rate |
Actual |
Predicted |
Difference |
Germany |
52 |
24.6 |
24.6 |
0.0 |
France |
49 |
24.4 |
25.4 |
1.0 |
Italy |
41 |
19.2 |
28.3 |
9.1 |
Canada |
44 |
22.2 |
25.6 |
3.4 |
United Kingdom |
45 |
25.9 |
24.0 |
-1.9 |
Japan |
25 |
29.8 |
35.8 |
6.0 |
United States |
40 |
23.5 |
26.4 |
2.9 |
Source: "Why Do Americans Work So Much More Than Europeans?" Federal Reserve Bank of Minneapolis Research Department Staff Report 321. |
As for the outliers, Italy and Japan, Prescott suggests that other factors may be significant. In Italy, cartels may have played a role in depressing labor supply below its predicted value. In Japan, significant measurement errors in actual hours worked could account for the overly high prediction by the model.
And what seems another anomaly is very likely an indirect confirmation of the importance of marginal tax rates on labor supply, according to Prescott. In the United States, actual hours worked per person increased by 10 percent from the 1970s to the 1990s, though the marginal tax rate remained at 40 percent. Prescott argues that U.S. tax reforms in the 1980s changed the effective marginal tax faced by married couples—dropping the rate in half for the second earner's income—even though it remained nominally at 40 percent.
"In the 1993-96 [period]," he writes, "the marginal income tax on the labor income associated with switching between a one-earner and a two-earner household is only 20 percent, not 40 percent." The issue warrants more attention, he says, and indeed, his colleagues Larry Jones, Rodolfo Manuelli and Ellen McGrattan have recently released a paper on this exact question. (See "Wives at Work.")
On the whole, Prescott states, the results show that "people are remarkably similar across countries" and not only for these relatively prosperous and homogeneous nations, but for Chile, Mexico and Argentina, as well, where other economists have found similar relationships. "Apparently, idiosyncratic preference differences average out and result in the [representative] household having almost identical preferences across countries."
The policy implications are enormous for high-tax countries. If France were to lower its effective tax rate from 60 percent to 40 percent, estimates Prescott, its people would work more (taking 6.6 percent less leisure) and—remember their high productivity?—would generate considerably more output. Tax revenues wouldn't diminish, because the 40 percent rate would be levied on a higher base. And overall French "welfare gains," as economists put it, would increase nearly 20 percent. In the United States, reducing marginal tax rates would have a more modest impact, according to the model: A 10 percent rate reduction would produce a 7 percent welfare gain. But even in the United States, Prescott's findings have huge implications for the viability of the Social Security system. (See "Shrinking a deadweight loss.")
Foreign affairs
In recent months, Prescott has traveled widely, presenting his findings not only to American audiences but to economists and policymakers in London, Berlin, Toulouse, Tokyo and elsewhere overseas. And in fact, says Prescott, Europeans tend to be more receptive than Americans. "The economists there understand that there is a problem," he said after returning from France in mid-September. "I got some excellent suggestions when I presented the paper, the best so far." But at all venues, he observes, the common denominator is surprise.
Prescott is the first to admit that he, too, thought the results were startling, unexpected. "I find it remarkable that virtually all of the large difference in labor supply between France and the United States is due to differences in tax systems," he writes in his Ely lecture. "I expected institutional constraints on the operation of labor markets and the nature of the unemployment benefit system to be more important."
Moreover, he concedes that cultural explanations might carry the day in a few settings. "Scandinavians seem to be a little bit different," he said recently, referring to research by Richard Rogerson, an economist at Arizona State University. "My theory is when one of those Swedes looks at you when you're not working, it's pretty intimidating." More seriously, he allows that in small, homogeneous cultures, social pressures can be quite strong.
But even in large, heterogeneous nations, tax wedges don't always tell the whole story, according to Prescott. "Taxes are not the only reason that the labor factors differ," says Prescott's Ely lecture.
Unemployment benefits and housing subsidies—not taxes—distorted labor mobility in the United Kingdom between the first and second World Wars, contributing significantly to that country's interwar depression. New Deal policies supporting cartels in America's heavy industries distorted wages and employment in the last half of the 1930s, contributing to the depth and duration of the Great Depression in the United States. Similarly, cartels in 1970s Italy may have suppressed employment there. Prescott relies on work by University of California, Los Angeles economists Harold Cole and Lee Ohanian in making these conjectures.
Still, while taxes aren't the all-powerful explanatory factor for all nations and eras, Prescott contends that in major developed countries in the time period under consideration, the labor supply impact of tax wedges is a powerful and undeniable fact.
Other academics
As befits the work of any prominent scholar, Prescott's theory has attracted close academic scrutiny—beyond the initial reaction of surprise—from both adherents and critics. In one recent paper, Peter Lindert, an economist at the University of California, Davis, refers to Prescott's study as dependent upon "a theoretical model heavily laden with assumptions. It is educated, intelligent, plausible fiction—but fiction nonetheless."
On the other hand (as Lindert points out) Prescott's model and findings are cited quite favorably by Nobel Laureate Robert Lucas in his 2003 presidential address to the American Economic Association.
Lindert calls for empirical tests. Steven Davis at the University of Chicago Graduate School of Business and Magnus Henrekson of the Stockholm School of Economics oblige with a careful econometric analysis of the impact of labor income and consumption taxes on employment and work activity. In their study of rich countries in the mid-1990s, they find that a 12.8 percentage point difference in tax rates is associated with 122 fewer market work hours per adult per year and nearly a 5 percentage point decrease in employment—population ratios—an indirect affirmation of Prescott's theory.
A very different perspective was presented earlier this year in a series of lectures by British economist Richard Layard, co-director of the Centre for Economic Performance at the London School of Economics. Layard takes issue with GDP itself as a satisfactory measure of human welfare—or utility, as Jeremy Bentham and subsequent economists have termed it—noting that "happiness has not increased, despite huge increases in living standards."
To summarize a lengthy argument, Layard's idea is that a tax wedge on labor income could actually increase utility by decreasing a sort of pollution: overwork brought on by the inherent human desire to do better than our peers, regardless of our absolute level of income. Keeping up with the Joneses, in other words, leads to overwork, ill health and unhappiness—rivalry distorts the leisure/labor decision. Appropriate public policy should diminish this pollution by taxing it. "In an efficient economy," Layard writes, "there will be substantial levels of corrective taxation ... 60 percent would not seem inappropriate, and that is in fact the typical level of marginal taxation in Europe—if you allow for direct and indirect taxes."
Prescott responds
Prescott's reactions to these ideas vary widely. Sitting in his seventh-floor office at the Minneapolis Fed, he reads through the first pages of Lindert's paper, then drops it on his desk. "It doesn't seem to be coherent," he says.
Davis and Henrekson's study, on the other hand, intrigues him. That might seem predictable given its broad support of Prescott's findings, but Davis and Henrekson employ a technique Prescott generally scorns: statistical regression. "Progress, don't regress," he says with a smile, quoting the slogan featured prominently on his Internet home page. Regardless of their method, Prescott is drawn to the findings and has invited Davis to Minneapolis to get a closer look at their work.
But Prescott's response to Layard's argument—more complete and nuanced—conveys a sense of Prescott himself. He begins by summarizing Layard's case in a phrase: "I'm happy if I have a lot more income—than you," he says, grinning and quite aware that he does. As to the overwork such rivalry might cause, "that just says there's a consumption externality."
Then he conveys the concept with a story. "I always tried to create a positive externality in Pittsburgh for my neighbors who had these beautiful lawns," he jokes of his grad school days at Carnegie Mellon University. "By my having a messy lawn, their lawns looked so much better. I mowed it, but I didn't do much else with my lawn. And it gave me utility to see them happier." He tells the story with a verbal wink, acknowledging silently that his Pittsburgh yard care externality may well have been less than zero.
The conspiratorial smile changes to professorial zeal as he begins to dissect Layard's reasoning: "Suppose everybody cares about relative consumption as well as own consumption. You work out the equilibrium, it's not Pareto optimal. Let's deal with the case where everybody enters symmetrically. So it's simple to make the ordering. Well, you can make everybody better off by just putting a tax on consumption so that they work less. That's a very standard model. Now what would be the empirical evidence for and against that?"
In under five minutes, Prescott has crystallized an argument, communicated it to a visitor in plain language and personal anecdote, then converted it to the idiom of economics and laid out steps for its confirmation or refutation. It's vintage Prescott: analytically brilliant, unexpectedly funny and several beats ahead of everyone else. That last bit is the essence of a conversation with the economist. When you ask him a question, it sometimes seems that his reply is off-topic; then it dawns on you that Ed Prescott is answering the question you should have asked.
A pattern of surprise
Prescott's willingness to entertain alternatives, to listen to critics, to incorporate the unexpected is deeply characteristic of his work. That flexibility is, in fact, the paradoxical outcome of a rigid research discipline. In setting model parameters, for instance, or reporting research results, "the investigator has no degrees of freedom," he says. "You have to tie your hands and if there's a deviation from your predictions, you report it. You can speculate on why, but you've got to be totally honest."
Intellectual honesty also means allowing findings to modify, even subvert initial hypotheses. It happens frequently, says Prescott. Much of the work for which he's best known—theories on time inconsistency, real business cycles, the equity premium and growth theory—has been developed in an ongoing process of research and revelation.
"When I work out the implications, I'm quite often surprised: The findings change my views quite dramatically," he says. "When I did the real business cycles work with Finn Kydland, I was certain that monetary shocks were the reason the economy fluctuated with the business cycles. Our findings were just the opposite. When I did some work with Rajnish Mehra on the equity premium puzzle, I was certain that the reason for the high historical difference in the return on equity relative to debt was just a premium for bearing aggregate, nondiversifiable risk. We found it wasn't." For time inconsistency and the impact of taxes on labor supply, as well, surprise has been an intrinsic part of the process.
Future direction
As striking as his labor supply findings are—and though many aspects of it remain unresolved—Prescott senses that the big theoretical questions in economic growth lie elsewhere, and he is now turning his attention to them. "I think I've had my say on labor supply," he concludes.
In his Ely lecture, he lays out three sources of economic growth: capital, labor and productivity. The first two are important in understanding why some nations remain poor while others prosper, but the central question, contends Prescott, is what determines productivity? "Given productivity, our macro models are great," he says. "But we treat it as exogenous. We've got to have a better understanding of mapping between policies and productivity."
In other words, what can governments do to enhance productivity? Prescott's main candidates are efficient financial markets, competition among producers and trading clubs. And currently, the last is his major focus.
"What is a trading club?" he asks rhetorically. "Well, first, free movement of goods between the member states. But it's much, much more than that. ..."
Prescott continues at length, with a discourse ranging from Toyota factories in Wales to trade among the U.S. states in the 19th century. He speaks quickly, and as he does there is a sense that each research question he asks leads him to a dozen more, each more interesting than the last.
He will travel soon to Warsaw and then Bogotá to explore these ideas with other economists and policymakers. "It's going to be fascinating to see what's happening in Poland," he remarks. In Colombia, "the president is trying to do some good things there, and we have to go down and help out."
He's not a policymaker himself. "I leave that to other people," he says. "I'm no good at it. My comparative advantage is working out implications of theory." And in so doing, it seems there is just one constraint: Even for Ed Prescott, a scholar who understands labor supply dynamics as well as anyone on earth, there are only 24 hours in a day. "Time," observes the economist, "is the most valuable resource."
1 The two assumptions: (1) that people decide between leisure and consumption based on their relative prices, at the margin, and (2) that in a competitive market, wages are equal to their marginal product of labor. The "key equilibrium relation" also depends on the share of a nation's output due to capital.