Behind the Numbers: Capital's Gain

The income squeeze on the middle class is the big economic story of this decade, but record-setting stock prices and soaring executive pay remind us that not everyone is experiencing a squeeze. The stock market boom and the executive windfalls, in turn, reflect growth in corporate profits. And the contrast between spectacular profit trends and the disappointing wage growth could not be more dramatic.

This disparity reopens the rather impolite question of whether we are observing a significant redistribution of income from labor to capital. Most economists have insisted that we are not. But the latest data strongly show that a nontrivial redistribution of what economists call "factor incomes" has indeed taken place in the 1990s, with the foundation having been laid in the 1980s.

The primary evidence of this shift is that the current rate of return to capital—profits and interest income per dollar of assets—is very high by historical standards and is high relative to similar returns in other advanced countries. If profit rates in the 1980s and 1990s had achieved only their long-term historic norm, then hourly worker compensation would be 3.6 percent higher than it is. A comparable estimate can be obtained by examining changes in capital's share of income. A 3.6 percent wage loss might not sound like much, until you realize that the entire growth in hourly compensation over the 1989-96 period has been just 2.8 percent. In other words, had this shift in income from labor to capital not occurred, recent wage growth would have been more than double its actual rate.

Yet many analysts continue to blandly insist that no such trend is taking place. For instance, in a review of William Greider's new book, One World: Ready or Not, MIT economist Paul Krugman claims that the book

evidently involved an immense amount of hard work, yet over and over again gets easy things—facts that could have been checked in a few minutes in any good library—utterly wrong. At one point, for example, Greider makes the astonishing assertion that international comparisons of labor's share of the returns from production are difficult, and then quotes some misleading data that seem to show a massive redistribution from labor to capital in the United States and other advanced countries. I say astonishing, both because such information is very easy to find, and because anyone even slightly familiar with U.S. data knows that the share of wages and benefits in national income has been remarkably steady (at about 73 percent) for the past generation.

But Krugman should have persisted in that library. If labor's share of national income is the yardstick, as Krugman suggests, then it rose from 68.5 percent in 1959 (the earliest available data) to 73.2 percent in 1979 and has been relatively stable thereafter (at 72.9 percent in 1994) according to the Commerce Department's Bureau of Economic Analysis. However, documents in our library from the Organization for Economic Cooperation and Development (OECD) show that, among the advanced countries as a whole, labor's share of income (in the business sector) has fallen from 68 percent in the 1970s to 65.1 percent in 1995. In the European Union, labor's share fell from 69.2 percent in the 1970s to just 62 percent in 1995.

There is an even stronger case that labor's share of income changed dramatically if one focuses on the manufacturing sector, which bears the most impact of trade and globalization. According to the Bureau of Labor Statistics (in unpublished data not found in a library—you have to call), labor's share of income in the manufacturing sector fell from 74.8 percent to 70.6 percent over the 1979-89 period, reversing the rise in labor's share that occurred over the 1950s, 1960s, and 1970s. Oxford University professor Andrew Glyn's research shows a similar decline in labor's share in manufacturing income over the 1980s in Europe and in the advanced countries as a whole (down from 73.4 percent to 67.3 percent). So Greider is definitely on to something.

What Krugman glibly dismisses, others at MIT see as a subject worth investigating. For instance, Krugman's colleague Olivier Blanchard delivered a paper at the January 1997 meeting of the American Economic Association that used OECD data and concluded that it is a "basic fact" that profit shares and profit rates in advanced countries have increased substantially in the last 10 to 15 years:

We know that real interest rates on bonds have increased substantially since the 1970s. . . . It is generally believed that weaker unions, together with higher unemployment in most OECD countries, may have led workers to accept lower wages. Thus, a higher cost of capital [that is, high real interest rates] and a lower cost of labor can plausibly explain a shift of firms away from capital and towards labor, and thus an increase in the profit rate.

Subscribe to The American Prospect

Blanchard's preliminary conclusion is that low inflation is causing higher profit rates, which, in turn, are shifting income shares from labor to capital.

Another of Krugman's MIT colleagues, James Poterba, organized the January 1997 session at the American Economic Association meeting and gave a paper of his own, "Recent Trends in Corporate Profitability." The purpose of the paper is to explain why "the mid-1990s have been marked by higher corporate profitability than any period in the last twenty-five years." Poterba's "very preliminary" results are that current high rates of profit reflect more than a cyclical recovery and may, therefore, be expected to persist. He also attributes a fall in capital per worker as the major explanation for higher profit rates.

Moreover, trends in national income do not fully capture shifting income shares. One measurement problem is how to count the growth of government over time. Since our national statisticians assign no profit to the capital employed in government, the public sector is deemed to generate only labor income. Labor's share of national income, therefore, automatically expands as government expands. This factor helps explain the rise in labor's share over the 1959-79 period and should make curious minds wonder why labor's share did not rise further after 1979. After all, the growth in the government/nonprofit sector was from 18.3 percent of national income in 1979 to 19.4 in 1989 and 19.7 percent in 1995, which, all things being equal, would have expanded labor's share of national income by nearly 1.4 percentage points. Another problem is accurately determining the capital and labor shares among "sole proprietors and partnerships."

One can avoid such technical problems by examining trends in the corporate private sector. Using information readily available in the National Income and Product Accounts, one finds that labor's share of corporate-sector income rose from 79.2 percent in 1959 to 83.9 percent in 1979 and then declined to 82.4 percent in 1989 and to 81.6 percent in 1995. The latest data show a drop in labor's share to 81 percent in 1996. As noted earlier, this shift in income is equivalent to a 3.6 percent reduction in hourly compensation.

An examination of labor and capital income shares, however, does not tell the whole story of income distribution. For example, several trends suggest that, "other things equal," capital's share might have been expected to decline and labor's share to rise. One such trend is a rapid growth in education levels and "labor quality" that ought to raise labor's share. Another is the rapid decline in the capital-output ratio since the early 1980s. That is, the capital stock has not grown as quickly as total economic output, which would suggest that capital's shares, if anything, should fall.



My colleague Dean Baker has computed a historical series on the return to capital using a methodology employed by Martin Feldstein and Lawrence Summers in an article 20 years ago. There has been a rapid growth in the return to capital, before and after taxes, starting in the late 1980s and continuing steadily through 1995. Before-tax capital income has grown to its highest levels since the mid-1960s, while the after-tax return on capital is as high or higher than in any year since 1959 (the earliest year for which a measure is available). Poterba's measurements are similar and show the same trends. This growth in the return to capital has allowed capital's share of income to rise despite the decline in the capital-output ratio.

This large increase in profitability does not correspond to the growth of productivity in the economy. Productivity has grown at about 1 percent each year in the 1980s and 1990s, a pace no better than that attained in the stagflation era of the 1970s. [See Jeff Madrick, "Spin Cycle," TAP, March-April 1997.] This suggests that the higher profitability of the 1990s is not the payoff for enhanced private-sector efficiency. Rather, higher profitability in the current context implies a larger flow of income to owners of capital out of the same steady stream of income (productivity) that the economy has been generating for several decades.

In short, this growth in profits in the 1990s mainly reflects business success in restraining wage growth. If the return to capital in 1994 and 1995 (10.66 percent) had been instead at the average of the cyclical peaks during the 1959-79 period (8.37 percent), then average hourly compensation would have been 3.6 percent higher. This loss of wages from the shift of income to capital is comparable in size to the lost wages of the typical worker due to factors such as the shift to services, globalization, deunionization, or any of the other prominent causes of growing wage inequality.

Consider another yardstick: The Congressional Budget Office estimates that the gain from balancing the budget in seven years will be a 0.5 percent increase in hourly compensation in the seventh year. Thus, reversing the recent redistribution of income from labor to capital would yield benefits to workers that are seven times as great as the expected gain from balancing the budget.



Some argue, however, that higher profitability has no distributional consequences. Newsweek columnist Robert J. Samuelson wrote last year that

[American capitalism] increasingly focuses on maximizing profits. But this may not pre sage a new class war, because the clamor for higher profits often comes from managers of pension funds and mutual funds, whose constituents are mainly middle-class.

This sentiment has been voiced repeatedly in the business press, suggesting that what workers are not getting in their paychecks they are getting in their pension plans. There is a surface plausibility to this argument, since the use of 401 (k) plans has clearly widened the extent of stockholding. The Federal Reserve Board, for instance, recently reported that 41.1 percent of all families in 1995 held some stock, either directly (stocks) or indirectly (mutual funds or retirement accounts), up from only 31.7 percent of families in 1989. As impressive as this trend is, it also proves that roughly 60 percent of families held no stock whatsoever. Therefore, it is improbable that most workers who are experiencing real wage reductions are somehow reaping their reward in their stock portfolio.

Of course, owners of a small amount of stock do not make much money from a stock boom. According to James Poterba and Andrew Samwick (of Dartmouth), although 36.8 percent of families held some stock, directly or indirectly in 1992, only 28.9 percent of families owned more than $2,000 worth of stock. Stock ownership is highly skewed: Half of all stock held by U.S. families is owned by the best-off 5 percent. In contrast, the bottom three-fourths of households own less than 20 percent of all stock, with the bottom half owning less than 5 percent. I wonder whom Samuel son considers "middle class."

There may be millions of people who own stock, but the gains from higher profitability and the stock market boom primarily benefit the best-off families and not the typical working family. Ordinary people still depend mainly on salaries and wages, and capital income has soared at the expense of their livelihood.

You may also like