BUSINESS WEEK RESTATES THE NINETIES
By Dean Baker[1]
April 22, 2002
A recent issue of Business Week featured a provocative cover story, which offered a new interpretation of the economy's pattern of growth in the nineties ("Restating the 90s," 4-1-02; p 50). The article claims that the big gainers from the nineties were actually workers, not corporations and shareholders. One representative comment asserts that "workers received 99% of the gains from faster productivity growth at non-financial corporations."
It is easy to demonstrate that this was not the case. Data from the Commerce Department and the Labor Department show quite clearly that there was a redistribution from labor's share to capital's share (profits plus interest) during the decade. This is the first time that this has been the case since the Vietnam War. During the rest of the post World War II period, labor share has either increased, or at least held constant. The fact that the redistribution went from labor to capital -- and not in the opposite direction, as implied by this quote -- means that capital received more than its share of the faster productivity growth in the nineties.
The basic issue on distribution can be settled easily by examining the commerce department's data on corporate income. This is graphed in the figure below.[1] The starting point is 1988, because that
![]() |
was the peak profit year of the last business cycle. The profit share dipped in the early nineties recession, as it always does during a recession. However, it recovered strongly in the mid-nineties, reaching a peak of 21.6 percent in 1997, a level exceeded only by the Vietnam War era profit peaks. The profit share trails off slightly over the next three years, but even in 2000, it is still above the peak share of the last cycle. This is true regardless of whether the broad measure of capital income is used -- combining profits and net interest, or a more narrow measure that just takes the profit share directly. (From the workers' standpoint, it doesn't matter whether their wages are reduced due to growth in the profit share or the interest share, in either case the rising capital share implies a decline in labor's share.)
It is possible to tell a slightly different story by focusing more narrowly on the non-financial corporate sector. Profit shares in the non-financial sector also peaked in 1997 at a level well above the eighties high point, but they fell more sharply in the last three years of the decade, as shown in the graph below.[2] As a result, the broader measure of profit share in 2000 was somewhat below
the peak of the eighties cycle. The main reason for the sharp falloff in the profit share in the last part of the business cycle was a 20.5 percent decline in profits in the manufacturing sector. This in turn was attributable to the flood of low cost imports resulting from the run-up in the dollar following the East Asian financial crisis. While this does support the claim of a shift away from capital at the end of the decade, this is only due to the fact that post-peak years are compared to the peak of the eighties or nineties cycle. At the peak on the nineties cycle, the profit share was almost a full percentage point higher than it peak in the eighties cycle (19.2 percent compared to 18.3 percent).
It is also worth noting that the non-financial sector is not very well defined. Profits can shift from non-financial to financial as firms change the way they conduct their business. For example, if a store extends credit directly to customers, and charges interest, these earnings would appear as profits in the non-financial sector. On the other hand, if they opt instead to rely on credit cards, then the interest earned from their customers will appear as profits in the financial sector. If more financial services are out-sourced in this manner, thereby shifting profits from the non-financial sector to the financial sector, it would be misleading to characterize the development as a shift to labor. The data from the broader corporate sector is unambiguous -- the profit share increased in the nineties, and remained above its previous business cycle peak (1988) until the onset of the recession in 2001.
One factor that may explain some of the conclusions in the business week article is that it uses data for the recession year of 2001. Profits always fall sharply in a recession, and 2001 was no exception. For example profits fell by 19.2 percent in the 1970 recession year and 11.6 percent in the 1980 recession year. The path of profits during the recession that ends a cycle reveals nothing about the course of profits during the cycle itself.
A different way of looking at the same issue is to examine the pattern in productivity growth and labor compensation. The main difference in these approaches is that productivity uses gross, rather than net output, and is divided by labor hours. Also, data is only available for the entire non-farm business sector and the non-financial corporate sector, not the corporate sector alone.[3]
The table below shows the average rates of productivity growth for the cycle as a whole, and the first and the second half, for both the non-farm business sector and the non-financial corporate sector. It also shows the rate of growth of real labor compensation, deflating compensation by the output deflator.[4] As can be seen, there is no evidence of a shift towards labor over the course of the
Average Annual Growth Rate |
||||
|
Non-Farm Business |
Non-Financial Corporate |
||
|
Productivity |
Real Compensation |
Productivity |
Real Compensation |
1989-2000 |
2.0% |
1.9% |
2.1% |
2.1% |
1989-1995 |
1.5% |
1.1% |
1.6% |
1.6% |
1995-2000 |
2.6% |
2.9% |
2.7% |
2.7% |
business cycle. For the non-farm business sector, there is actually a small shift toward capital over the cycle as a whole, with a shift towards labor in the second half of the decade not fully offsetting the shift toward capital in the first half of the cycle. In the non-financial corporate sector, labor compensation keeps even with productivity growth for the decade as a whole and through both sub-periods. There clearly is no evidence of a shift towards labor in this data.
It is important to recognize that all labor income, even the bonuses and stock option of CEOs are included in the labor share. The average hourly wage of the typical worker grew much less rapidly than the data in the table imply. For the business cycle as a whole, from the previous peak in June of 1990 to the beginning of the recession in March of 2001, the real hourly wage of a typical worker rose at an average annual rate of 0.5 percent. During the first half of this period, 1990-1995, the real hourly wage fell at an average rate of 0.4 percent annually. From 1995 to 2001, the real wage rose at an annual rate of 1.5 percent.[5] This led to a modest rate of real wage growth for the decade as a whole, but a considerable slower pace than for the decades of the fifties, sixties, and even somewhat slower than the 0.8 percent rate for the decade of the seventies. (Approximately 0.5 percentage points of the gap between the 1.9 percent annual growth in average real hourly compensation shown in the table, and the 0.5 percent annual growth in the real median wage, is explained by the use of different deflators. The rest of the gap is explained by the upward redistribution of wage income to high wage earners over the course of the business cycle.)
While both the income and productivity data show that the capital share at least held constant over the course of the nineties business cycle -- if not actually rising -- it is worth noting that holders of capital did even better when capital gains are included in the picture. At its peak in the first quarter of 2000, the stock market was valued at more than 180 percent of GDP.[6] By contrast, the market was valued at just 64 percent of GDP at the end of the last cycle in 1989. Measured as a share of GDP, the stock market generated an unprecedented amount of wealth for shareholders in the nineties. While this will be reversed in the coming decade as the stock bubble deflates, shareholders who are able sell their stock near the peak will garner an enormous windfall.
In short, the nineties were a very good decade for corporations and holders of capital, as is clearly shown when examining profit shares at the peak of the business cycle. There was some shift back to labor in the last three years of the business cycle, but by most measures, this did not offset the shift in the other direction in the first half of the cycle. The nineties cannot be accurately portrayed as a decade in which labor gained at the expense of capital.
[1] This data can be found in table 1.16 in the Commerce Department's national income and product accounts (NIPA). The measure of profit includes the capital consumption adjustment and inventory valuation adjustment (line 9), interest is taken from line 17, and the denominator is domestic income in the corporate sector (line 5).
[2] The data in the figure are taken from table NIPA table 1.16, lines 23, 27, and 35.
[3] There are two technical issues that complicate the effort to examine distribution through the productivity data. The first is that, although they are definitionally equal, GDP measured on the output has grown less rapidly than GDP measured on the income side. This is clearly due to measurement error, but it means that the data would show real labor compensation (an income side measure) rising relative to measured productivity growth, even if they were actually growing at the same pace. The second issue is that the depreciation share of GDP has risen at the rate of approximately 0.3 percentage points annually. This should lead to a slower rate of growth of labor compensation, since labor cannot be paid out of depreciation.
[4] The data for the non-farm business sector can be found in the Economic Report of the President, table B-49. Data for the non-financial corporate sector is available on the Bureau of Labor Statistics website [http://data.bls.gov/cgi-bin/dsrv].
[5] Data for median hourly wages can be found on the website of the Economic Policy Institute (www.epinet.org). To get real wage growth for this period, the CPI-U-RS was used as the deflator.
[6] This data can be found in the Federal Reserve Board Flow of Funds table L.213, line 20.