Did Wages Reflect Growth in…

                                Did Wages Reflect Growth in Productivity?

                                            Martin Feldstein*


      The level of productivity doubled in the U.S. nonfarm business sector
between 1970 and 2006. Wages, or more accurately total compensation per hour,
increased at approximately the same annual rate during that period if nominal
compensation is adjusted for inflation in the same way as the nominal output
measure that is used to calculate productivity.

      More specifically, the doubling of productivity represented a 1.9 percent
annual rate of increase. Real compensation per hour rose at 1.7 percent per year
when nominal compensation is deflated using the same nonfarm business sector
output price index.

       In the period since 2000, productivity rose much more rapidly (2.9 percent a
year) and compensation per hour rose nearly as fast (2.5 percent a year).

      The relation between productivity and wages has been a source of
substantial controversy, not only because of its inherent importance but also
because of the conceptual measurement issues that arise in making the
comparison.

       The relation between wages and productivity is important because it is a key
determinant of the standard of living of the employed population as well as of the
distribution of income between labor and capital. If wages rise at the same pace
as productivity, labor's share of national income remains essentially unchanged.
This paper presents specific evidence that this has happened: the share of national
income going to employees is at approximately the same level now as it was in


         *
        Professor of Economics, Harvard University, and President and CEO of the National
Bureau of Economic Research. This paper was prepared for presentation at the annual meeting
of the American Economic Association on January 5, 2008. I am grateful to Mark Shepard for
research assistance.

WagesAndProductivity.01052008                      -1-
1970.

        Two principal measurement mistakes have led some analysts to conclude
that the rise in labor income has not kept up with the growth in productivity. The
first of these is a focus on wages rather than total compensation. Because of the
rise in fringe benefits and other noncash payments, wages have not risen as rapidly
as total compensation. It is important therefore to compare the productivity rise
with the increase of total compensation rather than with the increase of the
narrower measure of just wages and salaries.

      The second measurement problem is the way in which nominal output and
nominal compensation are converted to real values before making the comparison.
Although any consistent deflation of the two series of nominal values will show
similar movements of productivity and compensation, it is misleading in this
context to use different deflators for measuring productivity and real
compensation.

       A quick review of what economic theory says about the relation between
productivity and compensation will clarify the correct choice of price deflator for
making this comparison and will also indicate how productivity and
compensation would be expected to move in a competitive economy. In the
classroom we often abstract from differences in prices by assuming an economy
with a single product and therefore summarize the basic wage determination
condition by saying that a competitive firm pays a wage equal to the marginal
product of labor. But when we recognize the multiproduct nature of the economy,
we say that the competitive firm pays a nominal wage equal to the marginal
revenue product of labor, i.e., to the marginal product of labor multiplied by the
price of the firm's product. The key real relation must therefore be between
changes in productivity in the nonfarm business sector and changes in the
compensation paid in that sector deflated by the product price and not by some
consumer price that also reflects goods and services produced outside the domestic
nonfarm business sector.

       What does theory say about the relation between productivity growth and
the rise of wages? Labor productivity is defined as the output per hour of labor
input, i.e., as the average output per unit of labor. In the special case of Cobb-
Douglas technology, the marginal product of labor is proportional to productivity.

WagesAndProductivity.01052008           -2-
In that case, the wage paid by a competitive firm should rise at the same rate as the
rise in productivity. With a more general technology, however, the marginal
product of labor is not necessarily proportional to productivity. For example, if
capital deepening causes a rise in productivity and the elasticity of substitution
between capital and labor is greater than one, the marginal product of labor will
rise proportionately less than productivity. The implication of this is that since the
actual technology may not correspond to the Cobb-Douglas specification, we
should not be surprised if compensation and productivity do not move exactly in
proportion to each other.

       Whatever the technology, the competitive firm pays a nominal wage equal
to the marginal revenue product of labor. As I noted a moment ago, this implies
that the real marginal product of labor should be compared to the wage deflated by
the product price and not by some consumer price index. The CPI differs from the
nonfarm product price in several ways. The inclusion in the CPI, but not in the
nonfarm output price index, of the prices of imports and of the services provided
by owner occupied homes is particularly important.

       Between 1970 and 2006, the implicit price deflator for the output of the
nonfarm business sector rose at an annual rate of 3.8 percent while the consumer
price index (the CPI-U series with the research series for the years after 1978) rose
at 4.3 percent. If one uses the CPI data to calculate real compensation, its growth
is reduced to 1.2 percent a year and therefore substantially less than the 1.9 percent
annual rise in productivity. Since the productivity tables in the Economic Report
of the President and in Economic Indicators use the CPI to define real
compensation, it is not surprising that many analysts have concluded that the
official data do not support the theoretical relation between productivity growth
and compensation. This is, as I have emphasized, a misreading of the data.

      There are of course other questions for which using compensation deflated
by the CPI or some other consumer price index is appropriate, including
measuring changes in the standard of living of wage earners and in the incentive to
supply labor. But the nominal compensation deflated by the CPI is not appropriate
for evaluating the relation between productivity and compensation.

      The second error that some analysts make is to compare productivity growth
with wages rather than with total compensation. Because of the rapid growth of

WagesAndProductivity.01052008            -3-
health insurance benefits and other fringe benefits, wage and salary payments
declined from 89.4 percent of total compensation in 1970 to just 80.9 percent in
2006. As a result, the annual rate of increase in wage and salary payments was
0.3 percent less than the rate of increase in total compensation. Note that these are
figures for the entire economy and not just for the nonfarm business sector .

       The practice of focusing on wages and salaries also leads to a mistaken view
of how the shares of national income have evolved. A very misleading number ­
the ratio of wages and salaries to GDP ­ fell from 53 percent in 1970 to 46 percent
in 2006, appearing to imply a sharp change in the distribution of income. But here
both the numerator and the denominator are conceptually wrong. Factor incomes
should be compared to national income rather than to GDP. Total employee
compensation as a share of national income was 66 percent of national income in
1970 and 64 percent in 2006. This measure of the labor compensation share has
been remarkably stable since the 1970s. It rose from an average of 62 percent in
the decade of the 1960s to 66 percent in the 1970s and 1980s and then declined to
65 percent in the decade of the 1990s and again from 2000 until the most recent
quarter.

      Another useful way to examine changes in the compensation share is to
focus on the nonfinancial corporate sector (as presented in table B14 of the 2007
Economic Report of the President.) This eliminates some of the very highly
compensated individuals in the financial sector. It also avoids the problems raised
by separating the capital and labor incomes of sole proprietors . Comparing the
compensation paid by the nonfinancial corporations to the net value added of the
nonfinancial corporations reinforces the conclusions based on the larger scope of
industries. In 1970 compensation was 74 percent of the value added of the
nonfinancial corporate sector. In the year 2006, it was 73 percent. The decade
averages rose from 70 percent in the 1960s and were very stable after that: 73
percent in the 1970s and 1990s, 74 percent in the 1980s and 75 percent since
2000.

       Changes in productivity are not immediately reflected in compensation. The
concurrent (i.e., within year) effect of a rise in productivity is a less than a one-for-
one rise in compensation. More specifically, a regression relating the change in
the logarithm of annual nominal compensation in the nonfarm business sector to
the change in the logarithm of annual nominal productivity (where nominal

WagesAndProductivity.01052008              -4-
productivity is defined as the product of the real productivity level and the
nonfarm business implicit price deflator) shows a coefficient of 0.79. When this
specification is extended to have two lagged annual changes in productivity, the
sum of the coefficients rises to 0.94 and is not significantly different from one (p =
0.53). This regression is estimated with the full sample of years for which official
productivity data are available, i.e., 1947 through 2006.

       The relatively long lags make it difficult to interpret short run movements in
productivity and compensation. In particular, when an economy is coming out of
a recession productivity and profits both tend to rise more rapidly because labor
that had been on the payroll gets used more effectively. The result is that
compensation does not rise as rapidly as productivity. The reverse happens when
an economy slows, causing productivity to decline or to grow more slowly.

       The period since 2000 appears to be a time in which productivity has grown
more rapidly than wages. The average rise in real productivity was 2.9 percent
while the average rise in real compensation as conventionally measured using the
consumer price index averaged just 1.7 percent. This difference of 1.2 percentage
points is however misleading because of the use of the consumer price index. If
we instead compare the average rise in nominal productivity (5.4 percent from
2000 through the third quarter of 2007) with the average increase in nominal
compensation (5.0 percent), the gap is reduced to 0.4 percent. In real terms,
productivity rose 2.9 percent while the corresponding measure of compensation
rose 2.5 percent.

      This measure of a small difference in recent years is confirmed by
examining the residuals of the period since 2000 in the regression of nominal
wage increases on the current and two annual lags of productivity change for the
period since 1949. The mean value for those residuals was -0.43 percentage
points in the period from 2000 through the third quarter of 2007, the same as the
raw difference between productivity growth and compensation growth.. For
comparison, the residuals averaged -0.03 percentage points in the 1990s, -0.23
percent points in the 1980s and a positive 0.34 percent in the 70s and a positive
0.07 percent in the 1960s.

     Although nominal wage increases in recent years lagged slightly behind
productivity gains , this may be a natural result of the rapid rise in capacity

WagesAndProductivity.01052008            -5-
utilization (from about 75 percent in 2001-02 to nearly 82 percent in 2006-07) or
of the capital deepening that occurred during these years. Those explanations
deserve further analysis.

        In summary, basic theory reminds us that real compensation should be
measured using the same price index that is used to calculate productivity. When
this is done, the rise in compensation has been very similar to the rise in
productivity.



END




WagesAndProductivity.01052008           -6-