Services Productivity in the United States: Griliches’ Services Volume Revisited

September 19, 2003

Zvi Griliches (1992) reviewed services sector productivity trends, as well as issues in measuring
services productivity, as these matters stood in the early 1990’s (see also his American Economic
Association presidential address, Griliches, 1994). In this paper, we analyze the rapid post-1995
productivity growth in services industries, which as we show have contributed greatly to the strength of
U.S. productivity growth in recent years. We also review some of the major measurement issues that
Griliches addressed, from roughly a dozen years on.

The contexts of the early 1990’s and early 2000’s are very different, yet at the same time similar.
Griliches wrote in the context of the post-1973 U.S. productivity slowdown, which was the big puzzle of
that day. He pointed out that services were crucial to the post-1973 slowdown, because productivity in
services industries grew much more slowly than productivity in goods-producing industries after the late
1960s. Services, therefore, acted as a brake on U.S. productivity growth, a conclusion that was
particularly unsettling because services have represented an increasing share of U.S. economic activity, a
pattern that is also evident in Europe and other advanced economies.

The post-1973 puzzle was never resolved, just abandoned by economists when they were
confronted with a new problem—the acceleration of U.S. productivity after about 1995. We find, in this
paper and in our previous one (Triplett and Bosworth, 2002), that accelerating productivity in services
industries played a crucial part in post-1995 U.S. productivity growth. Indeed, in recent years services
industry labor productivity has grown as fast as labor productivity in the rest of the economy, which is
why we have previously said that “Baumol’s disease has been cured.” In this, our findings are a mirror
image of the conclusions emphasized by Griliches: Both the post-1973 slowdown and the post-1995 acceleration in U.S. productivity growth—both labor productivity and multifactor productivity—is
located disproportionately, though not entirely, in services.

In Griliches’ time and now, services industries are the industries that are the most intensive users
of information and communication technology (IT) capital equipment. But unlike Griliches, who
complained that the IT effect on services productivity was invisible in the data of his day, we find that IT
investments now make a substantial contribution of labor productivity growth in services-producing
industries. This, of course, is another change from the early 1990’s, when lagging services productivity
seemed a stifling problem for economic growth.