Academic journal article Federal Reserve Bank of New York Economic Policy Review

Productivity Measurement Issues in Services Industries: "Baumol's Disease" Has Been Cured

Academic journal article Federal Reserve Bank of New York Economic Policy Review

Productivity Measurement Issues in Services Industries: "Baumol's Disease" Has Been Cured

Article excerpt


It is now well known that after 1995, labor productivity (LP, or output per hour) in the United States doubled its anemic 1.3 percent average annual growth between 1973 and 1995 (see chart). Labor productivity in the services industries also accelerated after 1995.

As we documented in a longer version of this paper (Triplett and Bosworth forthcoming), labor productivity growth in the services industries after 1995 was a broad acceleration, not just confined to one or two industries, as has sometimes been supposed. Using the 1977-95 period as the base, we showed that fifteen of twenty-two U.S. two-digit services industries experienced productivity acceleration. Both the rate of LP improvement in services after 1995 and its acceleration equaled the economywide average. That is why we said "Baumol's Disease has been cured." (1)

We also examined the sources of labor productivity growth. The major source of the LP acceleration in services industries was a great expansion in services industry multifactor productivity (MFP) after 1995. It went from essentially zero in the earlier period to 1.4 percent per year, on a weighted basis. As MFP is always a small number, that is a huge expansion. Information technology (IT) investment played a substantial role in LP growth, but its role in the acceleration was smaller, mainly because the effect of IT in these services industries is already apparent in the LP numbers before 1995. Purchased intermediate inputs also made a substantial contribution to labor productivity growth, especially in the services industries that showed the greatest acceleration. This finding reflects the role of "contracting out" in improving efficiency.


In the now standard productivity-growth accounting framework that originates in the work of Solow (1957)--as implemented empirically by Jorgenson and Griliches (1967) and extended by both authors and others--labor productivity can be analyzed in terms of the contributions of collaborating factors, including capital and intermediate inputs, and of multifactor productivity. To analyze the effects of IT within this model, capital services, K, are disaggregated into IT capital ([K.sub.IT]) and non-IT capital ([K.sub.N]), and the two types of capital are treated as separate inputs to production. Thus, designating intermediate inputs--combined energy, materials, and purchased services--as M:


A number of researchers have calculated the contributions of IT and MFP to the post-1995 acceleration of labor productivity growth at the aggregate, economywide level (at the aggregate level, of course, the intermediate inputs net out, except for imports, which typically are ignored). The most prominent examples are Jorgenson and Stiroh (2000), Oliner and Sichel (2000), Gordon (2000), and Council of Economic Advisers (2000). Although there is broad agreement among these studies, a major issue concerns the degree of MFP improvement in IT-using industries, on which the aggregate-level studies reach different conclusions.

Because the most intensive IT-using industries are services industries, the impact of IT on IT-using sectors and the extent of MFP in IT-using sectors provide part of the motivation for our focus on services industries. In addition, we have been leading a Brookings Institution project on the measurement of output and productivity in the services industries (an earlier report on this subject is Triplett and Bosworth [2001]). Clearly, services industry productivity remains a challenging issue with many unresolved puzzles.

We explored the impact of IT and of MFP on services industries by estimating equation 1 separately for each of twenty-seven two-digit services industries. Although our study uses the same level of two-digit detail employed by Stiroh (2001) and Nordhaus (2002) to examine LP, and also begins from the Bureau of Economic Analysis (BEA) database that they use, our research approach is most nearly similar to that of Jorgenson, Ho, and Stiroh (2002), who estimate labor productivity, MFP, and IT contributions for thirty-nine sectors. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.