A sectoral perspective on economic stability

Business Economics, Oct, 1998 by Gail Fosler, Kevin J. Stiroh

The strength and stability of the U.S. economy has spawned a recent debate about the death of the traditional business cycle. GDP growth through 1997 shows the economy expanding at a strong rate with few signs of inflationary pressure. Many question whether this unexpected combination of continued growth and relative price stability represents an end to traditional business cycle and macroeconomic models.(1)

Are these apparent changes simply due to the fortuitous absence of negative shocks, or to changes in economic policies and institutions, or has there been a fundamental change in the dynamics of the U.S. economy? This paper explores these business cycle sectors that comprise the private U.S. economy. By examining output growth patterns at three levels of aggregation-economy, sector, and industry-we show that important features of the growth process are missed when one looks only at aggregate data.

The aggregate economy, for example, shows little variation in the severity of the business cycle since 1958, but a more disaggregated analysis suggests a shift in the cyclicality of key sectors that comprise the economy. In particular, manufacturing has become somewhat less cyclical, while services show larger output fluctuations. In the past, manufacturing showed a higher movement with aggregate shocks, e.g., shortages, interest-rate increases, credit crises, etc., while today both services and manufacturing are likely to decline during a recession. In spite of the growing similarity, however, manufacturing remains more volatile.

More detailed data for individual industries show substantial differences in growth rates across industries within the broader sectors. For the entire period 195896, the computer and office equipment industry grew faster than any industry at 16.1 percent per year, while footwear grew the slowest at-2.9 percent per year. Both of these industries are in the manufacturing sector, which emphasizes the importance of looking beyond aggregate and sectoral data to get a complete perspective on the economic growth process.

On a year-by-year basis, the variation in growth rates across industries is even more extreme. For all 163 industries in the sample, the range between the fastest growing industry and the slowest growing industry in each year averaged an extraordinary 56 percentage points between 1958 and 1996. Furthermore, individual industries showed growth patterns that were quite insulated from aggregate fluctuations. In every year from 1958 to 1996-during periods of expansion and recession alike-at least one industry in the U.S. economy grew faster than 13 percent while another industry contracted by more than 5 percent.

The substantial differences in growth patterns reflect the idiosyncratic evolution of firms and industries as they respond to the changing economic environment. An important consequence of this heterogeneity is the reallocation of scarce resources towards fast-growing, high-productivity industries, a process that plays an important role in aggregate growth (See Baily, Hulten, and Campbell (1992)). By analyzing disaggregated data, we show that these industry-level differences are indeed large and evolving over time, but we find little evidence that these changes are reflected in aggregate business cycle data.

DATA AND SUMMARY STATISTICS

This paper uses real gross output data from the Bureau of Labor Statistics (BLS) for the period 1958 to 1996. Gross output estimates (the total value of producer's shipments) are provided for 169 private industries that comprise ten private industry groups or sectors. The ten sectors - agriculture; mining; construction; manufacturing; transportation; communications; utilities; trade; finance, insurance and real estate (FIRE); and services - in turn comprise the aggregate U.S. economy. Note that the sum of gross output across sectors exceeds GDP due to double counting of intermediate inputs and is used only as an indicator of the business cycle. (See the Appendix for details.)

At the most detailed level, real gross output for each industry is calculated by deflating current dollar gross output with the BLS price index for each of the 169 industries. Aggregate gross output for a sector is calculated as a chain-weighted Tornqvist index of the component industries. Aggregate economy gross output is likewise computed as a Tornqvist index of the ten component sectors.

Table 1 defines the ten sectors and reports the number of industries and the corresponding SIC codes [TABULAR DATA FOR TABLE 1 OMITTED] for each sector. Note that the manufacturing and services sectors have the largest number of component industries; these two sectors will be the focus of our subsequent analysis. Table 1 also reports real gross output, measured in 1992 dollars, for 1958, 1975, 1990, and 1996 for the aggregate economy and for each of the ten aggregate sectors. Manufacturing is the largest individual sector, but the three service-related sectors - trade, FIRE, and services - are increasing in relative size.