Introduction and summary
When academic economists talk about business cycles, they have something more general in mind than persistent fluctuations of gross domestic product (GDP) about its trend, which is the definition typically used by business economists. For an academic economist, the business cycle describes the way that cyclical fluctuations of GDP typically relate to cyclical fluctuations of other economic time series (such as consumption and investment) from the same economy. One of the most striking findings of the vast academic business cycle literature is that irrespective of the time period or particular country, business cycles are all alike. This means that the typical relationship between cyclical fluctuations of GDP and cyclical fluctuations of other economic time series of the U.S. economy is similar to the typical relationship between cyclical fluctuations of the same time series in all other market-based economies. As Lucas (1977) notes, this finding is both important and challenging for the study of business cycles, since it suggests the possibility of a unified explanation of business cycles based on general laws governing market economies.
So, we know that national business cycles are alike in important ways. What do we know about subnational business cycles? Given that subnational economies, such as those of U.S. states, are as large as some national economies, one would expect their business cycles to have been well studied. In contrast to national business cycles, little is known about subnational business cycles. The goal of this article is to expand our knowledge of subnational business cycles by testing whether the proposition that all business cycles are alike extends to U.S. states. We limit our analysis to the business cycles of the U.S. and the five states (Iowa, Illinois, Indiana, Michigan, and Wisconsin) of the U.S. Federal Reserve's Seventh District, home of the Federal Reserve Bank of Chicago. (1)
Our approach follows that of international business cycle studies, such as Backus and Kehoe (1992), by conducting a detailed analysis of the way in which activities within a regional economy relate to the region's aggregate business cycle and the way in which regional aggregate business cycles relate to one another. The main limitation on subnational business cycle research stems from a deficiency of state-level data analogous to the national income and product accounts data that are typically used in the analysis of national business cycles. We overcome this problem by finding suitable state-level proxies for national account aggregates (please see the appendix for further details). Consumption expenditure is proxied by real retail sales taxes, investment is proxied by the number of residential home sales and housing permits, while output is proxied by real personal income, various measures of labor input (including nonfarm payroll employment and average manufacturing hours) and capital utilization. One of the by products of this analysis is that we uncover new leading variables that could serve as useful indicators of the future direction of District state business cycles.
We find that District state business cycles are like the national business cycle along a number of dimensions. Turning to the within-region analysis, which explores the way activities within a regional economy relate to the region's aggregate business cycle, we find that state-level analogues of consumption and residential investment have similar business cycle characteristics to their national counterparts. Moreover, they tend to be strong leading indicators of the national and state business cycles. Our labor market analysis yields much stronger results. Here, we find that the business cycle characteristics of national and state-level measures of labor market activity are virtually identical, especially with regard to whether they are procyclical or countercyclical, as well as leading or lagging. …