Trends, Random Walks and Persistence: An Empirical Study of Disaggregated U.S.Industrial ProductionRobert KrolThe Review of Economics and Statistics, Vol. 74, No. 1. (Feb., 1992), pp. 154-159.Stable URL:http://links.jstor.org/sici?sici=0034-6535%28199202%2974%3A1%3C154%3ATRWAPA%3E2.0.CO%3B2-PThe Review of Economics and Statistics is currently published by The MIT Press.Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available athttp://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtainedprior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content inthe JSTOR archive only for your personal, non-commercial use.Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained athttp://www.jstor.org/journals/mitpress.html.Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printedpage of such transmission.The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academicjournals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers,and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community takeadvantage of advances in technology. For more information regarding JSTOR, please contact [email protected]://www.jstor.orgSat Jun 30 16:05:36 2007NOTES TRENDS, RANDOM WALKS AND PERSISTENCE: AN EMPIRICAL STUDY OF DISAGGREGATED U.S. INDUSTRIAL PRODUCTION Robert Krol* Abstract-Unit-root and variance-ratio tests are used to ex-amine the trend properties and degree of persistence of industrial production in U.S. industries and comparable ag- gregates during the post World War I1 period. The evidence from unit-root tests suggests that less than one-half of these industries have output which may be characterized as a ran- dom walk. The variance-ratio test results generally support this conclusion. Consistent with standard economic theory, fluctuations in durable-goods industries are less persistent than in nondurable goods industries. Finally, tests find rela- tively greater persistence in the aggregate industrial produc- tion data. I. Introduction Economists have become increasingly interested in the trend properties of macroeconomic time-series variables (see, for example, Nelson and Plosser (1982), Campbell and Mankiw (1987, 1989)). The main issue is whether these variables, such as GNP or industrial production, contain a unit root or a deterministic lin- ear time trend. The difference has important implica- tions for both economic theory and econometric esti- mation. When a time-series variable contains a deterministic linear time trend, the variable can be characterized as a trend stationary process. Fluctuations around trend for a trend stationary process are considered to be mostly temporary. Until recently, this was a fairly stan- dard belief in the business-cycle literature (see, for example, Blanchard (1981)). This approach was first challenged by Nelson and Plosser (1982). Nelson and Plosser were unable to reject the hypothesis that most U.S. macroeconomic variables contain a unit root. In other words, these variables appear to follow a random walk and can be characterized as a differenced station- ary process. The principal implication of a differenced stationary process is that at least part of the change in these variables is permanent. This suggests that when Received for publication June 19, 1989. Revision accepted for publication February 6, 1991. *California State University, Northridge. I would like to thank Lee Ohanian, Shirley Svorny, and two anonymous referees for very helpful comments. The author is responsible for any remaining errors. This research was sup- ported in part by CSUN and the CSUN School of Business Administration and Economics. [ 154 1 the variable is shocked, it never completely returns to its trend. In this setting, emphasis is generally placed on taste and technology shocks (real factors) rather than monetary shocks as the source of these perma- nent fluctuations.' In the unit root case, the distinction between long-term economic growth and the short-term business cycle becomes blurred (see Beveridge and Nelson (1981) for a discussion). This paper extends the analysis of the trend proper- ties and degree of persistence in time-series data to disaggregated or individual industrial production in- dices in the United States during the post World War I1 period. Disaggregated industrial production may provide us with a different picture of output trend properties and persistence. For example, how compa- rable are the trend properties across U.S. industries? Durable goods industries are generally considered to be highly cyclical (Black (1982)). As a result, durables should tend to be trend stationary and show less persis- tence compared to nondurable goods industries. In addition, the literature investigating the trend proper- ties of economic time-series variables has focused al- most entirely on variables such as GNP or aggregate industrial production. Since this paper investigates the trend properties of disaggregated and comparable ag- gregate industrial production indices, it will be possible to compare evidence using standard econometric mea- sures of persistence at different levels of aggregation. Augmented Dickey-Fuller and variance-ratio tests are used to investigate the trend properties and degree of persistence in individual and aggregate industrial production indices. We find that in only eight of the twenty-two industries investigated does the industrial production index appear to follow a random walk. All durable goods industries appear to be trend stationary. Furthermore, aggregate industrial production mea-sures show greater persistence than disaggregated in- dustrial production indices. This last result is consis- tent with the idea that aggregating nonstationary and stationary processes results in a nonstationary aggre- gate process (see Granger