How Long are Business Cycles? Reconsidering Fluctuations and Growth

I reconsider the empirical relationship between business cycles and economic growth and pose the question: How long are business cycles when we try to distinguish between business cycles and economic growth? There is a common held view - a myth - that the cyclical component has a period up to eight years but the empirical evidence underlying that myth is the tradition of empirical business cycle research of Burns and Mitchell (1946) and NBER. That is not a good measure of the duration of the cyclical component in macroeconomic time series since Burns and Mitchell do not distinguish between business cycles and economic growth and they measure economic growth as a part of the expansion phase of a business cycle and consequently overstate the duration of the cyclical component. The Burns-Mitchell concept of business cycles differs from the modern view where business cycles are defined as movements about trend in GDP. I measure the duration of business cycles using a computer algorithm for determining turning points in classical business cycles, recovery cycles, and growth cycles for 11 OECD countries for the postwar period, and I conclude that business cycles are for most countries shorter than six years and that this should be the basis for a distinction between business cycles and economic growth. I derive the implications for, first, optimal business cycle filtering of macroeconomic time series and, secondly, I compute business cycle stylized facts for the U.S. economy and for a standard RBC model