Long Cycles In Growth: Explorations Using New Frequency Domain Techniques With US Data

In his celebrated 1966 Econometrica article, Granger first hypothesized that there is a 'Typical' spectral shape for an economic variable. This 'Typical' shape implies decreasing levels of energy as frequency increases, which in turn implies an extremely long cycle in economic fluctuations and therefore in growth. Spectral analysis is however based on certain assumptions that render these basic frequency domain techniques inappropriate for analysing non-stationary economic data. In this paper three recent frequency domain methods for extracting cycles from non-stationary data are used with US real GNP data to analyse fluctuations in economic growth.