Towards a financial cycle for the U.S., 1973–2014

Abstract With this paper, we suggest a new approach to estimating financial cycles in terms of interactions of real-sector and financial-sector sentiments. We will apply this to U.S. financial indicators from 1973 to 2014. Based on Schumpeter’s and Minsky’s financial cycle concepts, we arrive at a selection of six indicators that capture finance and real sector linkages: the slope of the yield curve, a Purchasing Managers’ Index, real-estate price returns, the S&P stock price index, and leverage ratios of households (consumer spending) and non-financial corporations. We estimate lead-lag relations and apply principal component analysis to aligned series in order to construct factors. Our conclusion is that two factors, capturing corporate and consumer sentiments, account for over 60% of the cumulative variance in our data. Corporate optimism peaks before crisis episodes, while household/consumer sentiment is more persistent and follows corporate sentiment with a lag.