Economists have long been involved in the search for a few key indicators that predict the behavior of market economies. For Canada, it has been shown that the yield curve reliably tilts down and that real M1 growth declines before economic contraction, but this has been demonstrated almost exclusively in the context of single estimation equations or atheoretical VARs. This paper offers an alternative approach to the study of economic turning points. To qualify as a business-cycle indicator, a variable must behave differently when an economy is approaching or in recession than it does during economic expansions. That simple logic admits a variety of parametric and nonparametric tests of a variable’s usefulness, in forecasting. We examine the behavior around recessions of sixteen Canadian and U.S. time series. In the end we find that only the slopes of the Canadian and the U.S. term structure meet the prespecified criteria; the change in the nominal MCI and in real M1 follow behind.
