The Past as Prologue? Business Cycles and Forecasting since the 1960s

This paper examines how successfully economic forecasts as covered in public media foreshadow major real estate and financial crises. Building from existing research on the accuracy of economic forecasts, we focus on real estate crises and the portrayal of forecasts in the press. We utilize textual analysis to identify articles on specific topics and to track coverage over time, both in the print media, as well as the larger internet-based news world; we also use the Google tool, “Insights for Search,” to gauge public “forecasting” and popular beliefs. We find that the use of real-time data limits the accuracy of forecasts, that there appears to be a reluctance in the press to forecast bad news, and that when bad news is forecast, the report is often cushioned with counterbalancing observations. Preliminary results based on review of specific sets of articles show that articles reporting on academic forecasts have tended to be more accurate in hindsight than those relying on other sources. Reviews of two periods of real estate crises, the mid-1980s and the 2007-2009 crisis show that many of the serious problems were described in the press as early as two years in advance, but that interest and concern tended to taper off before the crisis occurred, leading to a sense of surprise. Our results suggest while the recent past is not always a good predictor of future trends, it can be very useful to be familiar with long term history that spans several cycles