forecastingprinciples.com Reviews of Important Papers on Forecasting,
1985-1995 Reviews
Review of:

Dominguez, Kathryn M., Ray C. Fair, and M.D. Shapiro (1988), "Forecasting the depression: Harvard versus Yale," The American Economic Review 78, 595 – 612.


Could the Great Depression have been forecasted if the time series forecasting techniques that are now available had been available in the 1920s? To address this question, Dominguez, et al. used only extrapolation techniques. That is, no causal methods were considered. What would you expect? Well, an alternative subtitle for this paper could have been Harvard 0, Yale 0. The sophisticated techniques did not help. Optimism about the economy would still have prevailed had these sophisticated methods been used. In light of research that has compared sophisticated with simple extrapolations, I do not think it is reasonable to expect that these sophisticated methods would have forecasted more accurately. This result is consistent, then, with prior research on the value of sophisticated versus simple extrapolations; sophistication, beyond a certain point, does not improve the accuracy of forecasts. Armstrong (1984) summarizes this research. It would have been more interesting to see if an econometric model would have provided more accurate forecasts. Existing research suggests that it would (Fildes, 1985), assuming that sufficient data would have be available. Finally, the authors claim that the actions of the Federal Reserve Board could not have been predicted. I believe that role playing techniques may have been able to provide useful predictions of how this group would have acted (Armstrong, 1987).

References

Armstrong, J. Scott, 1984, "Forecasting by extrapolation: Conclusions from 25 years of research," Interfaces, 14, 52-66.

Armstrong, J. Scott, 1987, "Forecasting for conflict situations," in: George Wright and Peter Ayton, eds., Judgmental Forecasting (Chichester: Wiley) 157-176.

Fildes, Robert, 1985, "Quantitative forecasting – the state of the art: Econometric models," Journal of the Operational Research Society, 36, 549-580.