Forecasting with Econometric Methods: Folklore Versus Fact

Abstract

Evidence from social psychology suggests that econometricians will avoid evidence that disconfirms their beliefs. Two beliefs of econometricians were examined: (1) Econometric methods provide more accurate short-term forecasts than do other methods; and (2) more complex econometric methods yield more accurate forecasts. A survey of 21 experts in econometrics found that 95% agreed with the first statement and 72% agreed with the second. A review of the published empirical evidence yielded little support for either of the two statements in the 41 studies. The method of multiple hypotheses was suggested as a research strategy that will lead to more effective use of disconfirming evidence. Although this strategy was suggested in 1890, it has only recently been used by econometricians. CommentsPostprint version.

Forecasting residential construction demand in Singapore: a comparative study of the accuracy of time series, regression and artificial neural network techniques

“…3 Based on the applied data, the simple univariate BJ technique has been found to be as accurate as the sophisticated ANN, while having performed outstandingly better than the multiple regression. These findings support published empirical studies dealing with accuracy(Slovic 1972;Armstrong 1978; Makri- dakis & Hibon 1979) which concluded that 'more complex or statistically sophisticated methods are not necessarily more accurate than simpler methods'…”

Economics as a Hard Science: Realistic Goal or Wishful Thinking?

“…Since I am dealing with hypothesis testing rather than with forecasting, it is not necessary to consider the question whether cvmometric models forecast better than naive models. For evidence on this, see Armstrong (1978b) and Zarnowitz (1979). .4ny superiority of econometric models need not result from their being an accurate representation of the economy, but could result from their picking up various autoregressive features of the economy, as well as from the ad hoc adjustments that the managers of models usually make to the raw output of their models.…”

Building State and Local Government Analytic Capacity: Using Regional Economic Models for Analysis of Public Policy

“…Most importantly, expertise must be available to operate the models appropriately. In the forecasting field, having a sophisticated model alone has not always been associated with increased forecasting accuracy (Armstrong 1978;Mahmoud 1984).…”