Why you should always use automatic model selection in empirical studies
David Hendry - Oxford University
Professor Sir David Hendry, Co-Director of the Economic Modelling Programme at The Institute for New Economic Thinking at the Oxford Martin School (University of Oxford) is presenting a paper on empirical macro-econometric modeling titled: "Why you should always use automatic model selection in empirical studies".
Macroeconomic time-series data are aggregated, inaccurate, non-stationary, collinear and rarely match theoretical concepts. Macroeconomic theories are incomplete, incorrect and changeable: location shifts invalidate the law of iterated expectations and `rational expectations' are then systematically biased. Empirical macro-econometric models are non-constant and mis-specified in numerous ways, so economic policy often has unexpected effects, and macroeconomic forecasts often go awry. The authors provide general critique of DSGE models for explaining, forecasting and policy analyses at central banks, and suggest new directions for improving current empirical macroeconomic models based on empirical modelling broadly consistent with better theory, rather than seeking to impose simplistic and unrealistic theory. In place of selecting a model using just one of theory, empirical evidence, policy relevance and forecasting, the authors propose automatic selection where available theory is retained, so its parameter estimates are unaffected despite searching over many candidate variables, longer lags, functional forms, outliers and location shifts. As theory is not imposed, it can be jointly evaluated against a wide range of alternatives, accepted when it is congruent, but a better model discovered when the theory is incomplete.