Generalization of Information, Granger Causality, and Forecasting

Purpose ndash; This paper analyzes forecasting problems from the perspective of information extraction. We study circumstances under which the forecast of an economic variable from one domain (country, industry, market segment) should rely on information regarding the same type of variable from another domain even if the two variables are not causally linked. Thus, Granger causality linking variables from different domains may be the rule and should be exploited for forecasting.Design/methodology/approach ndash; The article applies information economics, in particular the study of rational information extraction in order to shed light on the debate on causality and forecasting.Findings ndash; It is shown that the rational generalization of information across domains can lead to effects that are hard to square with economic intuition but are nevertheless worth being taken into consideration for forecasting. Information from one domain is shown to affect another domain if there is at least one common factor affecting both domains which is not (or not yet) observed when a forecast has to be made. The analysis suggests the theoretical possibility that the direction of such effects across domains can be counter-intuitive. In time-series econometrics such effects will show up in estimated coefficients with the ldquo;wrongrdquo; sign.Practical implications ndash; This study helps forecasters by indicating a wider set of variables relevant for prediction. The analysis offers a theoretical basis for using lagged values from the type of variable to be forecast but from another domain. E.g., when forecasting the bond risk spread in one country we suggest introducing in the time-series model the lagged value of the risk spread from another country. Two empirical examples illustrate this principle for specifying models for prediction. While we limit the application to risk spreads and inflation rates the approach suggested here is widely applicable. Originality/value ndash; The present study builds on a probability theoretical analysis to inform the specification of time-series forecasting models