A survey on the weak evidence of non-fundamentalness in economic applications
Time series analysis with economic data has unduly relied on the use of causal and invertible models, especially when it comes to dealing with autoregressive moving average (ARMA) models. Procedures that deal with non-fundamental models have slowly developed under the claim that causality and invertibility should not be taken for granted. It is believed that the conclusions of a model may be highly dependent on these assumptions. Empirical evidence of these non-standard features is not vast, casting doubt upon their economic importance or on the usefulness of new identification procedures. In this survey, we present some representative examples of non-fundamentalness found in economic literature but argue that these do not make a compelling case to affirm with certainty that non-fundamentalness is empirically important. Although there are some hints that these features are in fact relevant, empirics need to be performed more carefully to dissipate any doubt.
Autor:
Eduardo Laguna-Müggenburg e Ignacio Lobato
Número de revista:
34
