Predicting Returns and Changes in Real Activity in Emerging and Developed Economies

Abstract

Listen

Translate article

This paper investigates predictions of changes in real activity and stock returns in 24 developed and emerging economies. A simple general equilibrium version of an intertemporal capital asset pricing model (I-CAPM) motivates the analysis. The model implies that share prices will be proportional to real activity, i.e. real activity and share prices are driven by the same time-series process. Furthermore, in an efficient market where investors have a desire to smooth out predictable changes in consumption, and consumption is linked to real activity, returns will be predictable when changes in real activity are predictable, i.e. the same instruments that predict changes in real activity can predict returns, and the predictions of returns should be proportional to the predictions of real activity. From the I-CAPM, three empirical questions are derived and tested: (i) are the series for real activity and share prices cointegrated and thus driven by the same common stochastic trend?, (ii) do the same variables predict both changes in real activity and returns?, and (iii) are the predictions of returns proportional to the predictions of changes in real activity? The empirical evidence in support of (i) cointegration and (ii) common predictability is strong, whereas the evidence against (iii) proportional predictions is equally strong. Finally, special attention is paid to the ability of the deviations from the cointegration relations to predict changes in real activity and returns.