Analisi econometrica di modelli finanziari a variabili latenti: un'applicazione al mercato italiano

Abstract

Several linear asset pricing theories, including the Arbitrage Pricing Theory (APT) and Capital Asset Pricing Models (CAPM), imply a factor structure in expected asset returns, usually characterized by different numbers of factors. In this paper it is shown that it is possible to nest this class of theoretical models within the reduced rank regression model, which allows to test for the number of latent factors and to analyse directly the relationship between asset returns and macroeconomics variables, when the latter are assumed to be linearly related to the unobservable factors. Within these models it is thus possible to discriminate between alternative models, like the APT and the CAPM. The empirical application to Italian Stock Market returns suggests fewer significant factors than the ones implied by standard factor analysis techniques; moreover no significant relation is found between stock returns and a real production indicator, while both financial and monetary variables seem to be linked to the latent factors underlying stock returns.