LATENT CLASS ANALYSIS FOR FINANCIAL DATA

This paper deals with optimal international portfolio choice by developing a latent class approach based on the distinction between international and non-international investors. For thirty years financial literature has been looking for convincing explanations of systematic low international investments: share of foreign assets held by domestic investors is greatly lower than expected by risk-return efficient portfolios: this puzzle is usually called equity home bias. On the basis of micro data, we analyze the effects of many social, demographic, economic and financial characteristics on the probability to be an international investor. We specify a latent class model which allows to test the existence of two groups of investors: the sub-group of investors who are completely precluded from investment in foreign assets, and the sub-group of investors who are not prevented from investing in foreign assets. Our results shows how traditional measures of equity home bias are upward biased because they do not allow for the existence of international investment rationing operators. On the contrary, by resorting to latent class analysis it is possible to detect the unobservable distinction between international investors and investors who are precluded from operating into international financial markets and, therefore, to obtain an unbiased measure of equity home bias.