Home Bias in Portfolios and Taxation of Asset Income

There is now extensive evidence that individual investors have a strong tendency to invest in domestic rather than foreign equity. This “home bias in portfolios can potentially have important implications for economic behavior and economic policy. For one, it suggests that extra savings in a country will be invested primarily at home, consistent with the evidence for a lack of international capital mobility reported in Feldstein and Horioka (1980). In addition, the implied lack of capital mobility may explain the observed taxation of the return to domestic capital. In particular, when capital is fully mobile internationally a tax on domestic capital in a small country does not affect the net–of–tax rate of return available to capital owners and instead would be borne by immobile factors, primarily labor. In this setting, Diamond and Mirrlees (1971) show that such a tax would be dominated by labor income taxes (or consumption taxes) even from the perspective of workers. If capital were not so mobile, however, then capital should bear part of the tax, so that the tax might well be chosen for distributional reasons. These presumed implications of home bias can only be judged, however, in the context of some particular model that generates home bias. The objective of this paper is to choose a plausible explanation for the observed home bias, and then to use a formal model based on this explanation to explore whether the above two implications of home bias necessarily follow. Since the conventional forecasts from portfolio models are that investors will hold a fully diversified portfolio of equity issued worldwide, some other considerations must be included in the model so that it can explain home bias, Several of the approaches that have been explored, including the one we use, presume that domestic equity helps domestic residents hedge against other risks they inevitably face. For example, Hartley (1986) hypothesizes that the return on publicly traded equity will be negatively correlated with the return on nontraded domestic assets. In fact, however, Pesenti and van Wincoop (1996) find little correlation between the returns on equity and nontraded assets in the data. Similarly, Eldor, Pines, and Schwartz (1988) hypothesize that the return to domestic equity will be negatively correlated with domestic labor income. The empirical evidence for this explanation is mixed. While Bottazzi, Pesenti and van Wincoop (1996) report a nontrivial