Working Paper Series no. 266: Stockholding: Does housing wealth matter?

Abstract

The existence of an equity premium puzzle has been largely emphasized in the empirical literature: one observes low risky assets holding that the standard portfolio choice model (with expected and discounting utility and homothetic preferences) is not able to explain. Besides modifications of rationality assumptions, other changes have been introduced in the standard framework to account for this equity premium puzzle: the existence of transaction costs, liquidity constraints or other risks (e.g. risks associated with employment, income, health,). These background risks (i.e. risks that are unavoidable, exogenous and independent from the financial market risk) may lead households to reduce their stockownership in order to limit their global exposure to risk. Recent developments examine the role of housing on stockholding as another background risk. This paper analyses the empirical link between stockholding and housing wealth within this framework. We use the French wealth survey (Enquête Patrimoine 2004, Insee) that gives us detailed information on households portfolio, socio-demographic variables, and several measures of preference and exposition to various risks (income, unemployment, health, and business). We find that an increase in the housing to net wealth ratio crowds out stock market investment for a given total financial wealth: when facing real estate exposure to risk, households tend to moderate their global exposure to risk by limiting the share of their financial wealth invested in stocks. Among the other significant determinants of the equity premium puzzle, we emphasize the role of transaction and information costs, the attitude toward risk and the exposition to various risks in hampering investments in stocks.

Luc Arrondel and Frédérique Savignac
December 2009

Classification JEL : D91, G11, R22, E21.

Keywords : Portfolio choice, background risks, housing demand, life-cycle model.