Real estate ownership: what influence on the household portfolio?

Summary Real estate is often the most important financial asset of households. But it is also a durable consumer good, for which transaction costs, both at the time of acquisition and at the time of sale, are high. This dual characteristic invites us to assess the influence of real estate investment on the composition of households' financial portfolios. To do this, we attempt to identify two effects of opposite sign: a so-called "wealth effect" and a real estate risk effect. The increase in real estate wealth leads to a more risky financial portfolio, while real estate risk tends to reduce the ownership of risky assets. To separate these two effects, an indirect method proposed by Chetty and Seizdl (2010) is used: the wealth effect is captured by changes in housing wealth net of outstanding loans and the risk effect is captured by changes in its gross value. The hypothesis is that an increase in net wealth at a given level of gross wealth corresponds to an increase in the household's housing wealth without changing its degree of exposure to housing risk. Its effect will therefore measure a pure wealth effect. However, the approach poses a problem of endogeneity: the households that own the largest real estate assets and have the least need to incur debt are generally the richest, and therefore those that take the most financial risk. This problem is solved by instrumenting net and gross wealth with the price changes observed at the departmental level. In this way, we obtain the two expected opposite effects: the wealth effect tends to increase the share of risky assets in the portfolio, while the real estate risk effect tends to decrease this share. According to our estimates, the wealth effect is more important than the property risk effect.