The influence of homeownership on portfolio choice is a complex issue because housing is both an investment good and a consumption good. That said, a deeper analysis of the influence of homeownership is crucial, at least if we consider the recent high fluctuations in the housing market and the substantial increase in mortgage debt both in the US and some European countries.
Distinguishing between mortgage debt effect and home equity effect
The debate on the influence of homeownership on households’ portfolios is, first, a debate between theory and empirics.
Most theoretical papers try to explain why housing lowers the demand for risky assets. For instance Grossman and Laroque (1990) show that this effect is driven by the illiquidity of housing. For Flavin and Yamashita (2002), housing choice is overdetermined. In other terms, the level of real estate ownership which is optimal from the point of view of the consumption of housing services may differ from the optimal level of housing assets from a portfolio point of view. The optimisation of financial portfolio is therefore constrained by the predetermined housing investment. The authors show that for many households, the constraint binds, so they hold less assets than the non-constrained optimisation would have predicted.
By contrast, most empirical studies find no systematic relationship between housing and stockholding. Heaton and Lucas (2000) and Cocco (2005) estimate that the effect of housing property on stockholding is positive. Yamashita (2003) finds a more complex relationship: the proportion of stocks in financial portfolios exhibits a hump-shaped pattern against the house value to net worth ratio. In other words, when housing represents either a very small part or an important part of the total household wealth, household's stockholding is low.
However, these findings are potentially biased, since housing and stockholding result both from endogenous choices. The more shares the household owns, the more likely the household is to be rich and to therefore have an important housing wealth.
Reconciling theory and evidence
In response to this controversy, Chetty and Szeidl (2012) set forth a framework that reconciles theory and evidence. They separate the effects of mortgage debt and home equity to characterise the effect of housing on portfolio choice. More precisely, Chetty and Szeidl (2012) derive an approximate expression for optimal portfolio shares from a theoretical stylised two-period model. This approximate expression takes the form of a linear regression model in which the household's stock share depends on housing property value and home equity. It therefore seems crucial to distinguish between the effect of the current market value of the house (i.e. its property value) and the effect of the home equity (which is the net housing wealth, i.e. the property value minus the outstanding mortgage debt).
Our econometric approach
To tackle the endogeneity of these two regressors, we have followed the approach set forth by Chetty and Szeidl (2012) by using the same two instrumental variables, namely the average price of houses in the household's area of residence in the year in which portfolios are measured and the average price of houses in the household's area of residence in the year in which the house was bought (cf. Fougère and Poulhes 2012).
For obtaining statistical observations on portfolios of French households, we use the 2010 Patrimoine survey collected by the French Statistical Agency (INSEE, Paris). This survey concerns a nationally representative sample of 15,000 households for whom detailed information on personal holdings is available.
We estimate the magnitude of the effects of property value and home equity on stockholding by two econometric methods; a Two-Stage Least Squares (2SLS) procedure, and a Tobit model. The latter specification allows us to take into account both the extensive and the intensive margins.
Our main finding is that an increase in the household's property value reduces both the stock share of liquid wealth and the probability of owning stocks, while a decrease in home equity increases the stock share of liquid wealth and the probability of owning stocks. These qualitative results obtained with French data are similar to those obtained by Chetty and Szeidl (2012) with US data.
Wealth effect versus risk effect
However, we find some quantitative discrepancies.
For France, the estimated effect of home equity is roughly twice the estimated effect of the property value while in the US these two effects have the same order of magnitude.
In the US, the wealth effect of holding more home equity is cancelled out by the risk effect of owning a more expensive house (the sum of the two effects is around zero), in France the wealth effect dominates the risk effect.
Explaining the discrepancy
To explain, we exploit the theoretical predictions of the housing and portfolio choice model built by Cocco (2005), then used and generalised by Chetty and Szeidl (2012).
Numerical simulations show that an increase in the level of adjustment cost lowers the effect of property value compared to the effect of home equity on the stock share of liquid wealth. The two effects are of the same order of magnitude when there is no adjustment cost. When the adjustment cost is equal to 10% (i.e., when the share of property value that households must pay when they move is equal to 0.1), the property value effect is around 1.6 times smaller than the home equity effect. When the adjustment cost is equal to 20%, the property value effect is around 17 times smaller than the home equity effect.
Besides, according to several previous studies, it is reasonable to think that fixed adjustment costs are higher in France than in the US. These costs include a number of different types of costs and fees, such as transfer taxes (e.g. stamp duties, acquisition taxes, etc.), fees incurred when registering the property in the land registry, notary or other legal fees, and real estate agency fees. Andrews et al. (2012) calculate that housing transaction costs represent 10.65% of the property value in France, and 4.25% of the property value in the US. In another study, Laferrère and Leblanc (2007) estimate the average transaction costs at around 14% for France, compared with 10% in the USA.
It is these different levels of fixed adjustment costs in housing that might help explain the quantitative discrepancy of the effect of housing on portfolio choice between France and the US.
Andrews, Dan, Aida C Sánchez, and Åsa Johansson (2011), “Housing and the Economy: Policies for Renovation”, in Economic Policy Reforms 2011: Going for Growth, OECD Publishing.
Calvet, Laurent E, John Y Campbell, and Paolo Sodini (2009), “Fight or Flight? Portfolio Rebalancing by Individual Investors”, Quarterly Journal of Economics, 124(1), 301-348.
Chetty, Raj and Adam Szeidl (2010), “The Effect of Housing on Portfolio Choice”, NBER Working Paper, 15998, National Bureau of Economic Research.
Cocco, Joao F (2005), “Portfolio Choice in the Presence of Housing”, Review of Financial Studies, 18(2), 535-567.
Flavin, Marjorie and Takashi Yamashita (2002), “Owner-Occupied Housing and the Composition of the Household Portfolio”, American Economic Review, 92(1), 345-362.
Fougère, Denis and Mathilde Poulhes (2012), “The Effect of Housing on Portfolio Choice: A Reappraisal Using French Data”, CEPR Discussion Paper 9213, Centre for Economic Policy Research.
Grossman, Sanford J and Guy Laroque (1990), “Asset Pricing and Optimal Portfolio Choice in the Presence of Illiquid Durable Consumption Goods”, Econometrica, 58:1, 25-51.
Heaton, John and Deborah Lucas (2000), “Portfolio Choice in the Presence of Background Risk”, Economic Journal, 110(460), 1-26.
Laferrère, Anne and David Le Blanc (2006), “Housing Policy: Low-Income Households in France”, in Richard Arnott and Daniel P McMillan (eds.) A Companion in Urban Economics, Blackwell Publishing Ltd, 159-178.
Yamashita, Takashi (2003), “Owner-Occupied Housing and Investment in Stocks: An Empirical Test”, Journal of Urban Economics, 53(2), 220-237.