Working Paper: CEPR ID: DP3890
Authors: Loriana Pelizzon; Guglielmo Weber
Abstract: In this Paper we argue that standard tests of portfolio efficiency are biased because they neglect the existence of illiquid wealth. In the case of household portfolios, the most important illiquid asset is housing: if housing stock adjustments are costly and therefore infrequent, we show how the dynamic optimization problem produces optimal portfolios in periods of no adjustment that are affected by housing price risk (through a hedge term). When the housing stock is not adjusted, we argue that tests for portfolio efficiency of financial assets must then be run conditionally upon housing wealth. In our application, we use Italian household portfolio data from SHIW 1998 and time series data on financial asset and housing stock returns to assess whether actual portfolios are efficient. We first consider purely financial portfolios and portfolios that also treat the housing stock as another asset. We then consider the consequences of treating the housing stock as given and test for efficiency in this framework. Our empirical results support the view that the presence of illiquid wealth plays an important role in determining whether portfolios chosen by home-owners are efficient.
Keywords: efficiency; housing; portfolio choice
JEL Codes: D91; G11
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| housing price risk (R31) | financial asset allocation (G11) |
| illiquid assets (housing) (G59) | optimal portfolio choice (G11) |
| housing stock adjustments (infrequent and costly) (R21) | dynamic optimization problem (C61) |
| housing returns correlation with financial asset returns (G19) | mean-variance efficiency (D61) |
| presence of illiquid wealth (G19) | portfolio efficiency of homeowners (G59) |
| correlation between housing and financial asset returns (G59) | efficiency of household portfolios (G59) |
| housing wealth considered in portfolio optimization (G59) | mean-variance efficiency (D61) |