Working Paper: CEPR ID: DP4789
Authors: Massimo Massa; Andrei Simonov
Abstract: We exploit the restrictions of intertemporal portfolio choice in the presence of non-financial income risk to design and implement tests of hedging that use the information contained in the actual portfolio of the investor. We use a unique dataset of Swedish investors with information broken down at the investor level and into various components of wealth, investor income, tax positions and investor demographic characteristics. Portfolio holdings are identified at the stock level. We show that investors do not engage in hedging, but invest in stocks closely related to their non-financial income. We explain this with familiarity, that is, the tendency to concentrate holdings in stocks to which the investor is geographically or professionally close or that he has held for a long period. We show that familiarity is not a behavioural bias, but is information-driven. Familiarity-based investment allows investors to earn higher returns than they would have otherwise earned if they had hedged.
Keywords: asset pricing; hedging; portfolio decision
JEL Codes: G11; G14
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Investors do not hedge against nonfinancial income risk (G19) | Investors tilt their portfolios towards stocks correlated with their nonfinancial income (G11) |
| Familiarity (Y20) | Investors tilt their portfolios towards stocks correlated with their nonfinancial income (G11) |
| Familiarity (Y20) | Investors earn higher returns (G11) |
| Investors tilt their portfolios towards familiar stocks (G11) | Overall portfolio choice is skewed towards familiar stocks (G11) |
| Less informed investors (G14) | More influenced by familiarity (C92) |
| Familiarity-based investment strategies (G11) | Higher returns compared to hedging (G19) |