Working Paper: NBER ID: w1625
Authors: Stanley Fischer; George Pennacchi
Abstract: Optimal portfolios differ according to the length of time they are held without being rebalanced. For the case in which asset returns are identically and independently distributed, it has been shown that optimal portfolios become less diversified as the holding period lengthens.We show that the anti-diversification result does not obtain when asset returns are serially correlated, and examine properties of asymptotic portfolios for the case where the short term interest rate, although known at each moment of time, may change unpredictably over time. The theoretical results provide no presumption about the effects of the length of the holding period on the optimal portfolio. Using estimated processes for stock and bill returns, we show that calculated optimal portfolios are virtually invariant to the length of the holding period. The estimated processes for asset returns also imply very little difference between portfolios calculated ignoring changes in the investment opportunity set and those obtained when the investment opportunity set changes over time.
Keywords: Optimal Portfolios; Asset Returns; Serial Correlation; Financial Markets
JEL Codes: G11; G12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Holding period (C41) | Optimal portfolio composition (G11) |
| Serial correlation (C29) | Optimal portfolio composition (G11) |
| Holding period + Serial correlation (C41) | Portfolio diversification (G11) |
| Holding period + Serial correlation (C41) | Riskiness of asset returns (G17) |
| Holding period (C41) | Portfolio diversification (G11) |