Most TI discussion papers can be downloaded at http://www.tinbergen.nl.

This paper analyzes the optimal investment strategy for loss averse investors, assuming a complete market and general Ito processes for the asset prices. The loss-averse investor follows a partial portfolio insurance strategy. When the investor's planning horizon is short (less than 5 years), he or she considerably reduces the initial portfolio weight of stocks compared to an investor with smooth power utility. The empirical section of the paper estimates the level of loss aversion implied by historical U.S. stock market data, using a representative agent model. We find that loss aversion and risk aversion cannot be disentangled empirically. © 2004 President and Fellows of Harvard College and the Massachusetts Institute of Technology.

We derive empirical tests for the stochastic dominance e⁄ciency of a given portfolio with respect to all possible portfolios constructed from a set of assets. The tests can be computed using straightforward linear programming. Bootstrapping techniques and asymptotic distribution theory can approximate the sampling properties of the test results and allow for statistical inference. Our results could provide a stimulus to the further proliferation of stochastic dominance for the problem of portfolio selection and evaluation. Using our tests, the Fama and French market portfolio is signi¢cantly ine⁄cient relative to benchmark portfolios formed on market capitalization and book-to-market equity ratio.

No abstract