We develop an algorithm to compute asset allocations for Kahneman and Tversky's (Econometrica, 47(2), 263-291, 1979) prospect theory. An application to benchmark data as in Fama and French (Journal of Financial Economics, 47(2), 427-465, 1992) shows that the equity premium puzzle is resolved for parameter values similar to those found in the laboratory experiments of Kahneman and Tversky (Econometrica, 47(2), 263-291, 1979). While previous studies like Benartzi and Thaler (The Quarterly Journal of Economics, 110(1), 73-92, 1995), Barberis, Huang and Santos (The Quarterly Journal of Economics, 116(1), 1-53, 2001), and Grüne and Semmler (Asset prices and loss aversion, Germany, Mimeo Bielefeld University, 2005) focussed on dynamic aspects of asset pricing but only used loss aversion to explain the equity premium puzzle our paper explains the unconditional moments of asset pricing by a static two-period optimization problem. However, we incorporate asymmetric risk aversion. Our approach allows reducing the degree of loss aversion from 2.353 to 2.25, which is the value found by Tversky and Kahneman (Journal of Risk and Uncertainty, 5, 297-323, 1992) while increasing the risk aversion from 1 to 0.894, which is a slightly higher value than the 0.88 found by Tversky and Kahneman (Journal of Risk and Uncertainty, 5, 297-323, 1992). The equivalence of these parameter settings is robust to incorporating the size and the value portfolios of Fama and French (Journal of Finance, 47(2), 427-465, 1992). However, the optimal prospect theory portfolios found on this larger set of assets differ drastically from the optimal mean-variance portfolio.
