A BAYESIAN MODEL FOR PORTFOLIO SELECTION AND REVISION

摘要: IN PORTFOLIO ANALYSIS, the basic setting is that of an individual or a group individuals making inferences and decisions in face uncertainty about future security prices related variables. Formal models for decision under require inputs such as probability distributions to reflect maker's events utility functions preferences among possible consequences [30]. Moreover, when series interrelated be made over time, maker should (1) revise his new information obtained (2) take into account effect current on decisions. In terms formal decision-making process, revision can accomplished by using Bayes' theorem interrelationships taken consideration dynamic programming determine optimal Since portfolio selection involves should, insofar possible, incorporate these features. A search extensive literature concerning indicates, however, have ignored one both Markowitz [18] developed original model selection, considerable amount work has been conducted area mathematical analysis, much this summarized Sharpe [31] Smith [33]. Although emphasis analysis primarily single-period multiperiod are investigated Tobin [35], [32], Mossin [21], Pogue [22], Chen, Jen, Zionts [3], Hakansson [13, 14]. addition, general consumption-investment [10, 11, 12], Merton [19], Samuelson [29], Fama [6], Meyer [20]. However, it generally assumed interest completely specified they unaffected information, implying do not involve time. Bayesian received virtually no attention literature. Mao Siirndal [17] present simple, discrete, which returns from securities level business activity conditions. Kalymon [16] develops