假设金融市场中有两种风险资产，并且每种资产的收益中均含有不可观测项、对应的风险既有系统性风险又有自身特有风险，具有幂效用函数的投资者运用贝叶斯学习方法最优地选取自己的动态投资组合。理论模型与数值分析显示，在一定的投资期限内，对风险资产的投入是风险厌恶程度的减函数；但超过某一时刻，则相反。当风险厌恶程度不变时，对不可观测因素的了解使得长、短期的投资策略不同。而且这个转折点随投资者的风险厌恶程度的增加而减小。风险资产与不可观测因素的相关性与风险的大小成反比例关系。
In this paper, the optimal dynamic portfolio choice is analyzed under Bayesian learning. Assuming that there are two kinds of risky assets, the returns of each asset are predictable by unobservable predictor, and risks of each asset have both systematic and own risks. The empirical results highlight that the proportion of the investor's wealth invested in every risky asset is decreasing with the degree of risk aversion over short horizon, but it is opposite when the investment horizon is long. When the degree of risk aversion is constant, the Bayesian learning makes different about the long and short-horizon investment strategy. This critical point decreases with the degree of risk aversion of investor. The correlation between risk assets and the unobservable predictor is inversely proportional to the size of the risk of asset.
[1] Kolm P N,Tutuncu R,Fabozzi F J,60 Years of portfolio optimization:Practical challenges and current trends[J].European Journal of Operational Research,2014,234(2):356-371.
[2] Xia Yihong. Learning about predictability:The effects of parameter uncertainty on dynamic asset allocation[J]. Journal of Finance,2001,56(1):205-244.
[3] 袁子甲,李仲飞.参数不确定性和效用最大化下的动态投资组合选择[J].中国管理科学,2010,18(5):1-6.
[4] Brandt M W, Goyal A, Santa-Clara P, et al.A simulation approach to dynamic portfolio choice with an application to learning about return predictability[J].The Review of Financial Studies,2005,18(3):830-874.
[5] 杨朝军,陈浩武.参数不确定性对投资者最优资产组合的影响:基于中国的实证[J].中国管理科学,2008,16(3):37-44.
[6] 陈志英.状态变化和学习行为下的最优资产组合选择[J].管理科学,2013,26(2):81-90.
[7] Ando T,Bayesian portfolio selection using a multifactor model[J].International Journal of Forecasting,2009,25(3):550-566.
[8] Virbickaite A, Ausin M C, Galeano P. A Bayesian non-parametric approach to asymmetric dynamic conditional correlation model with application to portfolio selection[J].Computational Statistics and Data Analysis,2016,100:814-829.
[9] Zhu Shushang, Fan Minjie,Li Duan.Portfolio management with robustness in both prediction and decision:A mixture model based learning approach[J].Journal of Economic Dynamics & Control,2014,48:1-25.
[10] Fama E F,French K R.Permanent and temporary components of stock prices[J]. Journal of Political Economy, 1988,96(2):246-273.
[11] Boudoukh J,Michaely R,Richardson M,et al.On the importance of measuring payout yield:Implications for empirical asset pricing[J].Journal of Finance, 2007,62(2):877-916.
[12] Drechsler I,Yaron A. What's vol got to do with it[J]. Review of Financial Studies,2011,24(1):1-45.
[13] Ang A, Bekaert G.Stock return predictability:Is it there?[J].Review of Financial Studies,2007,20(3):651-707.
[14] Lettau M,Van Nieuwerburgh S.Reconciling the return predictability evidence[J]. Review of Financial Studies,2008,21(4):1607-1652.
[15] Brennan M J.The role of learning in dynamic portfolio decisions[J].European Finance Review, 1998,1(3):295-306.
[16] Branger N, Larsen L S,Munk C.Robust portfolio choice with ambiguity and learning about return predictability[J].Journal of Banking & Finance,2013,37(5):1397-1411.
[17] Liptser R S,Shiryaev A N.Statistics of random processes[M]. New York-Hedelberg:Springer-Verlag,1978.
[18] Jacquier E. Polson N,Bayesian methods in finance[M]//Geweleke J,Koop G,Van Dijk H.The Oxford handbook of Bayesian econometrics.Oxfor:Oxford University Press,2011.
