Central Limit Theorem to Approximate Aggregate Risk of Portfolio: Using the ModelRisk Software

In this note, the non-sampling information in portfolio management is considered. These information may be the past belief of investor about a special asset. They are characterized as the correlated binary random variables. Then, the Monte Carlo is applied to derive the posterior distribution of binary variables given the past returns which indicates the tendency of investor to keep or drop a portfolio via using the non-sampling and sampling information simultaneously. The posterior distribution of belief of investor and the accuracy of Bayesian method are shown via plotting histograms.