Content area
Abstract
Processing trading signals or views on the market to compute an optimal allocation is one of the main challenges in quantitative portfolio construction. Similarly, embedding stress tests in a risk model in a statistically sound way is key to a healthy risk management process. The generalised Bayesian approach known as entropy pooling, which is laid out in full generality in Meucci (2008), is a flexible framework for processing views and embedding generalised stress tests. The authors introduce an efficient numerical method called factor entropy pooling (FEP), which reduces the dimension of the asset correlation structure using a factor model methodology and selects coordinates such that the optimisation target becomes unconstrained. FEP can be used in large-dimensional problems typical of portfolio construction. An additional area of application of FEP beyond portfolio construction is heavy stress testing, where the market is subject to disruptive potential scenarios and their effect on the portfolio losses is observed.