MSCI Multi-Asset Class (MAC) Factor Model Validation
categories: Research Paper, general, HAYES Michael, VAJDA Balázs, Factor and Risk Modeling, Portfolio Management Analytics, Risk Management Analytics, BarraOne, Multi-Asset Class
tags: validation, mac, multi_asset_class,
The MSCI Multi-Asset Class (MAC) Factor Model introduces several major advances in risk modeling, including systematic MAC strategy factors, a next-generation fixed income model, and improved equity models. This document demonstrates the value of the MSCI MAC Factor Model in forecasting risk, based on (1) visual inspection of the risk forecasts and realized returns, and (2) statistical tests. The MSCI MAC Factor Model is evaluated on an absolute (stand-alone) basis, and is also compared with benchmark models. The context of the validation is the traditional asset allocation problem of multi-asset class investors / asset owners such as pension funds, endowments, insurance companies, and sovereign wealth funds. Previous validation research has analyzed the stand-alone fixed income model performance (Gladkevich and Tokura, 2018) at a level of detail more targeted to asset managers. The stand-alone equity model performance has been studied with more granularity in each component equity model’s research paper (e.g. Morozov et al., 2015, Wang et al., 2018). This paper takes a broader view that spans multiple asset classes and builds from individual market portfolios, with the goal of evaluating the multi-asset class correlations that are at the core of the MSCI MAC Factor Model.