KEY POINTS

• Multifactor equity risk models provide detailed insight into the structure and properties of portfolios. These models characterize stock returns in terms of systematic factors and an idiosyncratic component. Systematic factors are generally designed to have intuitive economic interpretation and they represent common movements across securities. On the other hand, the idiosyncratic component represents the residual return due to stock-specific events.
• Systematic factors used in equity risk models can be broadly classified under five categories: market factors, classification variables, firm characteristics, macroeconomic variables, and statistical factors.
• Relative significance of systematic risk factors depends on various parameters such as the model horizon, region/country for which the model is designed, existence of other factors, and the particular time period of the analysis. For instance, in the presence of industry factors, macroeconomic factors tend to be insignificant for short to medium horizon equity risk models whereas they tend to be more significant for long-horizon models. On the other hand, for developed equity markets, industry factors are more significant as compared to the country factors. The latter are still the dominant effect for emerging markets.
• Choice of the model and the estimation technique affect the interpretation of factors. For instance, in the existence of a market factor, industry factors represent industry-specific movements net of market. If there is no market factor, their interpretation is very close to market value-weighted industry indexes.
• Multifactor equity risk models can be classified according to how their loadings and factors are specified. The most common equity factor models specify loadings based on classification (e.g., industry) and fundamental or technical information, and estimate factor realizations every period. Certain other models take factors as known (e.g., returns on industry indexes) and estimate loadings based on time-series information. A third class of models is based purely on statistical approaches without concern for economic interpretation of factors and loadings. Finally, it is possible to combine these approaches and construct hybrid models. Each of these approaches has its own specific strengths and weaknesses.
• A good multifactor equity risk model provides detailed information regarding the exposures of a complex portfolio and can be a valuable tool for portfolio construction and risk management. It can help managers construct portfolios tracking a particular benchmark, express views subject to a given risk budget, and rebalance a portfolio while avoiding excessive transaction costs. Further, by identifying the exposures where the portfolio has the highest risk sensitivity it can help a portfolio manager reduce (or increase) risk in the most effective way.
• Performance attribution based on multifactor equity risk models can give ex post insight into how the portfolio manager's views and corresponding investments translated into actual returns.
• Factor-based scenario analysis provides portfolio managers with a powerful tool to perform stress testing of portfolio positions and gain insight into the impact of specific market events on portfolio performance.