Chapter 9Factor Models

This chapter introduces factor models—equations that establish links between security returns and their lagged values or exogenous variables. The latter variables are referred to as factors. The theoretical foundations of factor models stem from the idea that a part of the security's risk is shared by groups of other securities (referred to as systematic risk), while another part is unique to the individual security (referred to as idiosyncratic, nonsystematic, or specific risk).

Factor models are widely applied in the practice of investment management, and this chapter outlines some of their main uses. First introduced in the context of solving practical problems with portfolio optimization, factor models simplify portfolio optimization routines and make their input estimation more stable and computationally efficient. They have now broader applications in helping portfolio managers isolate and manage sources of risk (beta) and excess return (alpha) for the portfolio, and are used in portfolio performance attribution and stress testing.1 They are also used in forecasting for implementing active portfolio management strategies.

In order to provide context, we begin this chapter by introducing the classical pricing models from the financial literature. We then discuss practical aspects of the estimation and use of factor models in industry. We review statistical techniques for building factor models as well as the process of selecting the right type of ...

Get Portfolio Construction and Analytics now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.