Factor models and factor-based investing are changing the way institutional investment managers construct portfolios and analyze risk. This Insight discusses how factor models enable better portfolio risk assessments and how they are implemented in Charles River’s Portfolio Analytics solution. In a future Insight, we’ll explore the growing adoption of factor-based funds and how factor models help asset managers construct smarter portfolios.

Risk factors are granular attributes of one or more asset classes that explain risk and return. Examples of equity factors are company size (large cap vs. small cap) and style (growth vs. value). Fixed income factors include interest rate, credit and prepayment risk. Factors also include macroeconomic variables such as inflation and commodity prices that impact multiple asset classes.

Investment portfolios are typically constructed using asset classes as building blocks. Intuitively, fixed income should produce a different risk and return profile than equities or commodities. By combining multiple asset classes in a portfolio, managers try to ensure risk diversification.

But in reality, asset classes often share significant overlapping risk exposures. Diversification in name only (DINO) has been a long running criticism of traditional asset-class based portfolio construction and risk management. The DINO effect becomes especially apparent during market crises, when seemingly disparate asset classes suddenly exhibit significantly correlated price movement.

By constructing portfolios based on risk factors instead of asset classes, managers can potentially build more efficient portfolios that require less risk to achieve competitive returns. This also helps minimize overexposure to a particular risk factor within the portfolio.

A factor based approach removes the artificial constraints of asset class definitions, helping managers focus on risk drivers across their entire portfolio. This ensures greater flexibility when making de-risking and hedging choices. For example, multiple assets (commodities, airline stocks and the Russian Ruble) can all be impacted by a particular risk exposure (oil prices). The manager can lower overall portfolio risk by selling the most liquid of the three assets, rather than having to sell a less liquid asset at fire sale prices.

Factor models are an effective and versatile tool for analyzing current portfolio risk exposures, projecting future (ex-ante) risk, and attributing manager performance.

Risk decomposition helps explain the types of risk in a portfolio and quantify how much of each risk is present in the portfolio. The ability to decompose risk exposures into factors is especially useful when vetting managers for a fund-of-funds offering, where portfolio level transparency is lacking or investment strategies are not disclosed due to confidentiality concerns. Factor based decomposition provides an independent validation of managers claiming to offer uncorrelated strategies, and quantifies the actual degree of risk exposure overlap with other funds.

Factor models are also used to create risk forecasts that help guide a portfolio manager’s de-risking and allocation decisions. Forecasts can be generated for portfolio volatility and tracking error. In a third use of factor models, portfolio stress testing, managers can incorporate the most relevant model-derived factors in order to create more realistic and robust stress tests of their portfolios.

Performance attribution explains the sources of active portfolio returns.  A Return-based attribution framework (such as Brinson) is an industry standard approach that measures the sources of active return due to allocation decisions when investing in any single category, such as sector, industry or country. It also helps managers understand the impact of their security selection decisions within each of the allocated categories.

Factor-based attribution extends the analysis, by further assessing the impact of securities’ fundamental characteristics/factors such as momentum and style on overall active return of the portfolio. It provides insight into how each asset characteristic contributed towards their security selection decision. The factors used for performance attribution can also be used for risk attribution, allowing managers to analyze portfolio returns on a risk-adjusted basis.

Charles River supports three complementary approaches to factor modeling that portfolio managers can use to assess risk exposures at the security, asset class, and portfolio level. Models can be constructed for both single- and multi-asset portfolios and for investment horizons ranging from days to years.

Fundamental: This approach is based on decades of academic research into asset price behavior.  Fundamental models decompose risk using a few easily understood elements such as the previously mentioned equity and fixed income factors. Intuitively appealing, fundamental models are ideal for performance attribution and portfolio risk decomposition. However, since fundamental models are defined in terms of known factors, they can become less accurate when new, unknown factors emerge.

Statistical: Statistical models use a numerical technique called principal components analysis (PCA), to calculate risk exposures, without requiring any assumptions about what those factors represent.  PCA based models adapt readily to changing trends and correlations in the underlying markets, unlike fundamental models that utilize fixed factors. This makes statistical models especially useful during times of market stress when historical asset correlations break down. However, the factors derived from statistical models lack the easy interpretability of fundamental models.

Macroeconomic: Charles River’s macroeconomic model calculates risk exposures using observable economic time series such as inflation, industrial production, and the U.S. Dollar exchange rate as factors. This model augments fundamental and statistical factor models by gauging the impact of economic conditions on a portfolio. Macroeconomic models are primarily used for portfolio stress testing and performance attribution.

Charles River’s open architecture enables firms to incorporate bespoke and third-party factor models.  Unlike black box systems that limit investment managers to a particular model, Charles River’s approach lets them assess and manage risk using the models best suited to their mix of asset classes and investment styles. As factor modeling evolves to cover new asset classes, instruments and risk factors, this flexibility will enable asset managers to incorporate new models with minimal effort.

The growth of factor-based investing parallels the increasing importance of risk management throughout the investment process. The ability to decompose complex portfolios into intuitive risk factors is an important innovation that provides managers, investors and regulators with a clearer understanding of risk exposures and risk-adjusted performance. Charles River’s extensive and flexible support for factor models enables firms to implement more rigorous risk management practices, understand and communicate their performance in the context of risks taken, and construct more robust portfolios, a topic we’ll discuss in our next Insight.