In this Research Insight, we use the Barra Emerging Markets Model (EMM1) and the Barra Global Equity Model (GEM3) to attribute the returns of a representative set of emerging market portfolios.  We show that by aligning the estimation universe with the investment universe, the EMM1 model provides a more accurate and meaningful description of emerging market portfolios.

Using the lens of the Barra US Equity Model (USE4S), this Research Insight provides a practical guide to constructing investable factor portfolios. This paper begins by discussing the general concept of a factor portfolio. We then explore the role of optimization in making a 'pure factor portfolio' investable. We assess how investability constraints impact the performance of factor-replicating portfolios. Finally, we discuss how MSCI Market Neutral Barra Factor Indexes can be used in an...

In this Research Insight, we present a methodology for efficiently combining multiple sources of alpha when constructing a portfolio. The first part of our study shows that the most efficient implementation for a single source of alpha is the minimum-volatility factor portfolio, which has the lowest risk for a given level of expected return and, therefore, the maximum expected information ratio.  &In the second part of our study, we examine how to efficiently combine multiple sources...

In the MSCI Japan Equity Model (JPE4), we include some well-known Systematic Equity Strategies as risk factors (SES factors, for short). Incorporating these SES factors can help identify and measure risk in investment strategies typically used by fundamental and quantitative managers.  In this paper, we find that models including the SES factors produced more accurate risk forecasts for portfolios tilted toward those investment strategies. Furthermore, for optimized portfolios tilting on...

In an introductory paper, we explained Systematic Equity Strategies (SES) and how they can be used as factors in a risk model.  In this paper, we use data from the Japan equity markets to define seven new SES factors and study their empirical behavior.  Our findings illustrate the important role that these factors play in portfolio construction and risk management. Our study also shows problems associated with omitting these factors from a risk model, and explain why models that...

In this Research Insight, we outline the building blocks essential to constructing an effective standard risk model. We then turn to how risk models are used in the investment process — that is, constructing efficient portfolios and attributing their risk and return. Finally, we describe the best practices of proprietary model construction, including an empirical investigation of the economic impact of using proprietary models in portfolio optimization.

For years, practitioners have debated the benefits of using fundamental versus statistical models. In this Research Insight, we argue that the two approaches to risk modeling are complementary, not mutually exclusive. To support our reasoning, we provide a case study that demonstrates how the Barra North America Stochastic Factor Model (NAMS1) and the Barra US Equity Model (USE4) can work in concert to uncover hidden sources of risk.

Portfolio managers have long worried that discrepancies between risk and alpha factors may somehow detract from the performance of their optimized portfolios. This paper presents a comprehensive overview of alpha-risk factor alignment and its consequences, showing how penalizing the residual alpha may help reduce the unintended bets resulting from misalignment. However, we also illustrate that correcting for misalignment may not always be necessary and can sometimes be...

Following the “quant meltdown” of August 2007, market observers became concerned that quant strategies were leading to crowded trades. This paper analyzes the impact that a risk model used in portfolio construction has on manager crowding by identifying the drivers of crowding and by illustrating their impact.  A risk model’s effect on manager crowding depends, in part, on how alphas used by different managers are related to each other, and to the risk model factors. We...

Many portfolio managers use multi-factor models, but not all factor models are equally effective in forecasting risk.

Portfolio managers have long suspected that the risk forecast of an optimized portfolio tends to be optimistic. Many have identified the culprit as estimation error in the covariance matrix. Forecasts based on historical asset covariance matrices are particularly sensitive to this error. The bias is reduced dramatically by using a factor model. Even so, factor models still tend to under-forecast the risk of optimized portfolios, especially the risk coming from factors. In this paper, we show...

Portfolio managers have long suspected that the risk forecast of an optimized portfolio tends to be optimistic. Many have identified the culprit as estimation error in the covariance matrix. Forecasts based on historical asset covariance matrices are particularly sensitive to this error. The bias is reduced dramatically by using a factor model. Even so, factor models still tend to under-forecast the risk of optimized portfolios, especially the risk coming from factors. In this paper, we show...

The analysis of a possible positive relationship between economic growth and stock market returns is interesting both theoretically and practically. Investors often wonder if they should assign higher weight to countries with higher economic performance, hoping that economic growth will eventually show up in equity returns. Although this relationship seems quite intuitive, historically long-run stock price growth has fallen short of GDP growth in many countries. In this bulletin, we use...

In this study, our goal is to adapt mean-variance optimization to produce active portfolios with less exposure to extreme losses than normal optimized portfolios. Specifically, we illustrate how extreme risk can be incorporated into portfolio construction in a straightforward way by constraining the shortfall beta of the optimal portfolio. Our simple empirical examples suggest that constraining shortfall beta may offer some downside protection in turbulent periods without sacrificing...

When there is noise in a covariance matrix, portfolio optimization tends to produce portfolios for which the risk forecasts are underestimates of the true risk. In this paper, we take a closer look at the connection between estimation error and the underestimation of the risk of optimized portfolios. We pay special attention to the case in which returns have a known factor structure. There, the bias in optimization can be reduced dramatically by using a covariance matrix based on a factor...

This paper analyzes the impact of constraints on portfolio return and risk, extending the insights of previous research in this area. We show that constraints move a manager's portfolio away from the optimal unconstrained portfolio in two ways. First, they may rein in or increase the risk of the portfolio without impairing its information ratio. Second, they may force the portfolio to take unwanted bets that incur risk but yield no return. As a result, a constrained portfolio consists of...

Misalignment between alpha and risk factors may create unintended bets in optimized portfolios, as shown analytically in Lee and Stefek (2008).  In a March research insight, we introduced a way to mitigate this issue by penalizing the portion of the alpha not related to the risk factors, the 'residual alpha.' Here, we further illustrate this method with two signals commonly used by portfolio managers. The potential improvement from this method depends on the strategy in question, in...

The misalignment of alpha and risk factors may result in inadvertent and unwanted bets that may hamper performance. Lee and Stefek (2008) show that better aligning risk factors with alpha factors may improve the information ratio of optimized portfolios. They propose four ways of modifying a risk model to reduce misalignment. Here, we discuss one way to mitigate these problems by modifying the optimization process, itself. The objective function is modified to include a penalty term on the...

Portfolio managers worry that discrepancies between risk and alpha factors may create unintended biases in their optimized portfolios. We analyze the ramifications of using different factor models of risk and alpha in portfolio optimization and show that aligning risk factors with alpha factors may improve the information ratio of optimized portfolios, even if doing so lowers the overall accuracy of risk forecasts. We discuss ways of modifying a risk model that may help remedy these problems.