Controlling Idiosyncratic Risk in Value Strategies

• Over the last decade, much of the underperformance of value strategies, as proxied by the MSCI Enhanced Value Index, can be attributed to idiosyncratic — i.e., stock-specific — risks.
• By disconnecting the weight of each stock from its market cap, we removed the outsized impact of a handful of mega-cap stocks with large weights in the benchmark. This reduced the stock-specific risk and increased return.
• Using optimization to control stock-specific risk, we were able to mitigate some of the negative stock-specific impact on value’s performance without negative effects on value-factor exposures or contributions to the returns.

What accounted for the vastly different performance of value strategies over the past two decades? More specifically, why did they underperform from 2011 to 2020? In previous research, we detailed the value factor’s poor performance over the period and the negative impact of exposure to nonvalue factors. In this blog post we focus on the third and final culprit: stock-specific impact.

Stock-Specific Return Was a Drag for Value Strategies over the Last Decade

From 2011 to 2020, value strategies, as proxied by the MSCI World Enhanced Value Index,¹ underperformed the MSCI World Index by 4.1% annually. A significant portion (2.3 percentage points) was attributed to idiosyncratic, or stock-specific return. This means 2.3% of the underperformance of the index could not be explained by any of the factors in our factor model and was mainly due to the stocks that were included or excluded from the index and their active weights compared to the benchmark.

Performance Attribution of MSCI World Enhanced Value Index

Monthly data from Dec. 31, 2010, to Dec. 31, 2020.

Was Underperformance a Tech Issue?

As of the end of 2020, the 10 largest U.S. companies accounted for about 25% and 17% of the weight in the MSCI USA and MSCI World Indexes, respectively. In How Diversified Are US Equity Investors? we saw that this handful of mega-cap, mostly information-technology stocks made an outsized contribution to the risk and return of the MSCI USA Index. Of these 10 stocks, eight were never part of the MSCI World Enhanced Value Index over the last decade, and the other two appeared for short periods. The questions then are: Did the absence of these (relatively high-valuation) stocks with their large size and outsized performance contribute to the underperformance of value strategies over the last decade? And does this show up in the attribution analysis as stock-specific return?²

To investigate, we remove the outsized impact of these mega-cap stocks by using the MSCI World Equal Weighted Index as the benchmark and as the parent index to construct a hypothetical value index. The MSCI World Equal Weighted Index has all the constituents of the MSCI World Index, but they are given the same weights.

The exhibit below breaks down the return of this simulated value index to its underlying contributors. The stock-specific return and risk were lower (-1.4% and 1.6%) compared to the MSCI World Enhanced Value Index (-2.29% and 2.46%). By disconnecting the weight of each stock from its market cap, we removed the outsized impact of the handful of mega-cap stocks, which reduced the stock-specific risk and improved return.

Performance Attribution of World Enhanced Value on Equal Weighted Index

Monthly data from Dec. 31, 2010, to Dec. 31, 2020.

One Way to Control Stock-Specific Risk

Having established that some of the negative stock-specific contribution to the return of the MSCI World Enhanced Value Index stemmed from the massive weight of mega-caps in the parent index, how we can limit the impact of stock-specific risk and return?

To control the stock-specific risk/return contributions, we first need to measure the stock-specific risk. We used the MSCI Global Equity Model for Long-Term Investors (GEMLT), which allows us to directly access factor exposures as well as stock-specific risk, and MSCI’s Barra^® Optimizer to formulate the hypothetical index’s construction.

In the optimization used to construct this hypothetical value index, which we call the World Optimized Value Index, we aim to maximize exposure to the value factor and minimize stock-specific risk. There are various ways to structure optimization to achieve these two objectives simultaneously. We use an intuitive and simple implementation and simulate an index that maximizes the ratio of the portfolio’s value exposure to its stock-specific risk.

The exhibit below compares the characteristics of this simulated index to its parent, the MSCI World Index, as well as to the MSCI World Enhanced Value Index. Over the longer history the optimized value index’s performance was in-line with the MSCI Enhanced Value Index, but during the more recent decade, where value strategies underperformed, the simulated index outperformed the MSCI Enhanced Value Index.

The number of constituents for the optimized index has been about twice the number for the MSCI Enhanced Value Indexes, suggesting that the optimized index has tried to achieve some of its stock-specific risk reduction by targeting a higher level of diversification.

Impact of Controlling Stock-Specific Risk

* Gross returns annualized in USD ** Annualized one-way index turnover over index reviews *** Monthly averages

It is important to understand whether the difference in performance between the two value indexes was the result of the control applied to the stock-specific risk. The table below breaks down these indexes’ risk/return and shows which factors contributed to the performance of each over the last decade. The stock-specific contribution to both risk and return were lower in the optimized index, which highlights the impact of controlling stock-specific risk. Additionally, we were able to lower this risk without a negative effect on the exposure and contribution of value factors.

Stock-Specific Performance Attribution for the Optimization Case

Monthly data from Dec. 31, 2010, to Dec. 31, 2020.

Idiosyncratic Risk Reduced but Not Eliminated

While value strategies, as proxied by the MSCI Enhanced Value Index, remained an effective way to capture value, the results above suggest that some of the negative stock-specific impact on value’s performance over the past decade could have been mitigated by directly controlling stock-specific risk (though its contribution remained negative and significant) without negatively affecting value factors’ exposures or contributions.

¹The MSCI World Enhanced Value Index follows a methodology that selects a subset of its parent index (MSCI World Index) with the best value scores (cheapest) and weights each stock according to its market capitalization as well as its valuation score.

²Attribution for the MSCI Enhanced Value Index is analyzed relative to its market-cap benchmark, the MSCI World Index, which included these mega-cap stocks. They were not part of the MSCI Enhanced Value Index, which, relative to the benchmark, was underweight these stocks.

Further Reading

Purifying Value

The Theory of (Value) Relativity

Value-Performance Anxiety

Bringing Value to the 21st Century