By Scott Warner
Since the financial crisis of 2008, active equity investors have bemoaned the persistently high pairwise correlation between stocks. This market condition has become a common excuse for lackluster alpha generation and has caused many talented fundamental stock pickers to throw their hands up in frustration. While we think high correlations are a significant challenge for equity investors, we believe there are a number of useful approaches that can help mitigate the negative impact of high correlations.
As a quick refresher, pairwise correlation between stocks measures the degree to which prices of stocks move together. In recent years the average correlation between all S&P 500 constituents, and within many other international equity markets, have been unusually elevated. Despite a decline earlier this year, correlations have risen dramatically in recent weeks and are close to historically high levels. This dynamic has adversely impacted many active equity managers, as gains from positive security selection have been less pronounced than in other market environments. Correctly forecasting which stocks will perform well or badly has been less rewarding than in other market environments.
|S&P 500 Correlations - 50 Largest-Cap S&P Stocks (June 2011 to June 2012) |
Source: Credit Suisse Derivatives Strategy. Note: correlation is for top 50 stocks in SPX.
Investment research supports the view that correlation levels matter for generating returns from stock picking, as evidenced in a white paper by Morgan Stanley.1 Morgan Stanley found strong empirical evidence that managers showed idiosyncratic "stock picking" skill in the HFRI Equity Hedge strategies during what Morgan Stanley defined as “low correlation states.” During this type of market environment, stock picking skills were additive and amounted to approximately +0.87% per month. This contrasts the additive nature of the same "stock picking skill" during their self-defined "high correlation states," which amounted to a much lower return of approximately +0.25% per month.
These findings are similar to those found by the Credit Suisse Equity Derivatives Strategy Group.2 In a variant study, Credit Suisse attempted to quantify the opportunity set (i.e., the return generated by a perfect stock-picker) by calculating the spread between the 50 best and 50 worst performing stocks within the S&P 500 on a monthly basis. Upon mapping the volatility-adjusted opportunity set against realized correlations, they found that there is materially less reward for stock picking when correlations are high. Their findings showed that, on average, the availability of stock-picking alpha in high-correlation markets is about one-third of that in low correlation markets.
Our main conclusion from both of these research papers, and our own experience in evaluating managers, is that in high-correlation environments the reward for positive security selection appears to be reduced but not eliminated.
Toolkit for a Strategic Response to Higher Correlations
While many investors have greeted each year since the crisis with cautious optimism that correlations would decline, we at PAAMCO are a bit less optimistic. Timing a prolonged decline in pairwise correlations is a difficult task at best, and while we would be quite happy to usher in a new regime of lower correlations, we are not counting on it. As the expression goes, "hope is not a strategy."
Instead, the following concepts could be incorporated into a strategic response for institutional investors to mitigate the negative impact of investing in a high correlation environment.
Ditch the Benchmark: The amount of active risk traditional long-only managers can take is often dictated by the constraints of their mandate. Long-only managers are simultaneously tasked with trying to minimize tracking error while still producing profits from security selection. This constraint can handicap a manager’s ability to incorporate idiosyncratic risk in the portfolio. This is particularly acute in high correlation environments. Our view is that long/short equity managers who are less benchmark-constrained than long-only managers have the ability to more freely take idiosyncratic bets and amplify any profits from security selection.
Focus on Events: Another way to break the “correlation curse” is to focus on events, as corporate actions are often helpful in breaking the correlation regime. Merger arbitrage, spin-offs, asset sales, restructurings, and other hard events are inherently highly idiosyncratic in nature.
Care about Structure: Custom mandates and co-investments can be utilized to add highly idiosyncratic exposures to a portfolio mix. Finding ways to amplify the idiosyncratic risk of a portfolio in a fee-efficient manner may improve the ability of an investor to capture returns from security selection.
Diversify: While correlations may be elevated in one market, diversification may be a useful tool to seek out exposure to markets or market segments that are moving in a less correlated fashion. As an example, style factors such as value and growth showed little correlation last year despite elevated correlations in broader equity markets.
Be Patient: High correlations can often mask highly divergent business fundamentals. As a result, after a period of high correlations, valuation spreads are often quite stretched. Having patience and conviction on fundamental analysis can help to eventually realize security selection alpha as these stretched spreads converge to more normal levels. In essence, high correlations can create an "alpha backlog" which may take time to realize.
Given our view that high correlations may persist for the foreseeable future, we believe that investors should be prepared to deal with them head-on. While high correlations are clearly a headwind to security selection-driven "alpha," a carefully thought out strategic response can help mitigate its impact. As long as markets continue to be headline-driven, investors should be ready to take a hard look at how they are allocating their active equity exposure and consider steps to better deal with persistently high correlations in equity markets.
1. Baesel, Jerome, et. al., “The Effect of S&P 500 Correlation on Hedge Fund Alpha,” Morgan Stanley Alternative Investment Partners, April 2011.
2. Tom, Edward K, et. al., “Volatility Outlook: Quantifying the Cost of Correlation,” Credit Suisse Equity Derivatives Strategy, November 2010.
Scott Warner is a director and sector specialist for the long/short equity strategy at PAAMCO where he is focused on the evaluation and management of North American long/short equity managers. He also serves as the portfolio manager and main point of contact for certain institutional investor relationships and is responsible for portfolio construction of equity-focused customized solutions.
This document contains the current, good faith opinions of the authors but not necessarily those of Pacific Alternative Asset Management Company, LLC (“PAAMCO”). The document is meant for educational purposes only and should not be considered as investment advice or a recommendation of any type. This document may contain forward-looking statements. These are based upon a number of assumptions concerning future conditions that ultimately may prove to be inaccurate. Such forward-looking statements are subject to risks and uncertainties and may be affected by various factors that may cause actual results to differ materially from those in the forward-looking statements. Any forward-looking statements speak only as of the date they are made and PAAMCO assumes no duty to and does not undertake to update forward-looking statements.