Lost In The Crowd? Identifying And Measuring Crowding Strategies And Trades by MSCI
Mehmet K. Bayraktar, Stuart Doole, Altaf Kassam, Stan Radchenko
The “quant meltdown” of August 2007 and the subsequent unfolding of the global financial crisis highlighted the risks of crowded investment strategies. The rapid growth of “smart beta” indexes and their use in ETFs has added to the need for scrutiny. Accounting for crowding risk is necessary for any investment strategy because it may explain a substantial portion of strategy risk and performance during certain periods, especially during times of excessive market volatility.
In this paper, we propose a set of four key metrics — our “MSCI Crowding Scorecard” — for monitoring and detecting the crowding risk of any investment strategy. This work builds on our innovative analysis of historical behaviors of investment strategies and MSCI’s next generation equity risk models which incorporate Systematic Equity Strategies (SES):
MSCI Crowding Scorecard
- Mutual Fund Trading Activity
- Hedge Fund Trading Activity
- Pair-wise Correlations (“co-momentum”)
- Valuation Dispersion
The first two measures help capture crowding in the trading activity of various market participants, such as Value and Growth managers, and pinpoint the overlap in trading activity of otherwise heterogeneous investors. The last two measures capture the pricing and valuation impacts of such trading activity. Both sets of measures are essential in developing a Crowding Scorecard.
Finally, using the Crowding Scorecard, we find there were reasons to be moderately concerned about crowding in the U.S. Momentum factor as of the end of 2014.
The MSCI Crowding Scorecard can also be applied to single stocks, indexes and active strategies, making it an important tool for investment and risk managers following both quantitative and fundamental strategies — including recently popular factor index approaches. Using this approach can help managers understand the risks of overlap in trading strategies that may not be apparent by focusing on one of these metrics alone.
Our next areas of research are to extend the analysis to a global universe and to build stock-specific crowding scores.
Lost In The Crowd? Identifying And Measuring Crowding And Trades – Introduction
Since the “quant meltdown” of August 2007 and the subsequent unfolding of the global financial crisis, interest in measuring and monitoring crowding of systematic investment strategies has grown substantially. The rapid growth of smart beta indexes and their use in ETFs has only added to the degree of scrutiny.
A prevailing consensus for the cause of the quant meltdown is that similarities in quantitative equity managers’ holdings and trading styles together with their collective need for liquidity during the crisis led to sharply negative returns for many popular quantitative strategies. These outlier returns confounded the average correlations between strategies that had been historically observed and upon which many quant models’ construction was predicated.
Perhaps an even more striking example of a crowded trade took place during the global financial crisis. This time, the wider investment community “ran for the exits,” pushing the stock prices of financial firms towards zero. The bounce-back of these securities during March 2009 was so significant that it resulted in one of the worst-ever historical performances of the Momentum factor and Momentum-based investment strategies.
By analyzing the historical behaviors of investment strategies, in particular around these events, as well as drawing on existing academic and empirical research, we have developed a set of four crowding metrics, which together we call our “MSCI Crowding Scorecard”:
- Mutual Fund Trading Activity (using mutual fund holdings and trades)
- Hedge Fund Trading Activity (based on short-interest)
- Pair-wise Correlations (“co-momentum”)
- Valuation Dispersion (using price-to-book spreads)
MSCI’s Crowding Scorecard, which can also be applied to single stocks, indexes and active strategies, is an innovative tool for investment and risk managers following both quantitative and fundamental strategies — including recently popular factor index approaches.
The MSCI Crowding Scorecard can help managers understand the risks of overlap in trading strategies that may not be apparent by focusing on one of these metrics alone.
What Are Crowded Trades And Positions?
Crowded trades refer to trading activity involving a significant number of market participants with large pools of capital who trade in and out of stock positions in order to pursue the same, or very similar, investment strategies. A crowded position happens when there is a significant overlap of portfolio positions and allocations as a result of crowded trades which, in total, add up to a significant share of a stock’s free-float market capitalization.
Crowded trades generally result (at least in the short- to medium-term) in improved market efficiency. As a result, the forward-looking (expected) risk-adjusted return of a strategy declines as it becomes more crowded. Crowding thus reduces the future effectiveness of a given investment strategy in predicting stock returns. Depending on the extent of frictions, such as shorting constraints and transactions costs, this overlap of positions among managers may result in extreme levels of risk when those investors experience negative shocks in other parts of their portfolios, forcing them to liquidate their positions (selling what they can, rather than what they would necessarily like to). These “fire sales” may then cause losses for other investors following the same strategy and result in further liquidations, driving stock prices into a downward spiral. The “quant meltdown” is now a classic example of a crowded trade that resulted in significant performance drawdowns.
Crowding risk also affects a wide range of so-called “unanchored” strategies, such as Momentum or Quality, that do not rely on a consistent or independent estimate of fundamental value (Hong & Stein (1999), Stein (2009)). Investors tend to employ reasonable capacity assumptions in pursuing their own strategy, but they may underestimate the aggregate amount of capital following similar strategies. In this case, stock prices may over- or under-shoot their fundamental value and experience a sharp correction in subsequent periods as prices adjust to reflect fundamentals.
See full PDF below.