Quantitative vs. Fundamental
Answering quant's critics with empirical analysis.
BY Josef Lakonishok and Bhaskaran Swaminathan | August 23, 2010
The performance of quantitative money managers in the recent market and economic downturn has come under severe criticism.1, 2 This criticism has focused on three points: (a) the quant space is too “crowded” because everyone employs more or less the same alpha factors (value, momentum, quality, etc.), (b) quantitative managers rely too much on similar historical data and statistical methodologies to identify new sources of alpha which therefore tend also to be similar, and (c) as a result, their excess returns have become highly correlated and have been arbitraged away.
This criticism raises three important research questions:
- Is there more crowding among quantitative managers compared to fundamental managers?
- Has the active return-to-active risk ratio (information ratio) of quantitative managers declined over time relative to fundamental managers?
- How different are the style exposures and portfolio characteristics of quantitative and fundamental managers?
We use data from eVestment Alliance to address these research questions and focus only on long-only institutional managers. eVestment Alliance collects monthly data about performance and portfolio characteristics on active money managers who self-report the data. It classifies managers as quantitative or fundamental and by investment style. In determining style classifications, eVestment relies both on self-reporting and its own analysis. The quantitative vs. fundamental distinction, however, is based solely on self-reporting. The database was launched in mid-2000 but data is available as far back as the early 1980s.The database is widely used in manager search and performance measurement by consultants and institutional investors. There is no survivorship bias in the database after 2000 but there is a back-filling bias (managers might back-fill historical data once they start reporting) and a self-reporting bias. There is no reason, however, to believe that these biases affect the data collected for quantitative and fundamental managers differently. Thus, the relative performance between quantitative and fundamental managers is likely to be bias free. Given the various data issues, for this study, we focus on the 1995-2009 time-period although the key results focus on the 2001-2009 time-period, which is likely to be the least problematic.
The study focuses on the U.S. large cap space (value, core, growth, and enhanced index (against S&P 500 and Russell 1000)) and EAFE (value, core, and growth). We pay particular attention to value-weighted performance results, which are representative of the performance of large (by assets under management) money managers, which in turn is more reflective of institutional investor experience. Our main findings are as follows:
Finding 1: The evidence is inconsistent with the crowding argument:
- Assets under management (AUM) by quantitative managers as a percent of total assets under management in the U.S. large cap space (as defined above) have remained stable over the last ten years at around 16%. The number of quantitative firms as a percentage of all firms (quantitative plus fundamental) has also remained stable at around 29%. There is no evidence that there was a huge increase in assets managed by quantitative managers in the U.S. (See Table 1). In EAFE, the AUM by quantitative managers doubled from 12.4% in December 2000 to 26.6% in September 2009. However, this increase is likely due to many U.S. quantitative managers launching EAFE strategies at a later date than fundamental managers. In any event, concerns about crowding in the quantitative space are mostly directed at U.S. strategies not EAFE strategies.
- Average pair-wise correlations in the monthly excess returns (in excess of appropriate benchmarks) of quantitative managers are low and similar to that of the fundamental managers. Moreover, there is no evidence these correlations increased more than that of the fundamental managers during the crucial 2007-2009 time-period (See Table 10).
- There is as much dispersion in the alphas of quant managers as there is among the alphas of fundamental managers both during the 2001-2009 time-period and the more recent 2007-2009 time-period (See Table 11). Note that the generally lower dispersion for quant managers over both full and sub-period is matched by lower median tracking error.
Finding 2: Performance findings based on short time periods (and the resulting high standard errors) have to be interpreted with caution. Nevertheless, we find performance differences between quantitative and fundamental managers depend on the time-period and the investment style:
- There are no noticeable differences in performance between quantitative and fundamental managers in the value space (See Table 2). The evidence, if anything suggests that larger (based on assets under management) quantitative value managers achieved higher information ratios than larger fundamental managers during the 2001-2009 time period as well as during the 2007-2009 time period (value-weighted (VW) results are representative of large manager performance and equal-weighted (EW) results are representative of small manager performance). The differences, however, are unlikely to be statistically significant.
- In the core, growth, and enhanced index spaces, quantitative managers experienced difficulties in the 2007-2009 time-period primarily due to their poor performance in 2009 (Table 3, Table 4, Table 5). These difficulties are due to their somewhat more positive exposure to momentum (see point 3 below) and the extreme negative performance of the momentum factor in 2009 (-83% for the simple price momentum factor which is the worst in its history).3 Quantitative managers in the EAFE space also experienced difficulties in 2009 although they had better performance during the longer 2001-2009 time period (Table 6).
Finding 3: The style exposures based on a four-factor model consisting of market, size, value, and momentum factors (all from Ken French’s website) indicate that there are some small differences between quantitative and fundamental managers in their exposure to momentum. Quantitative managers tend to buy somewhat higher momentum stocks than the fundamental managers. However, in a year with such large negative realization for the momentum factor, even small differences can translate to substantial differences in performance (see Table 7 for style exposures and Table 8 for the style factor returns from 2000 to 2009). In evaluating how quantitative managers in the core, growth, and enhanced index spaces would perform in the future, we have to consider the fact that the chances of similar extreme negative performance for the momentum factor in the near future are likely remote. There is also very little difference in the portfolio characteristics (P/E, P/B, dividend yield, etc.) of quantitative and fundamental managers within each investment style (See Table 9). Differences across style are more significant than differences within style between fundamental and quantitative managers.
A minimalist interpretation of our findings is that not all quantitative managers are the same. There is as much heterogeneity among quantitative managers as there is among fundamental managers. Among the institutional long-only managers that we study there are clear differences in the performance of quantitative managers depending on the time period and the investment style. More broadly, there are quantitative managers with investment horizons ranging from a few seconds to a few months and there are long-term, buy-and-hold quantitative managers who have investment horizons of up to three years.4 It is erroneous to group them all together. (Download all charts here)
Josef Lakonishok and Bhaskaran Swaminathan are with LSV Asset Management.
1. This article is a summary ofa paper that will be submitted to an investment journal.
2. See Petroff and Center (2009) “What are you really getting? A philosophical and practical reexamination of single market (domestic) quantitative strategies” of WURTS & Associates and Scott Patterson (2010): “The Quants: How a new breed of math whizzes conquered Wall Street and nearly destroyed it.”
3. Factor data from Ken French.
4. For a popular perspective on high-frequency quants “who typically hold stocks for 11 seconds” see May 16, 2010 New York Time article by reporter Julie Creswell titled “Speedy New Traders Make Waves Far From Wall St.”.