How to Use Academic Research

The proliferation of publicly-available academic research concerning investments over the last two decades has changed investment practice in many ways.

It used to be rare for working papers of academics to get much attention other than from fellow academics.  The Social Science Research Network, now just SSRN, was created in 1994 and soon became the prime repository of papers across a variety of disciplines, including investments.  There are other sources, but SSRN is still the dominant one.

Prior to the broader availability of those working papers, academic research was limited to peer-reviewed journals, most of which were (and still are) quite expensive.  (The Financial Analysts Journal is an exception, at least for members of CFA Institute, to whom it is available for free.)

Using the research

Most papers follow a format something like this:

Abstract

Introduction

Relationship to previous research

Methodology, data, and formulae

Results

Conclusion

Bibliography

Appendices, including charts and statistical analyses

If you are a statistical maven, you may find the meat of a paper of interest (roughly, the methodology, data, formulae, results, and appendices), but you can get the essence of it by reading the abstract, introduction, and conclusion.  For additional perspective, the relationship to previous research and bibliography are almost always thorough, providing avenues for further investigation.

All of it should be viewed in light of critiques of finance research from within the academic community, including:

Using statistical inference to analyze a complex adaptive system sets up challenges that are often overlooked (Bailey and López de Prado).

Economic incentives can distort both academic and practitioner finance research (Harvey).

The “file drawer problem,” where only research that shows statistical significance receives attention, is evident in both the composition of academic journals (Morey and Yadav) and conferences (Morais and Morey), giving a distorted view of the body of research that has been performed.

Much of the research output is directly related to the work of those in quantitative finance, but it also can be of value for those who have other kinds of analytical, portfolio management, and advisory roles — which is the perspective taken in looking at the examples below.  Does the research pass the “smell test” of representing the world as you know it?  If there is a discrepancy between your experience/beliefs and the research, what is the nature of the disconnect?  Often the research can prompt new insights and a rethinking of old notions and practices, although sometimes you realize that the researcher’s assumptions, design, or data account for the disconnects.

Two other notes:

Disproportionate amounts of research are done on equity markets, mutual funds, and the United States, because that’s where the information is most complete.  That’s changing at the margin, as other asset classes, vehicles, and geographies are studied, but it is still the case.

In this posting, no attempt is made to evaluate the techniques or quality of the examples of research that are provided.

Pairs trading

A pairs trading strategy is executed by going long one security and short another, in the expectation that their relative values will converge over time.  Such strategies have been the subject of a fair amount of research in the past, much of it (like a paper reviewed in a 2009 piece from The Research Puzzle) related to what makes for a good pair, one of the topics addressed in “In Search of Pairs using Firm Fundamentals: Is Pairs Trading Profitable?” from Sungju Hong and Soosung Hwang.

The abstract states that “portfolios of pairs that have higher fundamental similarity outperform those that are fundamentally less similar by minimizing the non-convergence risk.”  Given that many pairs trades put more emphasis on the correlation of the two assets (which can change abruptly), the research prompts questions about the nature of risk management that should be involved.

The authors provide a long list of firm characteristics that were tested and rank them in their importance in determining the similarity of two equity securities for the purposes of the analysis.  But it is the broader conclusions that are striking:

Despite the outperformance of the pairs with fundamental similarities, these pairs do not show a significant alpha any more in the 2010s.  As in Gatev et al. (2006), the profitability of pairs trading continues to decrease because of active arbitrage trading and the changes in trading environment and information (Green et al., 2016), and pairs trading does not seem to be profitable any more in the US market.  The results indicate that temporal price deviations that can be exploited by arbitrage trading are less likely to arise, and if any, these pairs may be spurious.

Two charts provided illustrate a stark change in the overall historical pattern of returns to the strategy.

Questions:

If you are a manager who implements pairs trades, does this match your beliefs or prompt you to consider rethinking your approach?

If you are an investor, has your due diligence allowed you to understand the nature and prevalence of pairs trading among your current and prospective managers — and how they have adjusted to a less fertile environment?

How much of the change in alpha opportunity is related to the massive flows into passive investment products?  Would a reversal of that trend revive the trading strategy?

Pairs trading is common in other asset classes (notably fixed income) and even for multiple-security vehicles (increasingly, ETFs); has the compression of alpha proceeded in like fashion among those strategies?

Fund size matters

A couple of “stylized facts” from the body of academic research come into play in a paper from Gelly Fu, “Small Fund Size Matters.”  Namely, because investors place “a disproportionately large amount in the top performing mutual funds,” firms “have high incentives to produce outperforming funds” and align incentives for those managing funds accordingly.  Plus, “mutual funds exhibit decreasing returns to scale, that is, mutual funds perform worse as they grow larger.”

Based on those assumptions, Fu studies cases where portfolio managers run more than one fund to see if the smaller funds have better performance than they otherwise would have when controlling for size and other factors.  They do, but only when there is a significant difference in size among the funds managed.  (The paper also looks at changes in the relative size of funds overseen by a portfolio manager as a result of fund mergers — and references research by Agarwal, et al. which finds that “managers temporarily divert their attention” to a newly-assigned funds, which often come to them because performance has been poor under a previous manager.)

Fu uses the phrases “preferential attention” and “priority funds” to describe managers’ approach, which many involved would probably dispute, although:

I examine the trading activity of priority funds using quarterly holding data and find that priority funds exhibit less herding behaviour compared to the non-priority funds, in other words, managers tend to lead the institutional crowd with their trades in the priority fund.  This implies that when managers identify a good investment opportunity, they first trade on this investment with their priority fund before moving on to their non-priority funds.

Questions:

If you work at an asset management firm, does this research reflect what you have observed among managers with multiple portfolio responsibilities?

Are there practices at your firm that foster the observed performance differences, such as trade allocation policies, especially regarding IPOs?

If you are an investor, what are your opinions/operating procedures regarding portfolio managers with multiple responsibilities?

What are the incentives that exist for the asset management firms you use — and for the individual investment decision makers within them?

Selecting managers

Discerning the “why” of manager selection has been the subject of a number of analyses, among them, “Picking Partners: Manager Selection in Private Equity,” a paper by Amit Goyal, Sunil Wahal, and Deniz Yavuz.  (It uses an expansive definition of “private equity,” encompassing partnerships that could more narrowly be described as “buyouts, direct lending, distressed equity, growth, infrastructure, mezzanine financing, natural resources, real estate, and venture capital.”)

The abstract summarizes the findings regarding the general partners (GPs) who manage the funds and the institutional investors who are the limited partners (LPs) within them:

In addition to chasing GPs with high prior performance, LPs have large propensities to select first-time or young GPs without a performance history.  LPs also have tendencies to follow their peers’ investment decisions, to reinvest with the same GP, and to invest with GPs domiciled in the same state/country.  These selection criteria, however, do not provide information material for future performance, and in the case of first-time GPs are associated with lower future performance.

There’s much there to parse.  It is mostly as expected, but the conclusion regarding favoring first-time GPs is surprising.  However, there are some possible explanations, each of which is addressed in more detail.  It may be that the lack of access to established funds has forced organizations that are new to private equity (or have had small exposures) to ramp up their portfolio weights by targeting newer funds where they can get in.  Also, the authors acknowledge that some first-time funds may not have “true rookies” at the helm, but veterans of other firms who could have been observed by LPs previously.  In addition, newer kinds of investment strategies often are the province of newer kinds of firms.  Or, it may be that young funds offer fee discounts to attract investors.  Finally, there could be a belief “that first-time GPs may deliver superior performance.”

That last point represents a philosophy that is quite common (even among those who don’t invest in nascent firms because their governance precepts require them to wait for some performance evidence).  Therefore, the conclusion that first-time funds as a class do not meet that expectation is particularly notable.

Questions:

If you are an investor, how do the findings of this study match up with your beliefs and practices?

What are your “tendencies” as to following your peers’ investment decisions, reinvesting with the same GP, and investing with GPs domiciled in the same state/country?  Do you track them?  Which do you think add value and which detract from it?

What are the general principles that guide your consideration of emerging managers (of any kind)?

Summary

These are but three examples of the thousands of papers that are produced every year, some of which become better known because they shine a new light on areas of investment practice and/or they receive attention from the press or leading firms.

A regular diet of academic research will help you to reconsider your assumptions and to prompt new angles of pursuit.  It’s a mistake to look to the papers and articles for answers, but they can be a great source of important questions to ponder and investigate.