A First Glimpse Into The Short Side Of Hedge Funds
Jaewon Choi, Neil D. Pearson, and Shastri Sandy
June 22, 2016
Continued from part one... Q1 hedge fund letters, conference, scoops etc Abrams and his team want to understand the fundamental economics of every opportunity because, "It is easy to tell what has been, and it is easy to tell what is today, but the biggest deal for the investor is to . . . SORRY! Read More
We provide direct evidence about the profitability of hedge fund short trades in equities. We identify the opening and closing of equity short sales (and long side trades) by hedge funds and other institutional investors by combining data on the detailed transactions and holdings of the investors. Hedge fund short sales covered within five trading days are highly profitable, earning an average abnormal return of 14 bps per day, but short positions kept open longer than five days are not profitable. In contrast, non-hedge fund institutional investors suffer losses on short sales that are covered within five trading days but earn average abnormal returns of 2.6 bps per day on short trades covered between 21 and 63 days after being opened. Additional evidence suggests that some of the profitability of short trades is due to information and some stems from liquidity provision, and that short selling profitability is persistent.
A First Glimpse Into The Short Side Of Hedge Funds – Introduction
Short interest and other proxies for short-selling activity predict stock returns.1 However, there is little evidence about the identities of most short sellers and the profitability of their short trades. For example, while it is reasonable to think that a large fraction of short-selling is done by institutional investors, particularly hedge funds, little is known about the short-selling of these investors.2 Existing research on the trading of equity hedge funds comes primarily from “snapshots” of long positions disclosed in their quarterly Form 13F filings. For example, Brunnermeier and Nagel (2004), Griffin and Xu (2009), and Agarwal et al (2013) use Form 13F data to examine the long side holdings of hedge funds, but are not able to study the funds’ short positions because the 13F forms do not report them. The lack of evidence regarding hedge fund and other institutional short selling, combined with its likely importance in the price discovery process, motivates its study. The limited evidence of positive abnormal returns to hedge fund long equity positions in Griffin and Xu (2009) and to the positions disclosed in the funds’ original 13F filings in Agarwal et al (2013) also raises the question of whether equity hedge funds are able to generate positive abnormal performance on their short trades.
This paper provides direct evidence regarding the profitability of hedge fund short trades in equities. As explained below, we identify the opening and covering of equity short sales (and also long side trades) by hedge funds and other institutional investors at the level of individual investment management companies by using a novel algorithm to combine data from two different datasets on the detailed transactions and holdings of institutional investors. Based on our identification of trades, we examine the profitability of hedge fund short trades, including whether hedge funds have skills in both short opening and covering short positions. Hedge fund short sales covered within five trading days are highly profitable, earning an average abnormal return of 14 basis points per day, which translates to about 35% per year. For short sales kept open longer than five days the estimates of average abnormal returns are not significantly different from zero and the point estimates are close to zero. In contrast, non-hedge fund institutional investors suffer losses on short sales that are covered within five trading days but earn average abnormal returns of 2.6 basis points per day (approximately 6.6% per year) on short trades covered between 21 and 63 days after they are opened.
We also present evidence about some of the sources of the abnormal returns. Hedge fund short positions opened in a short window prior to earnings announcements predict negative earnings surprises, and their short trades that are open during the earnings announcement and covered within five trade dates of being opened are profitable, consistent with a role for private information. In contrast, short selling by non-hedge fund investors does not show any correlation with firms’ earnings surprises. Some of the hedge fund short sale profitability appears to consist of profits from liquidity provision, as short positions opened on trading dates with positive abnormal stock returns and closed with five trading dates are profitable on average while those opened on trading days with negative abnormal returns are not profitable. Hedge fund shortselling abnormal performance is persistent, consistent with it being due to skill, but that of non-hedge funds is not. In addition, we present corresponding results about the profitability of the hedge fund and non-hedge fund investor long side trades. Our result contrast with existing studies that are unable to identify short sales at the manager level but rather examine whether short sales (by unidentified investors) reflected in short interest and other proxies for shorting activity predict future stock returns.
The transaction data we use are from ANcerno, a trade execution cost consulting firm, and include on a daily basis all transactions by a set of institutional investors. Anand, Irvine, Puckett, and Ventakaraman (2012), Puckett and Yan (2011), and Jame (2015) have used ANcerno data to study institutional investors’ trading ability. The version of the dataset we use includes the names of fund management companies and both their long and short transactions, including the transaction prices. We combine these data with institutional holdings obtained from 13F forms, which report only long positions. The idea of our algorithm is to combine quarterly snapshot of long positions from the 13F data with the ANcerno data on all trades to determine the long or short position of each investment manager at the time of each trade, enabling us to identify the trades that open short positions. We also identify three other types of trades, these being purchases that cover short positions (short buys), purchases that establish long positions (long buys), and sales that close long positions (long sells). After identifying their trades we categorize investment management firms into hedge funds and non-hedge fund investors using information in the investment managers’ Form ADV filings.
Through this algorithm, we identify short sales (short sells), short buys, long buys, and long sells for 53 hedge fund management companies and 141 non-hedge fund management companies from January 1999 through September 2011. The data include a total of 2,967,918 daily hedge fund net trades, of which 213,337 are trades that either open or close short positions. To check whether the algorithm correctly identifies short sales, we examine the identified short sales during September and October of 2008 when short sales of the stocks of a set of financial firms were banned by the SEC. There is a dramatic decrease in hedge funds short sales of financial company stocks on the date of the ban and an even more dramatic increase on the termination of the ban, verifying that our algorithm successfully identifies short sales. Similar changes in short sales around the period of the short sale ban are reported in Boehmer, Jones, and Zhang (2013).
Having identified the short sales and other trades, we first provide evidence about the profitability of short sales by hedge funds in comparison to the other non-hedge fund institutional investors. To this end, we exploit the unique feature of our data and examine the profitability of the closed hedge fund short sales. We find that hedge fund short sales in our sample are profitable, especially for short horizons. Specifically, hedge fund short trades that are covered within five trading days on average earn an abnormal return based on characteristic-matched portfolios (Daniel, Grinblatt, Titman, and Wermers 1997) of 14.1 basis points per day, which translates to greater than 35% per year. For short sales kept open longer than five days the estimates of average abnormal returns are not significantly different from zero and the point estimates are close to zero. The returns on short sales by non-hedge fund investors are quite different. Their short sales covered within five trading days actually generate on average a loss of 10.8 basis points per day. Longer-term non-hedge short sales covered within one to three months are only profitable, providing an abnormal return of 2.6 basis point per day or about 6.6% per year. We obtain similar results by examining the abnormal returns or “alphas” of calendar-time portfolio return regressions using the Carhart (1997) four-factor model.
See full PDF below.