Alpha From Short Positions: Myth or Reality?
Corte, Kosowski, and Rapanos recently measured the impact of an investor's short positions on portfolio alpha.
What is the difference between a mutual fund and a hedge fund? A professional could rattle off a dozen things, but it really boils down to two: an incentive fee structure and an ability to short securities. The incentive fee is there to promote alpha generation by aligning the incentive (added fees) to the desired outcome (alpha generation). The ability to short is meant to both hedge systemic risk (something investors are not willing to pay active fees for) while also removing constraint from expressing market views. If a mutual fund hates a security, the most the manager can do is not own it. A hedge fund can short it with the intent of profit. Removing systemic risk by reducing her net exposure is a bonus.
Every hedge fund manager takes pride in his or her ability to short, but this can be a difficult skill to prove. Yet prove it they must. Without it, the proverbial “2 and 20” fee structure is under attack. Novus has made a name for itself by providing tooling to describe investment management skill, both for managers wishing to promote (and improve) their rare talents, and for capital allocators seeking truth. We write a lot on this blog about how to unpack skill through the lens of public ownership data. Unfortunately in the US, there is very little short information disclosed to draw these insights. In Europe, we have a different situation.
A Unique Window Into Hedge Fund Skill: Conviction
In the past, we’ve covered the European Union’s 2012 regulation requiring Short Disclosure of investors for public consumption. We were one of the first to systematically parse this data, going back to 2012. (Check out our most recent insights on this data.) Now that there are over six years of disclosure data, research about hedge fund managers’ ability to effectively short stocks is gaining momentum. Recently, a group of researchers (Pasquale Della Corte, Robert Kosowski and Nokolaos Rapanos, or ‘CKR’) performed a study to analyze the efficacy of hedge fund shorting through European disclosure requirements. Their findings align with much of our thinking about teasing out investment management skill: look at conviction.
First, a quick refresher on what European Short Disclosure is. Any investor who trades in EU marketplaces (either through equities, swaps, CFDs, or other derivatives) must disclose short positions to local regulators when their position equal 0.2% of the outstanding shares (delta-adjusted when dealing with derivatives) of that issuance. They must update their position with any 0.1% incremental move. That disclosure is not made public until the position reaches 0.5% of outstanding shares, at which time it is then published the day after the notification is made.
This data is unique in two ways: First, it is contemporaneous! Unlike 13F disclosure, it is provided the day after the level has been breached. Second, it is on short positions! 13F disclosure is 99% long-biased, only providing short information on notional exposure of put options. Here we have a unique dataset to analyze investment management skill in a way that truly differentiates hedge funds.
CKR build upon findings of prior researchers, most notably Jank and Smajlbegovic (2017). These prior researchers found that aggregate short position disclosure generates a 5.5% annualized alpha term when measured against a traditional Fama French 3-factor alpha, but that the alpha dissipates to almost 0% when including more contemporaneous factors such as momentum, betting-against-beta, and quality. There are two ways to view this. The optimist would say that hedge funds are good at timing factors for their shorts, whereas the pessimist would say that hedge funds do not add value beyond more modern style-factors. CKR add to this research by adding the important wrinkle of conviction. Can we draw a different inference if we dissect managers’ shorts based upon the size of the position (as a proxy of conviction)?
What they find is startling. When quintiling all managers’ shorts by their definition of conviction, they find far more alpha in the higher conviction names. This chart below from their working paper clearly demonstrates the difference in alpha between higher and lower conviction shorts.
Naturally, a skeptic would ask a few questions at this point. First, what happens when you factor in the liquidity / availability to short a name, and what about the short borrow fee? There are securities where there is consensus bearishness, but this is often arbitraged by the cost of borrow that the investor pays to the counterparty who provides shares to borrow (often a prime broker).
CKR tackle this issue by incorporating aggregated short-interest data provided by Data Explorers (a data package by IHS Markit), which provides aggregate borrow rates for securities daily. What we see is a noticeable impact on the alpha that can be generated by following high-conviction European short disclosure, but there is still alpha! The following chart demonstrates how a market neutral portfolio that shorts high conviction hedge fund shorts while going long low conviction shorts behaves with and without transaction costs. Each of the simulations demonstrates positive cumulative returns (or alpha, given the portfolio is market neutral).
What was a 1.23 sharpe ratio dissipates to a mere 0.55 sharpe ratio, yet there is still statistically significant alpha (and quite a large magnitude) when measured against a six-factor alpha model. That’s quite a bit of alpha!
It’s no wonder that Novus clients (both institutional allocators as well as hedge funds themselves) have requested European short disclosure data from Novus. In addition to raw data, we also provide tooling that helps public module clients explore this data, as well as customized back-testing and investment strategy design based on said data. Our European Short Observer page in our Public Ownership Module provides t+1 vizualization tools that allow users to track the aggregate short flows at the single name and single-manager level. Users can filter their choices by country of origin, or GICS sector.
This piece adds to the growing literature that demonstrates the value of public disclosure. What is unique about CKR is the insight given to the two distinct characteristics of hedge funds: short exposure, and the connection between incentive structures and alpha generation. By focusing on conviction, they can separate the wheat from the chafe. Investors should take note.