Investors tend to be bottom line-oriented; they want to know how their investment performed and tend not to be overly concerned with the details. Yet, how an investment achieves its return is one of the most important questions an investor can ask. The focus of this article is on the following question: can a long/short fund actually generate short side alpha?
In its purest form, alpha is merely a piece of statistical output that is generated from a regression analysis between the returns of some investment and an underlying benchmark. Assuming the regression is statistically meaningful, alpha measures the portion of the investment's returns arising from specific (non-market or non-benchmark) risk. In other words, alpha is a measure of a fund manager's skill level in adding some sort of value beyond a simple benchmark investment. For example, an alpha of 1.1 indicates that an asset, based on historical data, is projected to rise in price by 10% in a year when the return on the benchmark and the beta of the asset are both 0.
An interesting side note is that many in the alternative investment community incorporate various definitions of alpha to fit their distinct viewpoints of the hedge fund landscape. Consider, for example, the secretive world of hedge funds. In many cases, focusing solely on a hedge fund's return series can lead to erroneous conclusions regarding a manager's ability to produce alpha. Simply put, a manager has not necessarily added alpha just because he has outperformed his peer group or has produced positive returns.
Now that alpha has been defined, we are prepared to tackle the main issue: can short side alpha be generated in long/short equity strategies. Specifically, we will examine anecdotal evidence, the barriers to effectively sell short, whether short side alpha can even be calculated and whether such a calculation is relevant when applied to varying sub-classes of long/short strategies.
Do managers actually offer any hard evidence that they generate short side alpha?
Many managers claim the ability to add short side alpha. On the surface, this assertion may seem plausible. Yet most hedge fund managers offer little evidence in this regard. Managers do not provide investors with an audit detailing the timing and performance of each short position. Instead, investors receive an annual audited performance figure. How many investors run the proper analysis to determine if alpha was generated on the short side? If they did, I believe the results might surprise them. Indeed, an analyst from a well-regarded fund of funds told me that he could count on one hand the number of long/short managers that actually added alpha on the short side of their strategy.
What impediments do short sellers face?
Right from the start, the odds are stacked squarely against the short seller. What long managers take for granted, short sellers cannot. Four of the biggest issues facing short sellers are the ability to borrow, execution impact, capacity limitations and quantitative modelling.
Ability to Borrow: In order to short a stock, short sellers must first borrow the security via their prime broker's stock loan department. When shorting, the seller receives short credit rebates equal to some percentage of the fed funds rate. However, if a stock is "hard to borrow," instead of receiving short credit rebates, the short seller may actually have to pay interest in order to borrow the security. This situation is a minor problem in comparison to a short squeeze. Furthermore, 100% of a company's stock is never available for shorting. In reality, 25% is considered to be a rather large number. On this basis, there is much more inventory for long managers than there is for short managers and monumentally more liquidity in the large-cap universe than the small-cap universe. Generally speaking, the issues surrounding the borrowing of stocks can severely hinder a manager's ability to short by limiting stock selection and position sizes. This fact becomes amplified when considering small- to mid-cap stocks, lower priced stocks and moderate to highly illiquid issues.
Execution: Due to the up-tick rule, entry into a short position presents more difficulty than entry into a long position. In addition, orders are marked as "short sales." If a savvy specialist, market maker or programme trader becomes aware of a short seller's trading pattern, executions will be severely impacted. This point is particularly true with small- to mid-cap stocks, lower priced stocks, and moderate to highly illiquid issues. In general, trading slippage is much higher on the short side than on the buy side. As an aside, some institutional brokerage houses have designed clever methods to avoid the up-tick rule. However, additional maintenance costs are typically associated with these methods.
Capacity: Not any stock can be shorted. In fact, an analysis of the Russell 3000 suggests that perhaps as few as 50% of the names are available for borrowing without any obvious limitations. In addition, as fund assets grow, managers must pay closer and closer attention to liquidity and market impact issues, especially on the short side of their portfolio. When considering these issues in conjunction with borrow-ability and execution, it becomes obvious that asset accumulation severely reduces the universe of shortable stocks for any fund manager. Asset growth will eventually force managers to invest the bulk of their short capital in the ultra large-cap arena. As a result, a manager's historical short performance will have little relevance in forecasting his future returns.
Quantitative Modelling: When a unique factor is applied to a universe of stocks, prediction accuracy tends to be higher on the long side than the short side. In essence, an analysis of the regression of that factor to historical stock returns would illustrate that the error of the regression on the short side is many times higher than that on the long side. In addition, simulated results tend to dramatically underestimate the cost and market impact of shorting, particularly when dealing with smaller cap stocks. Generally, most market impact models are geared towards large-cap investors and simply are not calibrated to capture the execution impact of shorting mid- and small-cap names. In addition, simulations cannot factor in a stock's historical ability to be borrowed.
Is short side alpha measurable? Is it relevant?
We analysed three of the most common types of long/short strategies with these questions in mind.
Biased Long/Short Equity strategies can take huge bets ranging from 100% long to 100% short exposure at any given time. Moreover, leverage ratios can change on a daily basis to maximise the best absolute return scenario possible. Even those managers utilising some constraints often take huge sector bets. In any of these scenarios, it is nearly impossible for an investor to quantify how the returns are being generated. Did the manager time the market correctly, own the right sectors, short the proper individual stocks, and/or utilise effective leverage? To efficiently address these questions, an investor would need daily audited return and trading summaries. Also, these issues do not even begin to address the subject of what benchmark to use.
Market Neutral strategies are designed to assume zero beta exposure. From a quantitative standpoint, the long portfolios and short portfolios are identical. Capitalisation, sector, liquidity, and execution biases are eliminated. The objective of these strategies is to produce small, but stable monthly returns over time. As long as the longs outperform the shorts, incremental return is generated. Thus, short side alpha generation is a non-issue in these types of strategies.
Relative Value strategies are a "watered down" version of the pure market neutral strategy. Many are sector specific. For example, take the situation of a former high profile Wall Street software analyst who opens a relative value software hedge fund. The manager attempts to buy a group of stocks that are undervalued and short a group of stocks that are overvalued. In a best-case scenario, the manager's longs outperform his software benchmark and the shorts under perform the benchmark. However, what if the shorts outperform the benchmark, but the manager's net return is still positive? In this case, the manager produced a positive net return, but failed to add alpha on the short side relative to the benchmark. Should a relative value manager be penalised for this? Furthermore, is the utilised benchmark the proper performance yardstick to begin with?
In a nutshell
As expressed earlier, the question of whether alpha can be generated on the short side is not easily answered. However, we can summarise with the following:
- Shorting individual stocks is incrementally more difficult than buying individual stocks due to the up-tick rule and borrow-ability issues.
- Market impact is much greater when shorting stocks than when buying. As a result, execution costs can be quite high on the short side, often much higher than simulations might suggest.
- In reality, only a portion of a company's total capitalisation is available to short sellers. Consequentially, this creates a capacity issue for these sellers.
- The above reasons are amplified when attempting to short sell in the small- and mid-cap stock universe. Unfortunately, most inefficient security pricing occurs in this universe.
- As assets under management climb, managers must resort to putting their short capital to work in the ultra large-cap universe.
Based on the evidence, short side alpha seems to be quite elusive or, at the very least, very difficult to measure. Perhaps institutional investors have been focusing on the wrong issue and perhaps they have been marketed an unattainable ideal. Short side alpha sounds great, but is it really attainable? Maybe the principal focus should be on a manager's ability to produce stable and attractive returns while being able to skilfully navigate through periods of extreme volatility. What's more, investors may want to home in on the liquidity, capacity and scalability of a strategy. These topics certainly seem more tangible than short side alpha and, in the long run, are probably more important to long-term capital preservation and appreciation.