This elusive group of ‘short’ investors have taken the blame for many of the crashes of the last 400 years; from the East India Company of 1609, the Great Wall Crash of 1929, Black Monday in 1987 and the bursting of the dot-com bubble in the early-2000s. Short sellers have been labelled as “stock-bashing rumour mongers” (Albert et al, 1997) and the “assassins of Corporate America” (Business Week, 1996).
Even in the last few weeks we have seen the saga involving HBOS unravel. The owner of Halifax, Britain’s fifth biggest bank, accused short sellers of “market abuse…by spreading false rumours to profit from short selling” (Market Oracle). The Daily Telegraph’s Ian Cowie said in his article entitled ‘Hedge Funds behind HBOS rumours,’ 26 March 2008 that: “Selling property you do not own might sound like a criminal activity anywhere outside the City….never mind, for now, that most stock sold short is borrowed from banks – such as HBOS, the owner of Halifax, whose reputation was temporarily tarnished by the rumour-mongers this week.”
But would it not be fair to say that in many of these instances, namely the bursting of the dot-com bubble in the early 2000s as well as ABC and HBOS, there is a wider picture here. For one, share prices fall because companies simply under-perform – short sellers realise this and invest accordingly. The US sub-prime crisis that continues to penetrate world markets was perhaps a cause of banks being over-zealous with their lending rather than individuals betting that this would happen. Yes, short sellers have benefited, but they are perhaps a symptom rather than a cause of the problem.
Enough time has elapsed since the dot-com ‘burst’ and any crises before it for some interesting academic literature to evolve.
Various academic debates and essays stem from notorious incidences when short selling has made headline news. In the case of dot-coms, academics asked why investors failed to short overpriced internet stocks and thus bring them back to an equilibrium price. Ofek and Richardson (2003) amongst others, claim this was down to the existence of short sale constraints. For example, they found that internetstocks were relatively expensive to short because of high stock borrowing costs.
However, Geczy et al (2002) find little support for the notion that short selling constraints made it difficult for arbitrageurs to short ‘dot-coms’. Brunnermeier and Nagel argue that hedge funds were “riding the technology bubble,” rather than short selling apparently over-valued stocks.Academic studies have typically found that allowing short selling improves market efficiency. In a recent paper1, published on 10 June 2007, Pedro Saffi and Kari Sigurdsson from the London Business School were able to examine the impact of a liquid securities lending market on market efficiency at the level of individual securities. They concluded that short-sale constraints lower price efficiency. Stocks with limited lending supply and/or high borrowing fees responded more slowly to news in the public domain, and they did not exaggerate market-wide shocks, either.
Moreover, in 2002, the Financial Services Authority (FSA) said that short selling is a “legitimate investment activity which plays an important role in supporting efficient markets…Hence we see no case for prohibiting short selling…and the introduction of specific regulatory constraints would not be warranted.”
If unrestricted, say ISLA, short selling can transfer information from informed investors to the less informed, allowing market prices to reflect a fairer value of the securities. Another concern relates to what some call ‘news arbitrage,’ describing certain influential individuals who encourage false rumours (as in HBOS) in order to affect share prices. Regulators fear that it may be a ‘useful tool for those with intention to abuse a market’ (ISLA, 2008). This of course could be said about any other form of trading.
“Thomas (2006) suggests that the most persuasive evidence so far is provided by Woolridge and Dickinson (1994), who find that short sellers sell as stock prices rise and reduce short positions as they fall. In this way, short sellers act as ‘stabilising liquidity providers’, giving neither bullish nor bearish signals. This view is in agreement with Albert et al (1997) as mentioned previously, who found Nasdaq short sellers added liquidity by selling into rising markets.”
The relevance of short selling as a tool in a portfolio manager’s armoury was open to question during the lengthy bull run that only came to an end during the summer of 2007. Until that point shorting shares was seldom fruitful. Underperforming companies were few and far between and even those that could be identified could be subject to a takeover. The removal of the so called ‘private equity put’ has made it safer to profit from this activity, but there are still risks. James Clunie researched shorting for many years when a member of the faculty of Finance at Edinburgh University. He is now in charge of a 13030 UK fund at Scottish Widows Investment Partnership. Clunie’s latest paper addresses risks from ‘crowded exits’.
This paper fills a gap in the academic literature because our understanding of how liquidity constraints affect short sellers is limited. Clunie puts meat on the bones of an issue close to every potential short seller – that of timing a successful exit. He uses stock borrowing data from Euroclear in combination with information from Data Explorers to study 681 LSE listed securities.
“Crowded exits arise in stocks where short sellers hold large positions relative to normal trading volume, and when a catalyst prompts short sellers to cover their positions rapidly and simultaneously… The idea is akin to the audience in a crowded theatre rushing to a narrow exit door once the fire alarm sounds…only so many can leave the building in any given interval of time.”
“The temporary excess of demand for stock relative to normal trading volume leads to upward pressure on the stock price and these events are associated “The temporary excess of demand for stock relative to normal trading volume leads to upward pressure on the stock price and these events are associated
It will not come as a surprise to those people who have recent experience of shorting in the UK that the main findings are that crowded exits are associated with positive abnormal returns (ie. losses to short sellers) of up to 27% over a 60 day period. In situations where a short position is defined as large (above a certain % of the market cap on loan) and where liquidity is constrained (by calculating the number of days it would take to cover all outstanding shorts using average turnover) investors can surrender up to 27% of their profit by being the last – or forced – to close their short position.
In another paper, “Anatomy of Short Squeezes”2 Clunie looks at upwards of 350 UK listed stocks over a 45 month period and tries to define apparent short squeezes. “In the three day run-up to such events, stocks experience significant positive abnormal returns of 3.45% and hence short sellers would suffer significant losses when covering in the market stock recalled by lenders.”
Implications from these studies are interesting. Should stock borrow data become widely tracked, certain short term investors may seek to exploit these opportunities. That said, a truly ‘crowded exit’ is such that new investors cannot opportunistically buy shares when liquidity is so tight.
As outlined here, the literature shows that heavily shorted stocks perform poorly. It therefore follows that people could come up with imitation strategies if following the herd is profitable. However Clunie has the following warning, “the act of imitation changes the market dynamics and can lead to unexpected consequences. With imitation, short positions become more crowded, and the risk of ‘crowded exits’ increases.”
Two academics also unravelling the information content of short sales are Michael McKenzie and Olan T Henry. Taking advantage of publicly available information on shorting in Hong Kong, as published by the HKSE, they published a paper that appeared in the Journal of Business3 in 2006 called, “The impact of Short Selling on the Price-Volume Relationship”.
Their angle is novel and interesting. They believe daily cash equity trading can be broken down to reveal an informative signal. Nonlinearities in the volume-volatility relationship are tested using short sales information so that it is possible to identify the activity of short sellers on a daily basis to predict share price returns. One of the clues that led them down this path was a common finding in the literature on stock market volatility that negative shocks cause more volatility than positive stocks of equal magnitude.
“In contrast to previous findings, however, this paper documents a new form of asymmetry in the dynamic process that determines stock return volatility. The source of this asymmetry is the trading activity of short sellers. Short sales are motivated by bad news about a company’s future prospects. The trading activity of short sellers (volume) reveals their informational advantage to noise traders and, as markets typically overreact to bad news compared to good news, elicits a larger response in volatility compared to a day in which short sellers are absent from the market.
This is not to suggest, however, that volume drives prices as much as the theoretical literature suggests and many empirical studies implicitly assume. Rather, the parameter estimates of our model clearly indicate the presence of a strong nonlinear bidirectional relationship between innovations to volume and returns volatility.”
In other words, short selling volume tells you more about the behaviour of share prices than just straightforward volatility.
The plot thickened in 2007 when McKenzie and Henry went to work on the Data Explorers’ Hong Kong dataset. By combining shifts in the demand (to short) curve, as captured by DEL, they hoped to further refine the information content of short sales information. In general they found, “strong evidence that short interest is a major channel for the transmission of information about prices.4” At a higher level, they were glad to pursue the anatomy of trading volume that clearly originates from any number of agents, each of whom may hold a different information set. Some trades initiate from trend followers so are ‘noisy’. Others may be based on the need for liquidity to cover redemption and the like. Some, however, are based purely on information. “It is an interesting empirical issue to consider the extent to which the volume associated with different types of trading reveals information to the market. Especially interesting given the literature suggests that short sellers may be better informed than other traders.
There can be 4 different motivations to a trade. ‘Buying’ shares can be for two reasons: opening a ‘long’ position (therefore buying a share expecting it to go up); closing a ‘short’ position (ie. buying shares to make a profit having sold them at a higher price previously. ‘Selling’ shares can be for two reasons: closing a ‘long’ position (to take profits or cut losses or for liquidity needs); opening a ‘short’ position (in expectation of future price falls or for hedging).
With HKSE data and Data Explorers’ info, McKenzie and Henry can distinguish between short selling and all other trading volume. As McKenzie explained to the audience of the Spitalfields Advisors London Forum in March 08: “Short sales volume has significant explanatory power. All other volume does not. Higher short sales means lower returns in the next period.”
This is not the end of the story. Some might contend that the next most informative piece of volume to identify is the buying and selling being carried out by active investors. There may be a way to do this using the DEL dataset – so watch this space.
Other academics who have reported on this subject include: Asquith and Moelbroek (1996), and Angel et al (2003). Surely short sellers are better informed than other traders because they identify highly shorted stocks and measure their risk-adjusted returns.
“Chen and Singal (2003) argue that speculative short sellers may be the cause of the Monday effect. Angel et al (2003) assess the frequency of short selling by weekday for their sample. Little day-to-day variation is shown for the percentage of short trades and percentage of shorted shares” (ISLA, 2008).
The literature [described above] documents that short-sellers, unlike long only investors, face potentially unlimited losses, but does not describe how short-sellers manage this risk. There is a need to understand the importance of this risk, to what extent it acts as a short-sale constraint, and how short sellers manage the risk. Furthermore, the impact of this risk and related risk management techniques on asset pricing is not yet understood.
For as long as short selling takes place behind closed doors, there will always be an element of mystery surrounding it.
As you can see there are academics who are actively campaigning for the other side of the story to be heard and understood. Some see short selling as a convenient excuse to assign inexplicable price movements to the activity. After all, no one can question those that hang their hat on a coat stand that no one can see, as we mentioned with HBOS and ABC. Journalists assumed the rumours these stocks were in difficulty were initiated by traders set to profit from large short positions – but the data has proved otherwise.
However, others are seeking to profit as the lights start, slowly but surely, to illuminate the world of shorting. Both academics and active investors are seizing the chance to learn about an area of the market that is poorly understood in the hope of being ahead of the chasing pack.
Data Explorers provides its global client base with a quantitative measurement of securities lending performance and risk. It runs a daily blog, ‘Short Stories,’ on topical stocks. Data Explorers’ sister company is Spitalfields Advisors Limited, a consultancy business specialising in securities finance.