Were The Short Sellers Routed? Does It Matter? (Beware The ‘Gamma’)

The “twin peaks” pattern in GME’s share price (January and February/March) has shocked the markets.  

There is no obvious fundamental driver for a 1700% gain in the GME share price in less than a month. GME’s sales have contracted by 30% in the past two years, with losses of over $1 billion. Bloomberg columnist Matt Levine summed up the company as “a money-losing mall retailer in a dying business during a pandemic.” Maybe the future will be brighter, or maybe not. But it is clear that right now Price and Value are out of joint. 

Beyond that, however, the pattern make no sense technically, at least according to the standard view of how the market works and how prices move. Normally, when a stock moves up, we expect a certain steadiness. There are mechanisms – psychological, procedural – that guide-rail the process. A triggering event catches the market’s attention and prompts a reassessment. Some investors move quickly; others lag, then join in. Momentum develops, trend-followers pile on, the media talk up the story, the chat rooms get going, and investor sentiment ripens to a golden glow of positivity. The equilibrium-seeking tendencies of the market modulate price movements and smooth out the fluctuations. 

Tesla’s recent rise is a good example. Its shares doubled in the 4^th quarter of 2020 following the announcement that the company would join the benchmark S&P 500 index. The move was powerful, but not violent.  

A case can be made that Tesla is overvalued. Its price-earnings ratio is well over 1000. It could reasonably be seen as a “bubble.” But the pattern is familiar. 

The GME move is something else. The price runs right off the chart. The word “bubble” does not do justice to the violence of the GME price explosion. 

Single spikes like GME’s January surge are not unknown. They are rare, but explicable, usually as technical adjustment by investors compounded by some sort of crowded trade. But a double spike, without fundamentals, with outlandish volume statistics (see below)… is a new chimera. And yet there it is.

Is it a freak, an anomaly like the Flash Crash of 2010? Or is this a sign of something truly new, an important change in market behavior? Something that may become a regular part of the game going forward?  

The Gamma Process

Previous columns [see for details] made the case that this event arose from a novel process called a “gamma squeeze.” A rise in demand for GME call options led the option-sellers to hedge their risk by buying GME shares, driving the price upwards. [“Gamma” refers to an obscure variable used in option pricing math. It can be viewed here simply as a convenient label for this trading maneuver and does not need to be formally defined.] 

In the January surge, this gamma hedging action helped squeeze unwary short sellers, who were caught out and forced to cover [i.e., buy back the shares they had sold short], adding to the “melt-up.” In the February surge, it appears that the short interest played a much smaller role. Other observers came to the same conclusion. 

• “The original trade that sent GME stock soaring back in January looks like a well-executed short squeeze. It nearly bankrupted one hedge fund, Melvin Capital, and caused huge losses at many others. The coordinated effort led (largely) by traders on Reddit was brilliant. GME has rallied again, gaining more than 400% from a Feb. 19 close just above $40. But there are no short sellers of size in the market who didn’t prepare for such a move.”

This implies that in February the “gamma squeeze” alone was able to drive up the price. 

If this view is correct, it has broader significance. Short squeezes are well-understood, but are hard to engineer (for reasons discussed in my previous column). A “gamma squeeze” is much easier to mobilize. Options are much cheaper than the underlying shares. Social media can quickly enlist large numbers of retail players. If high demand for low-cost options can drive a corresponding demand for high-cost shares (for hedging by the option-sellers), this new tactic may have a much wider applicability, and could become a more common occurrence. 

The Reddit “apes” (yes, that is what they call themselves) have focused their attention until now on stocks with a significant short interest, which was clearly the case for GME prior to the January surge. But if the February surge was effected without the extra “fuel” of a large short interest, using the gamma mechanism alone, it may mean that this phenomenon could spread to many other companies, regardless of the short position. 

Some readers thought I got this wrong. They questioned whether the short interest really was reduced substantially (as officially reported) between the first and the second surge. They questioned my conclusion that short squeezing played a much smaller part in the February event. 

If there is a difference between these two events, a close examination of the two peaks may shed light on these matters.

The Official Short Interest Figures

The data on the short interest is available only biweekly. Prior to the January surge, the short interest in GME was very large — over 100% of the public “float.” [The float refers to the number of shares trading freely in the market, available to the public. It excludes shares held by insiders.]

Following the January surge, the short interest had been reduced by almost 80%. The straightforward interpretation is that short sellers were squeezed and most were forced to cover. 

The short interest going into the February surge was thus much lower, and did not change much during the event. 

If these figures are accurate, the proposition is clearly established. Short-covering played a large part in the January surge, and a much smaller part in the February surge. 

The Volume Shock

What is most striking about this episode is not the surge in prices, but the unprecedented explosion of trading volume. 

GME trading had averaged a few million shares a day, for many years. Then on the morning of January 13 2021, orders for 20 million shares hit the market in the first hour of trading. 93 million shares traded before lunchtime – more than twice the public float. By the closing, almost 150 million shares had changed hands. 

It was just the beginning. Waves of buy orders began to sweep through the market over the following weeks.  

To put this in perspective, the typical “hot stock” — like Tesla, or Apple – may trade a few percent of the float daily. On Apple’s heaviest trading day this year, shares equal to 1.7% of the float changed hands. Tesla’s trading has averaged 5% of its float daily over the last three months. Its heaviest volume occurred on Dec 18 (the day before the company joined the S&P 500). Tesla set a record for the dollar value of shares traded in the stock of a single company on a single day: 222 million shares at an average price of about $750 a share, 28% of Tesla’s public float.  

Since January 13, GME has traded a daily average of 118% of its float. On several days, the turnover has exceeded 300% of the float.  

During the February surge (Feb 24-26) and the January surge (Jan 22-27), the turnover ratios were 7:1 and 14:1 respectively. 

Where is this volume coming from? With the prices rising, the transaction momentum was clearly driven from the buy-side. Who was in control of the game? Who was buying? 

Was it the short-sellers buying to cover their short positions? Certainly some of the demand, in the January episode at least, derives from this source. But the entire short interest, as overdrawn as it was, could have been taken out in a few hours once the surge began. From January 13 to the end of the month, over 1.2 billion GME shares traded – about 20 times the short interest. 

By the time of the second surge in February, the short interest was small enough that it could have been cleared out in the first 60 seconds of trading at the open on February 25. We may be justified in concluding from this that the short interest was of negligible significance for the 2^nd peak. 

This big-picture perspective on the enormous trading volumes establishes that the short interest can have played a most a minor supporting role in January, de minimis in February.  

The fine structure of these two events may reveal a bit more, and confirm the big-picture conclusions.

[Caveat: The following discussion is rather detailed. It may not add a great deal to the understanding of these events, and may be of less interest to some readers.]

The January Surge – Close-up

In January, the volume shock preceded the price movement by a few days. This is unusual. The surge in volume was clearly coming from the buy-side, and should have moved the price up right away. 

Even more surprising is that the volume dropped off sharply as the price finally jumped. This is also unusual, if not abnormal. The raw correlation between the share price and the trading volume the 10 days prior to the week of the surge was 84%. The correlation during the week of the January surge was negative 75%. As the price soared, the volume fell.

There are various interpretations. But keep in mind that “volume” must include an equal number of buyers and sellers. If the price is not moving up in the face of high volume – it did not rise much until the week of January 25 – it must mean that sellers were matching buyers. Then came the market break.

In just 2 days (Jan 26-28), the share price rose by a factor of 3 – while the volume dropped by factor of 3. This is consistent with a scenario in which short sellers fought the buy-side volume up until Jan 26 — even as the buy-side pressure reached extraordinary levels – at which point the shorts could no longer hold on, and capitulated. They started covering — buying back the shares they had shorted. That could have allowed the “gamma buyers” applying the squeeze to step back, while the short-covering (along with the absence of new short-selling) helped propel the price upwards. (Probably some of the smart buy-side players became sellers at that point, feeding shares to the now-desperate shorts.)   

The February Surge

In February it was different.  

The volume and price now match up when the surge takes place, as they should in a more normal market. The raw correlation between price and volume in the two weeks prior to the surge was only 24%. But the correlation during the week of the 2^nd surge (Feb 22-26) was positive 95%.

The implies that the buy-side volume drove the demand in a more conventional fashion. It is consistent with the idea that the short-sellers were now on the sidelines, largely out of the line of fire. Burned badly in January, they stayed out of the market, which allowed the buying volume (generated by the “gamma squeeze” mechanism) to move the price up unimpeded. Hence the nearly perfect correlation of price and volume. 

Options Data

The options picture supports this. In the run-up to the January explosion, average call option volume increased by 300-400% in the last two months of 2020 – with spikes as high as 1000% above “normal.” Call options outnumbered Put options on every day but one between Nov 1 and January 22. This initiated the hedging process underlying the “gamma squeeze,” and slowly mopped up the “normal” sell-side interest. Some short sellers may have seen what was coming and exited their positions. Others may have increased their short positions to try to keep the lid on the stock.  

On January 22, the call options volume jumped to over 13 times the Nov-Dec average. This broke the “resistance” in the market. 

Interestingly the put options now rose above the call options for the first time. I believe this was partly opportunistic, and partly a shift by short sellers to maintain a principled negative view (that the stock was now overvalued) while exiting the dangerous short interest in the shares. 

I will also hypothesize that some of the call option sellers – prior to Jan 13 or Jan 22 at the latest – may not have realized what was happening, and may not have hedged – or not right away. When the stock took off, they were perhaps in the same position as the short sellers. They would have had to cover their naked calls by belatedly buying shares. Hard to verify, however. 

The February surge looks rather different, rather more “normal.” Volumes of Puts and Calls are essentially equal, which portrays a more balanced view in the options market.  

The options-sellers were now hip to the gamma game. Ready to respond quickly, by placing the hedges they needed. On Feb 24^th – when the surge began abruptly at about 2:36 pm in the afternoon – the hedging buys lagged the spikes in call options by just a few minutes. Here is the figure from the previous column. 

This rendered the gamma surge more surgical and more effective, producing a quicker response in the share price. The short sellers were no longer in the middle of the gamma process, selling into the demand, as they had been doing (I believe) in January. 

Short-selling is a technically complex topic. It encompasses a diverse set of trading and investing techniques and interests. Not all short-sellers share the same objectives, or respond in the same way to market movements. The data on short-selling is of a lower quality than many other types of market information. The applicable regulations are confused and ineffective (I speak in part from personal experience), as well as out of date (as are many regulatory policies in a fast-evolving high-tech financial system). I will address the subject in more detail in a subsequent column.

Nevertheless, I think the available evidence does support certain conclusions. 

Regarding the specifics of the GME event – 

• The “liquidation” of the overdrawn short interest in GME contributed to the January surge, but was not decisive
• The short interest played a minimal role in the February surge, which was a more pure example of the gamma process 

Regarding the broader significance of all this:: 

• The gamma process seems to be quite powerful. The option-cost-to-share-price leverage is real and significant
• The process does not depend strictly on the short interest to fuel the “squeeze”  
• The maneuver may therefore soon be extended to other companies that do not necessarily have a large short interest