For months, Twitter (TWTR) investors have anticipated a February 15, 2014 lockup expiration in TWTR shares as a significant event that could depress the share price. The expectation that Twitter's non-management employees would rush to dump their shares into the market caused many to believe the lockup expiration presented an easy shorting opportunity. But on the first trading day after the lockup expiration, TWTR shares opened higher and stayed that way for all but 2-3 minutes of the trading day, and traded at far below-average volume.
To any seasoned market observer, this should not have come as a surprise for several reasons, including widespread misunderstanding about the lockup expiration, the small size of the lockup expiration, the one-sided expectation of the trading opportunity, and what appears to have been a clever trick by brokers to use call options to mute the effects of the lockup expiration.
1. Many were mistaken about the February 15 lockup expiration.
Several news reports that discussed Twitter's February 15 lockup expiration (e.g., here, here, here, and here) may have left the investing public with an overly-simplistic belief that an event would happen on February 15 that would make an additional 9.9 million TWTR shares eligible for trading. But a more accurate description of the February 15 expiration is described in Twitter's IPO Prospectus, at page 161:
"beginning as early as February 15, 2014, up to an aggregate of 9,867,228 shares of our common stock that are held by our employees who are not executive officers may be eligible for sale in the public market in order to satisfy the income tax obligations of such employees resulting from the vesting and settlement of the outstanding Pre-2013 RSUs for which we expect the service condition will have been satisfied on such date and through the end of the lock-up period (or up to an aggregate of 3,282,859 shares of our common stock held by our employees who are not executive officers if we choose to undertake a net settlement of all of these awards to satisfy a portion of such income tax obligations)"
Parsing this statement closely shows that the February 15 expiration might not have been as simple as many believed. Consider the following:
First, the prospectus does not state that all 9.86 million shares would become eligible for trading on February 15, only that they will become eligible "beginning as early as February 15". This means that the lockup expiration might not have been a single event on February 15, but might instead be a rolling event that happens over many months. The Twitter statement specifically mentions that the timing depends on an expectation that a "service condition" related to the "vesting and settlement of Pre-2013 RSUs … will have been satisfied on such date and through the end of the lock-up period". By drawing a distinction between "such date" and the "end of the lock-up period", it appears that the shares will be released on some rolling basis. Thus, the so-called February 15 lockup expiration event might occur over the course of months. There appears to be no public information about the precise timing of when these TWTR shares become eligible for trading.
Second, the prospectus does not state that 9.86 million shares will become eligible for trading. Instead, it states that "up to an aggregate of" 9.86 million shares might become eligible. This means the release might involve any number less than 9.86 million shares, even a number significantly less. In fact, the prospectus indicates that instead of releasing shares to trade, Twitter might "choose to undertake a net settlement of all of these awards to satisfy a portion of such income tax obligations" and if it does, then only "up to an aggregate of 3,282,859 shares" will become eligible for trading. Thus, the 9.86 million share lock up expiration might only be a 3.28 million share lockup expiration. There appears to be no public information about what Twitter decided to do.
While many mistakenly believed that 9.9 million new TWTR shares would hit the public markets on February 18, in reality there may be as few as 3.3 million new shares that hit public markets between February 15 and May 6.
2. Whatever the lockup expiration, it was too small to impact TWTR trading.
In an article I wrote last week, I explained that major financial websites have presented a misleading picture of the short interest in TWTR shares. To briefly recap, Nasdaq reports that 33.8 million TWTR shares were sold short on January 31, 2014, a number that has risen with each short interest update posted since TWTR shares started trading publicly.
Yahoo finance has determined that Twitter has 362.6 million outstanding shares, while other sites such as Google Finance, MSN Money, and E*Trade represent it has 544.7 million outstanding shares. Using these figures, the sites claim TWTR has a 9.7% or 7.0% short interest. In practical terms, a short interest below 10% would not be expected to result in an enormously significant impact on daily trading levels.
But these short-interest numbers are grossly misleading because while it may be technically true that TWTR has 544.7 million (or more) shares, only 80.5 million of those shares traded before February 15, while the rest remained locked up. This means that before February 15, the effective short interest for TWTR shares was a staggeringly high 42.0%. In a Barron's report of the companies with the largest percentage of shares shorted, Twitter's 42.0% effective short interest would have put it in sixth place on the New York Stock Exchange for most actively shorted shares. Think of that: Only five companies on the entire New York Stock Exchange with a higher short interest. (Even if all Nasdaq stocks are added to the list, TWTR would still have been the 14th most heavily shorted stock.)
Assuming a new 9.86 million shares became eligible for trading on February 15, 2014, and that somehow none of those ended up in new short positions, it would have increased the effective float by about 12%, from 80.5 to 90.36 million shares. But that increase would have only reduced the effective short interest by an equivalent 12%, from 42.0% to 37.4%. Even with such a drop in short interest to 37.4%, Twitter would have only dropped five spots on Barron's list of highest short interest as percentage of shares, to 11th place. (If all Nasdaq stocks are added to the list, Twitter would have only dropped from the 14th to the 23rd most heavily shorted stock on the combined list.) But almost certainly, the February 15 expiration did not cause Twitter's effective short interest to drop to 37.4% because (a) Twitter's shares short have increased every reporting period, and there is no reason to think that trend has stopped, (b) all 9.86 million shares might not have become eligible for trading, and (c) even if they did, millions of those new shares likely got added to new short positions.
The table below is from information provided by Barron's. I have inserted Twitter where it would have appeared before February 15, and where it would appear today assuming all 9.86 million shares became eligible for trading without any new short positions. Regardless of what happened with the February 15 lock up period expiration, Twitter remains one of the most heavily-shorted companies.
Blyth, Inc. New Common Stock
ITT Educational Services, Inc.
Weight Watchers International
Veeva Systems Inc. Class A Com
NQ Mobile Inc. American Deposi
Twitter Inc. (before February 15, 2014)
Radioshack Corporation Common
Tri Pointe Homes, Inc. Common
J.C. Penney Company, Inc. Hold
Walter Energy, Inc. Common Sto
Twitter Inc. (after February 15, 2014)
Textura Corporation Common Sto
Nationstar Mortgage Holdings I
Cliffs Natural Resources Inc C
Lindsay Corporation Common Sto
Molycorp, Inc Common Stock
3. Brokers appear to have used options to mute the effects of trading on February 18.
I noticed something unusual with TWTR options this past weekend. Since December, I have had various TWTR options positions. One trade I used was to sell call options (which is similar to having a short position, except that I earn the option premium in exchange for limiting my gains to the strike price). Twitter options have generally had high premium, and the strategy has allowed me to earn income from premium deterioration while waiting for the share price to come down. I have always rolled options before expiration and I have never had a TWTR option assigned … that is, never until February 15.
On February 15, the day of the lockup expiration, all of the deep ITM February 22 TWTR calls that I had sold were assigned. On one hand, it was a pleasant surprise because I immediately recognized the entire premium for the options. But it also made me realize this could have been part of a clever tactic by brokers to sell newly eligible TWTR shares in a way that avoids a massive glut of supply during normal market hours on February 18.
If true, the tactic worked something like this. Twitter employees holding some part of the 9.86 million shares related to the February 15 lock up expiration had to place and sell their shares through a broker. These brokers likely had advance notice of approximately how many shares would hit the market as of February 15. Possibly fearing a glut of shares hitting the market on February 18 (the first trading day after February 15), these brokers could have accumulated call options in the days and weeks leading up to the expiration, i.e., buying the future right to purchase shares at a set price. I sold some of those call options. When February 15 arrived, the brokers who held call options could satisfy the sale of the newly eligible shares by exercising those options. As a result, the "up to 9.86 million shares" that became eligible for trading on February 15 did not create a glut of supply on February 18. (In fact, while many expected a deluge of trading on February 18, only an unusually low 10.7 million shares traded that day.) The tactic could have been particularly profitable for brokers who accumulated calls options when TWTR traded around $50-$54 and held them until they were assigned on February 15, when TWTR was valued at $57.44 per share at the most recent close of trading price. If these brokers held very large long positions, they could have caused events to inflate the price of TWTR shares leading up to the February 15 expiration.
If anyone reading this also had their Twitter options assigned on February 15, prior to expiration, please mention it in the comments. It would be interesting to see if my experience was typical.
4. Too many people on one side of a trade can ruin it.
It is common wisdom that if too many people take one side of an "obvious" trade, it is highly susceptible to go in the opposite direction. This is particularly true when trying to profit by shorting a heavily-shorted stock. A short squeeze can create truly absurd and irrational-seeming stock price movements. Of course, it's only absurd and irrational if one considers the share price to be the equivalent of a fair market value for the company it represents. In class short-squeeze scenarios, the share price seldom equates to a reasonable value of the company, but instead to the supply and demand of a limited number of shares. And with Twitter having an effective short interest of 37.4% to 42.0%, and with the company grossly-overvalued by traditional valuation metrics, its share price can be influenced by factors consistent with a classic short-squeeze play.
One important factor of a short squeeze is that disillusioned shorts who eagerly anticipated an event (such as the Twitter February 15 expiration) ultimately throw in the towel. In the coming days and weeks, many such disillusioned and impatient TWTR shorts who feel cheated by TWTR's post lockup price move might see no meaningful near-term catalyst to drive down TWTR shares, and they might cover their short positions, thereby creating further upward pressure on TWTR's share price.
In summary, it should not be surprising that TWTR's share price did not collapse after the February 15 lock up expiration. It is nearly impossible to know the precise timing or number of new shares that hit (or will hit) the market, but whatever the timing and number, it is not enough to fundamentally alter Twitter's massive short interest.
During the next several months, I expect TWTR share price to be highly volatile. Until the May lockup expiration, it seems reasonable to expect TWTR to trade between $50 and $75, not because that price represents a rational valuation of the company or its economic prospects, but because there are far too many people who have shorted TWTR shares. It would not shock me to see TWTR shares trade up to $100 or even much higher; again, not because it would represent fair value for Twitter the company, but because those moves can be orchestrated when a stock has such a high short interest. By the end of the year, I expect TWTR shares to drop back to the $20s or $30s, but the exact timing or catalyst for that happening is anyone's guess. (The most obvious catalyst might be the May 6 lockup expiration, but never underestimate the power of the masses to get disappointed by a seemingly sure-thing trade. And with another Twitter quarterly report within days of that expiration - the first expiration when employees and funds can cash in - it would not be surprising that Twitter will try to greatly impress the market with news at that time.) Until then, I'll try to keep a net short position on TWTR, but only in an amount where I can withstand the potential shock of a massive price spike.