During 2012, I have been caught leaning against not in one, but two giant short squeezes: Sears (SHLD) and the iPath DJ-UBS Natural Gas TR Sub-Idx ETN (GAZ). These are incredibly frightening events for anyone short, but also incredibly instructive. For those traders that ever contemplate taking short positions, I believe the lessons I am about to tell are hugely important and might make the difference between successful investing, and a full crash and burn. If you are one such trader, I'd go as far as suggesting you either print, or bookmark, this article.
A short goes against the long term bias in the market
The stock market has a positive bias. It goes up over time. It goes up because of economic growth, of population growth, and most significantly, because of inflation since the stock market is what I call "a nominal reality".
This means you can't marry to short positions, you can't be structurally short and you need very specific catalysts and reasons to be short. Do not become a permabear - such is a losing strategy over time. This is even truer today, with almost every central bank in the world willing to print money hand over fist to keep markets going north. If the stock market goes down 10-15-20%, always contemplate buying, always contemplate going net long. If you keep shorts, have them be just hedges and not some kind of permanent investment positions.
A short is structurally different
A short is structurally different from a long position. In a long position, the most you can lose is what you invested and unless you use margin or stick everything into the same stock, you can't really blow up the entire portfolio by means of a long position. Compounding works in favor of long positions, not against them - this means that as a long position goes against the trader, its weight in the entire portfolio grows smaller, and as it goes in its favor, it grows larger (compounds).
Not so with a short position - a short position, even a small short position, can produce almost limitless losses. The short position compounds against the trader - if the short position goes against the trader, its weight in the portfolio grows larger, and can, during a short squeeze, reach a size several times its original size. This means that even a small short position, say, of 10% of a traders' portfolio, can grow so much as to wipe out the entire portfolio, whereas in a long position the most the trader would lose would be 10% of his portfolio. This distinction isn't just academic - many short traders back during the dotcom bubble were wiped out by single trades on a single worthless stock in a matter of days (on this regard, I highly recommend the following trade story - the narrator is not me, it's really well written, hilarious).
For there to be a mispricing …
This lesson is that if there's a mispricing in the market, then it can always get larger. This applies even to a mispricing involving two identical equities - like Studioso Research once showed regarding the Bank of Ireland ADR (IRE) in his article "Bank Of Ireland Arbitrage: ADRs At More Than Double Value Of London-Listed Shares".
So if you are arbitraging something at a 3% discrepancy, you have to be ready for it to turn into a 30% discrepancy. And if you then get a 30% discrepancy, you need to be ready for a 100% discrepancy.
GAZ showed this in spades. I learned of its discrepancy to intrinsic value when it was trading at a 42% premium, and was able to open a position at around 60-70% premium. Yet, it got as far as an 170% premium, and even today it still trades at a 79% premium (though the nature of my position - I sold ITM calls that are now OTM - means I'm already well in the black).
This brings us to the next lesson …
No matter how irrational and absurd the mispricing is, no matter how certain you are of the fundamentals and the righteousness of the position you are taking, you need to have strict risk controls. Even if it's something that's certain to collapse, like the difference between an ETN and its intrinsic value, or the difference between an ADR and its stock in the original stock exchange. This difference can get as high as 100% or 170%, or famously with a Gold ETF a few years ago, 1000%. Such means that the position sizing even going against total irrationality must always be small. It's the very nature of short positions that demands it, or else you lose your main advantage versus irrationality: staying power to see the trade through to its inevitability.
Now imagine this: if in a short trade where you're nearly 100% certain of the outcome you still need to have a small position sizing to have staying power, what does that tell you about trades in stocks where the value is a much more subjective affair, like Sears ?
Buy-ins and option exercises
It gets weirder and harder. When you're shorting some stock that's being squeezed, and where there's few shares available to borrow, you can easily get bought in by your broker. This means that your short can be closed at the worst possible time. It happened to many shorts during the Sears short squeeze. So you need to plan for this eventuality. How can you do so? Well, you can steer clear of hard to borrow situations, or use options. If you use options, and are buying puts, those will probably be very expensive so you need some kind of catalyst to have the time exactly right. Or you can sell naked calls (like I did on GAZ), but this presents another problem.
Calls can be exercised well before maturity. This usually happens when the calls trade at, or below, their intrinsic value (the difference between the strike price and the price the stock trades at, for ITM calls). Again, this is something that happened both in Sears and GAZ. I actually had calls I had sold naked on GAZ be exercised twice on me. Once this happens, you have two choices - you either keep the resulting short position, which will probably be bought in within 3 days, or you can choose the timing to buy back the stock position, and short calls again (exposing yourself to early exercise again). Either way, it's not an easy position to be in.
The cost of staying short
The final lesson is something I have already talked about in my article "The Sears Short Squeeze Is Loosening Up". Shorting can have a cost, and that cost can be quite high in situations where the stock is hard to borrow. The cost is represented by a negative short rebate rate, and got as high as 83% (yearly rate) during the Sears short squeeze. Indeed, the drop in this rate is itself a sign of a loosening short squeeze since it represents a lower difficulty to borrow shares to stay short.
This cost might become relevant, especially in situations where you expect to need to be short for a long period - Great Northern Iron Ore Properties (GNI) comes to mind.
Including shorting in an investment strategy is not for the faint-hearted. Many an investor simply doesn't short - this includes, for instance, Warren Buffett. If you don't heed the lessons shown here, you're probably better off not shorting either.
P.S. I didn't learn these lessons with these two short squeezes - I had learned them before, so I easily managed these two positions even though these two short squeezes were rather brutal. Keep these lessons in mind, for as you can see, they can be useful even under extreme situations.