Short notice – The fundamentals of shorting on valuation are less sound than when going long

12th November 2014 by The Harried House Hunter

by Jamie Lowry

In recent months, The Value Perspective has found a few different ways to look at the idea of ‘shorting’ – that is, seeking to profit from the failure of a business or a decline in its share price. These have included the apparent reluctance of hedge funds to short at present (Short change) and the sort of lessons other investors can learn from short-sellers (Short shift).

We are therefore grateful to a blog by US value investor Alon Bochman that offers a fresh angle on the subject – not to mention a fresh opportunity for another bad ‘short’ pun. In Do valuation shorts work?, Bochman takes the unusual approach of suggesting there are only two broad reasons you might short a stock – either you think the company is a fraud or you think it is on the wrong valuation.

One way or another, we have touched on both these points – not least, with the opinion of Greenlight Capital founder David Einhorn that “It is dangerous to short stocks that have disconnected from traditional valuation methods. After all, twice a silly price is not twice as silly; it’s still just silly.” In other words, if something that is egregiously overvalued doubles, it is still egregiously overvalued.

But what Bochman has tried to measure is if shorting works in general and if one sort works better than another. The idea of shorting shares that stand on very expensive valuations does of course sound like the sort of thing we would do if we ran a short book – although, as Einhorn hints, there are behavioural finance reasons why value may not work so well on the short side as it does in the long-only world.

At this point, we should repeat that Bochman has written a blog, not an academic paper. The numbers he has crunched, which come courtesy of a new research service called Activist Shorts, cover 400-plus shorting campaigns conducted between 2002 and 2014, which is a pretty small sample and a comparatively brief time period.

Furthermore, the campaigns assessed are only ones that have been made public so there are undeniably significant biases within the data. Those caveats aside, however, the research throws up some interesting results, revealing that the average price change over a campaign – which, since dividends are not taken into account, does flatter the numbers on the short side – was -14%.

Some two-thirds (65%) of campaigns were successful – in other words, the target’s share price was lower at the end than it was at the start – while, in 4% of cases, the target’s share price dropped 99%, representing a big win for the short-sellers. Where it gets especially interesting though – at least to The Value Perspective – is when Bochman splits the campaigns between fraud and valuation.

We might question the precise methodology here but there is no doubting the stark beauty of the results. Of the 219 campaigns categorised as ‘valuation’ shorts, the average return was +3% – that plus sign being bad news for the short-sellers. Of the 229 ‘fraud’ campaigns, however, the average return was a striking -30%. That overall -14% appears to disguise disparate positive and negative numbers.

Taking these results at face value then, what conclusion may we usefully draw? Well, it would appear that shorting stocks – even on the valuation grounds we would naturally favour – may not always be the best strategy. The quotation that keeps recurring in our mind here is the one attributed to John Maynard Keynes that “The market can stay irrational longer than you can stay solvent”.

Also, to pick up on the point we made in For the best, while as long-only value investors we impose no explicit time limits on ourselves or our investments to pay off, shorting involves paying to borrow the shares you are betting will fall in value. As such, you have a much shorter window for things to come good for you and, given margin calls and so forth, the pain could be a whole lot worse if they do not.

Leave a Reply

Your email address will not be published. Required fields are marked *