Quantitative Short Selling, Why Bother

Fraser Dawson 25 Sep 2012

Many professionals sell short a stock in order to make a profit just as they do with their long portfolios. However, that is not the primary focus of our short portfolio.

Because the market (eg a long tracker) gives you 6+% per annum (if you stay in it long enough), a short portfolio has to generate a 6+% excess return just to break even. And that is not including trading expenses, which are greater for short selling than normal long trading. That is quite a headwind to sail against!

On the plus side, because most equity investors are long-only, there may be more mispricing opportunities to take advantage of.

So most investors should back off now and concentrate on their long portfolios.

But a very small number of mechanical equity investors may be interested in this portfolio for asset allocation and diversification purposes.

Take a quick look at some of the other asset classes.

  • Many view the bond markets as very expensive.
  • Commodities futures are very difficult for retail investors to get exposed to. ETFs are available but often come with a "Contango risk"
  • Currencies are a "zero sum game"
  • People will have their own views on property
  • Your cash savings are being quickly eroded by inflation.

So, not very enticing, agreed? But who wants to be long only with all of your assets with the macro environment we have just now?

For what it is worth, a backtesting exercise has been done to demonstrate the idea. A long strategy not dis-similar to Greenblatt's magic formula was backtested alongside a short strategy. Then the results were combined to approximate what would have happened if you had split your capital equally between the two. See the chart below.

The main caveat is that when a computer backtests a short strategy, it assumes all stocks are borrowable which they are not. In my experience, a short strategy that relies heavily on a value metric provides companies that are often not borrowable. This is also true for a short strategy that relies a lot on poor momentum to pick its shorts. This is probably because the hedge funds have already borrowed all available stock.

So the strategy tested below relies only a bit on value and momentum and instead picks companies with poor capital control. As well as providing companies more likely to be borrowable, these stocks tend to be lower beta than their poor value / poor momentum brethren.

As you can see, both the long and short strategies performed admirably on their own. They had only four and three negative years out of the 11 years backtest, respectively.

But the combined strategy had no negative years!

Another way of looking at this is via cumulative value of the portfolio over the years:

Short Selling in Practice

It is important to realize that your short strategy must not be the mirror image of your long strategy. Otherwise, you are actually making the same bet with both your longs and shorts. Take, for example, a value/momentum long strategy. Although over time this strategy provides alpha, periodically the market punishes these stocks. If your short strategy picks expensive stocks with poor momentum they will lose you money at exactly the time your longs are also losing! Hence, no hedge.

As it happens, value is a less effective KPI when picking shorts than longs. It will lead you to short lots of biotech/pharma/oil exploration/gold mining companies where historical profits are not what is driving the price. These companies are often expensive because of their pipelines or potential.

More effective KPI’s for picking shorts are those that highlight poor capital discipline and/or aggressive accounting. Rather get yourself a list of companies that are tapping the market for more credit, selling off their own shares or their cashflows don’t match their reported earnings. Only then is it a good idea to apply value and momentum KPIs.

We are thinking about launching a short-selling newsletter if there is enough interest.

If you are interested, please contact us.