Pairs Trading – A Market Neutral Strategy
In the wake of recent volatility, we look at a market-neutral strategy for active investing.
Hedge funds
The invention of the hedge fund is generally attributed to Alfred Winslow Jones (though Buffett insists his menor Ben Graham ran the first one).
- His big idea was to offset the market risk from shares that he expected to go up with short positions in shares that he thought were overvalued.
This is the classic long-short hedge fund strategy, and boiled down to its essentials it sounds pretty smart.
- Of course, this strategy evolved into the standard 130 / 30 structure (130% long, 30% short, to leave a net 100% in the market), which is not quite the same idea.
And then “hedge fund” became an umbrella term for all kinds of strategies that weren’t the plain vanilla “100% long equities”.
- And then hedge funds underperformed when investors needed them most (through and after the 2008 crash).
And then Buffett won a big bet with a hedge fund manager that the S&P 500 would outperform the hedgies over ten years, which meant that everybody knew that it had.
- So hedge funds are an unloved asset class these days.
But we aren’t here today to talk about hedge funds.
Pairs trading
Circling back to Jones’ central idea, it’s easy to see that we can build on this.
- We can match together pairs of stocks that share similar characteristics (are from the same sector, are roughly the same size, have similar PEs, etc).
- Buying one stock and selling the other will reduce our market exposure to the minimum.
This is known as a market-neutral strategy, since we can make money whether the general market moves up or down.
- It was “invented” at Morgan Stanley during the 1980s.
But after a couple of bad years, the group that discovered it was disbanded in 1989.
- Pairs trading was also associated with the ill-fated hedge fund run by Nobel Prize winners (Long Term Capital Management or LTCM) – so it has a chequered history.
You can use indices as well as stocks in pairs trading.
- For example, you could long the FTSE-100 and short the FTSE-250.
- Or go long the S&P 500 and short the FTSE-100.
You could also trade a single stock against the index, betting that it will outperform or underperform.
- The FTSE-100 is an awkward index for this, and a sector index might work better.
Other ways to use pairs trading include:
- Commodity pairs – gold and silver, for example.
- Currencies are traded as standard in pairs (as opposed to, say, trading a single currency against a basket of other currencies).
- You can also use a “pair of pairs” to create a synthetic currency pair that your broker doesn’t offer.
- Futures versus the cash (spot) price.
- Two classes of share in the same company.
- Two listings of the same stock on different exchanges.
- Use a call option for the stock you like and a put option for the stock you don’t.
- ETF pairs, or ETFs as one side of the pair.
Pairs trading is also a component of merger arbitrage, where investors sell predators and buy their targets.
Momentum or reversion?
If we buy the stock that is doing well (or will do well) and sell the stock that is doing less well (or will do badly) we are following a momentum or trend following approach to pairs trading.
The other way to do things is to look for the maximum divergence between the two stocks in a pair, and bet on that divergence reducing in the future.
- That’s the arbitrage / swing trading / reversion to the mean approach.
- This is actually the traditional approach to pairs trading.
The traditional approach depends on a historic correlation between two stocks, and looks for a temporary divergence between them.
- It assumes that the future will be like the (medium-term) past.
By contrast, the momentum approach assumes that the future will be like the recent past.
You can’t know in advance which approach will work, but one way to think about it is to look at your time scales for the positions.
- Many pairs traders are only looking at a few days or a couple of weeks, and they usually stick to mean reversion.
We’re looking at weeks to months (at the moment), so momentum might be best.
- Momentum usually dominates share prices over three months to a year while reversion to the mean usually wins out over longer time frames.
The way that you can use the strategy in two completely opposite ways reminds me of Bollinger Bands:
- Do you expect a breakout, through the bands, or are you expecting the price to “bounce” off the top or bottom band.
Indeed, you can use Bollinger Bands to examine the ratio of the prices of the two stocks in your pair.
- It would be particularly helpful if you were interested in the mean-reverting form of pairs trading.
- When the ratios hit the Bollinger bars, it would be time to open your pair of trades.
The downside
There are four issues with these strategies:
- You are doubling your costs (commissions and spreads).
- With spread bets (or CFDs) this is slightly offset by the opposing financing costs – one bet mostly funds the other
- If the stocks move in the opposite direction to the one you want, you have effectively doubled your position size on a losing bet.
- Reducing risk always comes with reducing rewards.
- If you are pairs trading in a bull market, you might make more money by simply backing the stock you believe in.
- Of course, the leverage in a spread bet or CFD can offset this factor.
- You are exposed to the risk of a bid from the stock you have bought for the stock you have shorted.
- Of course, things could go the other way and work in your favour.
Our approach
The hedgie 130:30 strategy will do well in a rising market, but if things get choppy (as they have begun to in 2018) then a pairs strategy could be our friend.
- After a blistering January, the Bonkers portfolio has run out of steam in early February, and pairs trading could be just the tonic it needs.
We’ll start with the momentum style of pairs trading, because things are choppy at present, and we need flexibility.
My plan (at this initial stage, at least) is quite simple and boring:
- Stick to the FTSE-100 as much as possible
- Work through the various industry groups
- Look for likely pairs of matching companies
- I’ll look for the largest companies where possible.
- Note that in classical pairs trading, I would be looking for stocks with high historic correlations – I won’t do this as I just want liquid stocks that are representative of their sectors.
- If three or four stocks look evenly matched, select all of them.
- Use trend following (moving averages) to identify the best and worst stock in each group.
- Bet on the good stock and short the bad one
- We’ll vary position sizes to make sure that we are market neutral.
If things go well with the FTSE, I’ll try to extend the system to the largest US stocks.
- But that’s a job for another time.
Conclusions
That’s it for today.
- Before the next post in this series, I’ll work through the 54 industry groups in the UK stock market.
My aim is to come up with around 50 groups of stocks that are potential pairs trades.
- Most will actually be pairs, but some groups will have three to six stocks in them.
In the next post I’ll add the momentum analysis to these groups and work out which trades we should place first.
- The goal is to build up a mini-portfolio of 10 to 20 open positions to sit inside the Bonkers Portfolio.
Until next time.