Last Updated on June 11, 2021 by Oddmund Groette
Pairs trading involves market-neutral strategies that aim for profits in any type of market, be it sideways, down, or up. Jim Simons’ The Man Who Solved The Market describes the origin of pairs trading: at Morgan Stanley in the 1980s. Small independent traders have used the same techniques, especially at proprietary firms because pairs trading needs leverage to efficient.
There are many aspects of pairs trading and many risks to consider. In this article, we (superficially) look at what pairs trading is and its advantages and disadvantages (pros and cons).
What is pairs trading?
Pairs trading could be done in a wide range of instruments, for example, gold and silver, but in this article, we look at pairs trading in the stock market.
To have a pair you need to have two stocks, of course, and look at their historical performance and co-movement. When the price difference between the two stocks weakens, for example, one stock rising more than the other, the idea is to short the strongest one of the pair and buy the weakest one. Of course, a pairs trade could also do the opposite. There is no definitive answer to what is the best method.
The price differential is called the “spread”. The spread converges and diverges and the idea is most of the time to go against those swings in the belief the spread will converge after it has diverged.
When the spread moves in a direction, it could be due to many reasons: filling of a big order, news, beta values, or whatever reason.
Examples of pairs could be GM/F, KO/PEP, and MS/GS. If GM outperforms F over a certain number of days, most pair traders would short GM and buy F.
Below is a chart of the share prices of KO and PEP (Coca-Cola and Pepsi-Cola). As is clear, over a 12 month period the share prices didn’t deviate many percentage points from one another. During brief periods of time, they diverged before they slowly converged and returned to “normal” again. These are the opportunities pair traders look for.
Jim Simons’ Medallion Fund and pair trading:
Jim Simons and his Medallion Fund use, to my understanding, mostly market-neutral strategies.
Pairs trading – a personal experience:
Before we go on here is a short anecdote about my own experiences as a pair trader at the beginning of the millennium:
When I first started trading full-time back in 2001, I started with pairs. Why did I start with pairs trading? It was one simple reason: my mentor traded pairs and merger arbitrage.
Pairs trading gave me a good starting point and pairs trading was highly likely much more profitable at that time than today. I traded only stocks listed on NYSE – none NASDAQ pairs. At the time, the specialist on the NYSE floor had a lot of power and most of the trading went through those firms. The specialist’s job was to create a fair and orderly market and to fill big positions.
The specialist system often meant I got price improvements:
If I had a bid at 50.25 and the specialist received a sell order for 40 000 shares, he might “sweep” all the bids and fill the order at 50.05. Thus, I managed to buy at 50.05 instead of 50.25. If I was filled I could either be naked long or short the other stock to make a “market-neutral” position. It’s unlikely that you get such price improvement anymore.
I can recall some of the most popular pairs at the time:
- RDC / ESV
- SII / BJ
- BHI / SLB
- FRE / FNM
- ETG / EVT
- RQI / RPF
Traders starting out today will notice that many of the tickers don’t exist anymore. To understand more of the flux of the stock market I recommend Hendrik Bessembinders study from 2017 called Do Stocks Outperform Treasury bills?. The average life of a public company was only seven years from 1926 to 2015 and just a small minority of the stocks beat Treasury Bills.
All pairs were traded intraday, ie. as day trades. I stopped pairs trading in 2005 because I found better strategies, and most likely I didn’t manage to adapt to the dynamics of pairs trading. However, even today, after pairs trading has been around for over three decades and the most obvious edges are “arbed” away, there are still profitable pair traders out there. I still believe Bright Trading has a pairs trading group.
Advantages with pairs trading:
The three most obvious advantages with pairs trading are these:
The market direction doesn’t matter:
You don’t need to make any bets on the direction of the overall market. You make a bet on the direction of the difference of the pair (the spread). You can make money if the market moves up, down, or sideways. However, due to the different betas of the stocks you’re trading the market direction might influence the movement of the spread.
You are market neutral:
If the market drops 15%, it’s unlikely you face a 15% loss in your pairs. But be aware that the position can be poorly hedged. Market neutral doesn’t equal low risk. It’s just as much risk in pairs trading as something else. The main reason for that is increased leverage.
Volatility is good:
Disadvantages with pairs trading:
Unfortunately, pairs trading has many disadvantages as well. The most obvious are these:
Pair traders take on too much leverage:
Pairs tend to wander away:
In his book Algorithmic Trading, Ernie Chan notes that pairs trading of stocks has become more difficult over time. Two stocks may cointegrate in-sample, but they often wander apart out-of-sample as the fortunes of the respective companies diverge.
Below is an example of the ratio between PEP and KO:
Pairs trading involves black swan events
Pairs trading is most of the time smooth sailing with many small winners. The win ratio is high. Because of this, you take on more risk. This is the optimism bias. Even worse, many are tempted to double up when a pair goes against you. Because of this, many traders lose months of profits in just one day, even in minutes. Moreover, because pairs tend to yield smaller unleveraged returns than buy and hold, traders use leverage to boost earnings.
Pairs trading involves cost for shorting
When you have a short position it costs money to borrow shares to sell short. If you are trading KO/PEP, this cost is negligible. However, if you trade some pairs that might not be as liquid as those, you need to pay to borrow the shorts. Even worse, some stocks might be impossible to borrow – so called hard to borrow stocks.
Pairs trading involves slippage and commissions
One trade involves four transactions. Perhaps needless to say, this increases trading frictions. This is not an insignificant amount.
Pairs have different betas, leverage, and fundamentals
There are a zillion reasons why two stocks shouldn’t move in tandem as time goes by. For example, different business leverage might influence the stocks a great deal. In a bull market, a leveraged stock might rise much more than one which has low leverage. Likewise, is there any reason why PEP and KO should move in tandem over time? It’s hard to say. Thus, most pair traders use a short time horizon.
Pairs trading is not really market-neutral
Despite being both long and short an equal amount, perhaps adjusted for beta and volatility, you are not really market-neutral. This is because you are essentially long and short two different stocks. This is a fundamental difference to being long and short the same asset, for example, to be long and short the same instrument with different calendar spreads.
How to pick pairs for trading?
The most important issue in pairs trading is the process of selecting pairs to trade. Which criteria would you use? What is the time frame? What is risk management? Do the correlations increase or decrease in bear markets?
Correlations are important for pairs trading
If two stocks move in the same direction, the correlation is positive. If the stocks move in opposite directions, the correlation is inverse – negative. Here is a plot of four different correlations for PEP/KO:
The chart above shows the correlations of PEP/KO for 10 days (red), 30 days (black), 50 days (blue), and 100 days (green). Clearly, the numbers vary depending on the number of days. The longest one, 100 days, has over the last 1.5 years been between 0.6 and 0.95, suggesting PEP and KO are pretty correlated.
Correlation vs. cointegration
Correlation is handy but might be a bad measure for the quality of a pair. Even a positive correlation can lead to wide variations in the spread ratio. Cointegration is often a better measurement for the quality of a pair. Why? Because cointegration measures the variability of the difference between two stocks.
Let’s take an example: If two stocks A and B today have a ratio of one (1) and the historical cointegration is perfect, we can expect the ratio to be one even after one year, even though the correlation might have gotten weaker. Any deviations from the base case of one imply that you have a perfect setup for a trade (a statistical abnormality).
It makes a lot of sense to pick stocks that are mostly influenced by the same factors. For example, both KO and PEP are in the same markets (soft drinks and snacks), they are subject to the same macroeconomic and industry factors, they are likely to be valued by the same financial indicators, and they have pretty similar earnings multiple.
Let’s test a pair trade: KO/PEP
The criteria are as follow based on end-of-day quotes:
- We buy if the three-day RSI of the pair ratio (spread) is below 10.
- We sell when the three-day RSI is above 60.
If the ratio of KO/PEP is below 10, then enter at close by buying KO and shorting PEP. This is the result from 2000 until March 2021:
Let’s test a pair trade: GS/MS (Goldman Sachs and Morgan Stanley)
The criteria are these:
- We buy if the two-day RSI of the pair ratio (spread) is below 5.
- We sell when the three-day RSI is above 60.
The result from the year 2000 until March 2021 is 187% on 310 trades: 0.6% per trade before commissions and slippage.
If we change the exit to RSI(2)>75 the total returns increase to 275% on 286 trades: 0.96%.
Another popular measure when I traded pairs was to use Bollinger Bands. Unfortunately, I couldn’t make that more profitable than the simple two-day RSI strategy.
Pairs trading is still popular among hedge funds and proprietary traders. However, we suspect the profitability is one of diminishing returns. Pairs trading offers no easy money and often involves “black swan” types of returns. However, this might be offset by diversification into many different pairs. We believe it’s paramount to understand the fundamentals of the stocks you’re trading, for example, earnings, ROIC, and leverage.
The backtests we did on KO/PEP and GS/MS yielded a return lower than buying and holding the S&P 500. This makes sense because you are short one leg of the pair, and we all know short faces a constant tailwind. You need leverage for this strategy to work. However, this might be offset by a higher Sharpe Ratio due to less swings in the equity curve.
Disclosure: We are not financial advisors. Please do your own due diligence and investment research or consult a financial professional. All articles are our opinions – they are not suggestions to buy or sell any securities.