Imagine two businesses operating in the same market, running the same strategies, and offering similar products. Although they’re competing with each other, their earnings and stock performance tend to move in sync. Both are highly correlated; it’s almost like they’re tethered together.

　　When one stock moves up or down, the other tends to follow suit. One stock may significantly jump ahead or fall behind the other (i.e., fall out of correlation), but such anomalies have, historically, turned out to be short-term blips. In a matter of time, their movements end up converging.

　　This high degree of correlation, suggesting that one stock will always (eventually) catch up with the other, is at the heart of a strategy called pairs trading.

　　The assumption is that highly correlated stock pairs will continue to move in a comparable, measurable, and hence predictable manner. And if their prices diverge, it can present a profit as prices converge back toward their correlated state.

　　But sometimes, two historically correlated stocks diverge and stay that way, due, perhaps, to a permanent fundamental change in one of the two companies. There’s a pairs trading strategy for that case as well.

　　An important note for this discussion: Selling stocks short requires a margin account, which typically has special rules and capital requirements.

　　Pairs trading, arbitrage, and market-neutral definitionA pairs trade is a market-neutral trading strategy in which you buy (take a long position in) a stock or other security and sell (take a short position in) another security. The typical goal of a pairs trading strategy is to find a pair that has fallen out of its historical correlation and seek to profit as the spread between them closes back toward historical correlation levels. It’s a type of relative-value arbitrage.

　　And what exactly is a market-neutral strategy? It’s one with no directional bias—it doesn’t matter if the prices of both securities in a pair go up or down. In the case of pairs trading, all that matters is whether their relative prices converge back toward their historical spread levels.

　　And if they don’t converge? That’s the risk of arbitrage-based pairs trading. (But it’s also an opportunity for anyone who’s been watching the fundamentals and is expecting a divergence.)

　　A simple pairs trading exampleImagine two stocks—ABC and XYZ—that usually move together by a comparable amount. So, if one stock gains, say 2%, the other stock also rises around 2%.

　　A pairs trader might expect this difference to correct, that is, for the prices to move back in sync sooner or later. The trader sells short ABC, expecting it to fall or stay flat, and buys XYZ, expecting its price to rise and get closer in range to the other.The trader profits if the stocks converge back to their historical price relationship.The trader wasn’t betting on both stocks moving in any given direction, up or down. All the trader cared about was that the stocks’ prices should revert to their usual price relationship—not jumping too far ahead or behind relative to the other.

　　Real market example: Coca-Cola and PepsiCoThe Cola Wars rivalry between Coca-Cola (KO) and PepsiCo (PEP) is a classic example of two similar companies that are locked in a perpetual battle over the same customers. Every now and then, one gets ahead of the other. But generally, their movements tend to shadow one another.

　　When the black line is near the top (green circles), and approaching the blue dotted line that marks 1.00 (or 100%), it means that KO and PEP are moving in sync with one another (almost 100% correlated).Compare the dotted line in the middle (0% correlation). When the black line reaches or falls below it (red circles), it means that the KO/PEP pair is no longer correlated or even negatively correlated.

　　Now, look at a price chart (the top half of figure 1) to see how these correlations come into play. Do you see how the pairs generally move together over time? When prices de-correlate (red circles and boxes), notice how the price spread between KO and PEP widens.

　　Trade A: In September 2023, PEP’s stock price soared well above KO; the correlation, typically near 100%, fell to zero.

　　A pairs trader might have sold PEP short while buying an equal (dollar) amount of KO.When prices converged near 100% correlation later that month, the trader would have closed the position at a profit.Note that both stocks fell during that period, but that gains from the PEP short trade would have more than compensated for the losses in KO, enough to generate a profit.Trade B: By the end of January 2024, price correlations shifted from above 75% to -10% as PEP share prices dropped significantly relative to KO.

　　A pairs trader might have bought shares of PEP while selling short KO.By the end of April, the trader would have closed the position at a profit.Although KO traded sideways during the period, possibly ending with a small loss, the bulk of the profit came from PEP, whose shares appreciated as the spread between the pair converged in price.It’s worth noting, however, that it took three months for this trade to come to fruition. For a pairs trader looking for an arbitrage-type convergence, that’s a lot of time between trade initiation and liquidation, considering a pairs trade requires a short stock position.

　　Decorrelation: The big pairs trading risk (and opportunity)The first two periods of noncorrelation between Coke and Pepsi resulted in a return to correlation. But what if you had repeated the process in June 2024 (trade C)? Notice that the spread didn’t stop at zero; it widened further to -25%.

　　Will prices converge, or has something fundamentally changed with either company that might cause the nature of the correlation to change? And at what point do you throw in the towel on this trade and close it out for a loss?

　　Pairs trading comes with an inherent risk that the price relationship between two stocks might change. This can happen due to significant financial challenges faced by an underperforming company, remarkable advances made by an outperforming company (making it difficult for its competitors to catch up), or other market-related conditions. For example:

　　In a competitive scenario between two burger joints, one shifts to a vegan menu while the other maintains its “beef and chicken only” strategy.One car manufacturer makes a massive shift from internal combustion engines, while the other doesn’t.A bank decides to go digital and close branches, while its competitor retains its brick-and-mortar model.A telecommunications company is charged by a regulator with “accounting irregularities,” while it’s business as usual for its chief competitor. The outcome of any of these situations will be dynamic and unknown, but the correlations are almost certain to change. And although that could spell trouble for the pairs trader who’s expecting a return of the old correlation, it could mean an opportunity for traders who follow fundamental shifts in the marketplace and position themselves for permanent decorrelation.

　　Other classic examples of stock pairsWhat other stocks have been traded as pairs? Here’s a short list:

　　Ford (F) and General Motors (GM)Visa (V) and Mastercard (MA)Apple (AAPL) and Microsoft (MSFT)ExxonMobil (XOM) and Chevron (CVX)Goldman Sachs (GS) and Morgan Stanley (MS)McDonald’s (MCD) and Yum! Brands (YUM)The bottom lineRemember: You can’t do a pairs trade unless you have a margin account—it’s required for anyone who wants to short sell stocks. And unless you’re an experienced trader who fully understands the risks of shorting stocks—and has the time, patience, and emotional wherewithal to monitor short positions—you might want to ease into pairs trading. Do your homework, study patterns, and learn the different types of pairs trading strategies you can use and the opportunities and risks they present.

　　Pairs trading assumes a consistent price relationship between two highly correlated stocks. But as the boilerplate financial disclaimer says, past performance does not guarantee future results.