- How a Coke-Pepsi Pairs Trade Makes Money
- What the Pairs-Trading Research Actually Found
- Why Coke and Pepsi Can Keep Moving Apart
- What Walmart-Target and Royal Dutch-Shell Add to the Story
- Why Waiting for the Gap to Close Can Be Expensive
- Author’s Take: Understand Why the Gap Opened
Suppose Coca-Cola stock races ahead while PepsiCo stock barely moves. Which one catches your attention?
The winner has a convincing chart. The laggard might look like a bargain. A pairs trader sees something else: a chance to own PepsiCo while betting against Coca-Cola, hoping the distance between them shrinks.
That is the idea behind pairs trading. It aims to profit when an unusually wide gap between two historically linked stocks narrows. Academic studies have found that versions of this strategy earned positive historical returns across portfolios of stocks.
The difficult part is figuring out why the gap opened. Sometimes the market is temporarily treating similar businesses differently. Sometimes one business really is doing better.
Coke and Pepsi make that distinction surprisingly easy to understand.
How a Coke-Pepsi Pairs Trade Makes Money
Start with two familiar companies: Coca-Cola, which trades as KO, and PepsiCo, which trades as PEP.
Both sell drinks around the world. They face some overlapping forces, including consumer spending, packaging costs and changing tastes. That makes them plausible candidates to investigate as a pair.
Now suppose a trader has tested their historical relationship and decides that Coke has moved unusually far ahead of PepsiCo. The trader opens two positions together: $1,000 of PepsiCo shares and a $1,000 short position in Coca-Cola.
Shorting means borrowing shares and selling them, then buying shares back later to return to the lender. Buying them back at a lower price produces a gain; buying them back at a higher price produces a loss.
For the pair to make money before costs, PepsiCo needs to outperform Coca-Cola after those positions open. A simple example shows why that can happen even when both stocks fall.
| What happens after entry? | PepsiCo shares owned: $1,000 at entry | Coca-Cola shares shorted: $1,000 at entry | Combined profit or loss |
| PepsiCo rises 10%; Coca-Cola rises 5% | +$100 | −$50 | +$50 |
| PepsiCo falls 5%; Coca-Cola falls 15% | −$50 | +$150 | +$100 |
| PepsiCo falls 5%; Coca-Cola rises 15% | −$50 | −$150 | −$200 |
Source: Illustrative calculations. Each row is an independent hypothetical outcome using unchanged share quantities and equal dollar positions at entry. Figures are in US dollars and exclude dividends, borrowing and financing costs, trading fees and taxes.
In the second row, owning PepsiCo loses money. But the gain from shorting Coca-Cola is larger, leaving the combined trade ahead.
The third row is the painful ending. PepsiCo keeps weakening while Coke keeps strengthening, and both sides lose money together. Having two positions does not guarantee that one will rescue the other.
The $50 gain in the first row is a dollar profit, rather than an account return percentage. There are $2,000 of combined positions, and the capital required depends on collateral and margin requirements. Proceeds from selling borrowed shares do not make the trade free.
There is another detail hidden in the phrase “Coke has run ahead.” The comparison needs the same starting date, adjustments for stock splits and dividends, and a defined historical relationship. Comparing the dollar price of one Coca-Cola share with one PepsiCo share tells you very little.
What the Pairs-Trading Research Actually Found
The idea has a serious academic history. In a paper published in The Review of Financial Studies in 2006, Evan Gatev, William Goetzmann and Geert Rouwenhorst examined US stock data from 1962 to 2002.
Their method first compared twelve months of stock histories, putting each stock’s dividend-adjusted performance on a common starting scale. It selected pairs whose paths had been particularly close, then traded them over the following six months.
An entry required a gap greater than two historical standard deviations: an unusually large separation relative to that pair’s past behaviour. Positions closed when the paths crossed again or the trading period ended.
The authors reported average annualised excess returns of up to approximately 11% for portfolios of selected pairs, before the trading-cost adjustments examined separately in the paper. This was a historical portfolio result under specified rules, rather than a return promised to someone trading Coke and Pepsi.
The international evidence is substantial too. Heiko Jacobs and Martin Weber’s 2015 paper, On the Determinants of Pairs Trading Profitability, examined more than 200 million stock-pair combinations across 34 international markets, alongside a separate US analysis. Its international sample covered 2000-2013.
The authors found historical profitability, but also found that the reason for a divergence mattered. Pairs that separated around company-specific news, such as earnings announcements, produced lower subsequent strategy returns than pairs on average. A fresh piece of information could justify a lasting difference.
Results also depend on costs and the rules used. Binh Do and Robert Faff’s 2012 study, covering the US market from 1963 to 2009, found much more modest profitability after trading costs; the contrarian strategies they examined were largely unprofitable after 2002. Different samples and methods can produce materially different outcomes.
Together, these studies support investigating a statistical pattern. They do not establish that familiar rivals must reunite, that every trade is likely to win, or that the approach always beats owning stocks.
Why Coke and Pepsi Can Keep Moving Apart
Here is the catch hiding behind the familiar logos: PepsiCo owns much more than Pepsi.
Its 2025 annual report shows that 58% of net revenue came from convenient foods and 42% from beverages. Its portfolio includes Lay’s, Doritos, Cheetos and Quaker, alongside its drinks.
Coca-Cola’s business centres on beverages, including the sale of concentrates and syrups to bottling partners. So a trade involving these companies also brings together different product mixes and operating models.
If snack demand weakens while demand for Coca-Cola’s drinks stays strong, PepsiCo can face a problem that its cola rival does not share to the same extent. Different exposure to countries, currencies and distribution costs can create further differences.
In that situation, a widening stock-performance gap might reflect changing expectations for profits. A historical chart alone cannot tell you whether those expectations are wrong.
This is where mean reversion, the idea that an unusual spread moves back towards its historical range, becomes a hypothesis to test. Stocks can move in the same direction often enough to look closely related while their relative values drift farther apart.
A serious analysis therefore needs more than a high correlation. It needs evidence that a properly constructed spread has repeatedly returned towards a reasonably stable range, and a business reason to believe that relationship could continue.
Position sizes matter too. Equal dollar positions make the example easy to follow, but a model may use different weights to account for how the two stocks move. Buying one share of each does not create a balanced pair.
Even a good historical relationship needs checking against later data that were not used to choose or tune the strategy. Otherwise, it is easy to discover a rule that fits yesterday beautifully and disappoints tomorrow.
What Walmart-Target and Royal Dutch-Shell Add to the Story
Walmart and Target: Similar Stores Can Have Different Fortunes
Walmart and Target sound like another natural pair. Both are major US retailers, but being exposed to the same consumer economy does not ensure matching results.
Their disclosures in May 2025 offer a useful historical example. Walmart reported 4.5% growth in Walmart US comparable sales, excluding fuel, for its first quarter of fiscal 2026. Target reported a 3.8% decline in comparable sales for its first quarter of 2025, ended May 3.
The companies use different fiscal-year labels and their quarter-end dates differ slightly. Even so, the releases show two retailers experiencing very different sales trends during broadly the same period.
Those are business results, rather than stock returns or a pairs-trading backtest. They explain why a trader would need to investigate a gap: customers can respond differently to two competitors, and the market may have a sensible reason to price them differently.
Royal Dutch and Shell: Even a Contractual Link Did Not Keep Prices Aligned
A stronger historical example involved Royal Dutch Petroleum and Shell Transport and Trading, the separately listed parents of the Shell group before its later unification.
Their arrangement divided the group’s economic interests in fixed 60:40 proportions. That gave their relative valuations a much firmer anchor than a shared industry or similar products.
Yet research by Kenneth Froot and Emil Dabora, published in 1999, documented deviations from the relationship implied by those cash-flow rights. Relative prices were also associated with the markets where the shares traded most actively.
This case makes the timing problem especially clear. Even when the economic link is unusually strong, the market does not promise to close a gap on a trader’s schedule.
Why Waiting for the Gap to Close Can Be Expensive
A historical chart can make a temporary loss look harmless once you know the ending. The trader living through it has to keep the position funded.
Stock borrowing can cost money, and a short seller must compensate the lender for dividends paid on borrowed shares. Margin requirements can force additional cash into the account, while a rising shorted stock can create theoretically unlimited losses on that position.
Trading fees and the difference between quoted buying and selling prices add further friction. If the expected gain from convergence is small, those costs can consume it. If the trade takes much longer than expected, ongoing costs keep accumulating.
Pairs trading is often described as “market neutral” because its two sides aim to reduce exposure to broad market movements. That neutrality is an objective to measure and maintain; the positions can still respond differently to a market shock.
Author’s Take: Understand Why the Gap Opened
The most useful lesson from Coke and Pepsi is a better way to examine relative performance.
When two familiar rivals separate, start with the businesses. Has one gained customers, changed its product mix or improved its earnings prospects? If the difference reflects a lasting improvement, the old relationship may have become a poor guide.
If the businesses remain closely linked, the next questions concern the evidence. Was the gap unusual under a rule chosen beforehand? Did that rule work on later data? Did the results survive borrowing costs, dividends and realistic execution prices?
Finally, examine what happens while the trade is wrong. A strategy needs a defined response to further divergence, a time limit and a funding plan. “They used to move together” cannot tell you how much loss is acceptable or how long to wait.
The research supports disciplined tests of relative mispricing, while the Coke-Pepsi rivalry alone provides very little evidence for a trade. The quality of the relationship, the reason it changed and the cost of waiting determine whether a promising chart has anything useful behind it.
A gap tells you where to investigate. Understanding why it exists is what makes the investigation worthwhile.