Pairs trading is a relative-value strategy: you go long one stock and short another that tends to move with it, then wait for the gap between them to close. The bet is not that the market rises or falls, but rather that the relationship between the two names holds well enough for the spread to mean-revert.
That sounds tidy in theory. In practice, the hard parts are picking a pair that is genuinely related, sizing the two legs so the position is actually neutral, and accepting that a broken relationship can cost you on both sides at once. This article covers selection, execution of the spread, and the risks that matter most.
What makes two stocks a plausible pair
The starting point is economic linkage, not a chart pattern. Two companies in the same industry—selling similar products to similar customers, with comparable cost structures and sensitivity to the same macro drivers—have a reason to move together. A regional bank and a money-center bank are not the same business, while two mid-cap industrial distributors competing for the same contracts usually are closer.
Sector membership alone is weak. Within energy, an integrated major and a shale producer respond to oil differently because their cost bases and hedging programs differ. Within retail, a discount chain and a luxury brand share a sector label and almost nothing else. You want overlapping revenue drivers, similar margin profiles, and comparable exposure to interest rates, input costs, or regulation.
A second filter is liquidity. Both legs need enough average daily volume for you to enter and exit without moving the price, and both need borrowable shares if you intend to short. Hard-to-borrow names carry financing costs that can quietly eat the edge. Check the short interest and the borrow fee before you model anything.
Corporate structure matters too, since dual-class shares, pending mergers, spinoffs, and index rebalances can break a historical relationship overnight. A pair that worked for three years can stop working the day one company announces a transformative acquisition.
Measuring the relationship before you trade it
The common approach is to compute the ratio or the log-price spread between the two series and study its behavior. If the spread oscillates around a stable mean, the pair is a candidate. If it trends, it is not.
A simple ratio (price A divided by price B) works when the two stocks are roughly comparable in price. A regression-based hedge ratio is more flexible: you regress one price on the other, and the slope tells you how many dollars of the second stock to short for each dollar of the first. The residual from that regression is the spread you would trade.
Stationarity is the property you are testing for. A stationary spread has a tendency to return to its average. A non-stationary spread wanders. Statistical tests exist for this, but they are backward-looking. They tell you the relationship held in the sample you fed them. They say nothing about whether it will hold next month.
Practically, most traders look at the spread’s z-score: how many standard deviations it sits from its recent mean. A high positive z-score suggests the first stock is rich relative to the second. A low negative z-score suggests the opposite. The z-score is a description, not a signal. It can stay extreme for a long time, and a widening spread is exactly what a losing pairs trade looks like.
Structuring the spread: dollar neutrality versus beta neutrality
A common mistake is matching the dollar value of the long and short legs and calling the position neutral. That only neutralizes exposure to a uniform market move. If one stock historically moves twice as much as the other for the same market move, a dollar-matched pair still carries net directional risk.
Beta-adjusted sizing corrects for this. You scale the short leg by the ratio of the two betas, so that a one percent move in the market produces roughly offsetting gains and losses. The result is closer to true market neutrality, though beta itself is an estimate that drifts.
There is a tradeoff, however. Beta-neutral sizing often leaves the two legs unequal in dollar terms, which means the position is exposed to the difference in their idiosyncratic volatility. You have reduced market risk and increased stock-specific risk: neither version is free of exposure. Market-neutral does not mean risk-free.
The mechanics of the short leg deserve attention. Shorting requires a margin account, and in the US, Regulation T governs initial margin while FINRA rules set maintenance requirements. In the UK, short selling is permitted but subject to disclosure thresholds and, in stressed markets, temporary bans on certain financial stocks. Your broker’s borrow availability and financing rate determine whether the trade is viable at all.
The trader’s view: weeks to a few months
Over a trading horizon, the spread’s short-term behavior dominates, so you are watching for a divergence that looks stretched relative to recent history. You enter when the z-score crosses a threshold, and you exit when it reverts toward the mean or when a stop is hit.
Position sizing has to account for the fact that both legs can move against you. A stop on the spread, not on each leg separately, is the cleaner approach. If the spread widens beyond a level that your analysis says is inconsistent with the historical relationship, the thesis is wrong and the trade should end.
Earnings dates are the biggest scheduled hazard. A single quarterly report can reprice one leg by a double-digit percentage while the other barely moves. Many pairs traders flatten before earnings or avoid pairs whose reporting dates are close together. Dividends matter as well: the short seller owes the dividend to the lender, so a pair with mismatched ex-dividend dates leaks cash.
Transaction costs accumulate faster than most people expect. Two commissions, two spreads, and borrow fees apply on every round trip, and a strategy that trades frequently needs a wider edge per trade to survive them. Capacity is limited by how much size the thinner leg can absorb.
The investor’s view: six months and beyond
Over longer holding periods, the question shifts from spread oscillation to whether the two businesses stay comparable. A relationship that held for years can decay as one company reinvests, takes on debt, loses a key customer, or gets disrupted by a new entrant.
An investor running a pairs position for months should re-examine the fundamental linkage on a schedule, not just watch the chart. Read both companies’ filings. Compare revenue growth, margin trends, and capital allocation. If one is gaining share at the other’s expense, the spread may be repricing a real change rather than a temporary dislocation.
Costs behave differently over long horizons. Borrow fees compound, and a hard-to-borrow stock can become expensive to stay short. Dividends paid on the short leg are a recurring drag. Meanwhile, the position ties up margin that could be deployed elsewhere, which is a real opportunity cost even when it does not show up as a line item.
Tax treatment adds complexity. In the US, tax treatment of gains and losses on the two legs may vary depending on position type, and short positions may have different rules. In the UK, stamp duty reserve tax generally applies to purchases of UK-incorporated shares but not to the short sale, which creates an asymmetry between the legs. These details vary by jurisdiction and by instrument, and they are worth confirming with a tax professional rather than assuming.
Risks that break the strategy
The central risk is that the relationship was never stable. Historical correlation is a summary of the past. Two stocks can be highly correlated for years because both were lifted by the same bull market, then diverge permanently when that tide goes out. Correlation is not causation, and it is not a guarantee.
Divergence risk is the second. When the spread widens, a market-neutral position can lose on both legs if the long falls and the short rises. The position is neutral to the market, not to the pair. This is the scenario that turns a small statistical edge into a large loss.
Short-specific risks are the third. A short squeeze can force the price of the borrowed stock sharply higher, while margin calls can force you to cover at the worst moment. Borrow recalls mean your broker can demand the shares back—potentially at a loss—and hard-to-borrow fees can rise without warning.
Finally, there is model risk. The hedge ratio, the mean, and the standard deviation are all estimated from a window of data you chose, so change the window and the z-score changes. A pair that looks stretched over sixty days may look perfectly normal over two years. Parameter choices are assumptions. They should be treated as such.
Conclusion: a relative bet with absolute consequences
Pairs trading offers a way to express a view on two companies without taking a broad market stance. The appeal is real: the position can profit in flat, rising, or falling markets if the spread behaves. The catch is that every assumption in the model, from the hedge ratio to the stability of the relationship, can fail at the same time.
Treat selection as fundamental work first and statistical work second, and size for the possibility that both legs move against you. Watch borrow costs, earnings dates, and dividends as carefully as you watch the chart, and keep in mind that market-neutral describes the exposure, not the outcome.
- Position Trading: The Long-Term Approach to the Markets
- Breakout trading: how to trade range breakouts
- Pullback Trading in Uptrends: Entries at Moving Averages
Frequently Asked Questions
What is pairs trading in simple terms?
It means buying one stock and shorting another that usually moves with it, betting the price gap between them will narrow. The profit or loss comes from the difference between the two positions rather than from the overall market direction.
Is pairs trading really market-neutral?
Only approximately. Dollar-matched or beta-adjusted sizing reduces exposure to broad market moves, but the position still carries risk from the two individual stocks diverging, from sector-specific news, and from changes in the historical relationship.
How do you choose which stocks to pair?
Start with companies that share revenue drivers, cost structures, and macro sensitivities, then check that both are liquid and that the short leg is borrowable. Confirm the price spread has historically reverted to a mean rather than trending.
What are the main risks of pairs trading?
The relationship breaking down permanently, both legs moving against you at once, short squeezes and borrow recalls, rising financing costs, and earnings or merger news that reprices one stock independently of the other.
Can retail traders run pairs trades in the US and UK?
Yes, but shorting requires a margin account and borrowable shares, and rules differ. Regulation T and FINRA maintenance requirements apply in the US, while the UK has disclosure thresholds and occasional short-selling restrictions on certain sectors.
AI Notice: This article was created wholly or predominantly with the assistance of artificial intelligence and was published without human editorial review.
This article is for general information only and does not constitute investment advice. Always do your own research before making trading decisions.