The Big Idea
Pairs trading is a market-neutral strategy where you go long one security and short another related security at the same time, profiting from changes in their relative price relationship. The classic example: you go long Coca-Cola and short Pepsi, betting that the historical relationship between these two similar companies will continue. If Coke rises faster than Pepsi (or falls less), you profit. If Coke and Pepsi both rise or both fall, you might profit minimally or not at all because the long and short positions hedge each other. Pairs trading is “market-neutral” because it doesn’t depend on overall market direction — only on the relative performance between the two securities.
Think of pairs trading like betting on a race between two horses, but only on which one wins, not on absolute time. If you bet horse A finishes ahead of horse B, you don’t care if both run fast or both run slow — you only care about the difference. Same with pairs trading: you don’t bet on the market going up or down. You bet that one stock will outperform another stock. This isolation from market direction is the appeal — pairs trading can theoretically work in any market environment as long as the relative relationships behave predictably.
For beginners, pairs trading sounds elegant and intellectually appealing. The market neutrality removes the stress of predicting market direction. The hedged positions feel safer. But pairs trading is actually one of the harder strategies to execute well. The relationships you’re trading on can break down. Both legs need careful sizing. Borrowing costs for the short side eat into returns. Many of the obvious pairs (Coke/Pepsi, Visa/Mastercard) have been arbitraged so heavily that opportunities are minimal. Pairs trading remains viable but typically as a smaller component of broader strategies, not as a complete approach for retail traders.
How Pairs Trading Works
The mechanics involve three steps:
Step 1: Identify Correlated Securities
Find two securities whose prices have moved together historically. Common pairs:
- Two companies in the same industry (Coke/Pepsi, Ford/GM)
- A stock and its sector ETF
- Two ETFs tracking similar indexes (SPY/IVV)
- Two share classes of the same company (BRK.A/BRK.B)
- A commodity producer and the commodity (gold miners and gold)
- Two currencies with related fundamentals
Step 2: Detect Divergence
Watch for moments when the historical relationship breaks down — when one moves significantly more than the other. The “spread” between them widens beyond historical norms.
Step 3: Trade the Convergence
Bet that the relationship will return to normal:
- Go LONG the underperforming security
- Go SHORT the outperforming security
- Wait for the spread to converge back to historical norms
- Close both positions when convergence happens
The Math Concepts
Pairs trading typically uses statistical concepts:
- Correlation: How closely two securities move together (-1 to +1)
- Cointegration: Whether their relationship is statistically stable over time
- Z-score of spread: How many standard deviations the current spread is from average
- Hedge ratio: How much of each side to make the trade truly neutral
The Hedge Ratio
Equal dollar amounts on each side don’t always create a neutral position.
Why Equal Dollars Aren’t Always Neutral
Two stocks might have different volatilities, different beta to the market, or different price levels. Equal dollar long and short positions can leave you with net exposure.
Beta-Adjusted Sizing
If stock A has beta 1.2 and stock B has beta 1.0, equal dollar sizing creates a portfolio that’s slightly long-biased to the market. To be truly neutral, you’d need more of stock B (or less of stock A) to balance.
Volatility-Adjusted Sizing
Higher volatility securities should have smaller dollar positions to balance their move size against lower-volatility securities.
Statistical Hedge Ratio
Sophisticated pairs traders calculate hedge ratios using regression analysis on historical price data. The ratio that minimized historical spread variance is used for sizing.
Practical Approach
For most retail pairs trades, equal dollar amounts work approximately. Fine-tuning the hedge ratio adds complexity that may not be worth the marginal improvement.
Types of Pairs
Same-Industry Stocks
Two competitors in the same business. Examples: Coke/Pepsi, Ford/GM, Visa/Mastercard, Home Depot/Lowe’s. The thesis: both face similar industry conditions, so divergent moves should reverse.
Challenge: industry leadership shifts. Coke has structurally outperformed Pepsi over decades. Pairs trading assumes mean reversion in the relationship; structural shifts break this.
Stock vs Sector ETF
Long a sector ETF, short a single stock that has outperformed the sector. Bets on the stock returning to sector-average performance.
Challenge: stocks regularly outperform their sectors for fundamental reasons (better products, better management). Betting against this is risky.
ETF Pairs
Two ETFs tracking similar things. SPY (S&P 500) vs IVV (S&P 500). They should track each other very closely; divergence creates short-term arbitrage-like opportunities.
Challenge: opportunities are tiny and quickly captured.
Share Class Pairs
Companies with multiple share classes (Berkshire Hathaway A and B, Alphabet A and C). The classes should trade in fixed relationships. Divergence creates opportunity.
Challenge: share class spreads have specific reasons (voting rights, liquidity) and don’t always converge.
Cross-Asset Pairs
Long gold miners ETF, short gold (or vice versa). Bets on miners’ valuation relative to gold.
Challenge: relationships are complex and influenced by many factors beyond gold prices.
International Pairs
Same company listed in different countries (ADRs vs ordinary shares). Should trade in fixed relationship adjusted for currency.
Challenge: spreads can persist due to capital controls, time zone issues, settlement differences.
The Cost Structure
Pairs trading has costs that can erode returns.
Borrowing Costs (Short Side)
Shorting requires borrowing the security. The borrow fee varies by stock and market conditions. For most large-cap stocks, fees are low (0.25-1% annually). For “hard to borrow” stocks, fees can be 10%+ annually, potentially exceeding the trade’s expected return.
Two Sets of Commissions
Each pairs trade has two sides — both incur commissions or implicit costs. Round trip = 4 transactions (open both legs, close both legs). At high frequency, costs add up.
Margin Requirements
Pairs trading typically uses margin for the short side. Margin interest applies if you’re net borrowing.
Slippage on Both Sides
Two trades mean two opportunities for adverse fills. Especially difficult for active pairs trading with frequent entries/exits.
The Cost Reality
Total costs (borrow, commissions, slippage, margin) can easily eat 1-3% annually for actively-managed pairs portfolios. The strategy needs to clear this hurdle just to break even.
The Statistical Reality
Modern pairs trading is largely statistical and quantitative.
Cointegration Testing
Statistical tests determine whether two securities have a stable long-term relationship. Without cointegration, divergences may not converge — your trade could lose forever.
Z-Score Entry/Exit Rules
Many strategies use spread z-scores:
- Enter when spread is 2 standard deviations from mean
- Exit when spread returns to mean (z-score = 0)
- Stop loss at 3 standard deviations adverse
Multiple Pairs Diversification
Single pairs are risky — relationships can break. Hedge funds running stat arb trade hundreds of pairs simultaneously, expecting roughly 60% to converge as predicted, with the diversification absorbing the failures.
Backtesting Issues
Pairs trading strategies that worked historically may fail going forward. The “obvious” pairs are often arbitraged away. Less obvious pairs may have fragile relationships.
Survivorship Bias
Backtests often show pairs that survived. Pairs that broke down (with one company going bankrupt or being acquired) are excluded. Real-time results often disappoint historical backtests.
Examples of Pairs Trades
Example 1 — Sarah’s KO/PEP Trade
Sarah notices Coca-Cola has outperformed Pepsi by 8% over the past month with no clear fundamental reason. The historical correlation is high. The spread is 2 standard deviations from average.
She enters:
- Short 100 shares of KO at $60 = -$6,000
- Long 100 shares of PEP at $170 = +$17,000
Wait, that’s not balanced. Equal dollar amounts:
- Short 200 shares of KO at $60 = -$12,000
- Long 70 shares of PEP at $170 = +$11,900
Better balanced. Three weeks later, the spread normalizes. KO drops $3 to $57. PEP rises $5 to $175.
- KO short profit: 200 × $3 = $600
- PEP long profit: 70 × $5 = $350
- Total: $950 on $24,000 capital deployed (about 4% return)
The pairs trade worked. The convergence she expected happened. Her 4% return came regardless of overall market direction.
Example 2 — Jake’s Failed Pair
Jake tries pairs trading with two electric vehicle stocks. Stock A has been a strong performer; Stock B has lagged. He bets the lag will reverse.
- Short Stock A
- Long Stock B
Three months later, Stock A continues its strong performance (it has better technology and growing market share) while Stock B continues struggling (their batteries have a defect). The “spread” widens dramatically against Jake.
His losses on both sides:
- Stock A short: -$3,000 (Stock A kept rising)
- Stock B long: -$2,500 (Stock B kept falling)
- Total: -$5,500
Plus borrow costs and commissions. The trade was a disaster.
Jake’s mistake: assuming statistical correlation without considering fundamentals. The two stocks weren’t really equivalent — one had structural advantages. The “correlation” was historical noise, not a stable relationship.
Example 3 — Maya’s Stat Arb Approach
Maya runs a small stat arb portfolio across 30 pairs. Each pair trade is sized at $5,000 long / $5,000 short.
Over a year:
- 20 pairs converged as expected, average gain 4% = $4,000
- 7 pairs partially converged, average gain 1% = $350
- 3 pairs broke down with major losses, average loss 15% = -$2,250
Net: $2,100 on $300,000 capital (about 0.7%). Plus costs. The strategy made money but barely covered overhead.
This shows the realistic returns for retail stat arb: positive but modest, with significant work required to maintain. Hedge funds make this work through scale and leverage; retail typically doesn’t.
The Modern Reality
Disappearing Edge
Computer programs scan thousands of potential pairs constantly. Easy opportunities are arbitraged within minutes. The retail trader who finds a “pairs trade” online has usually found something already exploited.
Hedge Fund Dominance
Major hedge funds run massive stat arb operations with:
- PhD-level quantitative researchers
- Massive computing infrastructure
- Direct exchange access
- Hundreds of millions in capital deployed
Retail traders compete against this for the same opportunities. The competition is unfair.
Retail Niches
Some niches remain accessible:
- Less liquid pairs ignored by hedge funds (small cap pairs)
- Slower mean reversion (multi-week holds rather than minutes)
- Cross-border pairs with regulatory friction
- Crypto pairs (if you understand the markets)
Pairs as Component, Not Strategy
Many successful retail traders use pairs trading as one component of a broader approach rather than as their entire strategy. Pairs trades selectively when good setups emerge, while other strategies fill the rest of the time.
Common Mistakes
- Trading uncorrelated “pairs.” Stocks that aren’t actually related won’t converge.
- Ignoring fundamentals. Statistical correlation doesn’t guarantee continued relationship.
- No stop losses. Letting “spreads will converge” become hope rather than discipline.
- Equal dollar without checking. Different volatilities or betas make equal dollars not actually neutral.
- Short side surprises. Borrowed-out stocks can be recalled; hard-to-borrow fees can spike.
- Single-pair concentration. One pair failure can wipe out diversified gains.
- Backtesting overconfidence. Past pairs don’t predict future pairs.
- Cost ignorance. Commissions, borrow, margin all add up to potentially exceed margins.
- News/event interference. M&A activity, earnings can break pairs relationships.
- Wrong scale. Hedge fund strategies don’t always work at retail size.
The Big Picture
Pairs trading is a real strategy with mixed prospects for retail traders.
Here’s what to remember:
- Pairs trading goes long one security and short a related one
- Profits from relative performance, not market direction
- Theoretically market-neutral
- Requires statistically related securities
- Hedge ratio matters for true neutrality
- Costs (borrow, commissions, margin) significant
- Modern markets have arbitraged most easy pairs
- Statistical pairs trading dominated by hedge funds
- Retail can find niches but margins are thin
- Best as component, not complete strategy
Pairs trading appeals because it sounds elegant. The market neutrality, the hedged positions, the statistical foundations — all feel sophisticated and intellectually pleasing. The reality is messier.
For a retail trader to make real money from pairs trading requires:
- Specific edge in finding less-obvious pairs
- Discipline to follow rules through losing trades
- Sufficient capital to diversify across multiple pairs
- Accurate accounting of all costs
- Acceptance that returns are likely modest
Without these, pairs trading often fails to deliver expected results. The retail trader who reads about Coke/Pepsi pairs and tries it usually finds:
- The opportunities they see have already been captured
- Their position sizing is wrong
- Their costs are higher than expected
- The correlation they thought existed isn’t reliable
- Returns are minimal or negative
This doesn’t mean pairs trading is bad. It means it’s not “easy money” as some marketing suggests. Real pairs trading requires real work, real capital, and real expertise.
If you want to explore pairs trading, start with paper trading or very small positions. Identify a few pairs with strong statistical evidence of cointegration. Track your trades meticulously. Calculate true returns including all costs. Be honest about whether the returns justify the work.
For most retail traders, pairs trading is more useful as concept than practice. Understanding the mechanics helps you understand market structure, hedge fund strategies, and risk management principles. Implementing it directly often disappoints expectations.
Some retail traders successfully use pairs trading as a small portion of their portfolio (10-20%) alongside other strategies. The diversification helps overall results. Used this way, pairs trading provides exposure to a market-neutral approach without dependence on it.
The honest assessment: pairs trading worked great in the 1990s and 2000s for early adopters. It became increasingly competitive as more traders entered. Today, it’s a viable but not exceptional strategy. Approach it as one tool among many, not as a magic formula. With realistic expectations and proper implementation, it can contribute to a diversified trading approach. With unrealistic expectations, it’s another path to disappointing returns.
Related Terms
- What Is Arbitrage? — Pairs trading is statistical arbitrage
- What Is Mean Reversion? — Foundation of pairs trading
- What Is Correlation? — Required for pair selection
- What Is Short Selling? — Required for short side
- What Is Algorithmic Trading? — How most pairs trading happens
← Back to the Complete Trading Terms Glossary
Focus on the process. Trust the stats. Stay consistent.