Pairs Trade Example

As with nearly any investment, taking a pairs trade involves more than just hitting the buy and sell button. Here we examine, in very broad terms, the steps required to enter and exit a pairs trade.

Assemble a list of potentially related pairs

Just as long-only stock traders scan the markets for suitable securities, a pairs trader must start with a list of potentially related pairs. This entails conducting research to find securities that have something in common – whether the relationship is due to sector (such as the auto sector) or to asset (for example, bonds). While any random pair could theoretically be correlated, it’s more likely that you’ll find correlation in securities that have something in common to begin with.

Determine the correlation level

The next step acts as a filter, or a means by which we can reduce the number of potential pairs in our quiver. One way is to use a correlation coefficient to determine how closely two instruments are related. The following chart shows a daily chart of the e-mini S&P 500 contract (in green) and the e-mini Dow contract (in pink). Below the price chart is an indicator that shows the correlation coefficient (in yellow). We can see from the chart that during this particular time period, the ES and YM are highly correlated, with values hovering well above 0.9. We’ll keep the ES/YM pair on our list of potential pairs candidates.
 

The ES (in green) and the YM (in pink) show potential as a pairs trade. Visual confirmation of price, backed by quantitative results from the correlation coefficient (in yellow), show that the two instruments are highly correlated. Image created with TradeStation.
 
Another chart (below) illustrates a pair that is not correlated. In this example, a daily chart of Johnson & Johnson (in green) and Proctor & Gamble (in pink) shows little correlation between the two instruments, despite the fact that they “have something in common.” Here, the correlation coefficient (in yellow) demonstrates that the relationship is scattered, ranging from high values of about 0.8 to values below zero, indicating a lack of correlation. In this case, we can remove the JNJ/PG pair from our list of potential pairs candidates.

This daily chart of JNJ (in green) and PG (in pink) shows that this is not an ideal pair. A visual review of prices, confirmed by results from the correlation coefficient (in yellow) indicate a lack of correlation between the two stocks. Image created with TradeStation.

It should also be noted that pairs can show negative correlation – a relationship between two instruments in which one’s price increases while the other decreases. This creates a type of “mirror image,” as can be seen in the following chart.

This daily chart of GE and Walmart show negative correlation. The prices of these two stocks tend to move in opposite directions.

Use modeling to determine specific rules

An ongoing component of the process is to research and test trading ideas and determine absolute methods of evaluating pairs and defining divergence. Traders have to answer questions like What constitutes “enough” divergence from the trend to initiate a trade? and How will this be evaluated? – for  example, using data from a price ratio indicator with standard deviation overlays. In general, it’s a good idea to focus on quantifiable data, such as “I will enter a pairs trade when price ratio exceeds two standard deviations.”

The following chart shows the e-mini S&P 500 (ES) and e-mini Dow (YM) on a daily chart. Below the price chart is a spread ratio indicator (in yellow), with a +/- one and two standard deviation overlay (dotted lines). The mean appears in pink.

A daily chart of the ES and YM e-mini futures contracts. A spread ratio indicator appears below the price chart, along with a standard deviation overlay. Image created with TradeStation. 

Determine position sizing

Traders can size their pair-trade positions differently, such as share-for-share and standard deviation weighted. Many traders use a dollar-neutral approach to position sizing when trading pairs. Using this method, the long and short sides of the trade are entered with equal dollar amounts. For example, a trader wants to enter a pairs trade with stock A, trading at $100 per share, and stock B, trading at $50 per share. To achieve a dollar-neutral position, the trader will have to buy two shares of stock B for every one share of stock A. For example: (See also: Optimal Position Size Reduces Risk.)

• Long 100 shares of stock A = $10,000
• Short 200 shares of stock B = $10,000

Buy the underperformer and sell the overperformer

Once the trading rules are met, the trader buys the underperforming security and simultaneously sells the overperforming security. In the following chart, the spread ratio has exceeded two standard deviations, and a trading setup has occurred in the ES/YM pair. Here, we go long the ES and short the YM.

A trade is opened in the ES/YM pair. The horizontal red and green lines at the top show the real-time P/L for each position. Image created with TradeStation.

Use sound money management principles to exit the trade

As with most investments, the timing of the exit is critical to the success of the trade. It’s important to apply money management principles to pairs trades (like any other trades), including the use of protective stop-loss orders and profit targets. Optimal levels are typically determined through extensive historical modeling. The following chart shows the ES/YM trade, exited using a conservative net profit level.

The ES/YM trade is exited with a small net profit. Image created with TradeStation.
 
Despite exhaustive research, modeling and testing, a pairs trading strategy may fail to live up to expectations. Two risks that traders have are model risk and execution risk, introduced in the next section.