# What is the proper capital split/allocation between the long and short in a pairs trade?

I am trying to understand the following. If I have \$100, how do I determine how much to allocate to the long and short of a pairs spread?

You might say "use the hedge ratio you calculate, as that will tell you the number of shares of stock B to long/short for every share of stock A". The hedge ratio, though, is basically the "average" ratio between the prices of A and B. Therefore, by following the hedge ratio, won't my allocation always be 50/50? I have heard this is not optimal.

Is there any other way of splitting the capital?

There are many different approaches to creating a portfolio comprising long-short pairs trades. Many take the approach of market neutrality. They attempt to create a portfolio that is insensitive to market direction and isolate the effect of the narrowing of the spread between the pairs.

Your description is an attempt to be market neutral by being dollar neutral. In other words, they take the same dollar value of the long position and the short position.

Others have a different definition of market neutrality, such as Beta neutrality. In this case they would take a Beta weighted value of each position such that the Beta of their long position would be offset by the Beta of their short position.

Others attempt to create a portfolio that is industry neutral, country neutral etc. And still others might take a volatility weighted off-setting positions to create a portfolio that is volatility neutral.

As you can see, the approach to creating a long short portfolio is dependent on what disparity of risks the portfolio manager is trying to exploit and what risks the portfolio manager is attempting to neutralize through the hedge, as well as that portfolio manager's definition of that risk.

• I greatly appreciate your answer and overview. This gives me a ton of keywords/concepts to look deeper into. If I understand correctly, if my objective is to be "dollar neutral", then 50/50 is, by definition, the only way I do that. Correct? So you're saying that the split between the two positions depends purely on what type of neutrality you want to achieve? Is achieving neutrality a prerequisite for a successful pairs trading strategy? For example, is there not some methodology to achieve optimal/"maximum-profit" bet-sizing that doesn't necessarily result in market neutrality? Mar 26, 2022 at 15:25
• To clarify my last question, I guess I'm imagining a situation like this - if you knew which stock in the pair was going to make the move to convergence, you would naturally go all in on that position (rather than splitting), right? Surely this is impossible, but coming from this idea, I'm wondering if there are any asymmetric betting schemes that don't necessarily guarantee neutrality but perhaps tend to increase profitability nonetheless? Mar 26, 2022 at 15:27
• Ah, but asymmetric betting would cease to make the trade market-neutral (in my case, "dollar neutral"), meaning you would inherently be betting that the market will move in one direction rather than the other in order to converge, which I suppose goes against the whole idea. Mar 26, 2022 at 15:37