Skip to content
All library documents

Calculating Long-Short Holding Period Returns with Portfolio Weights

Article Quant Q&A · Author: user546106

Summary

The document explains how to calculate the return of a long position in one stock and a short position in another over a holding period. It recommends calculating each asset’s ordinary return, then multiplying by its portfolio weight: positive for the long and negative for the short. The resulting weighted returns are summed. In the example, equal-sized positions in two stocks produce a 10% return when the long rises 20% and the shorted stock rises 10%.

Key ideas

  • Calculate each asset’s return independently of whether the position is long or short.
  • Represent short exposure with a negative portfolio weight.
  • Multiply each asset’s holding period return by its initial weight and sum the contributions when positions are not rebalanced.
  • If positions are specified as share quantities, choose a notional amount to express profit as a return.

Tags

Full text
# How to calculate holding period return of a long-short strategy?


# How to calculate holding period return of a long-short strategy?












I have daily close prices of two stocks, A and B. Suppose that we long stock A and short stock B. Assume that we do the long-short every day and hold that portfolio for some days. How to calculate each day's holding period return of this long-short strategy?

My idea is that we first calculate the holding period return for the long position and the short position seperately, then we subtract each day's long position holding period return by the short position holding period return. Is that right?

To calculate long position's hodling period return, I use $$R_A = \frac{\text{Final Price_A} - \text{Initial Price_A}}{\text{Initial Price_A}}$$

To calculate short position's holding period return, I use $$R_B = \frac{ \text{Initial Price_B} - \text{Final Price_B}}{\text{Initial Price_B}}$$

Then, to find day $k$'s holding period return, I use $$R_A - R_B$$

## Answer by Enrico Schumann (score 2, accepted)

https://quant.stackexchange.com/a/78311

The generic way to compute such returns is to multiply the asset returns over a specific period with the weights at the start of that period. For a short position, the weight is negative. If you do not rebalance over the k days, you should simply multiply each asset's k-day return with the initial weight, and then sum those weighted returns. The advantage of this generic approach is that it can easily handle rebalancing, i.e. changing weights.

To provide an example: Suppose you have a fixed notional amount of money to invest; 100 euros, say. You invest 100 euros (100% of your capital) in stock A, i.e. you buy `100/share_price_of_A` units of A. You sell short 100 euros (-100% of 100 euros) of stock B, i.e. you sell `100/share_price_of_B` units of B.

Suppose that over the holding period of k days, stock A rises by 20% and stock B by 10%.

So what is your return? `100% * 20% + -100% * 10% = 1*0.2 + -1*0.1 = 0.1 = 10%` (Or in monetary units, 10 euros on the 100 euros notional amount.)

In your description, you define the return for a short position different from (as the negative of) the return of a long position. In that case should indeed add those returns, not subtract them. The advantage of working with explicit weights is that the single-asset return computation does not depend on your position.

If you wish to work with shares as opposed to weights, you'll need to make some assumption regarding a notional; there is no generally-established "correct" way to compute returns in this case. You would then compute success in units of currency, and then divide by the chosen notional. (Or skip returns altogether and measure your success in units of currency only; futures traders often do that when the notional is not obvious.)

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.