Skip to content
All library documents

Choosing Raw or Excess Returns for Portfolio Optimization

Article Quant Q&A · Author: Dirty Dan
Read the original

Summary

The document addresses whether portfolio weights should be optimized using asset returns or returns measured relative to a benchmark. Its answer favors raw portfolio returns when the goal is to evaluate a complete investment strategy without a specific benchmark that must be beaten.

The rationale is that strategy evaluation typically includes portfolio selection and rebalancing, followed by review of risk and performance measures such as the Sharpe ratio, drawdown, and kurtosis. Using raw returns keeps the optimization and backtest focused on the realized portfolio outcome. This is a brief recommendation rather than a formal derivation or empirical comparison; when a benchmark is an explicit objective, benchmark-relative returns may still be appropriate.

Key ideas

  • Use raw portfolio returns when the objective is to assess an entire strategy without a required benchmark.
  • Include portfolio selection and rebalancing in the backtest when evaluating the strategy.
  • Assess portfolio outcomes with measures such as Sharpe ratio, maximum drawdown, and kurtosis.
  • Benchmark-relative returns can be relevant when outperforming a particular benchmark is a stated goal.

Tags

Portfolio constructionBacktestingRisk management
Full text
# Which returns to use for portfolio optimiaztion?


# Which returns to use for portfolio optimiaztion?












On the base of which returns do I have to derive optimal portfolio weights of an investment strategy? More specifically, do I have to use the excess returns or just the normal returns of the respective assets? I'd appreciate any help!

## Answer by Stéphane (score 2, accepted)

https://quant.stackexchange.com/a/51456

Unless you have a specific benchmark that you have to outperform, I would simply use the raw returns of the portfolio as opposed to any sort of excess returns over a benchmark. In the end, you will likely to be backtesting your whole strategy, portfolio selection and rebalancing included and you'd presumably be looking into metrics such as Sharp ratios, maximal drawdown, kurtosis, etc. It just seems like using only raw returns saves you a lot of trouble in that context.

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.