Does systematic use of stop-loss orders (automated position exits based on a cumulative loss threshold) improve net returns? Both the April 2008 paper entitled “Re-examining the Hidden Costs of the Stop-Loss” by Kira Detko, Wilson Ma and Guy Morita and the May 2008 draft paper entitled “When Do Stop-Loss Rules Stop Losses?” by Kathryn Kaminski and Andrew Lo address this question with theory and empirical tests. They conclude that:
“Re-examining the Hidden Costs of the Stop-Loss” finds that:
- Stop losses tend to reduce return volatility, but the opportunity cost of taking losses that subsequently reverse tends to offset the benefit of avoiding further losses.
- However, the effectiveness of stop losses varies with the price drift of the asset (see the table below).
- If price is drifting upward (bull market), use of stop-losses reduces expected return by missing rebounds after minor corrections.
- If price is drifting downward (bear market), use of stop-losses enhances expected return by avoiding further losses.
- Similar results hold for profit-taking stops (exit long position on achieving cumulative gain target).
The following table, taken from the paper, summarizes the implications of systematic use of trailing stop-losses, profit-taking stops and a combination of both for returns to a long-only position in an asset trending up, down and not at all. Use of stops tends to decrease return volatility in all cases. When asset prices are trending up (down), the use of stops tends to underperform (outperform) a simple buy-and-hold strategy. In other word, a trader who can predict the future price trend may reliably benefit from stop losses. These general results are exclusive of trading frictions, which degrade returns for any strategy that elevates trading activity.
“When Do Stop-Loss Rules Stop Losses?” finds that:
- Under the Random Walk Hypothesis, simple stop-loss rules always depress expected return
- However, in the presence of momentum and/or regime-switching, stop-loss rules can add value.
- Backtested against monthly returns for broad stock and bond indexes during 1950-2004, a simple stop-loss rule outperforms a buy-and-hold strategy by 0.5% to 1.0% per month during stopped-out periods (exclusive of trading frictions), indicating the presence of non-random behaviors in actual returns. In other words, the rule seems to be able to pick out periods in which long-term bonds substantially outperform equities.
In summary, systematic use of stop-loss orders may be beneficial, especially if one can project the general price trend and apply stop losses accordingly.
