Do stop-loss orders (automated position exits based on a cumulative loss threshold) enhance returns and reduce risk? In their 2008 paper entitled “The Value of Stop Loss Strategies” Adam Lei and Huihua Li investigate whether traders using stop-loss strategies to exit losing positions in individual stocks outperform a comparable buy-and-hold strategy. They test the following strategy alternatives: holding periods of three months, six months or one year; stop-loss thresholds of 5, 10 or 20 daily return standard deviations; reinvestment of stopped out positions in either the S&P 500 index or the one-month Treasury bill; and, a fixed stop price or a trailing stop price that follows stock price upward (but not downward). Using historical and simulated daily return data for a broad sample of NYSE/AMEX-listed stocks and random buy dates over the period 1970-2005, they conclude that:
- Using historical return data:
- Over the entire sample period, the level of returns for strategies using fixed and trailing stop losses are statistically indistinguishable from those for a buy-and-hold strategy.
- Fixed stop losses do enhance returns during some subperiods, and during other subperiods for stocks with high past return volatility. However, the ability of stock characteristics to predict the effectiveness of fixed stop losses declines over time.
- Over the entire sample period, trailing stop losses tend to reduce trading risk (standard deviation of returns). Fixed stop losses reduce trading risk only during the 1970s and 1990s.
- Using simulations:
- Neither fixed nor trailing stop-loss strategies affect the level of returns compared to a buy-and-hold strategy, whether daily returns are generated as independent, autocorrelated or trend-following.
- Fixed (trailing) stop losses do not (sometimes do, depending on stop levels) reliably reduce trading risk.
In summary, over the long term, systematic use of stop losses in trading individual stocks does not enhance the level of returns, but certain trailing stop losses reduce trading risk (standard deviation of returns).
