Truths About Stop-Loss Strategies That Nobody Wants To Believe by Tim du Toit, Quant Investing
This is mainly been because some testing we did came to the conclusion that a stop-loss strategy leads to lower returns even though it did reduce volatility (large losses).
Why we read all the time?
But you know I look at investment research all the time and I recently found three interesting papers that tested stop-loss strategies with very interesting results.
But before I get to how you can use the information to increase your returns first some detail on the research studies.
Research study 1 – When Do Stop-Loss Rules Stop Losses?
The first research paper I found is called When Do Stop-Loss Rules Stop Losses? and was published in May 2008 by Kathryn M. Kaminski and Andrew W. Lo.
The paper looked at the application of a simple stop-loss strategy applied to an arbitrary portfolio strategy in the US markets over the 54 period year period January 1950 to December 2004.
The strategy used a simple 10% stop loss value when exceeded the portfolio was sold and the cash invested in long term US government bonds. Investment would be moved back into the stock market once the 10% fall in the stock market was recovered (the 10% stop-loss was recovered).
What they found it worked very well
If the researchers excluded the technology bubble (used data from Jan 1950 to Dec 1999) when their simple stop-loss model got back into the stock market too quickly, they found that when the model was invested in the stock market it provided a higher return than bonds 70% of the time, and during stopped-out periods (when the model was invested in bonds), the stock market provided a higher return only 30% of the time.
Even applied to the whole 54 year period the study found that this simple stop-loss strategy provided higher returns while at the same time limiting losses substantially.
What they also found was that the stop-out periods were relatively evenly spread over the 54 year period they tested. This shows you that the stop-loss was not just triggered by a small number of large market movements (crashes).
Research study 2 – Performance of stop-loss rules vs. buy and hold strategy
The second research paper was called Performance of stop-loss rules vs. buy and hold strategy, published in 2009 by Bergsveinn Snorrason and Garib Yusupov.
What they tested
They compared the performance of following a trailing and traditional stop-loss strategy to a buy and hold strategy on companies in the OMX Stockholm 30 Index over the 11 year period between January 1998 and April 2009.
Investments were made on the first trading day of a quarter (starting January 1998) and at the end of a quarter the proceeds were reinvested.
When a stop-loss limit was reached, the company was sold and the proceeds held in cash until the next quarter when it was reinvested.
They tested stop-loss levels from 5 to 55%.
Trailing stop-loss results
The table below shows the results of the use of a trailing stop-loss strategy.
As you can see the highest average quarterly return (Mean = 1.71%) was obtained by trailing stop-loss with a 20% loss level limit. The highest cumulative return (Cumulative = 73.91%) was achieved at the 15% trailing stop-loss limit.
The only stop-loss level that delivered a worse result than the buy-and-hold (B-H) portfolio, with a negative average return of 0.12% and cumulative return of -8.14%, was from a trailing stop-loss strategy with 5% loss limit.
Returns from using a trailing stop-loss strategy
In the following chart which shows the results of the trailing stop-loss strategy.
Trailing Stop-loss, equally-weighted portfolio performance
The result of using a 15% and 20% loss levels give you about the same overall result with a 20% stop-loss level leading to higher returns most of the time.
Traditional stop-loss strategy
The following table shows you the results if you applied a traditional stop-loss strategy, which means that you would calculate the stop-loss from the purchase price.
Returns from using a traditional stop-loss strategy
As you can see all traditional stop-loss levels from 5% to 55% would have given you better returns than the buy and hold (B-H) strategy.
The highest average quarterly return (Mean = 1.47%) was achieved at the 15% stop loss level and the highest cumulative results of 57.1% at the 10% stop-loss level closely followed by the15% stop-loss level at 53.31%.
The chart below shows you the results of the traditional stop-loss strategy for all tested stop-loss levels.
Traditional Stop-loss, equally-weighted portfolio performance
In the chart you can see that the 15% loss level would have given you the best result over the largest part of the 11 year test period.
What is the best strategy?
So what is the better stop-loss strategy, you may be thinking?
To find out I deducted the results of the traditional stop-loss strategy from the trailing stop-loss strategy.
The results are summarized in the following table:
Trailing stop-loss minus Traditional stop-loss
Trailing better than traditional
Only at the 5% and 10% loss levels did the traditional stop-loss perform better than the trailing stop-loss. At all other loss levels the trailing stop loss out performed, most notably at the 20% loss level where it performed 27.47% better over the 11 year period.
Research study 3 – Taming Momentum Crashes: A Simple Stop-loss Strategy
The third research study I looked at is called Taming Momentum Crashes: A Simple Stop-loss Strategy by Yufeng Han (University of Colorado), Guofu Zhou (Washington University) and Yingzi Zhu (Tsinghua University) and was published in August 2014.
What they looked at
The researchers applied a simple momentum strategy of buying the 10% of companies with the largest price increase the past six months and selling short the 10% of companies with the largest price decline the past six months.
Once the stop-loss was triggered on any day the company was either sold (Winners) or bought (Losers) to close the position. The proceeds were invested in the risk-free asset (T-bills) until the end of the month.
Tested for 85 years
They applied this strategy over the 85 year period from January 1926 to December 2011 to all US domestic companies listed on the NYSE, AMEX, and NASDAQ stock markets (excluding closed-end funds and real estate investment trusts).
At a stop-loss level of 10%, they found that the monthly losses of an equal weighted momentum strategy went down substantially from ?49.79% to ?11.34%.
For the value-weighted (by the last month-end market size) momentum strategy, the losses were reduced from ?65.34% to ?23.69% (to -14.85% if August 1932 is excluded).
Not only did the application of the simple stop loss strategy reduce losses it also increased returns.
The stop-loss strategy increased the average return of the original