This simple but persistent anomaly shows the U.S. stock market has been far from efficient. The anomaly is present due to highly negative serial correlation in market returns and when used as a strategy, year-to-date has outperformed SPY buy and hold by 150 basis points. It’s interesting this simple anomaly has been so persistent.
The WR2 anomaly has been featured in several articles in this blog in the last six years. It’s a variant of an anomaly due to high negative serial correlation in U.S. stock market returns as described in this article and shows that the U.S. stock market is far from efficient.
WR2 anomaly: Buy after two back-to-back down days and sell after two back-to-back up days.
Below is the equity curve of this anomaly in SPY ETF since inception to 07/01/2021. The backtest includes $0.01/share commission for fully invested equity.
This simple anomaly has generated about 80% of the annualized return of SPY ETF since inception. In addition, MAR (CAGR/MDD) is 0.26 versus 0.19 for SPY buy and hold.
Year-to-date, SPY is up 13.8% but the WR2 anomaly gain is 17.5%. This is due to persisting negative serial correlation in daily returns, as shown in the chart below.
Since SPY inception, 1-lag, 252-correlation of daily returns has been negative other than during short periods of time but after the 2020 crash it became extreme and still remains low.
Note that the WR2 anomaly has also performed relatively well for stock portfolios, for example Dow 30 stocks.
For the backtest below we used Norgate data for Dow 30 index that include current and past constituents to remove survivorship bias. We highly recommend this data service to those who would like to remove survivorship bias from backtests (we do not have a referral arrangement with the company.)
The strategy misses buy and hold annualized return by about 100 basis points but outperforms based on Sharpe: 0.76 for the strategy versus 0.55 for buy and hold.
I believe this WR2 anomaly is “transitory.” In market this may take longer than most participants think. I don’t think this is a good strategy to use at any rate. The risk of “sudden death” is high. Like other simple anomalies in the past, when WR2 will stop working, it will be a race to the bottom with no recovery.
See also: Efficient Market Hypothesis: A Farce?
Charting and backtesting program: Amibroker. Data provider: Norgate Data