There has been a bit of buzz generated by the appearance of the golden cross, where the 50-day moving average crosses the 200-day moving average, experienced by the S&P 500.



Be contrarian? Or sheep?
The talk is that such golden cross events are bullish for the market. The premise is based on trend following models, which depend on persistence of price trends.

The question is, do you want to be a contrarian or do you want to a sheep (and jump on the bandwagon)?


Do price trends persist?
To state the obvious, if these trend following systems are to work, then price trends have to persist. In quantitative terms, these strategies are only profitable if there is auto-correlation in price movements.

To test those ideas, I analyzed the daily correlations of the returns of the S&P 500 with the previous day’s return to for the persistence of price momentum for the period from 1993 to the present. While the median correlation of one-day S&P 500 returns was -0.08, indicating a tendency for price reversion, I found that indeed there were periods of strong momentum. An optimized solution that forecasts those periods based on indicators of autocorrelation turned out to produce better risk-adjusted returns.

The chart below shows the results. The dark blue line shows the cumulative long-only returns of the S&P 500 from 1993. The red line shows the returns of a 50-200 day moving average crossing system, where the signals are defined as:

Long: 50 day MA > 200 day MA and the 50 day MA acts as trailing stop
Short: 50 day MA < 200 day MA and the 50 day MA acts as trailing stop

When a signal is generated, the action is taken the next day based on the next day’s closing price.