If you’ve ever wondered about the difference between technical and quantitative analysis, you’ve come to the right place.

Recently, one of my colleagues posted to our internal message boards an article suggesting that investors should be concerned about the fact that the S&P 500 has spent a record amount of days above its 200 day moving average.  The article suggested that when stocks spend too much time above the moving average it can be dangerous because:

1. Investors have the collective memory of a goldfish, so when there’s a protracted bull market…
2. …nobody can remember the last time stocks had a meaningful drawdown.
3. This leads to a complacent approach to risk management…
4. …priming the market for a crash.

It’s a compelling argument because people tend to make irrational investment decisions especially at inflection points in the market.  But a technical indicator such as days above a moving average is only useful to the extent that quantitative analysis suggests there is a statistically significant edge. So we decided to examine what the original post had left out: how well days above a moving average predict subsequent returns.

What we found is that – with the exception of the 30-day time horizon – there’s virtually no relationship at all.

When you think about it, there are two questions that need to be answered:

1. Whether you should expect lower returns then usual after the S&P 500 breaks it’s 200MA; and,
2. Whether there is any relationship between the length of the bull run and the subsequent performance.

The second row from the bottom shows the statistical significance of the conditional mean (after a crossover) relative to the full sample mean.  Anything below .05 would appear to suggest a statistically significant relationship.  Considering that the 30-day performance shows a p-value approaching 0, it’s reasonable to expect that when the market breaks below the 200MA you will experience below average performance.  However, that time interval was the only one showing statistically significant adverse performance.

But there’s something else to pay attention to: the original idea was that there was some sort of relationship between the amount of time that the market spent above the 200MA and the subsequent performance.  Specifically, the idea was that the longer the bull run lasted, the harder the fall would be.  And we can now confidently say that based on this study,  that’s rubbish.  In the bottom row of our chart we show p-values between .28 and .39, none of which meet the threshold of statistical significance for establishing a relationship between the length of the bull run and the subsequent conditional performance.

As quantitative investors, we don’t mean to disparage any other forms of analysis, technical or otherwise.  But we do enforce a strict requirement that conclusions drawn follow from the data.  In this case, there appears to be a short-term pullback in the S&P 500 when the 200MA is broken, but the length of the bullish streak simply doesn’t matter.