Sunday, May 28, 2006

When Emerging Markets Submerge: Another Way Markets Confound Human Nature

This past week, we saw the Emerging Markets index (EEM) record a sobering 16% drop on a 10-day basis. I was surprised to find that the correlation between ten-day returns on the EEM and on the Standard & Poors 500 Index (SPY) was a hefty .74. That led me to wonder if drops in EEM might represent the kind of fear among speculative traders that you would see at market bottoms.

Since late April, 2003 (N = 768 trading days), we've had 40 occasions in which EEM has been down by more than 5% on a 10-day basis. Ten days later, SPY has been up by an average 1.03% (30 up, 10 down)--much stronger than the average ten-day change for the sample overall (.43%; 459 up, 309 down). Big drops in emerging markets have thus, on balance, been bullish for American stocks in the near term.

How about when emerging markets are strong? We've had 114 occasions in which EEM has risen by more than 5% in a 10-day period. Ten days later, SPY was up by an average of only .08% (60 up, 54 down)--much weaker than its average performance. Strong emerging markets have thus led to underperformance among American stocks in the near term.

Readers of this blog will recall that, a while ago, I posted some findings regarding speculative vs. non-speculative stocks and the tendency of the former to lead the latter. EEM is one worthy proxy for those speculative issues. When the most speculative participants in the market are most bearish, it's best to step up to the plate. When those speculators are most bullish, it's time to take a few chips off the table.

It's another way in which markets confound human nature. We tend to extrapolate from the recent past to the near future, whether in predicting sports events, people's moods, or weather. Markets are one of the only spheres of life in which the recent past is an excellent predictor of the near-term future--in reverse!

2 comments:

sinnerMan said...

This is one of the key problems I have with "quant" trading. It often does not take the psychological aspect of the market into account. Having a very long investing horizon should hopefully help smooth out some of the bumps along with way.

Brett Steenbarger, Ph.D. said...

Yes, the performance of historical models of the market are always going to degrade as patterns change, which is why multiple models and time horizons become important. I do think many psych variables in the market can be quantified, but those patterns are subject to shifts as well. Thanks for the note--

Brett