Wednesday, April 21, 2010

Examining Traderfeed's EEM:SPY Sentiment Indicator Concept

In a Traderfeed post on Monday Dr. Brett Steenbarger presented the concept of using the EEM:SPY pair as a sentiment gauge. The idea is that when EEM is outperforming SPY, traders are more willing to take on risk and stocks as an asset class should benefit. When SPY outperforms EEM, then traders are seeking relative safety and the forward outlook for stocks isn’t as good.

Last year I used the Nasdaq vs. S&P 500 relative strength and showed a model that used basically the same risk seeking/aversion idea. A link to that post may be found here. You may also download the model (though you will need to update it with recent prices).

I thought it might be interesting to substitute EEM for the Nasdaq values using that model. (Note: my relative strength calculation, which originally came from Gerald Appel’s Book, “Technical Analysis, Power Tools for Active Investors” is a bit different than Dr. Steenbarger’s calculation. The model calculation examines relative strength over a 10-week period.) After doing so I noted the following observations:

  • Investing only when the EEM is leading the SPY would have resulted in being in a position about 2/3 of the time.
  • By placing your money in SPY, total return since 6/13/2004 would have been a little over 35% using the model vs. about 20% with SPY buy and hold.
  • In both cases all of the gains were made since the March 2009 bottom.
    Prior to that bottom, the model preserved capital quite a bit better than SPY buy and hold.
  • Over that same time period, EEM appreciated about 220%.
  • Buying EEM instead of SPY when the model was positive would have resulted in a gain of about 186%.

In summary, while history is short, Dr. Brett’s EEM:SPY pair seems to work well as a sentiment indicator. As we saw with the Nasdaq:S&P model last year, there appears to be an advantage not only in entering the market when the riskier index is leading, but also in trading the riskier index rather than the SPY.

1 comment:

keithpiccirillo said...

Thanks for the interesting post.
This seems to be an easier, albeit less time consuming and interesting concept, with it's large cap mix from diverse countries where future growth is expected.
Would I be fair to equate this as a bit thinner i.e., riskier than using the 200 day moving average timing of the 5 asset classes as set forth in Meb Faber's paper?
How would one bail out when the EEM fell below the SPY, and if they fell at discordant rates, what indicator do you have that could stop a big draw down?

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=962461