Wednesday, February 19, 2014

Emerging Markets Remain Relatively Submerged

Emerging markets as a whole just can't seem to get out of their own way. The MSCI Emerging Markets ETF (EEM) is down 6% YTD and has been underperforming the S&P 500 for years. What is the current outlook?

Ultimately I believe a view on emerging markets in isolation is not enough and instead should be framed emerging versus developed, or as more of an asset allocation decision. Should I be repositioning more funds out of developed (S&P 500) and into emerging markets (EEM), or vice versa?

I have found that over time the relative return of industrial metals versus a commodity index (CRB) does a very good job at leading the relative performance of the EEM vs. the S&P 500. The relationship makes sense since emerging markets tend to have a high sensitivity to global economic activity, and industrial metals tend to have a higher sensitivity to global economic activity than most other commodities.


The weekly chart above shows five signals (2 buys, 3 sells) since 1999, triggered by trend line breaks in the industrial metals vs. CRB relative return (lower inset). 

Note that a relative-return buy signal for EEM has yet to be triggered. Therefore, based on this one indicator, better to remain overweighted in developed (S&P 500) and underweighted in emerging.


  1. I don't think this chart is very accurate. The first trendline on the bottom doesn't even look like it could be identified in real-time and doesn't look like much of a trendline. Also, the first trade from 2000-2003 looks like it's a loser (at best, break-even). The second trade looks good. The third trade also has a questionable trendline (why not start at the peak? That's what you would probably do in real time.); and the third trade, depending on the entry point, could be a loser or close to break-even. The fourth trade is fine. Then, recently, the fifth trade is fine.

    In conclusion, for relative strength trades on a long-term timeframe, I would hope the success rate would be better then 3/5 (2/4 if you count completed trades). The trendlines look nice in hindsight, but it appears to me they would be very difficult to follow in real-time.


  2. I either don't follow all of your comment or disagree. The "first trendline on the bottom" -- which? 1999-2000? As for the 2000-2003 signal, even if break-even or a slight loser, as I wrote, I don't use the signal to go long/short but rather to help guide allocation decisions. So in this case, from 2000-2003, it favored OW developed amounting to break-even which I would not count as a "loser" but rather a push, in Vegas betting terms. Your comment on the third trade, "why not start at the peak?" -- I don't follow? The third signal is triggered by the break in trend line from the 2007 peak. I see that signal clearly as a winner, not sure what you see, and it further minimizes massive downward volatility thus improving Sharpe or reward/risk. I see the tally more as 4-0-1, and I tend to doubt the current signal will be negated as a loss. Also, as I have written many times, and cited in blog entry ("based on this one indicator"), this is just one indicator and is meant to be used in addition to other indicators to arrive at a more informed decision. My hope is readers have their own set of trusted indicators and what I provide can be potentially supplemental to that set, or if readers do not have their own indicators, then frankly better to have this one indicator than none. But again, 1) we differ on what I see versus what you see in the chart, and 2) I never pretend that one indicator is the Holy Grail, and yet that doesn't mean it's not worth highlighting in a blog post.

  3. Ed - thanks for responding. I understand this is just one indicator of many. I would hope as well that investors have their own viewpoints and use different tools and indicators, otherwise everyone would think the same thing and there could be no competitive advantage. ... Yes, the 1999-2000 trendline does not look like much of a trendline, and certainly not one that would be identified in real-time. I understand the decision is not whether to go long/short, but rather which one is showing better relative strength. I suppose "loser" is a bit negative of a term, but it's synonymous in this case for "the indicator picked the one with worse relative strength". Why not start at the peak - this refers to the 2007-09 trendline - I don't understand why this trendline did not start at the peak in the first half of 2007 but rather starts in the last half of 2007. This is why I think it would be difficult to identify in real-time.