Emerging markets: hedge or diversifier?


At what point does a diversifier become a hedge?  Asset allocation has been driven of late by the desire to access a wide range of alternative asset classes.  University endowments like Harvard and Yale have become the gold standard in investment management.  (See this post by Mebane Faber at World Beta on the subject.)  The entire drive for broad diversification is an attempt to access a diverse set of risk premia in addition to generating active alpha.

The case for including an asset class in a portfolio is often made on the correlation of that asset with other “core” asset classes.  Emerging markets have been touted for some time as an excellent portfolio diversifier due in large part to their low correlation with developed market equities.  (The folks at Bespoke Investment Group have a cool correlation matrix showing recent changes in asset class correlations.)

This question about diversification came to mind when reading an article by Agnieszka Baczyk and Herbert Mayo at IndexUniverse.com.  In it the authors examine the correlation coefficients of a variety of emerging market equity ETFs against the S&P 500 (SPY).  In their research they find that the correlations are higher (and rising) than that previously reported in the literature.  Where they go off track is in their conclusion:

To achieve diversification, the numerical values of the correlation coefficients relating the returns on the alternative investments need to be small or even negative. The correlation coefficients calculated in this study are relatively large and inconsistent with prior research. Based on these results, investing in emerging market ETFs traded on American securities markets no longer offers much potential to diversify a portfolio consisting of American stock.

That is indeed a high standard for a portfolio diversifier.  They seem to be saying that even mild portfolio diversification is not worth the effort.  For instance, most investors include both large and small cap equities in their portfolio despite their very high correlation.  It seems the authors are conflating the issue of diversifiers and hedges.  Hedges are explicitly designed to (or inherently) provide offsetting returns.  Applying that standard to portfolio diversifiers would in fact eliminate many asset classes from consideration.

Another point is that you cannot judge diversifiers simply by the results of a correlation matrix.  Portfolio construction involves assembling assets that generate different risk premia across time and space.  It is pretty clear that emerging markets, especially the frontier markets, are not simply S&P 500 wannabes.  Their returns come from a very different set of real economic factors.

Emerging market equities may (or may not) be the greatest asset class since sliced bread.  After a strong multi-year run, led by the BRICs, they have pulled back substantially this year.  But emerging markets do provide investors the opportunity to access risk premia that are different from typical core assets like the S&P 500.  The decision to include emerging market equities in your portfolio, especially via the raft of ETFs now available, is always your own.  But you should not be scared away from them simply because their correlations are at times closer to 1 than 0.


4 Responses to “Emerging markets: hedge or diversifier?”

  1. good analysis of the hedge/diversify distinction

    what’s even more interesting is that these benchmark funds you mention are really the “speculators” that the liberal press is blaming – it’s the big endowments and pensions who’ve diversified into alternate asset classes like COMMODITIES, which plays a role in their price increases. of course, there’s nothing wrong with that, but it’s interesting that the press has a stereotypical image of a speculator that i’m fairly sure does NOT include the heads of Harvard and Yale endowments.

  2. A corollary to the important second point is that correlation matrices are so often unstable (time-varying). Sometimes the apparent tight correlation is due to shared exposure by the two “correlated” asset classes to some other set of factors/regime, which is only historically relevant to the window. In other cases, correlation is temporarily elevated due to low volatilities, as we recently saw among certain hedge fund strategies. The CDO mess proves that default correlations are notoriously chimerical, but, only to a lesser extent, so to forward-looking return correlations. It’s impossible to eliminate sample bias (i.e., the one sample that recent history happens to, maybe accidentally, provide).

  3. A note from Herbert Mayo, co-author of the cited article (via e-mail):

    Thanks for the comment on our piece concerning emerging markets and
    ETFs. We initially thought the correlations would be low…hence
    the potential for diversification exists….But they were so high
    that the potential is marginal…that does not mean do not include
    emerging market ETFs….but it does suggest that brokers/advisors/
    etc. have oversold the diversification benefits.

  4. correlation drift and response to shock is a severly understudies area of finance. considering all the sloppy attention to mean-variance analysis this is shocking.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: