ETFs start to underperform

By Felix Salmon
February 21, 2010
Ian Salisbury has an important story this weekend, saying that the average ETF underperformed its benchmark by 125bp in 2009, and even the monster SPY underperformed by 19bp. That's more than twice its total expense ratio.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Is this the beginning of the end of ETFs as an asset class? Ian Salisbury has an important story this weekend, saying that the average ETF underperformed its benchmark by 125bp in 2009, and even the monster SPY underperformed by 19bp. That’s more than twice its total expense ratio.

The problem is that when ETFs become very big, they become lumbering and predictable, and nimble hedgies know exactly what they’re going to do and when they’re going to do it. As a result, the smart money front-runs the dumb ETFs, which end up underperforming, sometimes by a very large margin: the $40 billion iShares MSCI Emerging Markets Index ETF (EEM) lagged its benchmark by a whopping 6.7 percentage points in 2009. That’s over nine times its total expense ratio.

So while it’s a good idea to avoid small ETFs, and to avoid commodity-based ETFs as well, even the biggest, safest ETFs are beginning to look as though they might have reached a level of size and popularity that makes them suboptimal investments. That’s sad, if true, because they were great while they lasted, and because there’s no real alternative out there.

But the fact is that there’s no rule of investing saying that there is always an easy and obvious investment strategy for people of relatively modest wealth. Investing involves taking a large number of risks, some obvious, some less so. And if ETFs continue to underperform in 2010 to the same degree that they underperformed in 2009, their repo income notwithstanding, then ETFs — which looked for a while there as though they really might be that rarest of animals, a positive financial innovation — might well turn out to be a grave disappointment for millions of investors who thought they could make a handful of asset-allocation decisions and then sit back doing little if any more work from then on.

We’re not quite there yet: as Salisbury points out, EEM is still outperforming its benchmark since inception, and it ouperformed in 2008. And for long-term investors, a single year’s underperformance shouldn’t matter a great deal. But if this turns out to be something newly endemic to the asset class, there might well be no cure for the problem — and that’s worrying, given how popular ETFs have become.

Update: On the other hand, if EEM is good enough for the Harvard endowment to have $388 million in it…


We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see

I don’t think it’s so clear that the evidence points to size as being a substantial contributor to ETF benchmark underperformance. At the very least, the front-running claim should apply to index mutual funds as well. Why should ETFs be different in this respect? Furthermore, if front-running index changes is a cause of underperformance, why should it matter whether you’re invested in, e.g., a large S&P500 fund or a small one? The problem would be attempting to track a popular index.

Citing EEM as evidence seems particularly problematic. As has been discussed widely elsewhere, EEM’s index sampling opens it to tracking error. VWO more fully replicates the same index and performed much better. According to your front-running hypothesis, shouldn’t VWO have suffered similarly?

Finally, regarding securities lending, ETF providers do not all remit the same amount of lending revenue to their funds. Vanguard remits substantially all, SSGA (SPDRs) somewhat less, and Barclay’s (iShares) only half. Where tracking errors of tenths or hundredths of a percentage points are at issue, these differences can matter.

Posted by ginsbu | Report as abusive

I would presume these errors don’t apply to indices that attempt to cover the /entire/ stock or bond market, as opposed to just large caps like the SPY. If everything’s in the index already, you can’t front-run it.

Given that small investors of modest means probably shouldn’t be making sector or country bets, these total ones are or should be the most important ETFs.

If I’m not mistaken, Vanguard for several years has been steering investors away from the ETF/index funds that track the S&P 500 and toward its total market index for just this reason.

Posted by expatsp | Report as abusive

Your piece doesn’t really say avoid commodity ETFs, it just says that the price of commodities might be hit if commodity ETF investors decide to sell. Which is true, but rather different than saying the ETF itself is defective (they might or might not be, hard to make the case in gold where they are backed by a physical asset but perhaps more possible in others)

Posted by mjturner | Report as abusive

Felix, consider this:

EEM has an annual turnover of just 5%.

If there is no turnover, there is nothing to front-run. How could poor pricing on %5 of the portfolio lead to underperformance of 6.7%. That would suggest that the fund is overpaying and underselling by more than 50% on buys and sells respectively. Likely?

There must be some other cause for this difference, and it must balance out for other years, no?

Posted by DanHess | Report as abusive

Dan, read the first comment. EEM samples. Felix has a point (SPY has no business failing to track perfectly, full stop), but this column wasn’t his best work.

Expatsb, what? You can front-run anything, but especially smaller, less-liquid names.

Felix, please revise and resubmit. ETFs are a funny space, and it’s worth asking why SPY didn’t track and why EEM doesn’t replicate. But this column isn’t the one you wanted to write.

Posted by wcw | Report as abusive

WCW, Ginsbu, good points, thank you (scales falling). Index sampling is not insidious because while there may be deviations, it is not biased against the investor. The investor might just as easily beat the index as fall short.

Many mutual funds underperform because of front-running, a problem exacerbated by the fact that their managers like to trade all day long. ETFs, with low turnover, are the solution to the front-running problem! Few trades, little front-running!

Long Live ETFs!!

Posted by DanHess | Report as abusive

Felix, why isn’t the problem true with normal index funds?

Posted by ReutersRat | Report as abusive