Why ETF-Squareds Are a Bad Investment

|
 |  Includes: PAO, PCA, PTO
by: Felix Salmon

Floyd Norris today brings up the question of mutual fund fees, which gives me a good excuse to revisit the ETF-squareds I wrote about yesterday. By coming up with an extra service (quarterly rebalancing) they justify an extra 25bp in annual fees; I was dubious as to whether that was much of a deal for any investor. After all, there are S&P 500 index funds charging just 7bp which have to benchmark a very specific portfolio of 500 stocks.

The Condor Options blog agrees that 25bp seems a bit steep:

The sponsors (PowerShares) should set the rebalancing back to annual, and cut the fee to 5bp. We'd buy that, or at least might consider putting our cousins and neighbors in such a fund.

I also corresponded with Russ Koesterich of ETF giant BGI, who's recently written a book called The ETF Strategist. He was reluctant to talk about a rival (PowerShares) product directly, but he did say this:

The real value of the ETF is the ability to build multi-asset class diversification at a low cost, and with a great deal more ease than you ever could trying to assemble a portfolio of individual stocks. At the end of the day, I would still be suspicious of outsourcing periodic portfolio rebalancing.

There are three reasons why I would be suspicious of the PowerShares product, beyond the fact that I'm just not sure the rebalancing service is worth 25bp.

Firstly, the rebalancing only takes place between the pre-ordained constituent ETFs. Any sensible investor will take advantage of the rebalancing process to do a quick check on how the ETFs are doing: if they're underperforming their benchmarks by a large margin, or if they are truly atrocious, then they can be sold outright and replaced with something better.

What's more, there's a slow but steady increase in bargain-basement ultracheap ETFs. As these funds come on to the market, it makes sense during the rebalancing process to rotate out of your older, more expensive, funds, and into the cheaper ones. The ETF-squareds don't do that: You're stuck with what was available when the ETF-squared was formed.

Finally, the ETF-squareds will look at the balance between funds, but not the balance within funds. Here's Koesterich, again:

While investors should always be focused on minimizing unnecessary transaction costs, it is sometimes worth rebalancing if the composition of an ETF's underlying index changes dramatically. Consider the case of an investor who bought an ETF providing exposure to large cap stocks in the mid 90's. By mid-2000, Technology stocks made up over one third of a US large capitalization index. An investor in the SPY or other US large cap ETF now owns a fairly concentrated portfolio, heavily weighted to Tech, rather than a diversified portfolio of securities, which was probably their original intention. In that instance, it would certainly of been worth the time, effort, and expense to rebalance given that Technology stocks proceeded to lose 80% of their value over the next two years. The point is not to engage in market timing, but to adjust your portfolio when the instruments you own no longer reflect the exposures you were looking for in the first place ...

If you bought a broad commodity ETF a few years ago, and due to the runnup in oil the fund is now mostly weighted to energy, youi may want to rebalance so as to ensure that your portfolio has the desired exposure to a broad basked of commodities as opposed to a single commodity, such as crude oil.

This kind of rebalancing, where you look at the constituent parts of the funds you bought and you ask yourself whether that was what you wanted when you bought the funds in the first place, is something I think the ETF-squareds don't do. For them, a stock fund is a stock fund; a commodity fund is a commodity fund.

For all these reasons, then, I'd be wary of recommending the ETF-squareds to anybody, really: they're an expensive way of putting together a hard-to-change ETF portfolio. If you want an idea of a good mix of which ETFs to buy, then by all means check out the ETF-squareds for ideas. (Or go here.) But I don't see any good reason to actually pay PowerShares the extra 25bp for doing that for you.