The ETF-Squared: It Reallocates For You

Remember Ron Lieber's catchy little slogan? "Index (mostly). Save a ton. Reallocate infrequently." Turns out, here's a new ETF out there which doesn't just do the indexing for you, it does the reallocation as well. To be precise, there are three of these ETF-squareds, and they're listed on the Amex under the tickers symbols PTO, PAO, and PCA. All of them, interestingly, comprise nothing but other ETFs.

PowerShares Autonomic Growth NFA Global Asset Portfolio ETF (PTO) is the most aggressive, with 90% equity and 10% debt; PowerShares Autonomic Balanced NFA Global Asset Portfolio ETF (PCA) is the most conservative, with 60% equity and 40% debt. All of them have a 5% or 6% real-estate component (included in equity), with the rest of the equity split roughly 42-58 between US stocks and the rest of the world. Both the equity and the bond holdings are very diversified, and there's even a small 2% to 3% stake in commodities and currencies, included in the debt allocation.

All of these seem like pretty sensible allocations, depending on your risk profile. And the funds seem to reallocate sensibly too:

The underlying indexes are rebalanced on a quarterly schedule but can be rebalanced more often if market conditions cause the indexes to deviate from their targeted risk levels.

I have two issues with these ETFs. The first is that, since they're sponsored by PowerShares, PowerShares also gets first dibs on providing the underlying ETFs, even if those underlying PowerShares ETFs might not be the best or cheapest option.

And then there's the whole question of the value of rebalancing. The ETF-squareds charge a fee of 25bp on top of the fees charged by the underlying ETFs. Basically, it's a fee charged for the quarterly rebalancing service.

Now you don't need to reallocate quarterly: annually is just fine, especially if the reallocation involves actually selling some of your holdings, rather than just investing new savings so as to rebalance your portfolio. But just how important is reallocation, and is it worth paying PowerShares 25bp per year for them to do it for you? Has anybody tried to quantify the costs of not reallocating, or of doing it badly?

Comments

[billb]

I think studies have shown that quarterly is WAY too frequent. As a general rule, the less rebalancing the better. I believe the case was made that you'd mostly be rebalancing into bonds since stocks outperform and more money in bonds = lower return. So by rebalancing into bonds more frequently, your returns are pretty lousy over time.

I have not tried this myself, have only read of it.

[Josh77]

Claymore's UEM product (based upon one of our indexes) includes stock, bonds and cash and rebalances monthly.

[karchad]

I agree with Billb, and my own personal experience bears it out, he's exactly right. When i rebalanced too frequently, i was missing too much gain in sharp equity moves up. I do it once a year, February, hoping it's the end of the year end rally (where was it this year???)

[fingham]

My experience is that one should re-allocate whenever an asset class moves out of a targeted allocation collar. That can happen any time. A collar of around 1% in either direction has worked well for me, though, as always, do your own research. Though this collar might seem narrow, I don't actually need to touch my allocations that often - every couple months or so. Also, if I add or remove funds to my portfolio, I use those opportunities to buy the most underweight (when adding funds) and sell the most overweight (when removing), whether or not the class is outside the collar. This helps keep things in line too.

[Smart ETF]

Wake up world, yes rebalancing frequently is inefficient (IN YOUR MODELS) using mean-variance models such as Modern Portfolio Theory (circa 1950’s) and Arbitrage Pricing Theory (circa 1970’s), ala Black-Litterman. How can one year of new data possibly effect a change on more than 50 years of data? The real questions is who would want to rely on the average price, risk, return and correlation of a security or asset class based on the average of 40 or more years? Can you name one company whose 40+ year historical average is currently performing the same? Why do you think Fama, Mandelbrot and now Sharpe have all debunked MPT? Only Markowitz is holding onto his dream.

I find it hard to believe people still put stock in these theoretical models, but then again you also believe in Normal Distributions. May I suggest you read The (Mis) Behavior of Markets by Mandelbrot or any of Taleb’s books.