A reader was kind enough to pass along this Mark Hulbert story (free registration required) from The New York Times that indicates that all-ETF portfolios tracked by his service have "turned in disappointing results."Hulbert notes that his research sample size might be too small, but the results are what they are. The reader wanted my opinion on the outcome. I would tie this in to what I have been writing all along about all-ETF portfolios and investment products in general: They all have flaws that need to be understood and possibly mitigated through prudent portfolio management. I first touched on this idea about a year and a half ago on my blog, but it's worth repeating: An all-ETF portfolio is not a great way to go. The biggest flaw with ETFs is that they often yield less than common stocks. I said back then that the consequence is that an all-ETF portfolio has a much heavier reliance on price appreciation, and in a flat market like we have had recently, all-ETF portfolios will lag a stock portfolio that includes stocks with large dividends. Another issue that could account for Hulbert's results might be that most all-ETF portfolios use what I believe are the wrong ETFs. Most of the products I have seen rely too much on cap size and style ETFs and very rarely do they use sector and single-country ETFs. I have disclosed before that I use a lot of ETFs to assemble some client portfolios, but none of them are 100% ETFs, even if they do have a lot, and only in extenuating circumstances do I use any style or cap-size ETFs. For these accounts I use sector ETFs and weight them the way I weight sectors in regular accounts.
Because iShares has sector funds that are domestic only and also has sector funds that own foreign stocks blended with domestic, I am able to capture the sector weights I want and can get pretty close to the foreign weight I want as well. I can then use a single-country ETF to capture a little more foreign exposure, and possibly more yield. The following is an example of what this might look like. I have intentionally omitted the weighting so that no one blindly copies this without doing proper homework.
Financials: Dow Jones U.S. Financial Sector (IYF); Australian Bank; Irish Bank. Tech: Dow Jones U.S. Technology Sector (IYW); Goldman Sachs Semiconductor (IGW); MSCI Taiwan (EWT). Health: S&P Global Healthcare (IXJ); generic drug stock; medical device stock. Industrial: Dow Jones U.S. Industrial Sector (IYJ); Power Shares Water Resources (PHO); defense stock; foreign industrial stock. Consumer: Dow Jones U.S. Consumer Goods Sector (IYK); retail chain; coffee company; foreign consumer stock. Energy: S&P Global Energy Sector (IXC); oil sand name; Chinese company. Telecom: Vanguard Telecom Services Viper (VOX); foreign stock. Utilities: Dow Jones Utility Sector (IDU). Materials: StreetTracks Gold (GLD), timber REIT, emerging market stock. This is just an example and does not represent any one account. The above has 26 holdings, and it would cost about $260 to implement, which would be a 0.26% drag. I would use only a cap or style ETF for a very small account or for someone with an unusually low tolerance for volatility, of which we have a couple. I believe this approach is better than a lot of what I have read elsewhere, but this is not my first choice about what to do in a portfolio and, like all similar portfolios, it is flawed. But I feel I can better control the overweights and underweights with this type of mix. If I really wanted to tweak it, like to reduce the median cap size, I could add one of the micro-cap ETFs, but this is as close to all-ETF as I get.