Model ETF Portfolio Adds Foreign Funds to the Mix

    by James B. Cloonan

    Model ETF Portfolio Adds Foreign Funds To The Mix Splash image

    Last spring, AAII began an experimental research portfolio comprised solely of exchange-traded funds (ETFs) called the Model ETF Portfolio. Starting with this article, I will be covering the Model ETF Portfolio separately from the Model Mutual Fund Portfolio. In addition, I will review the exchange-traded funds in the Model ETF Portfolio each May and November in the AAII Journal. Since we began the Model ETF Portfolio on April 1, 2006, the new publishing dates will allow us to provide six-month and one-year portfolio results.

    The Model ETF Portfolio

    At this time it is important to point out the differences between the Model ETF portfolio and the Model Mutual Fund and Shadow Stock Portfolios. The latter two portfolios are based on criteria that are fixed and anchored in historical research, although these criteria may change gradually over time.

    On the other hand, the Model ETF portfolio consists of a class of investments that is new, as are the numerous individual funds that are open and are opening almost daily within this class.

    In trying to sort out an overall balanced portfolio of these funds, there are two separate decision issues:

    • First, what asset classes belong in an effective portfolio and,
    • Second, which of the many exchange-traded funds purporting to represent that class will actually do the best job of mirroring that class in terms of return and risk.

    The Model ETF portfolio at this point, as shown in Table 1, consists of the following asset classes:

    • Large-capitalization stocks
    • Mid-capitalization stocks
    • Small-capitalization stocks
    • Micro-capitalization stocks
    • REITs
    • International stocks

    However, because of the complexity and short history of ETFs, I want to emphasize that this is an experimental portfolio. While I believe the kinds of assets we are pursuing make sense based on history and theory, this is only partially substantiated by performance.

    In addition, how well the managers of the various ETFs will accomplish their objectives is still uncertain. So even if an approach is theoretically sound, the implementation also needs to be observed and evaluated over time. While we do not have specific fixed criteria for selecting ETFs at this point, we do have objectives as shown in the box labeled “Selection Rationale.” Our selections are based on an attempt to achieve risk reduction through diversification with minimum reduction in the rate of return.


    Unique Problems

    As an example of the kind of problem that might emerge within our experimental Model ETF Portfolio, think of the problem of weighting the various holdings in an individual fund.

    Many of the exchange-traded funds in the Model ETF Portfolio are not capitalization-weighted. That is, the percentage holding of the individual stocks in the fund are not based on the stock market capitalizations (number of common shares outstanding times the price per share); instead, the percentage stock holdings are based on some other criterion, or they are equally weighted (the same dollar amount is invested in each stock).

    I strongly believe that weighting by the market capitalization of each stock reduces the total return of the portfolio, and this is a topic I will devote much of a column to in the near future.

    But the advantage of using market capitalization as the basis for the stock weightings in the portfolio is that you never have to actively rebalance a portfolio. As the prices vary, the market capitalization changes to maintain the proper weight—in other words, the portfolio automatically rebalances itself, with no transactions necessary.

    Any other weighting system requires rebalancing, and rebalancing involves transaction expenses. So part of the gain from a better weighting system will be given up to transaction expenses.

    How this resolves itself is one of the challenges of an ETF portfolio. [For more on this issue, see the article “Explaining the Value Tilt and Its Implications for Investors” by William Reichenstein in the July 2007 AAII Journal.]

    All of the above explains why I think of the ETF portfolio as experimental and why I would suggest only committing limited funds to it initially—if any.

    But even though there may be problems with the portfolio, I believe following its development will provide an interesting and worthwhile learning experience. I certainly intend to share the experiences of the portfolio development with you.

    I also intend to spend time looking at some of the unusual ETFs being developed, even if we don’t put them in the portfolio. And when we have sufficient historical data, we will look at how various weightings of the individual funds can affect risk.

    The Portfolio's Performance

    The results to date for the portfolio are presented in Table 1. As you can see, we have fallen behind the benchmarks due to inferior performance over the past six months. This is almost entirely due to the REIT fund. While REITs have significantly outperformed the S&P 500 for the past three-, five-, and 10-year periods, they are having a difficult time this year—at least so far.

    REITs, which gave us superior performance last year, are not doing as well this year. This illustrates that the advantage of REITs long term is not just their returns but the reduction in portfolio volatility (risk) that they provide in the short term. Of course, most investors would prefer not to dampen the upside movement, but that is part of overall risk reduction.

    It is still too early to illustrate performance with a graph, but we will start it next year.

    Adding International

    As discussed in February, we studied the issue of diversifying internationally. We did add a foreign component to the portfolio as of the last week in June.

    Because we wished to diversify within the international market but did not want this asset class to be too large a portion of the portfolio, we needed to weight the international holdings differently than the U.S. holdings:

    • The five U.S. holdings will continue to be weighted equally, but now at 16% each rather than 20% of the overall portfolio.
    • Each of the new foreign holdings will be weighted at 5% of the total portfolio.

    This means that for every $1,000 you have in each of the five existing ETF positions, you should put $312.50 in each of the four new ETF positions, if you are already investing in this model portfolio.

    The New Funds

    The following are the four new international funds added. They are listed in the table but will not impact the portfolio results until the future:

    • SDPR S&P Int’l Small Cap (GWX): An index of small-cap (market cap under $2 billion) stocks of developed nations, and capitalization weighted;
    • SDPR DJ Wilshire Int’l Real Estate (RWX): Basically an index of non-U.S. REITs and other foreign real estate holdings;
    • Vanguard FTSE All-World Ex-U.S. (VEU): A cap-weighted index of the equities of 47 countries; and
    • Vanguard Emerging Markets (VWO): Focuses on emerging markets and follows the MSCI emerging markets index.

    As you can see, this combination of funds covers most sectors of the non-U.S. equity market. However, the newer approaches using weighting other than cap-weight and creative indexes are not yet available for non-U.S. equities. Therefore, these are primarily capitalization-weighted funds. Because of the cap weightings, they tend to be dominated by larger companies and there is less emphasis on smaller enterprises than I would prefer, but this may be safer for foreign firms.

    The two Vanguard ETFs have competitors that are more established. But since Vanguard is low cost, has extensive experience, and these are straight index funds, I chose the Vanguard ETFs.

    The Long-Term View

    It will be some time before we can make definitive judgments about the importance of diversifying internationally or the best way to do it. Many of the cycles that influence such judgments take five years or more, and in most cases we can’t even go back in history that far because the funds are new. But I have learned from the Shadow Stock Portfolio that five or 10 years goes by pretty fast.

    The Model ETF Portfolio, like the Model Mutual Fund Portfolio and the Model Shadow Stock Portfolio, is intended to have a long-term view. There is no—and will be no—effort to guess which geographic regions or market segments will do best in the short term.

       Model ETF Portfolio: Selection Rationale

    The rationale used in building the Model ETF Portfolo is to achieve diversification across the equity classes listed below while maintaining a weighting that, in our assessment of historical data, will provide the maximum opportunity for long-term rates of return. We have a bias toward smaller-cap and value stocks and so does history.

    Across national boundaries—U.S. versus foreign:

    We begin with an 80% U.S. and 20% foreign portfolio but this could change. Foreign stock returns involve currency relationships as well as the usual equity analysis. The initial weighting takes into consideration the fact that many U.S. companies have significant foreign involvement.

    In foreign investments:

    • Style will be diversified. We will seek emphasis on value stocks when it is possible.
    • We will seek a heavier weighting in the small-capitalization area than the typical portfolio.
    • We will diversify across equities and real estate, but will not use foreign bonds for risk reduction—at least not initially.

    In U.S. investments:

    • We will diversify across equities, real estate, and short-term bonds. Short-term bond ETFs will be included as an option for investors who need further risk reduction. However, they will not be in the actual Model ETF Portfolio.
    • Our style diversification will aim at a heavier emphasis on value than the overall market.
    • The capitalization weightings will provide considerably more emphasis on small-capitalization stocks than the overall market. We will seek to achieve this not only by including small-cap ETFs but by choosing larger-cap ETFs that do not weight solely on capitalization.

    Which specific ETFs?

    Although the above outlines the areas in which we will look for ETFs, it does not explain how we will choose specific ETFs when there are multiple ETFs in an area.

    It will be many years before we have enough history to develop a solid set of criteria as we have for the Model Mutual Fund Portfolio. Many of the sponsors of ETFs, however, have a history with other investment vehicles that can provide a guide, as can liquidity, expense ratios, and the philosophy espoused in prospectuses. Over time we should be able to harden our criteria.

→ James B. Cloonan