Product and Strategy Notes: Commodity Fund Fees, Retail Hedge Fund Products and New ETFs in USA

The Impact of Commodity Fund Fees

More than once we have been asked whether we take the relatively high expenses charged by commodity index funds into account when developing our model portfolios. The answer is we do not, because we have no way of knowing what commodity index funds may be introduced in the future (e.g., there are rumors that a commodity index ETF will soon be introduced with much lower expenses than the two existing U.S. commodity index mutual funds). Still, it is a question that we thought deserved further analysis. We took two approaches to this issue. First, we calculated the expected value of different share classes of the Oppenheimer Real Assets Fund after different holding periods, based our underlying asset class real return assumptions (expected annual return of 8.1%, and standard deviation of 18.3%). In our analysis, the fund's Class A shares had a front end load of 5.75%, and an annual expense charge of 1.49%. The Class B shares had no front-end load; their annual expense charge was 2.44% through year six, after which it was 1.49%. The Class C shares also had no front end load, but charged annual expenses of 2.40% throughout their holding period. Our analysis was undertaken from the perspective of a long-term investor. The following table shows the expected value of the different share classes (based on an initial $1,000 investment) after holding periods of different length.

This table certainly makes one thing clear: let it never be said that the folks at Oppenheimer don't have sharp pencils! As you can see, for a long-term investor, there is basically no difference between the expected value of the Class A and Class B shares. The Class C shares are another story, however, and seem better suited to people who don't expect to own the fund for very long.

Our second analysis looked at what would happen to our model portfolio asset allocations if we reduced the expected return on the commodities asset class to reflect the incremental expense (above that of a "typical" index fund) associated with the Oppenheimer fund. We chose to use our U.S. dollar 7% target real return portfolio in our analysis. While keeping the commodities asset class standard deviation unchanged, we reduced its expected return from 8.10% to 6.86%, reflecting the Oppenheimer Real Asset Fund's "extra" expenses of 1.24% (1.49% less a "normal" index fund expense ratio of .25%). We tested the impact of this change using both historical asset class returns and expected future returns as inputs into our simulation optimization model. In the former case, we found that the change had some impact, resulting in a 5% reduction in our allocation to commodities, and a 5% increase in our allocation to domestic equity. However, in the simulations based on expected future returns, the reduced commodities return had no impact on our asset allocation. Given the 67% weight put on the historical portfolio in determining our final model portfolios, the net impact of the higher commodities fund expenses was approximately a 3% reduction in our commodities asset class weight, and a three percent increase in domestic equities. However, this analysis comes with one important caveat: we deliberately used the most expensive commodities index fund in this analysis. Had we chosen another fund (say the PIMCO Commodities Real Return Fund institutional shares that are available through many fund supermarkets, and which have a relatively low .75% expense ratio), we suspect there would have been no impact at all on our asset allocations.

More on Retail Hedge Fund Products

A number of readers have written us asking us to write about a number of actively managed mutual funds that employ hedge fund-type strategies. The great majority of these are long/short equity funds in which the manager attempts to make money not only by being long undervalued companies, but also short overvalued ones. These funds include BPLEX, BNNIX, RMSIX, RMNIX, GFLAX, MLSAX, and CHLAX. While we don't know enough about any of these to prefer one over another, we have a couple of observations to make. First, the expense charges (and often front-end loads) on these funds are all quite high. Second, their performance is not very impressive (though many of them are new). Finally, as we have noted in previous articles on this subject, we are not big fans of the long/short equity subset of the hedge fund universe. As we have shown, the returns on the CSFB/Tremont long/short style index are quite highly correlated with returns on the equity market. Based on our analysis, a far better choice is a combination of the equity market neutral and global macro hedge fund style indexes. Whereas long/short equity still contains a substantial amount of equity market beta risk, equity market neutral managers typically hedge all of this away, leaving just their bets on company-specific risks and returns (also known as alpha). Logically, EMN has a lot of appeal: you should not pay the high fees charged by an active manager to obtain beta (i.e., pure asset class) risk and return, which you can obtain much more cheaply through an index fund. Rather, you should pay up only for alpha, which is what you're getting with the EMN style. Unfortunately, we do not know of any retail EMN funds available today.

Global Macro funds are altogether different. In essence, global macro managers pursue the age-old art of market timing, pure and simple, tactically shifting their asset class weights in line with their forecasts for future relative returns. While there is still no "pure" global macro style index fund available, there may be the next best thing. Pimco has recently launched its All Asset Fund, managed by Bob Arnott of First Quadrant Securities (one of the smartest people you will encounter in the investment business). The fund's brief is to maximize long term real returns by tactically shifting its allocations between the same broad set of asset classes we use in our model portfolios. The institutional class shares of this fund (which may be available through various fund supermarket programs) trade under the ticker PAAIX. Between June, 2003 and February, 2004, the funds compound nominal return was 0.86% versus 1.22% on the CSFB/Tremont Global Macro Hedge Fund Style Index (the funds in which probably use more leverage than the Pimco fund). The correlation of returns between the two (over an admittedly very short period) was .57. In sum, if one were trying to figure a low cost way to include hedge fund type vehicles in a portfolio, this looks like it would be a very good place to start.

New ETF's in the USA

Finally, a short note on two new ETFs that will soon be launched. The first will track a gold index, and the second will track an index of Chinese stocks. The attractiveness of both of these issues is contingent on your views regarding the alternative scenarios we described in this month's Economic Review. Clearly, if future changes indicate that the most dangerous situation is developing, shorting China (if you are inclined to make this type of active bet) and going long gold may make a lot of sense. In the absence of these conditions, however, we would avoid a tilt towards gold within the broader commodities asset class (after all, gold is included in both the Goldman Sachs and Dow Jones-AIG indexes), or towards China within the broader emerging markets asset class (again, China is included in the MSCI Emerging Markets Index).

| Economic Update | Model Portfolio Update | This Month's Letter to the Editor: Commodity Funds with Timber and Alternatives | Product and Strategy Notes: Commodity Fund Fees, Retail Hedge Fund Products and New ETFs in USA | Global Asset Class Returns | Equity Market Valuation Update |