The initial portfolio weights were determined by the relevant market caps at the end of 2001, as per the capital asset pricing model's idealized instructions. This is simple enough for the stocks, bonds and REITs. The challenge was figuring out how to weight commodities, which of course have no market capitalization.

Indeed, commodities are inherently a speculative asset since there is no formal cash flow, earnings, or other fundamental data available for analysis. Yes, supply and demand factors are relevant and so they ultimately dictate price. But the task of figuring out what commodities in the aggregate are "worth" in an accounting sense requires guessing in more than trivial doses. Ideally, the guessing is rooted in something approximating hard numbers and financial theory, which is why we settled on using the total dollar value traded [TDVT] figure, as calculated by Goldman Sachs. By our reckoning, that seems to be about as close to a market capitalization-equivalent in terms of what's available to the public. The TDVT number is calculated over several months for a given year and encompasses roughly two dozen futures contracts traded in Chicago, New York, and London.

Nonetheless, there are a number of caveats. First, a global market portfolio, even if it was calculated with absolute perfection as per financial theory (and GMP falls short of that idealized standard), is appropriate for the average investor. Any given investor will have particular financial objectives, risk tolerances, cash flow needs, etc., that call for some deviation from the true global market portfolio. There are also those investors who believe (rightly or wrongly) that they're in a position to correctly second-guess Mr. Market, in which case there's an argument for doing something different, perhaps radically so, relative to the global market portfolio. In addition, our index at the moment goes back to only 12.31.01, which is far from a robust sample.

Still, it's important to know what's available as a default investment portfolio, courtesy of the global capital and commodity markets. Such a benchmark reminds us what's available to everyone, at minimal cost. No skills are required. The GMP index, we believe, provides such a rough approximation of the true default portfolio.

Keep in mind that there are a number of moving parts to calculating a global portfolio index. Different researchers may reach different conclusions about building such a benchmark, and so different benchmarks probably will dispense different results.

This is why:

The choice of indices can alter results. Meanwhile, certain asset classes were left out of our index. Notably, we don't use foreign REITs or foreign inflation-indexed bonds.

Different start dates can produce different track records. We choose 12/31/01 because we have yet to find data for all the asset classes prior to that date. As we find older data, we'll recalculate going back further in time.

Perhaps the most important caveat is that turning the GMP index into a real world portfolio would incur costs in terms of fees, taxes, etc. — burdens that paper benchmarks can blithely ignore. So, yes, you can buy the relevant mix of ETFs and/or index mutual funds and create your own global market portfolio, with or without your own tactical overlays. But jumping from paper portfolios, regardless of the underlying strategic design, invariably introduces new and perhaps expected risks.

Then again, one has to begin somewhere when it comes to pondering the investment challenge and the global market portfolio seems like an obvious place to start if only to offer a bit more perspective on the money game.