April 8, 2008
INDEXING THE GLOBAL MARKET PORTFOLIO
Buying Mr. Market in all his various asset class flavors is easy these days, thanks to the proliferation of ETFs and mutual funds that mine all the major (and increasingly minor) niches in the capital and commodity markets. But what is Mr. Market offering exactly? And what does his track record look like?
That's a crucial question for strategic-minded investors, if only to catch a glimpse of the true global market benchmark, which by definition is diversification in full. Alas, there's no off-the-shelf index for the global portfolio, at least none that we've come across. The vacuum inspired your editor to put one together, and so today we unveil the Capital Spectator Global Market Portfolio Index (GMP), which is an approximation of the global capital and commodity markets and weighted as per Mr. Market's valuation. We'll be using the index in future posts to compare and contrast various trends in the financial markets.
The methodology behind the benchmark in discussed in some detail below, but first let's address the obvious question: how has GMP performed? The quick answer is in the chart below, which shows the relative total return performance of GMP against the S&P 500. As you can see, GMP handily beat the S&P 500, from the end of 2001 through March 31, 2008.
Looking at the trailing performance numbers, the GMP index generated a 13.4% annualized total return for the five years through the end of last month vs. 11.3% for the S&P 500. In addition to outrunning the S&P, the GMP index did so at about three-quarters of the S&P's volatility over those five years (measured by annualized standard deviation of monthly total returns). The practical evidence of this smoother ride is evident in recent history. From the S&P's peak back in October 2007, the stock market suffered a -13.8% tumble through the end of last month, vs. a mild -3.2% loss for GMP.
What's behind GMP's impressive risk-adjusted performance? It's rather elementary, actually, in that our index embraces modern portfolio theory and builds a benchmark that holds all the major asset classes in their market weight on the start date (Dec. 31, 2001). Make no mistake: this is passive investing writ large, with all its faults and benefits.
Accordingly, Mr. Market is the investment strategist with full authority over the portfolio and so he incessantly adjusts the weights as he sees fit, letting the financial chips fall or fly where they may. After the initial market cap settings, subsequent investment performance alone determines the rise or fall of any asset class's weight in the index. Meanwhile, because the portfolio holds some assets that post low and negative correlations with one another, financial intuition suggests that this index is superior in many ways compared to the stock market by itself, or any one asset class index for that matter, as a long-run proposition.
As for GMP's basic design, we recognize 10 asset classes for the index and use some familiar and not-so familiar benchmarks as proxies:
1. US stocks (Russell 3000)
2. US bonds (Lehman Bros. U.S. Aggregate)
3. Foreign stocks/developed markets (MSCI EAFE)
4. Foreign emerging market stocks (MSCI Emg Mkts)
5. Foreign bonds/developed markets (Citigroup WGBI)
6. Foreign emerging market bonds (Citigroup ESBI)
7. US REITs (DJ Wilshire REIT)
8. Inflation-indexed US Treasuries (Lehman Bros. US TIPS)
9. Commodities (DJ-AIG Commodity)
10. US high yield bonds (Citigroup High Yield)
note: the foreign bond and stock indices are in unhedged dollar terms
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 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.
A few examples for why that's so: 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.
Posted by jp at April 8, 2008 11:28 AM
The weightings are based on GMI's origin date of 12/31/97. Those weights are drawn from publicly available information based on the market values of the various indices. For instance, I'm using data from Standard & Poor's (www.sp-indexdata.com). It's a simple matter to add up the market values and passively allocate GMI. The tricky part is estimating market values for REITs and commodities.
Posted by: JP at November 5, 2011 7:21 AM
Where do you get the current weightings?
Posted by: Afs at November 4, 2011 7:43 AM
The market cap numbers come from S&P/Citigroup. For the latest data, see the previous post.
Posted by: JP at April 9, 2008 8:21 PM
Also curious, what are the weights per today's global market valuations? Where are you getting your numbers?
Posted by: SM at April 8, 2008 10:29 PM
As of 3/31/08, the broad weights work out to roughly half in stocks, one-quarter in bonds and one-quarter in commodities, with an insignificant amounts in REITs.
Posted by: JP at April 8, 2008 2:10 PM
But what weights did you arrive at?
Posted by: Lord at April 8, 2008 1:51 PM