January 23, 2013
Two More Betas For The Global Market Index
The Global Market Index (GMI) is expanding its asset class horizons. Starting with the numbers as of December 31, 2012, GMI's allocations will add foreign REITs and foreign high-yield bonds to the mix. Next week, when I publish the monthly update on asset class returns through January, these two betas will make their formal debut. (For new readers who are wondering what I’m talking about, here’s the latest monthly update of the major asset classes and GMI.) The reasoning behind this change is that the arrival of ETFs that track these markets makes it easy and cost-efficient to allocate to non-US REITs and high-yield bonds. Considering the history of these markets in context with other asset classes, there's also a strategic/tactical argument for adding foreign REITs and foreign high-yield bonds to the investment opportunity set. The correlations between the new additions and the usual suspects is still fairly high, but it's well short of perfect positive correlation. Accordingly, there's opportunity in these betas for enhancing the rebalancing bonus, if only on the margins.
In the grand scheme of the market-weighted GMI, which is never rebalanced, the new betas on the block will have a minimal effect on risk and return for the foreseeable future. Offshore REITs and junk bonds constitute a small slice of the global capital markets. The market value of foreign junk bonds is about 10% of the non-US global investment-grade corporate bond market as of 2012's close, according to data from Markit Economics and Citigroup. Foreign REITs represent a relatively larger share of the global market cap for real estate securities, according to data from Standard & Poor's: roughly 40%, with US REITs grabbing the remaining 60% slice of the pie, as of December 31, 2012. Of course, the fact that REITs represent a tiny piece of the global capital markets is an artifact of securitization—the true value of real estate proper is dramatically higher than the market cap for REITs. As such, adding foreign REITs based on market cap adds up to a marginal change as well for the market-value weighted GMI. Should you own a bigger slice of REITs in your portfolio to accurately reflect the world's property values? Possibly, although I'll leave that topic for another day.
For consistency, GMI and its counterparts will reflect the changes from December 31 onward. Alternatively, there's an argument for restating the entire GMI history by recalculating returns from the benchmark's 1997 start date. But this path is problematic for an index that is first and foremost designed as an investable measure of the major asset classes. In a bid to maintain its real-world, investable track record--as it currently exists--I've opted to add foreign REITs and foreign junk bonds from here on out, and leave the historical record as is.
In order to add the new allocations and maintain the continuity of GMI it's necessary to finance the extra investments from elsewhere in the portfolio. For simplicity, I'm simply taking the appropriate market-value share out of investment-grade foreign corporate bonds to finance the foreign high-yield addition; foreign REITs are equivalently financed from the existing domestic REIT allocation. The bottom line: GMI's net asset value is continuous and unaffected by the two extra betas.
The indexes that will represent the new asset classes in GMI: Markit iBoxx Global Developed Market ex-US High Yield and S&P Global ex-US REIT. Following the rules elsewhere in GMI's use of foreign assets, the unhedged-currency versions of the indexes will be used to calculate GMI's returns. In other words, the assets of the indexes are priced in local currencies and routinely translated back into US dollars at current market rates for generating performance data. That's fairly standard when it comes to pricing foreign assets in mutual funds and ETFs. For good or ill, it also exposes the portfolio to forex risk, which is arguably another layer of diversification, albeit embedded in the assets as opposed to carving out a separate allocation for currencies.
For GMI in its ETF version (GMI-F), which is designed as an investable benchmark, I'll use two funds as proxies for the asset class additions: iShares Global ex-USD High Yield Corp Bond (HYXU) and Vanguard Global ex-US Real Estate ETF (VNQI). Both of these ETFs have competition, and so no one should assume that these are the only product choices for the betas in question. But for our purposes here, this pair represents a reasonable solution for securitizing the beta additions to GMI when it comes to calculating GMI-F.
Although the influence of the new asset classes on GMI will be slight, due to the relatively small market caps, foreign REITs and high yield bonds will have a somewhat larger impact on the rebalanced version of GMI through time, aka GMI-R. An even larger influence is likely for the equal-weighted version of GMI (GMI-E).
The main incentive for adding these asset classes is that they deserve routine analysis and consideration for portfolio design. Indeed, buying and selling these betas is no more difficult than trading US equities or bonds writ large. If you have strong views on foreign junk or foreign REITs, you may decide to overweight, underweight or ignore the assets entirely. In fact, no less is applicable for all the asset classes. But when it comes to monitoring the broad set of betas in the world, it's no longer reasonable to ignore REITs and junk in foreign markets.
The opportunity set for GMI now stands at 14 in terms of defining the major asset class—15, if you include cash, which isn't used for calculating GMI or its counterparts. Yes, most if not all of the 14 asset classes can be further divided into a more granular set of markets, depending on the investment strategy and the assets under management. In addition, one could just as easily argue that the major asset classes should be categorized by risk factor—value vs. growth stocks, for instance, or momentum vs. capitalization. The sky's the limit for defining risk. But as a first approximation of the world's betas, the following list captures the lion's share of the primary drivers of risk factors available to everyone. What you do with these betas is something else entirely, but at the very least you should be aware of the menu if you're going to take a seat in the restaurant.
Ultimately, the goal in calculating GMI is one of tracking how a broad set of the major asset classes perform through time sans active management. As such, GMI and its counterparts are robust benchmarks, since anyone and everyone can replicate the results with minimal effort. In short, a Ph.D. in finance isn't required.
If nothing else, GMI provides a roadmap for thinking about how (or if?) to customize the passive mix of the world's primary betas. As discussed frequently on these pages, the record with a passive allocation to everything is competitive if not impressive vs. a wide array of strategies and funds intent on beating Mr. Market's portfolio (see my analysis here, for instance). With the addition of foreign REITs and junk bonds, the competitive aspect of GMI will likely rise a bit, particularly for the rebalanced and equal-weight versions. History shows rather convincingly that rebalancing a broad set of betas lays a solid foundation for earning a respectable risk premium through the years. More is better, up to a point, when it comes to rebalancing. Assuming, of course, that "more" is defined intelligently.
For my money, defining the world's major asset classes is a powerful starting point for surveying the possibilities. The fact that passive asset allocation tends to deliver average to above-average results only sweetens the deal. Adding two more betas—betas that deserve a spot at the asset allocation table--will likely add a bit more sugar to Mr. Market's cocktail.
Posted by jp at January 23, 2013 9:23 AM