Some Thoughts On Sizing Up Asset Allocation Funds

Morningstar’s Greg Carlson writes that “allocation funds are different beasts than most of the funds we rate.” In my view, the main challenge in this corner is finding a benchmark for analyzing portfolios that routinely hold multiple asset classes. If we’re talking about a single-asset-class strategy—large-cap US stocks, for instance—there’s usually an assortment of reference indexes to consider for crunching the numbers on risk and return. By comparison, asset allocation portfolios–whether home grown or in prepackaged in a mutual fund or ETF–are complicated for one key reason: there’s no obvious benchmark.

The Morningstar article inspires some comments about the task of dissecting multi-asset class portfolios and deciding how a given strategy stacks up to the competition. There are pros and cons to farming out asset allocation and rebalancing decisions to third parties, but let’s put that aside for another day. Instead, how should we think about analyzing asset allocation strategies in broad terms? The answer is complicated and worthy of a book… or five. I took a stab at the topic in Dynamic Asset Allocation, although there’s much more to say when it comes to the detailed work of dissecting any given fund or portfolio. I’m not going to go off the deep end here, but a couple of thoughts come to mind in matters of evaluating multi-asset class strategies.

I usually start with a quantitative review by comparing track records for performance and risk relative to some index. And there’s the rub: finding a relevant benchmark can be problematic. Asset allocation choices, after all, are usually the byproduct of subjective decisions, ideally for all the right reasons. In contrast with single-asset-class strategies, “passive” or equal weighting for a mix of asset classes is a rare bird if you’re looking for off-the-shelf indexes. This is largely due to the fact that the decision process for choosing asset classes, portfolio weights, and rebalancing rules is a reflection of specific objectives, expectations, etc. Accordingly, you won’t find a generic benchmark that’s perfectly suited to every investor’s strategy in such an amorphous realm.

As a starting point, however, it’s advisable to consider a passive strategy, even if it’s not directly relevant. The first issue with analyzing asset allocation decisions is developing perspective that minimizes subjective preferences and biases. Easier said than done, but we should try. My effort on this front is the Global Market Index (GMI), a proprietary benchmark that’s a regularly updated on these pages (see here and here for recent examples). Yes, there are some issues, so to speak, when it comes to building a genuinely passive benchmark. Nonetheless, GMI offers a reasonable approximation for measuring the optimal portfolio for the average investor with an infinite time horizon.

It’s no surprise that GMI’s record is encouraging in the sense that it delivers competitive results at a rock-bottom price (assuming we populate it with low-cost index funds). Indeed, GMI is designed to capture what you might call the global risk premium with minimal fuss and expense. The underlying strategy is simply holding a market-value weighted mix of all the major asset classes and letting the allocations drift with the tide. There are no rebalancing decisions because the weights are subject only to the market’s ebb and flow. It’s a set-it-and-forget-it strategy that holds everything in terms of relative market-value weights. As such, GMI is an obvious starting point for analyzing asset allocation funds. (As a side note, some quasi-passive asset allocation funds are well-suited to serve as benchmarks. One example is the Vanguard Star Fund, as I discussed last year.)

There are numerous ways to proceed when it comes to reviewing a specific fund or portfolio, but I prefer to begin with a returns-based regression analysis. You can learn quite a lot by running a factor-analysis test. That is, by performing a multiple regression on a fund against the various benchmarks that comprise GMI we can see which risk factors are driving performance, or not. It’s not unusual to discover that impressive returns are a byproduct of making relatively large or small bets on a handful of asset classes. That may or may not be productive, but it’s essential to recognize where risk and return is coming from in some detail.

As you might expect, quite a lot of what looks like brilliant investment management turns out to be basic decisions on overweighting and/or underweighting asset classes relative to a passive market-value-weighted mix. In turn, the question becomes: How much should you pay someone to allocate a portfolio across the major asset classes? Before you answer, consider a simple fact: It’s easy to replicate GMI for as little as 25 basis points, perhaps even less, depending on the funds you select. Should you pay someone 100 or 200 basis points in the hope of beating GMI? Maybe, although you’ll need to develop some confidence that the higher price tag is worthwhile.

A common mistake is assuming that a relatively high trailing return for an asset allocation fund is all the proof you need for thinking that the underlying strategy is the result of superior management. Sometimes that’s true, but it’s relatively rare, as the numbers suggest. Yes, there are genuinely talented managers in the field of allocating assets, and some of them are actually worth their fees. But locating this elite group requires a deeper dive than most folks realize. The reasoning starts with a basic calculus: it’s easy to achieve broad diversification across asset classes, but it’s not worth paying much for this service unless you’re sure that you’re getting something beyond closet indexing.