January 9, 2012
Equal Weighting: A Timely Antidote For Uncertainty?
Are investors irrational or just cautious when it comes to the fundamental stock/bond asset allocation decision? It’s a topical question in light of MarketWatch’s Jonathan Burton article that reviews the case for “Why stocks will beat bonds over the next 20 years." The basic outline is by now familiar: bonds have done well in recent years while stocks have performed poorly. For example, over the past five years, the stock market (Russell 3000) is flat while bonds (Barclays Aggregate Bond) are up an annualized 6.5% through the end of 2011 (for a summary of recent returns, see my latest update of the major asset classes). Assuming a dose of mean reversion, the outlook for equities is favorable and prospective returns for bonds look relatively mediocre if not negative.
“Using history as their guide, and weighing current stock valuations and interest rates, various investment pros believe stocks have a much better chance than bonds to beat inflation in the long run,” Burton reports. He cites several analysts who forecast equity returns ranging from 5% to 9% for the long run. Meanwhile, the benchmark 10-year Treasury yield is roughly 2%, which doubles as the expected return.
In short, the gap in favor of equities looks compelling. Meantime, history suggests that rebalancing a mix of broadly defined asset classes in favor of the laggard and cutting back on the leader tends to boost portfolio returns. By contrast, allowing the winner to dominate the asset allocation has a habit of delivering meager results. This is all familiar terrain, of course, as I discuss at length in my book, Dynamic Asset Allocation. The implication in the current climate: investors should dump bonds and load up on stocks… NOW!
If that is in fact the rational thing to do, it follows that those who don’t jump on the equity bandwagon are irrational. Maybe, but since the future’s uncertain it’s hard to say for sure if an equity-heavy allocation will be optimal. But doesn’t history tell us that it’s always wise to buy stocks when they’ve underperformed bonds? Yes, but history isn’t guaranteed to repeat.
Nonetheless, I think that a healthy dose of equities is compelling for investors with long-term horizons. But there’s always room for doubt. On that note, consider a recent study by professors Robert Weigand (Washburn University School of Business) and Robert Irons (Dominican University): “The Relative Valuation of US Equities at Bear Market Bottoms: A Perspective on the Equity Risk Premium." Part of the paper documents a widely recognized relationship between equity returns and market values. That is, buy cheap and sell dear. But the authors also find that the crowd’s habits evolve in pricing assets, and not necessarily for the better in terms of anticipating stock market returns these days.
After studying stock returns, earnings, interest rates and relative valuation in the U.S. in the context of bear markets and subsequent trends, Weigand and Irons discover that “bear market bottoms since 1950 have been associated with shorter bear markets, lower average market earnings yields and slower real earnings growth following the market bottom, but higher real stock returns over the next decade.” In other words, the market’s been pricing equities at higher levels and this increase in valuing corporations has been advancing at a faster pace than the underlying growth rate of earnings—a trend that’s been unfolding for decades. As for the implications, Wiegand and Irons advise:
Equity values growing faster than earnings for an extended period of time means that the stock return/earnings relation is significantly different pre- and post-1950. Before 1950, the market earnings yield predictably reverted to the mean (often overshooting at bear bottoms), and future equity returns were related to expansion and contraction of this key ratio. Post-1950, we find that the stock return/earnings relation becomes strained as stock prices grow faster than the long-term trend in earnings, eventually resulting in stock prices and earnings losing the cointegrating relation that drove mean reversion in the ratio. In the period following World War II, stock returns have become gradually disconnected from earnings to the point that the earnings yield is no longer reliably mean-reverting, and thus no longer predictive of future equity returns. US stocks' earnings yield and trailing earnings growth are unrelated to future long-term returns post-1950.
The bottom line:
Despite all the anxiety about the "lost decade" in stocks, US equity values in 2011 remain inflated compared with the fundamental earnings US companies will most likely be capable of producing, and long-term future expected returns remain low. Although we estimate the real equity risk premium to be only 0.5% below its post-1950 average, in a low inflation, low bond yield environment, our forecast is that US equities are priced to deliver real returns of approximately 3.5% per year for the coming decade.
None of this convinces me that equities should be shunned. But the paper suggests that the jury’s still out on deciding if bonds are in a bubble, or if stocks will be a slam-dunk winner in the years ahead. So, what’s an investor to do? If you have a high degree of confidence that stocks are indeed a great buy, or that bonds are priced for disaster, well, you’ve already made your decision. But if you’re not quite sure who’s right, it may be time to consider the model-free antidote to uncertainty: equal weighting.
As I wrote recently for Financial Advisor, “equal weighting looks good on paper and it performs handsomely in practice.” Why? It’s arguably the ultimate risk management technique, or at least after adjusting for the required effort/analysis for implementation. First, by diversifying equally across assets without making a judgment about prospective returns, you’re hedging your bets in the extreme. As such, you’re pre-emptively giving up big returns but you’re also minimizing the risk of suffering big losses. Second, maintaining the equal weighting insures that you’ll capture the mean-reversion effect, assuming it rolls on, via rebalancing. If mean reversion weakens, or evaporates entirely, the need for rebalancing will be muted, of course, in which case no harm done since you won't be trading much because relative weights are somewhat static.
An equally weighted portfolio of all the major asset classes returned roughly 4.3% a year over the last five years, by my calculation. That’s hardly spectacular, although it’s probably competitive in the grand scheme of clever managers trying to outsmart the markets. Naïve strategies that deliver decent returns sounds counterintuitive, but this is finance and so the risk of trying too hard should be considered.
In any case, here’s a prediction: Five years from now, an equal weighting of a broad mix of asset classes will look pretty good, just as it does now. This too requires a bit of faith, as does every other investment strategy. The difference is that equal weighting demands a substantially lower leap of faith than most if not all of the competing strategies. That’s no small selling point these days.
Posted by jp at January 9, 2012 6:34 AM
I think i agree with you...that equal weighting should be a starting point for asset allocation, which should be amended for future expectations based on one's metric(s) of choice...
Posted by: jeremy at January 11, 2012 12:00 PM
Jonathan, you can find my list of the major asset classes on these pages at the start of each month with an update on returns, including various proprietary indices. Here's the latest:
Posted by: JP at January 10, 2012 7:10 AM
"An equally weighted portfolio of all the major asset classes returned roughly 4.3% a year over the last five years, by my calculation."
What do you include in the "major asset classes", specifically?
Posted by: Jonathan at January 10, 2012 1:19 AM