April 20, 2011
The Power & Peril Of Rearview Mirrors
Investors shouldn't be slaves to market history, but neither can they afford to ignore the past. The thought comes to mind as I look through the new edition of the Ibbotson SBBI Classic Yearbook.
You can't see the future by looking in a rearview mirror, of course, but it's an exercise that's still essential. Historical context is a valuable ingredient for intelligently estimating risk premiums, but it's merely a starting point for making informed guesses about what's coming. The good news is that the Ibbotson SBBI Classic Yearbook makes this task a lot easier by presenting a large chunk of the key numbers clearly within a single volume.
You could spend the better part of a week poring over the statistical information in this book, although the data point that’s quoted the most is SBBI’s annualized total return of large-cap U.S. stocks (S&P 500) since 1926: 9.9% through the end of 2010. But as the book reminds, that return can and does vary dramatically over shorter periods.
For instance, in the decade that just passed (2001-2010), the annualized performance of U.S. equities was a spare 1.4%, a mere shadow of the long-run return for stocks. The comparison is even worse when you consider that the previous decade of the 1990s generated a stellar 18.2% annualized performance for U.S. equities.
Stocks may be winners in the long run, as Professor Jeremy Siegel argues in his best-selling book, but there's no guarantee that you'll receive a windfall (or even a positive return) from equities in shorter periods. And for those who think a decade is sufficiently long for defining "long term," the SBBI Yearbook is worth its price if only as a tool for outlook adjustment.
Meantime, recognizing the potential for trouble with an all-equity strategy encourages using asset allocation as a risk management tool. Here, too, the SBBI Yearbook offers some perspective with a simple benchmark that compares various combinations of U.S. stock/bond portfolios through time. As the table below shows, holding bonds and stocks delivered higher returns in the past decade compared with an equities-only strategy. A 50/50 mix, for instance, earned 4.7% a year, well above the pure stock portfolio’s 1.4%.
One of the key questions with asset allocation is estimating how the risk-return tradeoff will play out going forward based on varoius assumptions. Once again, history offers a clue, but it's not necessarily fate for the period ahead. Even so, it’s a useful place to start the analysis.
Using Ibbotson data, consider how portfolio volatility (annualized standard deviation of return) changes with performance, as illustrated in the chart below. The basic message: adding bonds to a portfolio of stock lowers volatility, or so the track record since 1926 shows. But in contrast to the last 10 years, which offered the unusual profile of lower volatility and higher return by holding bonds and stocks, the past eight decades suggest that lowering volatility also pares return. The 2001-2010 experience, it seems, was the exception. That inspires caution for assuming that adding bonds to a stock portfolio will always lower volatility and increase return.
Another variable to consider is rebalancing. Based on the Ibbotson data, rebalancing a stock/bond portfolio reduces volatility but at a cost of reducing return, albeit in varying degrees. For example, in the chart above, a rebalanced 50/50 mix of stocks and bonds earned 8.2% a year since 1926 with a volatility of 11.4. By comparison, an unrebalanced 50/50 mix (i.e., a portfolio that was evenly divided but then allowed to drift with the market) earned a modestly higher 9.0% a year, albeit at a slightly higher risk of 15.9.
Was the higher return worthwhile for the extra risk? There are no easy answers for the simple reason that each investor's asset allocation should be customized to satisfy a range of factors, including age, risk tolerance, net worth, investment horizon, etc.
Let's remember too that the rebalancing history in the SBBI Yearbook is merely a benchmark, and a simple one at that. Rebalancing is a critical factor for designing and managing investment portfolios, but its effects vary considerably, depending on the asset classes, the time period and the frequency of the rebalancing.
In fact, it’s best to take history with a grain of salt with rebalancing studies. It’s important to study the past and understand why and how it delivered a particular risk-return profile, but the real insight comes primarily from studying the individual asset classes. But even this productive road only brings you so far. The danger is thinking that the future will deliver something comparable. There’s a case for extrapolating the past as a forecast of the future, but only as a prediction for the very long run using broadly defined asset classes and asset allocation strategies. And even that prediction is questionable, although less so relative to the short run. Indeed, periods as long as 10 to 20 years can be surprising, as the last decade painfully reminds.
The only intelligent application of historical market data is using it as a foundation on which to build estimates of risk and return with a variety of tools. Indeed, looking at long-run summaries of the past imply that returns and risk are static. In reality, expected return and risk jump around, sometimes dramatically over short periods. The central issue in asset allocation is getting a handle on how those jumps will play out for the foreseeable future.
The good news is that there are lots of research and financial models to help develop relatively robust forecasts. The bad news is that history, intriguing as it is, is of limited value for looking ahead. That doesn’t mean we can ignore history. But those who depend heavily on history as an excuse for not doing the hard work of projecting risk premiums by using additional tools are probably destined for disaster.
Posted by jp at April 20, 2011 9:59 AM
A great question; an important question. The answer, of course, is complicated and far beyond the scope of this particular post. For now, let's just say that those variables are critical factors for real-world money management.
Posted by: JP at April 21, 2011 8:25 AM
The purpose of managing and earning returns is to eventually use the money for some purpose. This historical work assumes reinvestment of all income. How are portfolio returns impacted if we assume a 3% annual spending policy? How is that impacted by rebalancing or not rebalancing?
Posted by: david at April 20, 2011 11:36 AM