Tuesday, 25 September 2012

The Truth About Stocks for the Long Run

By Alex Dumortier, CFA September 19, 2012

Don't get me wrong. I'm convinced that equities are an appropriate and important component of a long-term strategy to build wealth. However, there are a certain number of popular myths regarding stocks that have taken hold and that can be potentially dangerous to your financial well-being. In this article, I'm highlighting two of them.
Myth 1: Stocks' expected return is 10% to 11% annually
I have seen financial writers and professional investors cite this range (or a figure contained in this range) for the historical average return for stocks innumerable times. That's fine, in principle; the trouble is that they often imply or even assert openly that this is a sound basis for future expected returns.
As far as I can tell, the source for this range may be Long-Run Stock Returns: Participating in the Real Economy (2003), in which Roger Ibbotson of Yale University and his co-author Peng Chen calculate that stocks produced an average return of 10.70% during the period 1926-2000. That's a historical observation (anomaly?) and investors should absolutely not anchor on it when they think about future returns.
That specific period was atypical in a way that can be dangerously misleading if you don't look beyond the "headline" number. Stocks started out at depressed multiples (a price-to-trailing-earnings multiple of 10.2 for the S&P Composite Index) and finished at inflated ones (26.0 for the S&P 500 Index (INDEX: ^GSPC  ) ). All told, the expansion in the multiple's girth alone fattened stocks' average return by a full 1.25 percentage points annually. Unless you have reason to believe that rising valuations will make the same contribution over your investing horizon, expecting the same average return going forward is wrongheaded.
A more realistic benchmark
Taking a longer observation period (January 1871 to August 2012) over which the change in P/E multiple was less dramatic, I found an average return of 8.61%, with the change in P/E contributing just 25 basis points, and inflation 208 basis points (100 basis points being equal to one percentage point).
8.61% - 0.25% - 2.08% = 6.27%
A reasonable historical benchmark with which to begin thinking about future returns is 6% to 7% after inflation. That range is consistent with the long-run stock return estimates in the 2007 edition of Jeremy Siegel's Stocks for the Long Run.
Incidentally, if you don't think a 1.25 percentage point difference is even worth trifling over, consider the end value of a dollar invested at 6.5% over a 30-year period: $6.61. At 7.75%, you'll have $9.38. If you were counting on the higher return and earned the lower one instead, you're now facing a 30% shortfall.
Myth 2: The longer the time horizon, the safer stocks become
This is an idea that has been heavily marketed based on Jeremy Siegel's observation that, historically, the standard deviation of stocks' average returns has fallen as you extend your time horizon. But Siegel himself is pretty cagey when it comes to the broader implications of that finding. This is what he told an audience of financial advisors in 2004:
One thing I should make very clear: I never said that that means stocks are safer in the long run. We know the standard deviation of [stocks'] average [annual return] goes down when you have more periods ... What I pointed out here is that the standard deviation for stocks goes down twice as fast as random walk theory would predict. In other words, they are relatively safer in the long run than random walk theory would predict. Doesn't mean they're safe. [emphasis added]
In a March 2011 paper, Lubos Pastor and Robert Stambaugh, respectively of the University of Chicago and the University of Pennsylvania, show that stocks are more, not less, volatile over long periods.
Siegel compiled historical return data going back over two centuries, and that is fine as far as describing how stocks behaved in the past (strictly speaking, there are problems with this data, to begin with). Pastor and Stambaugh's argument is that observing historical average returns and extrapolating them into the future leaves investors open to "estimation risk." In short, today's investors don't care what stocks did in the past; the only thing that counts is what stocks do in the future, and even two centuries of data does not allow us to know stocks' expected return with certainty. Once you take that uncertainty into account, Pastor and Stambaugh found that stocks are riskier over longer periods.
4 practical recommendations for long-term investors
Don't cling to investing myths such as the two I have highlighted above. Accepting that they are false means recognizing the seas you must navigate are more uncertain and less hospitable than you once thought. It does not mean that throwing up your hands is all that is left to do. Here are four practical recommendations:
  • Time horizon is not the only relevant variable in figuring out your allocation between stocks and bonds (and other asset classes). Your risk tolerance, current earnings, and career risk are all things you should consider.
  • Use conservative estimates for the equity returns you require in order to achieve your goals to account for "estimation risk."
  • Always remain cognizant of valuations -- particularly when they reach extremes (i.e., bubbles).
  • Always strive to keep your costs as low as possible; this makes an enormous differenceover time. In that regard, products like the Vanguard Total Market ETF (NYSE: VTI  ) , the Vanguard Dividend Appreciation ETF (NYSE: VIG  ) , and the Vanguard MSCI Emerging Markets ETF (NYSE: VWO  ) are all excellent choices.

No comments: