David Beutel

David Beutel
Contributor since: 2013
Company: Dynamic Allocators LLC
Thanks for your article. Excellent point supporting longer-term normalized PE ratios: "when a writedown occurs, we would spread it over prior periods, because prior accounting was too liberal."
Please provide more explanation/references for these longer-term valuation measures:
- Price-to-Resources
- Financial Stress indexes
- Eddy-Elfenbein’s Stock Market if valued like a bond measure
In assessing the degree to which the US stock market is overvalued, do you look at the implied forward equity risk premium of stocks versus bonds, or at any indicators measuring the attractiveness of stocks as compared with other asset classes? Many would argue that stocks may be overvalued relative to their own historic valuations but less so relative to the expected returns of other assets.
Thanks for comments, good points.
Regarding Bogle, I'd like to highlight his occasional statements sympathetic to valuation-based dynamic allocation--albeit in a much more constrained form than that practiced by Bill Fouse.
For example, late 1999 was another time when large US stocks were obviously overvalued, with the stock market trading at unprecedented valuation multiples. In this environment, Bogle appealed to “investors with conviction, courage and discipline,” suggesting that they "consider at least some modest leaning against the powering wind that is driving this great bull market” (Common Sense on Mutual Funds). Bogle imposed a +/-15% constraint on such shifts from baseline allocation (e.g. from a baseline of 60% stocks to as low as 45% stocks when overpriced, as in 1999, or as high as 75% stocks when cheap, as in 1982).
This is simply the conventional logic of rebalancing carried one step further. If a method that forces you to systematically buy low and sell high is good, then a method that forces you to systematically buy (more) low and sell (more) high is even better.
Anyone who followed Bogle's suggestion regarding valuation-based shifts in late 1999 is surely glad they did, even given his constraints and the impossibility of perfectly timing any such shifts. If they kept on by increasing stock exposure up to 15% during the subsequent 2000-02 bear market, they'd be even gladder--and gladder still if they repeated the process again by downshifting in 2007 and upshifting in 2008-09!
Thanks for comments Cranky.
BGI-TAA used just two assets (large US stocks and US bonds) from 1977-85, when cash was added as a third asset. I don't know too many details of their formulas (if anyone does, please comment!), but here's how I understand the method from what I've read:
--Fouse used a dividend discount model to estimate fair/intrinsic value of large US stocks in aggregate. From here, he calculated a expected return for stocks, probably the implied annual change to convert current price to fair value over a certain time period, probably about +/-10 years.
--he calculated an expected return for US bonds. Probably he used yield to maturity of 10y Treasuries, with some adjustments.
--Fouse used the estimated forward returns for stocks and bonds as inputs for a Mean-Variance Optimization (MVO) based on Harry's Markowitz's Modern Portfolio Theory. Note that Fouse did NOT use long-term historical stock and bond returns (like those from Ibbotson), as many naive users of MVO continue to do. (MVO also requires estimates of volatility for each asset and the inter-asset correlation of expected returns, but these have relatively small impacts on the result, especially if weighting constraints are introduced.)
--The output of MVO is an efficient frontier of asset combinations that maximize expected return along a spectrum of risk tolerance. BGI-TAA was designed as an alternative to a traditional 60-40 static portfolio, so presumably Fouse selected that asset mix along the optimized efficient frontier with the same volatility as the historical returns of a 60-40 mix (about 11% by my calculations).
--From 1977-1984, BGI-TAA followed a discipline of recalculating the efficient frontier once a month and shifting assets if the freshly-calculated optimal asset mix was sufficiently different from the portfolio's actual weights. By 1984, with the aid of computers, BGI-TAA began performing MVO calculations daily, with portfolio swifts implemented semimonthly if merited by a sufficiently large shift in market values.
Hope this reply has been useful!