Saturday, March 07, 2009

How Jim Glassman Got Skull-fucked By Angry Bear: A Tragiocomedy In 2 Parts.


Dean: I have a question for you: 80-plus years of meticulously documented stock market history (in fact, 200 yrs if you look at Jeremy Seigel's work) show: 1) large-cap stocks, on average return four percentage points more annually than T bonds, and 2) based on accepted measures of volatility, stocks over long periods (10-20 yrs) are not significantly riskier than bonds. This leads to two conclusions: a) investors should buy stocks for the long run (i.e., for retirement), and b) if the historic risk-return mismatch between stocks and bonds is reconciled -- that is, if the so-called "equity premium puzzle" is resolved -- stocks will be a great deal higher, say, 36,000 or maybe 18,000 or 40,000. You pick the number. The question is, What is wrong with this analysis, which was the guts of "Dow 36,000"? Forget the snide remarks about journalism. Just answer the question, which, to me, comes down to this: Is financial history a poor basis for making judgments about the investment future? Or is there some other reliable basis for making investment decisions? These are serious questions, not the stuff of cutesy-pie nasty remarks. All the best, Jim

Posted by: James Glassman


All right, let's treat it as a serious question.

From ca. 1967 to ca. 1984, the Dow "appreciates" from around 874 to around 875. (Op. cit. Warren Buffett.)

Let's treat that as the equivalent of the Swiss Franc ca. 1971--an undervalued asset.

We move to floating FX rates under Nixon, and by 1974 Galbraith's daughter (iirc) is saying she wants to be paid in SFR/CHF {pick your abbreviation of choice).

From 1998 to 1999 alone, the DJIA appreciates about 18% (using the monthly numbers), while the lowest investment-grade (BAA) Corporate Bonds are yielding ca. 7.25-7.50%, and 10-year Treasuries are running 5.5% at the beginning of the year and close to 4.5% at the end.

So even if we're assuming a ca. 4% standard Equity Premium--ignoring transaction costs, as well as some regulatory restrictions on USTs that may distort a lot of the data, as per Mehra and Prescott, Epstein and Zin, et seq.--we're looking at a situation in the late 1990s that is closer to the 1971 USD than the 1971 CHF.

Not, in short, a situation where "three to five years" is likely to produce a near-trebling of the Dow. (By your own Equity Premium, that would require Treasury rates running ca. 50% in the longest case [5-year] scenario.)

If the 4% EP holds--not the way I would bet, but your mileage clearly varies--and the U.S. grows at 3% with no inflation, we would see 36,000 from the current levels around March of 2034.

Those of us who describe the Equity Premium Puzzle as the Phillips Curve of Neoclassical Economists will take the Over, thank you, recalling Robert Heinlein's old dictum: TANSTAAFL.

No comments: