DOW 36,000 Redux

Enrique Diaz-Alvarez enrique at
Thu Aug 19 16:48:48 PDT 1999

Hi y'all,

Another two candidates for Doug's Economist Charlatans Hall of Fame. Glassman and Hasset, of DOW 36,000 fame, have a longish article in this month's Atlantic. First of all, you'll be relieved to know that somebody pointed out to them the difference between dividends and reported earnings, so they tone down their predictions a bit: down to 27,000. However, old habits die hard, and they still say (misteriously) that perhaps one should use reported earnings after all in comparing dividends to treasuries. This conveniently ignores the extremely low quality of current reported earnings, perhaps the lowest ever, but never mind that.

Anyway, this is their argument, in a nutshell: A financial instrument that pays X% is equivalent to another that pays Y% but increases this payout by Z% a year as long as X=Y+Z. This, of course, assumes that one holds on to both for all eternity, but never mind that. Now, last year T-bill yielded 5.5%. They adjust this (incredibly) by expected inflation _over the next 10 years_ of 2.6% (according to the CBO) to get a real yield of 2.9%, instead of the actual reported inflation of 1.6%. Since 1946, real dividend growth has been 2.3%. The last 50 years have been an extraordinarily favorable period for US capitalism, but never mind that. This means that, assuming no risk premium is accorded stocks, the proper dividend yield for stocks should be 0.6%. Since the DOW index currently pays out 1.0%, this would imply that the proper value for the DOW is around 18,000. Somehow they fudge the number to 36,000, by astonishingly assigning the "dividend yield reported by the Fed of 2%" (??!!) to the DOW back when it was at 9,000 (it was actually 1.2%), and adding a tenth of a point or two here and there - hey, this is Anumerica, who's counting - but never mind that.

Of course, if we use the actual real yield for long term T-bills for 1998, 3.9%, we get that the fair vaule for the DOW is 7,000. If we use the _current_ values, 6.0% T-bills and 2% inflation, 6,500.

But that's just gravy. As I see it, the real problem with this calculation is the abnormally low real rate of return assigned to long term T-bills, a result of last year's flight to safety. According to Vanguard Funds, the average real rate of return for US government bonds since 1926 has been 5.1%, not 2.9%. Of course, this includes capital gains and losses besides dividens, but since interest rates are not much different today than in 1926, the _average_ real yield should come close (correct me if I am wrong). Now, using 5.1%, a 2.3% real rate growth rate for dividends we get an equivalent dividend yield of 2.8%. This means that fair value for the DOW is 3,900.

Any thoughts?


More information about the lbo-talk mailing list