Monday, February 26, 2007
I usually watch the Consuelo Mack show; it is on Sunday at 1pm where I live. The topic this week was the importance of dividends with Ben Stein, Robert Kessler and David Winters.
Kessler and Stein got into a tiff over something said by Kessler that I too question or perhaps the result is a data anomaly.
Kessler said that over the last "20 odd" years the return of treasuries, specifically zero coupon bonds, has equaled stocks. I think Ben would still be telling Kessler he's wrong if Consuelo hadn't jumped in.
Kessler said one thing that was confusing for people new to zeros he said "over the last 20 years they have outperformed almost every asset class for only one reason; that they are consistently paying interest." Well zeros pay no interest they accrete higher in price yes but pay interest no.
At one point he says it works because over the last 20 or 25 years zeros have compounded at 5% or 6%. Well at no point in the 1980's (according to Yahoo Finance) did 30 year treasury yields go below 7%. He may be using yields as high as 14% for his work.
I am not positive what long term zeros were "yielding" 20 years ago but Yahoo has the yield of a regular 30 year treasury bond at 7.5% (about the low of the decade) in March 1987.
I did a calculation of a 20 year period ending 12/31/2006 and zeros bought 20 years prior compounding at 10% would have had a roughly 610% increase. The S&P 500 opened at 242.17 on January 1, 1987 and closed out 2006 at 1418 so the gain from that period in stocks works out to about 590% so he's right! Oh wait, no, dividends, dang!
According to BigCharts.com the yield for most of the time studied has been around 2% (you probably already know that) and while I did not take the time to calculate it through, 20 years of even 1.5% would put stocks over the top, so it seems, but I have not seen Kessler's data.
The thing that stands out about Kessler's comments is that he uses the 1980's as a starting point which had some of the highest yields (if not the highest) in the history of our country, certainly the last 100 years.
Later he said that stocks would, over the next 20 years, have to go up 300% to outperform zeros. Well over the last 20, which takes in a crash, cutting in half and many crises, the market went up 590%. Twenty years before that (so opening day 1967) the SPX opened at 80.38 and went up 305%-ish to the 242.17 and this period included a big stretch of "going nowhere" and yes we still need to add the dividend in. Twenty years before that (so opening day 1947) the SPX was at 15.30 so the the following 20 years the market was up better than 500%, uh plus the dividend. The numbers here are from the Stock Trader's Almanac.
The lowest closing price I can find for the S&P 500 after the tech wreck was on March 11, 2003 at 800.73. Twenty years earlier the S&P closed at 151.24 which is about a 520% gain, plus dividends.
So betting against 300% in 20 years seems like a bet I would not want to make.
Clearly Kessler has data to back this up but looking forward I don't see how it can help unless you really want to take the under on the US stock market and think 5% compounded will beat stocks. I have written before that I think we may be in for a period of below average returns which I think might be in the 7-8% range. If this turns out to be correct I think adding a couple of hundred basis points is possible with foreign investing.
If Kessler's model is built around double digit rates, well how useful is that today? It will be useful if rates go back that high to be sure.