All good things must come to an end, even the 20 year boom in the bond market.

If you had bought the longest 30-year “strip” of the then-new zero coupon Treasury bonds in 1982 (which pay no interest, but accrue principal to maturity) you would have bought an instrument that promised to pay you $100 at maturity in 2012 and you would have paid a price of around $3.80, which compounds to $100 at a yield of 11.5 percent (ordinary Treasury bond yields back then were over 13 percent, but the longest “strips” yielded less.)

Today, you could sell that now 9-year “strip” for $72.80, a 3.59 percent yield. You would have made 1,907 percent on your money, or 15.07 percent per annum, compared to a return on the Dow, bought at the low of that bear market year, of about 1,175 percent, or 12.44 percent per annum (though the Dow would also have paid you some dividends.) The bond return was entirely risk free, in nominal terms, in the long run — whatever happens, the “strip” will get you $100 in 2012.

As with stocks, there is no chance that this record will be repeated over the next twenty years, not even close. However, in the case of bonds, you can prove it mathematically. There are currently no 30-year “strips” because the U.S. Treasury gave up issuing 30-year bonds in 2001 (a decision I bet it now regrets, given the low prevailing yields and the size of current and future budget deficits.) However, a regular 28 year Treasury bond, the longest available, yields 4.45 percent, so one can postulate that a 28 year strip would sell on about a 5 percent yield basis, at a price of $25.51 per $100. Suppose in 20 years time, June 2023, money has become free (ignoring the inflationary effects of such an event!) Then you would get back $100 on that bond, which still had 8 years remaining and a 0 percent yield. The yield over the 20 years? 6.72 percent per annum, less than half the return of the 1982-2003 period. And that is, in every possible sense, a ceiling.

There are other reasons, too, to think that the 20 year bull market in bonds may be nearing a peak. Fed Chairman Alan Greenspan recently warned about deflation, and theoretically if we had a sharp deflation long term bond yields could fall arbitrarily far, since real yields would remain at an appropriate level. However, this consideration does not apply to Treasury Inflation-Indexed Securities (TIIS), the bonds issued by the U.S. Treasury since 1997 that pay interest and principal linked to the U.S. consumer price index. Yields on these securities should, subject to the vagaries of the CPI that I will discuss below, reflect fairly closely the yields on government bonds from the gold standard period before 1914, the last time investors were able to ignore inflation in their calculations (yields dropped somewhat, of course, at times of large new gold discoveries such as those in California (1849) and the Yukon (1896-97). The yield on British government Consols hit its all-time low of 2.21 percent in 1897, the year of Queen Victoria’s Diamond Jubilee.

Since TIIS were generally issued when their yields were much higher than they are today, mostly around 3.5 to 4 percent, almost all issues are trading at large premiums, giving rise to a tax effect for investors that distorts yield comparisons. The only fairly long term TIIS not trading at a huge premium, the 3 percent of July 2012 is today yielding 1.66 percent. Even the longest TIIS, the 3 3/8 percent due April 2032, trading at a 22 percent premium, is today yielding 2.28 percent, only a little above the all time record low yield on Consols. Thus, if the U.S. consumer piece index is not distorted, the real yield on U.S. government bonds is the lowest risk free rate the world has ever seen; money is exceptionally “easy” and it is extremely unlikely that rates will decline further.

However, the CPI, published by the Bureau of Labor Statistics, is not as bullet-proof as you may think. There are two major distortions in its construction, which prevent it from being a genuine index of the prices people pay. First, the index does not take account of house prices directly, but only of “housing costs”, reflecting either rents or monthly mortgage and tax payments. In a period such as the present, when interest rates are declining, house prices are rising quite rapidly, and rental charges are declining because of the low-interest-rate subsidy to home purchase, this results in the CPI housing sector price rise being considerably lower than the rise in an appropriate average of actual house prices. A true inflation measure would look at the price of homes, adjusted for size and facilities changes, but leave the financing decision, affected by interest rate levels, separate from the purchase decision. The choice is of course a political one; in 1980, when house prices were not rising rapidly but interest rates were, the CPI calculation was switched to a rental basis, ignoring home purchases completely and thereby again depressing the published CPI figure from the true level of inflation.

The other problem with the CPI calculation is the adjustment made by the BLS for quality improvements. In its 1996 study of the CPI, the Boskin Commission calculated that in 1995, the CPI would have risen by 3.9 percent without correction for quality improvements, whereas the actual price rise was 2.4 percent. However, the quality improvements correction is highly subjective, and in many cases does not reflect a true increase in functionality to the consumer. For example, a 2003 computer, the 2003 version of Microsoft Windows and the 2003 version of MS Word do very little for the consumer that is different from or superior to the 1995 versions of these items. The computer has more clock speed and more RAM, but it needs it to deal with the increasingly kludgy computer programs, modern versions of which will not run on perfectly adequate machines from 1995-97. Yet the BLS, through its “Hedonic model” has since 1998 every year made an adjustment for the increasing power and size of computers and software, even though the real change in functionality is minor.

The effect of these two factors since 2000 has been to suppress growth in the CPI. This has had two effects. First, it means that TIIS, whose return is based on the CPI, have a yield in true inflation-proof terms that is significantly lower even than the very low current published yields. Second, it has enabled proponents of easy money and fast money supply growth, notoriously including Alan Greenspan’s Federal Reserve, to raise the specter of deflation as an excuse to lower short term interest rates still further.

In fact, we are very far from deflation. Even on the published figures, the CPI increased by 2.1 percent in the 12 months to May 2003, or 1.6 percent excluding food and energy. The bond markets, too, are not expecting deflation; the 9 year fixed coupon Treasury note yields 3.34 percent, a premium of 1.68 percent over the yield on the TIIS of the same maturity, indicating that the market expects the CPI over the next ten years to increase at an average rate of 1.66 percent, well above zero. This is not surprising; M3 money supply has increased at an average annual rate of 6.6 percent in the last 12 months and 8.8 percent since November 2000, at a time when real GDP growth has been no more than 2 to 2.5 percent per annum. Since all that money has to go somewhere, the M3 growth suggests that, knocking out the average of 1.2 percent growth and 1.7 percent inflation over the period since November 2000, there’s a residual of close to 15 percent of inflation waiting to appear if and when money velocity gets back to its 2000 level.

It’s a pity, in a way, that we’re not getting back to some kind of mild deflation. The nineteenth century, in general, was deflationary and yet economic growth was satisfactory in both Britain and the U.S. throughout the century, with only moderate interruptions. Provided the rate of deflation does not exceed the rate of real GDP growth (as it did for a short period in the early 1930s) then deflation is no real threat to us, it simply means that consumers and workers get the benefit of economic growth mainly through falling prices rather than through rising wages. For certain people, particularly the elderly living on fixed incomes, that is greatly preferable — there is no moral or economic benefit in having a continual income redistribution from retired people and stable job-holders to the strongly unionized and the job-hopping young, which is what we’ve had with inflation since 1939.

As for the bond market, if real yields are at levels lower than has ever been seen, and inflation is likelier to rise than fall, then it must follow that the fixed coupon bond market is a disaster waiting to happen. The long bull market of 1982-2003 is over, and bond returns going forward will be minimal on short term bonds and negative on long term bonds. This in turn will cause a crisis in the housing market and, as the tsunami of home mortgage refinancing ceases, in the consumer debt market and the economy as a whole.

Sell! Everything! Maybe except a little gold.

-0-

(The Bear’s Lair is a weekly column that is intended to appear each Monday, an appropriately gloomy day of the week. Its rationale is that the proportion of “sell” recommendations put out by Wall Street houses remains far below that of “buy” recommendations. Accordingly, investors have an excess of positive information and very little negative information. The column thus takes the ursine view of life and the market, in the hope that it may be usefully different from what investors see elsewhere.)

This article originally appeared on United Press International.