|
|
Limiting Condition
The following prediction may not last out the interval between deadline and publication; but, no guts, no medals: long term rates, including mortgage rates, will fall no further.
Oh, they might, of course; but only in exceedingly unlikely circumstances, and such painful ones that no one should look forward to the event.
I pretend no magical ability to predict the future. However, the U.S. Treasury has given all of us a new tool with which to evaluate interest rates: the first American bond to be indexed to inflation.
This CPI-indexed bond provides an answer to one of the thorniest market problems, previously theoretical and unanswerable: "Of any given market interest rate, what portion is the 'real', non-inflation portion expected by investors?"
The real interest rate has for a long time been assumed to be around 4.00% -- "around" meaning somewhere between 2.00% and 5.00%. If we knew the real rate with precision, forecasting would be arithmetic, not Greenspan-watching. If we expect 3.00% inflation for next year, and we know that investors expect a 4.00% real return on their money, long term market rates should be 7.00%. If the real rate is lost in the haze between 2.00% and 5.00%, it has no predictive value. Better to spend time guessing Mr. Greenspan's mood.
But now, the new age, and precision. In the last three years, the Treasury has sold three issues of 10-year, CPI-indexed notes. The rates they pay are fixed, and the principal amount of the note is indexed: if inflation rises by .2% one month, so will the amount of money the Treasury has to pay you at maturity. The Treasury picked a fixed coupon rate ("coupon" is slang for the nominal rate on any bond; a remnant of physical bonds bearing coupons which owners clipped and cashed) in the range of theoretical real rates: the first issue pays 3.625%, the second 3.375%, and the most recent again at 3.625%.
Since issued, the market value of the two higher-coupon notes has never fallen below 99% of indexed value, and the lower coupon never below 97%. As all three trade at slight discounts from face value, the real yield-to-maturity for all three issues has hovered slightly above the coupon rates at about 3.75%.
That's it. End of debate. No more real rate metaphysics. For ten-year United States Treasury notes, investors expect a 3.75% real rate of return.
Now some forecasting. Start by predicting the present: if the real ten-year rate is 3.75%, and the inflation forecast is 2.00%, ordinary ten-year T-notes ought to pay about 5.75%. Miracle of miracles, this week they pay 5.70%.
Everybody knows that if forecasts for inflation rise, so will interest rates. But now we have a "real" rate basement to which to add inflation, and have a much better chance to identify too-high or too-low market rates. If a bond pays "too high", buy it; if "too low", don't buy it.
The "real" rate basement at 3.75% provides more help than estimating bond value: it also sets a limiting condition for how low rates are likely to go -- ever. Only in times of outright deflation -- sustained negative CPI -- would Treasury market rates fall below 3.75%.
Long before market rates can get to the real rate basement (last seen in the 1930's), resistance to further rate decline will become enormous. The closer rates get to the floor, the decreased present return increasingly fails to compensate for future inflation risk.
The limiting condition comparison goes like this: The investor asks, with increasing skepticism, should I buy a ten-year note paying 3.75%, completely, absolutely compensated for any future inflation, dollar for dollar? Or should I buy an ordinary T-note yielding 5.70%, and bear all the risk of future inflation during the next ten years?
Resistance, right there.
|
