As discussed in my earlier post, the yield on TIPS are being widely misinterpreted as signaling long-term deflation. Last week Greg Mankiw pointed out further data problems with TIPS and specifically with the Treasury’s constant maturity series. Part of the challenge in analyzing TIPS has to do with the imbedded option in their price. The maturity value of these securities is adjusted by an index ratio tied to the CPI; on issuance date, that ratio is 1.00. However, at maturity the investor can never get back less than face value. This means that there is much greater principal risk for an investor buying a seasoned bond with a high index ratio (e.g., a ratio of 1.20 could go to 1.00, with an associated loss of principal as the maturity value would go from 1.20 times face to 1.00 times face), while a new issue bond offers less risk (e.g., even if the ratio goes from 1.00 to 0.80, the investor still gets back face value, not face times 0.80).
This pricing complexity no doubt contributes to investor preference for conventional treasuries. Also related is the superior liquidity of conventional Treasury securities — at current low interest rates, liquidity and simplicity premiums play a larger role in pricing (e.g. a 30 basis point preference is a small component in a 5% yield environment, but a large one in a 2% yield environment).
The most compelling reason to distrust any deflation interpretation of TIPS data are those stated in my earlier analysis — the recognition that over the very short term TIPS, which lag CPI by three months, must incorporate the decline in commodity prices, particularly oil. My friend Lincoln Anderson was kind enough to send me this chart which shows the close correlation between short term moves in oil and CPI:
Current TIPS prices are reflecting the past drop in commodity prices, not future deflationary trends.