Arnold Kling  

My Understanding of TIPS

Depression Mania... Heretical Thoughts...

There has been a lot of commentary concerning Treasury Inflation-Protected Securities. One alleged anomaly is that the 10-year bond that expires on January 15 is reported to have a double-digit yield. How can that happen?

The overall return on a seasoned TIP consists of three things. First, there is the coupon on the security. Second, there is the appreciation or depreciation of price. Third, there is the change in the accrued inflation factor. The coupon on the 1/15/09 TIP is 3.875 percent, so interest accumulates at that annual rate between now and then. So you can expect about 0.5 percent in interest (taking the number of days between now and January 15th, dividing by 365, and multiplying by the coupon rate). The price this morning was roughly 98.35, and if you hold it until January 15th it will effectively go to 100. So, you have a 1.65 percent gain from that. Annualize that number, and there's your double-digit yield.

The catch concerns the accrued inflation factor. Because the bond has been outstanding for almost ten years, the accrued inflation factor as of December 1st is 1.33 and headed down, because of the decline in the Consumer Price Index in October, to 1.32 on December 31 or a little less than a one percent drop.

I am guessing that the inflation factor for January 1st through 15th depends on the November CPI. If the November CPI declines by, say, one percent, then my guess is that the inflation factor would decline by another 0.5 percent. If that happens, then by my calculation your yield will be 0.5 + 1.65 - 1.5 = 0.65, over a period of roughly 50 days. At an annual rate, it is something just over 4.5 percent. Nice for a short-term Treasury, but nothing like double digits.

On the other hand, if we get an enormous drop in the CPI (keep in mind that gas prices are certainly going to be lower), you could actually lose money. If I'm doing the arithmetic right, with a 2.3 percent drop in the CPI, the accrued inflation factor would go down another 1.15 percentage points, and your total yield would be zero.

The accrued inflation factor is never allowed to go below 1.0. That means that a brand new TIP offers protection against deflation over the life of the TIP. But there is no deflation protection whatsoever on a seasoned TIP. And my understanding is that the quoted yields on seasoned TIPS do not take into account even deflation that has already been built in to the accrued inflation factor.

That does not mean that there aren't some good buys out there in the TIPS market. But you do need to understand how the yield relates to the accrued inflation factor--which I don't necessarily understand correctly myself.

Do not make investment decisions based on anything I have written here. I am not just saying that as a formal legal disclaimer. I am saying that I genuinely do not have confidence in my understanding of the institutional details, and those details are critical.

Comments and Sharing

COMMENTS (2 to date)
John writes:

You might have used too many significant digits.
The CPI in October was 216.71, a 1% drop was 214.54. Divided by a base of 164 gives 1.31, not 1.32.

That would also mean YoY inflation would fall to 1.7%.

By my calculation, on Jan. 15th you will receive $131 of principal and $2.54 in interest. I don't know this market well enough to know exactly what the adjustment factors are. For instance, when you buy Treasury bonds, there are adjustment factors to take into account when you buy the instrument. That might be what is keeping the yield high.

Mike writes:

I too am a bit puzzled by TIPS. If there is anyone reading this that has a better understanding of the subtle nuances of TIPS your thoughts would be much appreciated.

In particular, economists use TIPS as a synthetic means of ascertaining inflation expectations. As we speak, people are making the observation that TIPS are yielding the same as comparable maturity Treasuries for say 5 year maturities which implies zero inflation expectation over the next 5 years. They then leap to the conclusion that deflation is near (and all that suggests vis the U.S. in the 1930s and the Japanese 1990s)!!

If TIPS have what Arnold is observing then there may be a subtle bias in the design of TIPS that may not allow us to draw that conclusion. This seems to be a very important issue for investment purposes as well as for determining the expectation of deflation.

Comments for this entry have been closed
Return to top