The Fed and the financial media have touted the $412+ billion TIPS (Treasury Inflation Protected Securities) as providing a guaranteed real yield. (See here, here and here (pdf warnings on last one).) Academics and professional investors also believe that market inflation expectations can be derived by subtracting the TIPS yield from a like term nominal bond. However, taxes falsify the first premise for taxed investors, and have serious implications for the second.
The principal invested in a TIPS is grossed up for the actual consumer price index (CPI) semiannually, and interest is paid on the total at the coupon rate. Federal current income or deferred taxes apply. Clearly the TIPS structure makes the investor whole for CPI and provides a real yield before, but not after taxes. While the idea of an after-tax real yield is not new (Darby-Feldstein, 1976), it causes TIPS to behave in ways that have not been fully articulated by academics, professional investors or the Fed; and explains yield “anomalies” such as why the TIPS yield is higher than expected and correlates with the yield on nominal Treasuries of similar yield to maturity. (graph)
Consider the 10/29/07 10 year TIPS yield of 2.04% and the 10 year nominal Treasury bond yield of 4.38%. The profession thinks that the market expected average rate of inflation is (4.38% - 2.04%) = 2.34%. The expected after-tax real yield for the TIPS is [(1- 35% assumed top tax rate) x (2.04% coupon + 2.34% expected inflation)] – 2.34% expected inflation = .5%; the same as for the nominal bond: (1- .35) x 4.38% - 2.34% = .5%.
What happens if actual inflation averages 5%? The math shows real after-tax TIPS annual yield of -.4% and -2.15% for the nominal Treasury at 35% tax rates. The TIPS does mitigate the after-tax impact of inflation; but does not remove it. In fact, the only way to guarantee a real bond yield is to exempt it from tax and offer the inflation gross-up. Then the yield will be real. The current TIPS yield is unquestionably nominal based on two risky sets of expectations: effective tax rate and inflation. This fact also means that the price or principal value of TIPS are at risk merely due to changes in inflation expectations – contrary to all media messages and Fed and Treasury literature that I have seen.
If the TIPS yield is real, why would it correlate (move up and down) with the nominal bond yield? This would only happen if the real yield required by the market was changing. It should not happen if only inflation expectations were changing. However, as shown above, the TIPS yield is nominal, just like the nominal bond yield; so a yield correlation is predicted. Another interesting possibility is that the real yield was in fact too low against its long-term mean since the crash of 2001 and has begun an upward trend – this too would explain a correlation in yields.
Now consider expected inflation and effective tax rate assumptions. It is implicitly assumed that TIPS and nominal treasury investors have the same inflation expectations (this part the Fed is aware of and has questioned; see above links). What if they have different effective tax rates? If TIPS investors are more sophisticated, richer, and therefore face a higher expected tax rate; any attempt at deriving inflation from the yield spread fails unless these tax rates are known. Further consider emerging upside risks to the top tax bracket due to the coming presidential election outcome uncertainty.
For example, at a 35% expected effective TIPS investor tax rate with 1% required real after-tax yield and 2% expected inflation, the nominal TIPS yield would be 2.61%. The nominal bond yield with the same 2% inflation expectation but 25% average effective investor tax rate would be 4%. The spread is 1.39%; far LESS than the 2% mutually identical inflation expectation.
Is the Fed itself using these spreads to guide its interest rate policy? If it is, and since it isn’t considering tax rate differentials, it could be looking at grossly understated market inflation expectations and thus inappropriately cutting interest rates if it is concerned about inflation expectations. If it is not relying on the spread; then why not – since the claim is that it embodies market inflation expectations? If it is incorporating some estimate of tax rate differentials, then why did the Fed help the Treasury sell these bonds as a guaranteed real yield with the financial media swallowing the story?
Since it is widely thought that TIPS yields are real, the fluctuations in TIPS yields are considered proof of a varying real required rate of return– this is very suspect based on the above tax effect analysis. In fact, a very strong case can be made that the required real long term after-tax rate is constant – a “Fisher Effect”: see my Seeking Alpha article “Peak Gold – Whither, Whence and Why” and my co-authored refereed journal publications and working papers available on the Internet by “Googling” my name or “Required Yield Theory”.
In summary: TIPS are risky due to uncertainty in both effective tax rate and inflation which can result in negative real after-tax yields to investors. The yield spread with nominal like term bonds does not necessarily reflect market inflation expectations – only if tax rates and inflation expectations for both investors groups are the same; which is unlikely with regard to the former. Any policy reliance on yield spreads is misguided.