As you may have noticed, in the past few weeks stocks are down, gold is down, oil is down… And Treasury yields are also down. (So much for my “bond crash”.) It sure smells like the market is pricing in a little deflation or disinflation or something. But how can we be sure? I mean, maybe oil demand is just down a bit, and gold is just being MAH-NIPPLELATED, and hordes of Russian teenagers are just front-running Goldman’s program trades from their parents’ basements.
Once upon a time, you could pull up nice charts of inflation expectations from the Cleveland Fed. These were derived from the TIPS spread, and the Cleveland Fed staff applied all sorts of corrections for liquidity and whatnot to derive accurate market expectations for future inflation. It was a wonderful site, but it went dark on October 31. (When a regional Fed bank is out stoking deflationary fears, my guess is that the home office is not amused. Just a guess.)
For about one month after that, you could take the TIPS yields published by Treasury and calculate your own spread. You might not know how to apply the liquidity corrections, but at least it was something. Then, on December 1, Treasury changed its methodology for calculating TIPS yields, in a way that just happens to bias the spread upward.
So, as interested observers, whatever are we to do? Is there any way left to get an accurate read on market-based inflation expectations?
Meet the zero-coupon inflation swap. The idea is very simple. You and I enter into a contract where, after N years, I will pay you “inflation” and you will pay me a fixed rate y. We exchange a single lump-sum payment after N years, but no payments along the way; hence the “zero coupon” in the name.
In this context, “inflation” means CPI, and “y” is the yield I demand from you in exchange for providing inflation protection. That is, after N years, you will pay me
(1 + y)N – 1
…and I will pay you
(CPIN / CPI0) – 1
…where CPI0 is the CPI today and CPIN is the CPI after N years. The “-1″ terms are because we are “swapping yield but not principal”. (Although in reality, we will net our payments; that is, whichever of us owes the other more will pay the difference. So the “-1″ terms are irrelevant because they cancel. But they always seem to show up in the term sheets for these gadgets anyway.)
As usual, these payments will be multiplied by the “notional value” of our swap, which we will state in our contract. Typically, “you” are some poor sod seeking inflation protection, and “I” am a commercial or investment bank.
For example, see ING’s fact sheet. (Note that they are missing a superscript on one of their N’s.) Unlike most references I found on this subject, this fact sheet makes it clear what happens if the CPI decreases; i.e., if we experience deflation. The answer is that you pay me on both legs.
Now, why do we care? Because by examining the market rate for inflation swaps (“y”, in the notation above), we obtain the pure, unadulterated market expectation for the rate of inflation for the next N years. Or to be precise, we obtain the market expectation for the rate of growth of the CPI. (The CPI is either the most accurate measure of inflation you will ever find, or it is a totally artificial figure produced by a deceitful government. Depends on whom you ask.)
Anyway, Bloomberg publishes these rates for 2-year inflation swaps, 5-year inflation swaps, and other maturities. For posterity, I have grabbed a screen shot of the 2-year summary and chart as of today (July 8, 2009):
I think it is fair to say that CPI expectations for the next two years are in the process of declining. Persistent rumors suggest that 2% is the inflation expectation the Fed would like people to have. (You gotta have faith!) I do not know exactly how they measure inflation expectations, but the 2-year and 5-year inflation swap rates are probably not too far off the mark. If these rates continue their downward trend — and especially if they go negative again — I would expect the Fed to react.
In the meantime… Be careful out there.
(Update 2010-09-01: I made one mistake in this post. Well, at least one. Correction here.)