Both private investors and policymakers pay close attention to estimates of expected inflation. If investors expect inflation to be high, long-term, fixed-rate securities become less attractive, and market prices of those securities fall. If central bankers see that investors’ inflation expectations are high, they may decide to tighten policy. The trouble is, neither investors nor central bankers can observe inflation expectations directly, and not all methods for inferring expectations from market data give the same result. Sometimes, as at present, different estimation methods even point in opposite directions.
Consider, for example, a recent post by Scott Grannis, titled Inflation Expectations are Rising. “For the past 15 months, the inflation expectations embedded in TIPS and Treasury bond prices have been moving higher, most likely in response to the Fed’s increasingly accommodative policy stance,” writes Grannis. Striking an optimistic note, he offers the opinion that rising inflation expectations are all to the good, as they signify a return of market confidence, which will soon manifest itself as a willingness to invest more in equities and other relatively risky assets.
A key piece of evidence that Grannis cites is the spread between the yield on ordinary 10-year Treasury securities and 10-year Treasury Inflation Protected Securities (TIPS). That is called the breakeven rate because, if inflation over the next 10 years were in fact to be equal to the spread, the inflation-adjusted yield on the two securities would be identical. For that reason, the breakeven rate is often used as an approximation of the inflation rate that participants in the bond market expect over the life of the securities in question.
The following chart shows that the 10-year TIPS breakeven rate has been trending irregularly upward since late 2010, with its rise of 14 basis points since the last week of November and 47 basis points since May providing the grounds for Grannis’s headine.
However, the TIPS breakeven rate is not the only way to measure inflation expectations. Opinion surveys and inflation swaps provide additional data. Furthermore, the breakeven spread itself is not a pure measure of expected inflation. Instead, it reflects both the expected future rate of inflation and a risk premium based on the probability that the actual rate of inflation over the life of the securities in question will turn out to be above or below the rate now expected.
The research staff at the Federal Reserve Bank of Cleveland publishes its own estimates of expected inflation that incorporate data from a variety of data sources and, at the same time, split out the inflation premium from the purely expectational component of the breakeven rate. (Links to their complete data and methodology can be found here.)
The Cleveland Fed’s estimate of the 10-year inflation rate, shown in the next chart, was just 1.52 percent as of the most recent observation, at the beginning of December, nearly a full percentage point lower than the breakeven rate. That puts it close to a 20-year low. It is also 33 basis points below its value in August 2010—the low point for the TIPS breakeven rate, which has since risen 97 basis points.
What, then, is really going on? Should our headline be “Inflation Expectations are Rising,” as Grannis has it, or would “Inflation Expectations are at a Twenty-Year Low” be more accurate? It is tempting to say that the Cleveland Fed’s numbers must be better, since they take more data sources into account and, in addition, separate out the risk-premium from the pure expected inflation rate. However, two considerations should make us cautious about reaching such a conclusion.
One is the fact that complex models do not necessarily give better answers than simpler ones; sometimes they just open the door to new sources of error. The other is that the Cleveland Fed’s reserearchers themselves conclude that outside special periods, their model by and large vindicates the use of changes in the breakeven rate as a gauge of changes in expected inflation. Even though their estimates attribute about 50 basis points of the breakeven spread to the risk premium, the premium is not very volatile, so that expectations dominate in the long run.
The bottom line is that we can’t really measure inflation expectations as accurately as we would like to. By historical standards, inflation expectations have clearly been moderate throughout the recent recession and recovery, despite the Fed’s aggressively accommodative policies. Changes over the past few months would appear to fall within the range of measurement error. It would be premature to interpret the latest wiggles in the charts either as a signal for the Fed to slam on the brakes before it is too late or as a hopeful sign that better days for the real economy are just around the corner.