Assessing the term structure of expected inflation using treasury inflation—protected securities: near real-time measures for those who need quick estimates

Business Economics, Jan, 2003 by Albert E. DePrince, Jr.

While a case is made that the TIPS provide a quick response to incoming information in the previous section, some may argue that, from a forecasting perspective, the case for their use must rest on their accuracy, particularly with respect to the SPF. Unfortunately, it will be several years before actual inflation data are available to compare with the intermediate-term TIPS-based forecasts (i.e., five and ten-year forecast horizons) in a rigorous manner. Also, while actual inflation values are available to compare with the shorter forecast horizon (two- and three-year forecast horizons), they are too few to provide a reasonable test. Undoubtedly, such tests should and will be conducted when more data become available with the passage of time.

In Sum

TIPS can provide real-time estimates of the term structure of inflation expectations over a variety of horizons. Over the intermediate- (five- and ten-year) and long-term (thirty-year) forecast horizons, TIPS-based forecasts can be regularly generated, and given the ease of generating these estimates, they could be regularly used by businesses, financial market participants, and policymakers as inputs into their decision-making process. From a practitioner's perspective, the simple spread can provide an approximate estimate of the capital market's estimate of inflation over a given horizon. For example, using the latest ten-year inflation-protected and conventional notes, this methodology produces a ten-year inflation estimate of 1.6 percent per year as of September 30, 2002. Using the latest inflation protected and conventional thirty-year bonds, the thirty year inflation estimate would be 2.2 percent per year as of September 30, 2002.

However, without the regular auction of five-year TIPS, estimates of expected inflation over short-term horizons must await the aging of older ten-year TIPS. Even so, it will still be several years before a time series of inflation forecasts over two- and three-year horizons could be assembled using aged ten-year TIPS.

Results reported above showed that the term structure of inflation expectations is usually positive; that is, inflation is expected to rise as the future time horizon is lengthened. This was uniformly true when comparing inflation forecasts over five-, ten- and thirty-year horizons. However, there were times when the term structure was negative between short- and intermediate-term inflation forecasts.

Admittedly, various factors could corrupt the estimates of expected inflation when using the spread between the inflation-protected and conventional Treasury securities of the same maturities. Thus, estimates of the term structure of expected inflation should be treated with some caution, but shifts in the shape can provide guidance on shifts in inflation expectations.

While the debate will continue, TIPS-based estimates of expected inflation provide a quick means to obtain near real-time estimates of changing views on inflation. Equally important, they offer the ability for practitioners to obtain forecasts over ranges not typically available from forecast surveys. Aside from the ease of calculation, TIPS-based estimates have the added advantage of being unfettered by personal forecasting biases or by forecasts constrained by corporate views, both of which might taint survey results. As the characteristics of TIPS-based inflation forecasts become better known to practitioners, the use of TIPS-based inflation forecasts will likely increase. It is hoped that this study is a step in providing increased understanding of the characteristics of such forecasts.