In Swanson and Williams (2013), we demonstrate a new way to measure how much the zero lower bound constrains Treasury securities of any maturity. We look at how Treasury securities of different maturities respond to major macroeconomic announcements, such as reports on U.S. employment, gross domestic product, or consumer price inflation. These announcements have important implications for the U.S. economy and future monetary policy, and thus tend to move interest rates of all maturities.
In particular, we estimate the time-varying sensitivity of a given Treasury yield to major macroeconomic announcements relative to a benchmark period when the zero lower bound was not a concern, taken to be 1990–2000. If a particular Treasury yield responds to news today by just as much as it did in the 1990s, then we would say that Treasury yield is unconstrained by the zero lower bound. Alternatively, if a given Treasury yield used to respond to macroeconomic announcements in the 1990s, but no longer responds to those announcements today, then we would say that Treasury yield is completely constrained by the zero bound. A yield could also be partially constrained, that is, respond to news significantly less than normal. This statistical approach provides a rigorous test of whether any given yield is constrained by the zero lower bound and a quantitative measure of how much that yield is constrained.