Abstract
During the 1930s and early 1940s U.S. Treasury bonds and notes had negative nominal yields as they approached maturity. But since an investor can always hold cash, this is impossible. Any bond must have a positive nominal yield. This paper poses a resolution to this puzzle: in addition to making coupon payments, Treasury securities were options that gave the owner the right to buy a new security on a future date. The paper proposes a method for valuing this 'exchange privilege' and computing the yield to the coupon bearing component of these composite bond/options. The case of the negative nominal interest rates demonstrates that the construction of accurate data requires close examination of the institutional environment, even when studying financial markets. The corrected bond and note yields are used to calculate new estimates of the term structure of interest rates from 1929 to 1949. These new data allow one to follow changes in the both the level and the shape of the yield curve during the Great Depression.