The US Bond Market before 1926 : Investor Total Return from 1793, Comparing Federal, Municipal and Corporate Bonds Part II: 1857 to 1926

From 1857 scholars have relied on Macaulay (1938) to track changes in interest rates during the period before the Ibbotson data begin. Holding period returns, where of interest (e.g., Siegel 1992a, 1992b), have been calculated from summary yield inputs such as those tabulated by Homer (1963), rather than observed prices of individual bonds. Here in Part II of the paper I explain how Homer got Macaulay wrong, misleading downstream compilers such as Siegel, and causing him to under-estimate 19th century bond returns. Values in Homer taken from Macaulay are not yields, but mathematical constructions erected on a (distant) foundation of observations. I correct Siegel's under-estimate by retrieving bond prices from Macaulay's sources and calculating holding period returns directly. I also correct a more general failure to treat Federal bonds properly during the greenback era. In the aggregate I find real bond returns in the second half of the 19th century to be about 150 basis points higher than Siegel. With this correction, in conjunction with corrected stock returns before 1871, I find that bond returns matched stock returns over the entire 19th century. The “stocks for the long run” thesis now appears to be a mistaken extrapolation from a few decades in the middle of the 20th century. No support for it can be found in the 19th century