This paper and accompanying data may be reproduced by any means mechanical, electronical, or photographical, provided only that this working paper is cited in any publication that uses this data.
The accompanying tables extend the McCulloch U.S. Treasury term structure data appearing in the Handbook of Monetary Economics (McCulloch 1990) to February, 1991, and at the same time provide much greater detail of maturities.
There are 12 data files in all. The 6 files ending in -1 represent the 465 months from 12/1947 to 8/1985. For these months, callable bonds had to be heavily relied upon for the longer-term maturities. These were treated as if they were noncallable and maturing on their final maturity date if selling above par, or on their call date if selling below par, with no adjustment made for the value of the imbedded option. The 6 files ending in -2 represent the 67 months from 8/1985 to 2/1991. For these months, callable bonds were not used at all, giving a purer term structure. The conversion month, 8/1985, is given both ways, both for comparison and so that monthly returns can be computed consistently. See McCulloch (1975, 1990: 672-673) for further details and definitions.
The files PARYLD1 and PARYLD2 give the par-bond yield curve. ZEROYLD1 and ZEROYLD2 give the zero-coupon yield curve implicit in coupon bond prices. FWDRATE1 and FWDRATE2 give the corresponding instantaneous forward interest rate curve. All rates are given as percentages per annum, and are on a continuous-compounding basis. All are derived from a tax-adjusted cubic spline discount function, as described in McCulloch (1975), with the modification that bond coupons are actually summed rather than proxied by an integral. All term structures represent the afternoon of the last business day of the month indicated.
Each curve is based on the following 56 maturities:
| 0.000 | 0.083 | 0.167 | 0.250 | 0.333 | 0.417 | 0.500 |
| 0.583 | 0.667 | 0.750 | 0.833 | 0.917 | 1.000 | 1.083 |
| 1.167 | 1.250 | 1.333 | 1.417 | 1.500 | 1.750 | 2.000 |
| 2.500 | 3.000 | 4.000 | 5.000 | 6.000 | 7.000 | 8.000 |
| 9.000 | 10.000 | 11.000 | 12.000 | 13.000 | 14.000 | 15.000 |
| 16.000 | 17.000 | 18.000 | 19.000 | 20.000 | 21.000 | 22.000 |
| 23.000 | 24.000 | 25.000 | 26.000 | 27.000 | 28.000 | 28.000 |
| 30.000 | 31.000 | 32.000 | 33.000 | 34.000 | 35.000 | 40.000 |
These are monthly from 0 to 18 months, then quarterly to 2 years, then semiannually to 3 years, then annually to 35 years, and finally a 5-year jump to 40 years. A copy of this table appears at the head of each data file. Missing values are entered as -88.888. The 56 data points for each month are preceded by 6 entries that give 1. the year, 2. the month, 3. the number (up to 56) of maturities that are not missing, 4. the number of observations on individual bonds, notes and bills used in the regression, 5. the standard error of the weighted regression, measured in the half-spreads that were used to weight the regression, and 6. the best-fitting tax rate (for ordinary income). The corresponding 6 positions preceding the intial maturity list contain strings identifying these entries. An attempt was made to make the files easily legible by both FORTRAN and GAUSS, at the cost of some cumbersomeness both ways. All numbers are written with an F8.3 format, including items 1-4 above, which are always integers. For ease of scanning on screen, or printing on an 80-column line printer, it is organized into records of length at most 80. Take a look at it before trying to load it. Standard errors for the 6 data files are contained in files PARERR1, PARERR2, ZEROERR1, ZEROERR2, FWDERR1, and FWDERR2. The zero-maturity interest rates (which are identically equal for all 3 term structure measures) and long-end forward rates are particularly prone to measurement error. The data precisely match those given in the Handbook of Monetary Economics prior to 8/1983, with an added digit of precision. A few minor errors (mostly constituting valid observations that had been effectively omitted) were corrected in the data from 8/1983 - 2/1987, which may cause very slight differences to appear for these months. Note that the HME gives only the no-call version for 8/1985, even though this date is best treated as a break in the time series. The regression standard errors indicate that the observations for each date prior to the Accord of March 4, 1951 are much less internally consistent than those afterwards. The user may therefore consider omitting the pre-Accord data.
The data after 1983 were collected by H.C. Kwon, who updated it to 1991 for his dissertation (1992). Others who have collected data include Krista Chinn, Steve Garasky, and Richard Harmon. The data from 1946 to 1966 is from CRSP files, and a large block of the data was contributed by Bob Hetzel of the Richmond Fed.
The Gauss program loads the suffix -2 files from a directory called \DATA\TSDISK, and draws nice plots of the three curves on your screen, which you can page through with your ESC key.
Kwon, Heon-Chul. "The Time Variant Term Premium of Interest Rates," Ohio State University Economics Department Ph.D. dissertation, 1992.
McCulloch, J. Huston. "The Tax-Adjusted Yield Curve," Journal of Finance 30 (1975): 811-830.
____________________. "U.S. Term Structure Data, 1946-87," Handbook of Monetary Economics Volume I, pp. 672-715.
410 Arps Hall
1945 North High Street
Columbus, Ohio 43210
All contents copyright (C) 1998. All rights reserved.
Revised: April 27, 1998
URL: http://economics.sbs.ohio-state.edu/