Successive radiometers have been flown on spacecraft for measurement of the radiation budget of the Earth and to date have provided data sets which overlap in time over a 25-year period. Together these data sets cover two cycles of any decadal oscillation and are useful for climate research. However, before multiple data sets can be used, it is necessary first to establish precisions and relative biases of the data sets. Although these instruments were all calibrated with great care, inevitably there are differences between the instruments. Also, instruments change while in orbit due to degradation of optical elements and changes of the detectors. A number of studies have compared measurements between pairs of radiometers in order to establish these differences. In this paper we compile these results and by use of a bar chart demonstrate the traceability of calibration of satellite radiation budget instruments over these two decades. Using this compilation the differences are computed for any two instruments listed. The instruments considered here include the ERBE scanning and nonscanning radiometers, ScaRaB I and II, and the CERES instruments aboard the TRMM, Terra, and Aqua spacecraft. These results are a prerequisite to use of multiple data sets for long-term studies of climate process. The largest biases are between the ERBS scanning and non-scanning radiometers, which at the "top of the atmosphere" differ by 5.4 W m-2 for shortwave fluxes and by 5.8 W m-2 for daytime longwave fluxes. The standard deviations of the biases, which define how well the data sets can be connected, are less than 1 W m-2 for longwave and between 1 and 2 W m-2 for shortwave fluxes.