Studies of sea-level change from altimetric missions require accurate calibration if systematic errors are not to be aliased into the solution. For ERS-2 there is now growing evidence that the altimetric range is biased. Global studies over the period of concurrent operation of ERS-1, ERS-2 and TOPEX/Poseidon show that the mean sea-level rise, including the steric effect, was -4.7 mm/yr and-5.6 mm/yr for ERS-1 and TOPEX respectively but 9.0 mm/yr for ERS-2. Further, repeat pass data between ERS-1 and ERS-2 yields a systematic difference equivalent to a 11 mm/yr slope. The results show that discontinuities in the repeat pass data correspond to the ERS altimeters being switched off after an instrument anomaly. These discontinuities occur despite utilisation of the correction data available to the user community. Corrections to the long-term systematic errors in the ERS altimeters have been determined by using dual satellite crossover residuals with TOPEX. These enhancements remove the spurious trend to the extent that the differential slope between ERS-1 and ERS-2 is reduced to 0.2 mm/yr. With most of the correction attributable to ERS-2 it is apparent that this satellite is more problematic than ERS-1. ¿ 1999 American Geophysical Union