Solar variability can influence surface climate, for example by affecting the mid-to-high-latitude surface pressure gradient associated with the North Atlantic Oscillation1. One key mechanism behind such an influence is the absorption of solar ultraviolet (UV) radiation by ozone in the tropical stratosphere, a process that modifies temperature and wind patterns and hence wave propagation and atmospheric circulation2, 3, 4, 5. The amplitude of UV variability is uncertain, yet it directly affects the magnitude of the climate response6: observations from the SOlar Radiation and Climate Experiment (SORCE) satellite7 show broadband changes up to three times larger than previous measurements8, 9. Here we present estimates of the stratospheric ozone variability during the solar cycle. Specifically, we estimate the photolytic response of stratospheric ozone to changes in spectral solar irradiance by calculating the difference between a reference chemistry–climate model simulation of ozone variability driven only by transport (with no changes in solar irradiance) and observations of ozone concentrations. Subtracting the reference from simulations with time-varying irradiance, we can evaluate different data sets of measured and modelled spectral irradiance. We find that at altitudes above pressure levels of 5 hPa, the ozone response to solar variability simulated using the SORCE spectral solar irradiance data are inconsistent with the observations.