Abstract. We add idealized clouds into a single column model and show that the cloud radiative effects as observed from satellites can be reproduced by a combination of high and either low or mid-level clouds. To quantify all-sky climate sensitivity we define a "fixed-cloud-albedo" null hypothesis, which assumes an understanding of how cloud temperatures change, but assumes no change in cloud albedo. This null-hypothesis depends on how clouds are vertically distributed along the temperature profile and how this changes as the surface warms. Drawing only distributions which match the cloud radiative effects of present day observations yields a mean fixed-albedo (also keeping surface albedo fixed) climate sensitivity of 2.2 K, slightly smaller than its clear-sky value. This small number arises from two compensating effects: the dominance of cloud masking of the radiative response, primarily by mid-level clouds which are assumed not to change with temperature, and a reduction of the radiative forcing due to masking effect by high clouds. Giving more prominence to low-level clouds, which are assumed to change their temperature with warming, reduces estimates of the fixed-albedo climate sensitivity to 2.0 K. This provides a baseline to which changes in surface albedo, and a believed reduction in cloud albedo, would add to.