We investigate geoengineering as a possible substitute for mitigation and adaptation measures to reduce damages from climate change. With the help of an integrated assessment model, we distinguish between the effects of solar radiation management (SRM) on atmospheric temperature levels and its side-effects on the environment. We account for the uncertainty in the magnitude of side-effects and their persistency over time, and show that this geoengineering option lacks robustness. The optimal climate portfolio represents a mix between mitigation, adaptation, and SRM, and no strategy fully substitutes the others. We then analyse the welfare consequences of basing the SRM decision on wrong assumptions about side-effects, and show that total output losses are considerable and increase with the error horizon. This reinforces the need to balance the climate portfolio in favour of mitigation.