There is currently no consensus among scholars on how to achieve a just low-carbon transition. This paper subjects a macrosimulation model to an extensive sensitivity analysis, and trains random forests on the simulation results to identify which policy combinations are most effective in reducing carbon emissions while improving the distribution of income. The results suggest a trade-off between inequality and emissions, which limits the extent to which inequality can be addressed through the growth of bottom incomes alone, and indicate that environmental and distributive goals can be met jointly only through a variety of coherent policies in different domains.