Climate policy design must balance emissions mitigation with concerns for fairness, particularly as climate change disproportionately affects the poorest households within and across countries. Integrated Assessment Models used for global climate policy evaluation have so far typically not considered inequality effects within countries. To fill this gap, we develop a global Integrated Assessment Model representing national economies and subnational income, mitigation cost, and climate damage distribution and assess a range of climate policy schemes with varying levels of effort sharing across countries and households. The schemes are consistent with limiting temperature increases to 2 °C and account for the possibility to use carbon tax revenues to address distributional effects within and between countries. We find that carbon taxation with redistribution improves global welfare and reduces inequality, with the most substantial gains achieved under uniform taxation paired with global per capita transfers. A Loss and Damage mechanism offers significant welfare improvements in vulnerable countries while requiring only a modest share of global carbon revenues in the medium term. The poorest households within all countries may benefit from the transfer scheme, in particular when some redistribution is made at the country level. Our findings underscore the potential for climate policy to advance both environmental and social goals, provided revenue recycling mechanisms are effectively implemented. In particular, they demonstrate the feasibility of a welfare improving global climate policy involving limited international redistribution.

As a common feature in many federal states grants-in aid are payed to jurisdictions ex post, i.e. after local policy measures have chosen. We show that the central government cannot offer grants ex ante in a federal states with informational asymmetries as well as inter-temporal commitment problems. Local governments’ incentives to provide public goods are distorted if they rely on federal grants-in-aid offered ex post. Furthermore it becomes obvious that local governments are apt to substitute tax revenue for higher grants-in-aid if relevant local data are unobservable for the central government. To which extend ex post transfers mitigate local governments’ incentives crucially depends on the information structure predominant in the federation.