We study a bank levy that funds government guarantees in a general equilibrium setting where banks intermediate between risk-averse households and state-contingent investments. We offer an analytical characterization of the optimal bank levy as a function of household risk-aversion and guarantees. We show that household risk-taking is increasing in guarantees, while it is decreasing as the bank levy and household risk-aversion increase. This allows us to establish a non-trivial relationship between the optimal bank levy and household risk-aversion: Higher risk-aversion optimally induces a higher levy when guarantees exceed a threshold; otherwise, a higher levy shall be observed in economies with less risk-averse households.

We study the political economy of loan guarantees within a credit-rationing framework. A government uses guarantees to decrease the borrowing cost, thus making more households incentive compatible. This shifts capital to productive projects (allocative effect). Backed by taxpayers, loan guarantees also shift consumption from nonborrowers to borrowers (redistributive effect). While a welfare-maximizing planner is only concerned about the allocative effect, vote-share-maximizing politicians are driven by the interaction of both effects. As a result, politicians may underprovide or overprovide guarantees compared to the welfare-maximizing solution, depending on the electoral setup, household risk aversion, income heterogeneity, and guarantees’ externalities.

In a two-country general equilibrium setting, we study competition between governments with two policy tools: capital requirements and a bank tax. Since banks raise equity and deposits from domestic and foreign households, governments face cross-country externalities the sign of which depends on the extent of positive spillovers, i.e., revenues from taxing banks, and negative spillovers, i.e., deposit guarantee costs. We show that regulatory competition yields the efficient allocation when governments have at their disposal policy tools that enable them to optimally internalize domestic distortions. Our first finding is that this is the case when governments are not restricted by supranational regulation. Our second finding is that supranational regulation may or may not impede efficiency. This conclusion is the result of a detailed analysis where we consider conceivable supranational regulatory schemes, derive their welfare implications and identify those that cause inefficiencies.