Who truly bears (bank) taxes? Evidence from only shifting statutory incidence

We show strong overall and heterogeneous economic incidence effects, as well as distortionary effects, of only shifting statutory incidence (i.e., the agent on which taxes are levied), without any tax rate change. For identification, we exploit a tax change and administrative data from the credit market: (i) a policy change in 2018 in Spain shifting an existing mortgage tax from being levied on borrowers to being levied on banks; (ii) some areas, for historical reasons, were exempt from paying this tax (or have different tax rates); and (iii) an exhaustive matched credit register. We find the following robust results: First, after the policy change, the average mortgage rate increases consistently with a strong - but not complete - tax pass-through. Second, there is a large heterogeneity in such pass-through: larger for borrowers with lower income, a smaller number of lending relationships, not working for the lender, or facing less banks in their zip-code, thereby suggesting a bargaining power mechanism at work. Third, despite no variation in the tax rate, and consistent with the non-full tax pass-through, the tax shift increases banks' risk-taking. More affected banks reduce costly mortgage insurance in case of loan default (especially so if banks have weaker ex-ante balance sheets) and expand into non-affected but (much) ex-ante riskier consumer lending, experiencing even higher ex-post defaults within consumer loans.