Bank Bailouts with On-Site Monitors: Evidence from a Supervisory Experiment

Abstract

I use unique supervisory intervention – discretionary installment of de novo Resident Examiners in Russian banks that received government liquidity and capital support funds – to study the causal effect of on-site monitoring on bailed-out banks behavior in crisis. By exploiting detailed monthly data on lending patterns in control and treated bank samples and the variation in the timing and duration of the resident examiners’ assignments in treated banks, I find that on-site monitoring leads to a significant increase of a bailed-out bank lending in crisis. I also find that this lending behavior is prudent: in the presence of a resident examiner, better capitalized banks tend to lend more while banks with higher accumulated bad loan portfolios lend less. Collectively, these findings reveal an examiner’s ability to align the credit supply and the risk-restricting agendas in bank monitoring. They also contribute to the ongoing academic and regulatory debate on the optimal design of a bailout policy and the microprudential supervision in crisis.