Connected Lending of Last Resort

Find all details here:

Abstract

Using previously unknown, hand-collected central bank data on daily discount window lending, we show that, during the worst banking crisis in French history (1930-1931), France’s central bank secretly lent 30–40% more to connected banks – commercial banks whose board members were voting shareholders of the central bank. The central bank did not pursue a classic lender of last resort policy response to contagious bank runs such as lending broadly to banks in need of liquidity. Instead, it prioritized shareholder value. However, connected commercial banks, on average, performed no better during the panic of 1930-31, in terms of either retaining deposits or survival. Our results are consistent with theories arguing that, without broad-based emergency lending, a LOLR cannot arrest a panic: connected banks did not escape the crisis because connected lending by the central bank failed to prevent negative spillovers. Despite initially lending in a connected fashion to prevent losses on its portfolio, the central bank ended up with massive losses on its loan book due to crisis-period lending, leading to a surreptitious government bailout of the central bank that was only revealed to the public months after the fact. Our analysis shows that, in case of connected central bank emergency lending, neither the borrowers nor the lender comes out ahead.