Robust Political Economy and the Lender of Last Resort

Of the leading versions of lender of last resort policy, which is to be preferred in a world of realistic incentive and information imperfections? The three most prominent versions of lender of last resort policy are: the Classical system of central bank lending on good collateral at a penalty rate, the Richmond Federal Reserve system of open market operations to prevent liquidity drains, and the New York Federal Reserve system of commitment to taking any and all action necessary to prevent the spread of financial contagion. I compare these policies to the mechanisms that developed in free banking systems. Free banking systems had no formal lender of last resort, but instead developed institutions that lessened the possibility of systemic panic in the first place. I find that free banking weakly dominates the Classical system. Free banking also outperforms the New York Fed and Richmond Fed systems on the incentive margin, but is weaker on the information margin. In addition, the paper discusses how the New York Fed doctrine is the only stable activist policy, since the limited responses necessitated by the Classical and Richmond Fed policies are not credible