The concept of Lender of Last Resort (LLR) was elaborated in the 19th century by the governor of the Bank of England, Thornton, and by the editor of The Economist, Bagehot. Their classical doctrine asserts that the Central Bank, the LLR, should lend to illiquid but solvent banks under certain conditions. Current theorists have argued that this view is now obsolete due to market efficiency. Professor Jean-Charles Rochet, at LSE, London, and Xavier Vives, The Portuguese Council Chaired Professor of European Studies and Finance at INSEAD, provide a theoretical foundation for rescuing Bagehots view in their recent working paper.
Banking crises have been recurrent in most financial systems. The problem of liquidity, which is created by the difference in demand and supply of cash, cascades into bank-runs, and its then that even well run banks can go bankrupt. The LLR facility and deposit insurance were instituted to provide stability to the banking system and eliminate traditional banking panics.
But modern liquidity crisis are different. They take the form of bank runs, where large well-informed investors refuse to renew their credit due to a specific event. (For example Certificates of Deposit). A case in point is Russias default in 1997 that led to further intermediary defaults.
In such a situation, one school of thought argues that the financial markets, especially interbank markets, are efficient and can resolve the liquidity problems without the help of LLR. This view also expands internationally, the expectation being that the joint action of the Federal Reserve, the European Central Bank and the Bank of Japan can take care of any international liquidity problem.
So does that mean that LLR are no longer necessary to have in case of a bank panic?
The authors of this working paper build a model of a banks liquidity crisis that arises with a unique combination of market forces. The model shows, that given precise private information, there is a risk that solvent banks could fail simply because a large proportion of depositors withdraw their money. Thus, in absence of LLR intervention, the bank would fail.
They explain the coordination failure can be avoided by appropriate solvency and liquidity requirements. However, the cost of doing this will typically be too large in efficiency terms. The model prescribes that the LLR aid should be provided to the defaulting banks at a higher interest rate (perhaps more appropriately referred to as the penalty rate).
However, a point to be noted is that the Central Bank usually has information about the banks that the market does not have. When providing aid, it is imperative that the Central Bank exercises caution in making this information public; an oblique statement by the banker is optimal as it would provide information without creating a common knowledge signal. Increasing the precision or detail of publicly available information can cause further coordination failure.
The authors conclude by pointing out that while the LLR can provide a lending hand the efforts at managing the crisis should however be concluded by early closure policies of the banks, prompt corrective action and orderly resolution of failures.