Basel’s Liquidity Coverage Ratio: Redux

Last summer I contributed a post about the Liquidity Coverage Ratio (LCR), a new regulation that is part of the latest international Basel Accords (Basel III) and that is being imposed on U.S. banks and other financial institutions. As I explained in that post, the LCR requires banks to hold “high quality liquid assets” (HQLA) sufficient to cover potential net cash outflows over 30 days. Both George Selgin and I have pointed out that the LCR probably contributes to the continuing desire of banks to maintain such a high level of reserves.

Two economists who have severely criticized the LCR are Gary Gorton, noted for his work on bank panics, and his co-author, Tyler Muir. Earlier this year they published online a short version of a much longer unpublished paper that scrutinizes the potential impact of the LCR. Whereas my post, appropriately entitled “Reserve Requirements Basel Style,” compared the LCR to the traditional but now largely abandoned reserve requirements imposed on banks, Gorton and Muir compare it to the bond-collateral (or bond-deposit) requirement of the national banking era, prevailing from the Civil War until creation of the Federal Reserve. They conclude that the LCR will cause the same sorts of problems that, ironically, the Fed was supposed to solve.

For those unfamiliar with the bond-collateral requirement, it was a feature of the system of nationally chartered banks established by Congress. Congress imposed a tax that suppressed the banknotes of state-chartered banks, but national banks could still issue banknotes, with a requirement that tied their quantity to the banks’ holdings of Treasury securities. Although nominally national banknotes were the liability of the issuing bank, they were in fact fully guaranteed by the government. The system thus created what was referred to as an “inelastic currency,” in which widespread demands to convert bank deposits into more liquid banknotes led to financial stringency and bank panics. This problem became the justification for the creation of the Federal Reserve as a lender of last resort.

Today the LCR counts both reserves and Treasury securities as the most highly liquid assets. So thinking about the LCR as either a modern modified reserve requirement and thinking of it as a modern modified bond-collateral requirement are complementary, particularly since reserves now pay interest and therefore have become like Treasury securities that only banks can hold. Gorton and Muir recognize this similarity in their online article when they state that “the LCR then requires that government money (Treasuries) must be used to back bank money – i.e., short-term debt, a kind of narrow banking [emphasis mine].”

Gorton and Muir’s concerns about the LCR can be briefly summarized in two points. First, despite the growing U.S. national debt, the worldwide liquidity demand for Treasury securities will severely impinge on their general availability. As they put it, “the LCR may result in a shortage of safe debt because too much of the Treasuries is tied up.” Second, this “shortage of safe debt” may encourage “other forms of [privately issued] short-term debt to emerge, making the financial system riskier.” Here they are drawing an analogy to the way restrictions on banknotes during the national banking era not only encouraged the increasing use of bank deposits but also contributed to financial instability.

I am not convinced by all parts of Gorton and Muir’s argument. Indeed, their characterization of bank deposits as the “shadow money” of the national banking era, by implication, gets the causes of the inelastic currency exactly backwards. The panics and financial stringency of this period did not result from too little regulation of deposits but instead from too much regulation of banknotes. Moreover, Gorton and Muir’s fear of a new form of shadow money seems to cut against their enthusiasm for the widespread use, prior to the financial crisis of 2007-2008, of repurchase agreements and asset-backed commercial paper. Nonetheless, their overall point is sound. Given the universally conceded failure of the bond-collateral requirement during the national banking era, why in the world are regulators going back to the same kind of system with the LCR? It did not work then; why should it work now?

HT: Bruce Tuckman

[Cross-posted from]