L Street: Bagehotian prescriptions for a 21st century money market.

• Author: Selgin, George
• Position: Lombard Street and Walter Bagehot - Essay

In Lombard Street, Walter Bagehot (1873) offered his famous advice for reforming the Bank of England's lending policy. The financial crisis of 1866, and other factors, had convinced Bagehot that instead of curtailing credit to conserve the Bank's own liquidity in the face of an "internal drain" of specie, and thereby confronting the English economy as a whole with a liquidity shortage, the Bank ought to "lend freely at high rates on good collateral." Bagehot's now-famous advice has come to be known as the "classical" prescription for last-resort lending.

Largely forgotten, however, is Bagehot's belief that his prescription was but a second-best remedy for financial crises, far removed from the first-best remedy, namely, the substitution of a decentralized banking system--such as Scotland's famously stable free banking system for England's centralized arrangement. Bagehot's excuse for proffering such a remedy was simply that he did not think anyone was prepared to administer the first-best alternative: "I propose to maintain this system," he wrote, "because I am quite sure it is of no manner of use proposing to alter it.... You might as well, or better, try to alter the English monarchy and substitute a republic" (Bagehot 1873: 329-30).

Like Bagehot, I offer here some second-best suggestions, informed by recent experience, for improving existing arrangements for dealing with financial crises. Unlike Bagehot, who merely recommended changes in the Bank of England's conduct, I propose changes to the Federal Reserve's operating framework. And although, like Bagehot, I consider my proposals mere "palliatives," I do not assume that we cannot ultimately do better: on the contrary, I doubt that any amount of mere tinkering with our existing, discretionary central banking system will suffice to protect us against future financial crises. To truly reduce the risk of such crises, we must seriously consider more radical reforms (see, e.g., Selgin, Lastrapes, and White 2010).

A Top-Heavy Operating System

Both the financial crisis and the ways in which the Fed felt compelled to respond to it point to shortcomings of the Fed's traditional operating framework a framework that relies heavily on a small number of systematically important financial firms known as "primary dealers," as well as on JPMorgan and Bank of New York Mellon in their capacity as "clearing banks" for the Fed's temporary open market transactions.

In theory these private institutions serve as efficient monetary policy agents--that is, as private middlemen or conduits through which liquidity is supplied by the Fed to the rest of the financial system. The theory breaks down, however, if the agents themselves become illiquid or insolvent, or if some agents fear being damaged by the liquidity or insolvency of others. In that case, the agents may cease to be effective monetary policy conduits. Instead, their involvement can undermine the implementation of ordinary monetary policy, denying solvent firms access to liquid assets. The Fed may for these reasons alone--and setting aside others that contribute to the agents' "systematic significance"--be compelled to bail out a monetary policy agent, further interfering with efficient credit allocation. The expectation that it will do so in turn enhances agents' "too big to fail" status, encouraging them to take excessive risks, and increasing the likelihood of future crises.

In what follows I explore the drawbacks of the Fed's top heavy operating framework, especially as revealed by the recent financial crisis. I then offer suggestions for making that framework both less top-heavy and more flexible. The suggested reforms should serve to reduce both the extent of the Fed's interference with an efficient allocation of credit and the extent of implicit guarantees in the financial system, while making it easier for the Fed to adhere to the spirit of Bagehot's classical rules for last-resort lending. More specifically, the changes I recommend seek to ground Fed operations more firmly in the rule of law--and make them less subject to the rule of men--by allowing the Fed to rely on one and the same operating framework to both implement normal monetary policy and meet extraordinary liquidity needs during times of financial distress.

Ordinary Monetary Operations

The Fed traditionally conducts monetary policy by means of a combination of "permanent" and "temporary" open market operations. Permanent operations involve outright purchases and sales of Treasury securities. Because permanent open market sales are relatively rare, purchased securities are usually held in the Fed's System Open Market Account (SOMA) until they mature. Permanent open market purchases are mainly used to provide for secular growth in the stock of base money, and especially in the outstanding stock of paper currency.

Temporary open market operations, in contrast, are aimed at malting seasonal and cyclical adjustments to the stock of base money, and are typically conducted, not by means of outright purchases and sales of Treasury securities, but by means of repurchase agreements or "repos" involving such securities. Although in name a repo is contract providing for the sale of a security with an agreement by the seller to repurchase the same security at a specified price within a relatively short period after the initial sale, in practice repos resemble collateralized loans in which the security to be repurchased serves as collateral. The Fed, having first introduced repos to the U.S. economy in 1917, shied away from them after the massive bank failures of the 1930s. They came back into favor as monetary policy instruments following the 1951 Treasury Accord. Eventually a private repo market developed in which repos, instead of being confined to Treasury securities, came to include a broad range of private debt instruments (Acharya and Oncu 2010: 323-30).

The self-reversing nature of repos, and the fact that the vast majority of them are overnight loans, make them especially fit for temporary open market operations, because the Fed has only to refrain from renewing its repos to absorb base money after a peak demand for it subsides. Repos come in handy, for example, during the Christmas season, when the Fed uses them to offset the decline in bank reserves that must otherwise result from heavy currency withdrawals. Repos also help the Fed to implement its federal funds rate target, because for banks overnight Treasury repos are a relatively close substitute for borrowing in the federal funds market. Arbitrage thus tends to cause the federal funds rate to track the rate for such repos. The Fed is consequently able to use repos to move the federal funds rate in whatever direction it desires, and move it more assuredly than it could do using outright Treasury purchases and sales.

Both permanent and temporary open market operations have traditionally been conducted with a limited number of counterparties known as primary dealers. Although the roots of this primary dealer system trace to 1935, when the Fed was first prevented from buying bonds directly from the U.S. Treasury, the system officially got started with 18 members in 1960. By 1988 the number had climbed to 46. But on the eve of the crisis it had dwindled to just 20, including a dozen foreign bank affiliates. Today, after the failure of MF Global--one of two post-crisis additions to the list there are 21. The Fed normally conducts its open market operations with these dealers only, arranging both outright Treasury security purchases and repos with them, and leaving it to them to channel funds to other financial firms mainly by means of private repos, with commercial banks in turn sharing reserves through the overnight federal funds market.

Two other private market agents also assist the Fed in implementing monetary policy. The failure of two major security dealers during the 1980s gave rise to so-called "tri-party" repos, in which repo counterparties, including the Fed, rely on third parties, known as clearing banks, to price and otherwise manage repo collateral. Today, as at the time of the crisis, there are only two such banks--JPMorgan Chase and the Bank of New York Mellon. Besides being conduits for the Fed's open market operations, the clearing banks also play a crucial role in allocating available liquidity among primary dealers.

Ordinarily, as Donald Kohn (2009: 6) observes, the primary dealer system "'allows the Federal Reserve to implement policy quite efficiently ... with minimal interference in private credit markets." Because it relies on the private market to price and direct funds, the system avoids any risk of credit being provided at subsidized rates, and so heeds Bagehot's classical prescription. The Fed nevertheless maintains a standing facility--the discount window--for the purpose of direct lending to illiquid financial institutions, partly in recognition of the possibility that open market operations, as ordinarily conducted, may prove inadequate for meeting "serious financial strains among individual firms or specialized groups of institutions" during times of financial distress (Board of Governors 1971: 19).

Generally speaking, the presence of efficient wholesale lending markets means that banks are unlikely to turn to the discount window unless they lack the sort of good collateral that would qualify them for classical last-resort loans. The Fed, for its part, appears unable to resist lending to insolvent banks. (1) Consequently, several economists (Friedman 1960:50-51 and 1982; Humphrey 1986; Goodfriend and King 1988; Kaufman 1991, 1999; Lacker 2004: 956ff.; and Hetzel 2009) have recommended doing away with extended discount-window lending altogether, and having the Fed supply liquidity solely through the open market. The crisis has, however, been regarded by some as proof that such a step would be imprudent. "A systemic event," Stephen Checchetti and Titi Disyata...