Central Bank Lending and Incentives

Goodfriend, Marvin, and Jeffrey Lacker. 1999. "Limited Commitment and Central Bank Lending." Federal Reserve Bank of Richmond Economic Quarterly 85, no. 4 (Fall): 1-27.

Marvin Goodfriend was a remarkably original thinker and did much to advance the science of monetary policy. We both benefited not only in key ways from Marvin's published work in economics, but also just as much from his contributions at seminars, conferences, and over lunch or dinner. Our essay primarily concerns Marvin's work with Jeff Lacker, in Goodfriend and Lacker (1999). It analyzes the roles of central bank lending and central bank credit policy, describes what can go wrong with central bank lending, and suggests how to fix these problems.

Central bank lending

Lending to the financial system has been a critical feature of central banking for a very long time. Typically, central banks are constrained to hold assets to back their liabilities, and those assets include loans to the private sector. Indeed, in some central banking systems, the principal mechanism by which the central bank controls the quantity of its liabilities in circulation is through lending to private banks. Notably, at its inception in 1914, the Federal Reserve System was constructed as an organization of semiautonomous regional banks that financed lending to member banks by issuing currency. These regional Federal Reserve Banks became the sole issuers of circulating currency. Each regional Reserve Bank in turn held the liabilities of its member banks to back the currency. Government debt was not initially a key asset in the Fed's portfolio, and open market operations were not an important component of Fed activities until the 1920s. Today, the European Central Bank intervenes primarily through central bank lending to banks in the euro area, by way of the ECB's main refinancing operations. So, one possible design for central banks includes the use of central bank lending in day-to-day or week-to-week financial market intervention. But — and this is the most important aspect of Marvin Goodfriend's research that we want to address — lending to financial institutions is a key component of central bank crisis intervention. The lender of last resort role for central banks was essentially invented by the Bank of England (BoE), as documented for example in Bagehot (1873). In the later 19th century, the BoE was a private institution, which had been granted a special place in the British financial system. It had a different financial structure from a typical private bank in the United Kingdom at the time and had been granted a monopoly on currency issue by the Crown through Peel's Bank Act of 1844.

The BoE had also built a reputation for safety by the late 19th century. So, during the recurring financial crises in the UK in the later 19th century, consumers and firms typically fled from bank liabilities, perceived to have increased in their riskiness, to BoE liabilities, and this inflow of funding at the BoE was then used to finance lending to Banks. But the BoE was presumably more well-informed than the general public about which banks were insolvent and which were merely illiquid and so could profit from judicious lending under Bagehot's (1873) principles: lend freely, at a high rate, against good collateral.

In the United States, a principal objective of the authors of the Federal Reserve Act of 1913 was preventing, or at least mitigating, the effects of the banking panics that occurred in the US in the late 19th century and early 20th century. Central bank lending to private banks was seen as the principal crisis mechanism at the Fed's disposal. This was a key element in research done under the National Monetary Commission leading up to the Federal Reserve Act legislation in 1913. But Fed policy during the Great Depression is typically viewed as a failure,¹ in part because the Fed did not lend adequately to the private banking system. Seemingly, Fed leadership absorbed the lessons of the Great Depression in subsequent years, and the global financial crisis was a quite different story. Indeed, Ben Bernanke has argued that much of the Fed's lending policy during the global financial crisis was motivated by Bagehot's principles.² The argument for central bank crisis lending generally rests on advantages the central bank might have relative to the private sector. For one example, the central bank could have superior information on banks' creditworthiness, due to its supervisory role in the banking system. For another, the deep pockets of the central bank are important, given the backstop of the federal government's power to tax.

There were at least two unusual elements in the Fed's lending strategy during the global financial crisis. The first was lending through the Term Auction Facility (TAF). An undesirable feature of lending through the Fed's discount window is that a bank's discount window borrowings can send a signal to the market that the bank on the receiving end of the loan is distressed. This "stigma" can deter borrowing, which works against the intent of central bank lending in a crisis.³ Stigma can occur because, even though the details of discount window lending are not public, it may be possible to infer which banks are borrowing at the discount window, particularly if the banks are large. So, the goal of the TAF program was to auction off discount window funds to member banks willing to post the appropriate collateral. The assumption was that both distressed and nondistressed banks would be on the receiving end of TAF funds, thus eliminating, or at least mitigating, the stigma effect.

A second example of unusual lending during the global financial crisis was the extension of Fed loans to financial institutions that were not commercial banks. In historical banking panics prior to the founding of the Fed, for example, those that occurred in the US in the late 19th and early 20th centuries, the key problem was massive outflows of deposits from commercial banks that caused the disruption of payments, bank failures, and the forced sale of bank assets. Such negative effects, where they occurred in solvent but illiquid banks, could in principle have been mitigated by central bank lending to commercial banks if there had been a US central bank during this period. The global financial crisis was different, however. Early on, panics appear to have occurred at the wholesale level, for example, when asset portfolios of asset-backed securities were funded by rolling over over-night repos. The perception that some asset-backed securities were much riskier than previously thought generated the withdrawal of funding for such "shadow-banking" entities, in part inducing fire sales of assets. Later, problems involved systemic risk, which had been latent until 2008. For example, the Fed's direct lending to American International Group (AIG) — primarily an insurance company — was intended to address a novel crisis problem not directly related to retail banking. That is, through the sale of financial derivatives, AIG had made itself highly sensitive to aggregate risk and had thus created a threat to the entire financial system. Whether the Fed has a legitimate role in lending to nonbank financial institutions is not clear. For example, it seems harder to make the case that the central bank has an information advantage in lending to financial firms that it does not regulate, or that the Fed somehow has an advantage relative to large private financial institutions in such lending.

Goodfriend-Lacker and central bank lending

Crisis lending by the central bank might on its surface appear straightforward. The basic nature of banking is captured nicely by the classic model of Diamond and Dybvig (1983). Banking, by its nature, involves the transformation of illiquid assets into liquid ones. Such transformation is socially useful, as retail payments work efficiently if consumers and firms can trade the widely acceptable liabilities of third parties — here, banks — for goods and services. These third-party liabilities are viewed as highly liquid, despite the underlying assets held by banks being difficult to exchange for goods and services. But in conducting liquidity transformation, banks leave themselves open to runs, as shown by Diamond and Dybvig (1983). If each depositor anticipates that all other depositors will run to the bank to withdraw their deposits, then the bank — which would otherwise be solvent — cannot satisfy all requests for withdrawal and it fails. Though a key motivation for establishing the Federal Reserve System was to prevent or mitigate banking panics through central bank crisis lending, the banking theory literature appears to have given central bank lending short shrift. For example, though the Diamond-Dybvig (1983) model has received much attention, it does not directly address the role of central bank crisis lending. Thus, Goodfriend and Lacker (1999), while not making a contribution in a formal modeling sense, use available theory to analyze the incentive problems inherent in central bank lending to private financial institutions and provide useful recommendations for improving this aspect of central bank policy.

Goodfriend and Lacker (1999) recognize that, while we know less than we would like about the pitfalls of central bank lending, we know a lot about how private loan commitments work. And in some ways private loan commitments are not so different from central bank lending. For example, discount window arrangements between the Fed and private commercial banks are essentially loan commitments. The Fed specifies the terms under which banks can borrow, including interest rates, admissible collateral, and collateral haircuts, and then commits to lending to banks on these terms, with some restrictions. The key difference, however, is that private banks making loan commitments are concerned with their own profits, while the central bank has public policy goals in mind, for example, the systemic implications of the failure of large financial institutions. But since the Fed's loan commitments to banks are in part insurance, Fed loan commitments are subject to the same moral hazard issues as private loan commitments, with no profit motive for the Fed to motivate structuring its lending to deal with this moral hazard.

But what could central banks learn from the structure of private loan commitments? First, as with borrowing from the central bank, borrowing from a private bank under a loan commitment may become desirable when a firm is under financial stress. For example, a large firm may normally have access to the commercial paper market, but if it becomes widely known that the firm's financial state is precarious, borrowing by way of commercial paper may become more costly or impossible. Under these circumstances, taking down a loan commitment may be desirable for the firm. Loan commitments may have im-plications for the commercial paper market, ex ante, as well. That is, a firm may obtain more favorable terms in the commercial paper market because it has a standing loan commitment with a bank. Subsequently, in the event of financial distress on the firm's part, loans under the commitment from the bank could serve to insure the holders of the firm's commercial paper.

Of course, the bank making a loan commitment does not want to be in such a position. So typically, as Goodfriend and Lacker (1999) point out, loan commitment arrangements have terms that give the firm incentives not to take out loans — or prevent the firm from taking out loans — if the firm has a high probability of defaulting. First, collateral can play an important role, both in insuring the bank against losses in the event of the firm's default and giving the firm the incentive not to default. Second, covenants in the loan commitment can allow the bank not to lend to the firm given the firm's financial state or limit lending in various ways. Third, the bank can have the right to monitor the firm's activities, which, in conjunction with covenants, could serve to limit lending in particular circumstances.

For central banks, the incentive problems are similar to those with private loan commitments, with some caveats. First, the central bank's lending behavior has implications for other bank creditors, though in the case of a regulated bank, this is muted by the existence of deposit insurance. However, lending by the central bank to a precarious commercial bank can provide a window of time when uninsured depositors can conveniently exit. Effectively, the central bank acts to provide implicit insurance to the uninsured depositors. But the central bank's lending behavior can ultimately change the payoffs to the stakeholders in important ways in the event of a bank default. The Fed protects itself by requiring that banks pledge good collateral against central bank loans, but that implies that this collateral cannot be used to pay off the bank's other creditors in such an event. Further, if the Fed lends to a bank that ultimately fails, the Fed becomes a senior creditor, which matters for the other regulator, the Federal Deposit Insurance Corporation (FDIC). Once the FDIC steps in to resolve the failed bank, it pays off the bank's debt to the Fed. Though the FDIC then retains the collateral pledged against the central bank loan, this may not cover the loss.

Second, the central bank, like private banks making loan commitments, may want to commit to limiting central bank lending under particular circumstances. It is well-accepted, for example, that lending to insolvent banks is a bad idea, and that central bank crisis lending should be limited to cases where there is just a liquidity problem. Of course, it may be difficult to tell the difference between an insolvent bank and an illiquid one, but the Fed is in a good position to discriminate, given the information it acquires through its supervisory role in the commercial banking system. But commitment to limit central bank lending may be difficult nevertheless, and Goodfriend and Lacker (1999) are particularly pessimistic about this. When push comes to shove, they argue, the Fed will typically opt for the path of generous lending except in the most egregious cases. The key issue is that moral hazard problems associated with borrowing by banks are accentuated as actual insolvency becomes more likely for an individual bank. But, in general, the probability of bank insolvency is related to aggregate activity, so many banks will be in a precarious state at the same time. This is exactly the type of circumstance where the Fed is likely to lend freely, rather than cutting off banks that are likely on the brink of insolvency.

We highlight a particularly prescient passage in Goodfriend and Lacker (1999, p. 15):

The financial stability mandate can create pressure to expand the scope of central bank lending to nonbank financial institutions. Nonbank financial intermediaries are capable of amassing sizable financial market positions. The liquidation of these positions could be seen as a threat to the stability of asset prices and the solvency of many other financial institutions, including insured banks. A central bank with no formal authority to lend outside a narrowly defined set of institutions is, of course, well positioned to resist influence. Otherwise, we might see a tendency to expand the range of institutions receiving central bank line-of-credit assistance.

This is an accurate prediction of part of what happened during the global financial crisis, when the Fed expanded its lending beyond the commercial banking sector. For example, after the Fed extended a large loan to AIG, Goldman Sachs and Morgan Stanley became bank holding companies and thus eligible for Fed lending. Of course, a key difficulty is that it can be easy to see the short-term gains but hard to see the long-term costs of central bank crisis intervention that goes beyond the commercial banking sector. In theory, we understand the implications of moral hazard in central bank lending for excessive risk-taking and the expansion of already-large, too-big-to-fail financial institutions. But these effects can be difficult to measure in practice.

Dealing with moral hazard in central bank lending

As we have argued, the view of Goodfriend and Lacker (1999) is that there are issues in Fed lending to the private sector that need to be ad-dressed. So, what to do about it? One possibility they consider is that the Fed could forgo lending to the private sector entirely and focus instead on facilitating the easing of liquidity problems in large financial institutions and on orderly resolution in the event of large financial failures. A few examples in the past two decades, however, suggest that this would be a radical alternative.

The role of the Fed in coordinating privately financed emergency lending and support in the financial sector is an aspect of the use of the Fed's "good offices." For example, in 1999, the New York Fed participated in the effort to prevent the failure of a hedge fund, Long-Term Capital Management (LTCM). At the time, LTCM was viewed as a systemically important financial institution that was encountering liquidity problems and thus faced a potential forced sale of assets. That is, LTCM was in a position that several large financial institutions would find themselves in during the global financial crisis in 2008. The Fed did not participate financially in propping up LTCM, but it helped facilitate an arrangement by which a group of private financial institutions recapitalized the troubled hedge fund. Was the Fed's intervention during this episode necessary? Was there risk to the Fed in injecting itself into negotiations over the potential failure of a private hedge fund that was well outside the Fed's normal supervisory purview?

An instance where the Fed's intervention with respect to a large troubled financial firm outside the commercial banking sector moved from mere facilitation to key financial participation was the Bear Stearns failure in spring 2008. As discussed in more detail in Goodfriend (2011), the Fed facilitated an orderly resolution in the Bear Stearns failure by setting up a limited liability company, Maiden Lane I, which then proceeded to purchase troubled assets then held by Bear Stearns, with funding coming from a loan from the Fed. This made it more attractive for JPMorgan Chase to take over what remained of Bear Stearns. The motivation for the Fed's Bear Stearns intervention was similar to that for the LTCM intervention, in that Bear Stearns was a systemically important financial institution and a disorderly failure would potentially have much wider effects in the financial sector. But this illustrates what happens when the Fed becomes involved in negotiations involving large troubled financial institutions. In such negotiations, the Fed is the elephant in the room and can end up more financially involved than it initially intends, left holding the bag by private financial institutions or passing on losses to taxpayers.

As Goodfriend and Lacker (1999) point out, the Fed may protect itself by requiring good collateral with appropriate haircuts to secure central bank lending, but safety for the Fed in this respect can be to the detriment of the FDIC and a bank's uninsured creditors. Potentially, there could be stricter capital thresholds for closing distressed banks, though this is problematic due to the subjective nature of asset valuation. Goodfriend and Lacker (1999) also suggest that the Fed might choose to be "constructively ambiguous." This means the Fed could be deliberately vague about the conditions and terms on which it will lend. Presumably, this could limit the quantity of lending, reducing moral hazard, and causing banks to bear less risk. But this would have to be balanced against the increase in perceived risk faced by the banking system due to the Fed's unpredictable behavior.

During the global financial crisis, the Fed not only intervened in important ways with respect to large nonbank financial institutions — Bear Stearns and AIG in particular — but it made large loans to large banks, Citigroup and Bank of America. The latter two banks certainly fall in the too-big-to-fail category and are therefore important examples of the moral hazard problems associated with central bank lending, as discussed by Goodfriend and Lacker (1999). Much of the lending to large financial institutions — nonbanks and banks alike — during the financial crisis fell under Section 13(3) of the Federal Reserve Act, which gave the Fed very broad lending powers. In line with Goodfriend and Lacker's (1999) emphasis on limits to Fed lending the Dodd-Frank Act of 2010 included provisions to constrain the Fed's powers under Section 13(3). However, Goodfriend and Lacker (1999) might suggest we worry about the array of Fed lending programs introduced during the COVID-19 pandemic, which included lending to nonfinancial businesses (the Main Street Lending Program), to state and local governments, and to money market mutual funds. In most cases, there were explicit arrangements for the Treasury to absorb potential losses, but this raises some of the key issues discussed in Goodfriend and Lacker (1999), relating to moral hazard and the redistribution of creditor losses.

Finally, using an analogy to inflation control, Goodfriend and Lacker (1999) argue that the Fed could establish a reputation for limiting its lending to financial institutions, and that this could ultimately curb the moral hazard problem associated with Fed lending. Unfortunately, inflation control as a central bank goal has the advantage of simplicity — for example a 2 percent inflation target — and it is relatively easy for people to evaluate the Fed's success in achieving the goal. With respect to Fed lending, it is much more difficult both to establish what the goal is and to evaluate the Fed's performance relative to the goal. Perhaps the only limits on Fed lending that can have force are those specified explicitly in the Federal Reserve Act.

Other approaches to dealing with moral hazard in central bank lending

It is possible that the key problems discussed by Goodfriend and Lacker, associated with moral hazard and central bank lending, could be solved by simple (though perhaps radical) changes in institutional structure. We will consider two: one that changes central bank interventions that we term "repos only," and another that reforms private financial institutions that is conventionally called "narrow banking."

One simple approach would involve a central bank lending policy that restricts intervention to the repo market. Under this setup, the central bank sets a target for the overnight interest rate and then achieves that through two standing facilities, a repo facility and a reverse repo facility, both involving fixed rate and full allotment auctions. As the Fed has learned since establishing a floor system for monetary policy following the global financial crisis, intervention in the overnight repo market — on either side of the market — proved important for achieving the Fed's overnight interest rate target. So, un-der the proposed system, the Fed would choose the size of its balance sheet and the target for the overnight interest rate, and then the two standing facilities would look after the rest. Thus, the Fed could con-duct balance sheet policy independent of interest rate policy.

Such a system appears to solve some of the problems discussed by Goodfriend and Lacker. In particular, discretion would be removed from central bank lending, which would be done at arm's length through third parties. Thus, lending would be limited, and any lending to troubled banks would be made on the same terms as by private repo market lenders. Perhaps a defect of such a system is that collateral would be restricted relative to what is normally acceptable for discount window lending, although this might be what Goodfriend and Lacker had in mind.

The Standing Repo Facility, established by the Fed in July 2021, is related to this proposal in that it increases the likelihood of regular central bank lending. However, lending through this facility is currently at the same rate as the discount rate, which is set above the interest rate on reserves, whereas our proposal would have lending at the policy rate. Note also that the Standing Repo Facility accepts only a restricted set of collateral — Treasuries, agency securities, and agency mortgage-backed securities.

Another simplified approach would be narrow banking. Proposals for narrow banking have existed since at least the 1930s, as put for-ward by the Chicago Banking School.⁴ A key proponent of narrow banking was Milton Friedman, who argued in 1960 that monetary control would be improved if all private transactions accounts were backed 100 percent by reserves. In general, a narrow bank is a financial intermediary that backs all liabilities used as means of payment with safe assets, typically central bank reserves or safe government debt. Such safe backing could be a legal restriction or it could be an unconstrained choice of the bank. A useful survey of narrow banking proposals is in Pennacchi (2012).

Generally, narrow banking separates money from credit. For example, under a narrow banking proposal requiring that all means-of-payment liabilities be backed 100 percent by central bank reserves and government debt, much of the structure of banking regulation could be eliminated. There would be no need for capital requirements, leverage requirements, or deposit insurance, for example, as all means-of-payment liabilities of banks would be essentially risk free. There would of course be other financial intermediaries holding risky asset portfolios but, according to narrow-banking proponents, the liabilities of such institutions would be efficiently priced and not subject to flights to safety.

There are two standard issues with narrow banking. One is that a narrow banking structure potentially increases the demand for safe assets, in that banking would no longer be about transforming risky assets into safe ones, but guaranteeing the safety of bank deposits by backing those deposits with safe assets only. This would be particularly problematic in the current environment, in which the world is suffering a scarcity of safe assets.⁵ The second issue is that narrow banking potentially causes disintermediation effects in crowding out the risky assets that are currently held by regulated banks.

In addition, narrow banking (as well as the simplified, repo market approach to monetary policy outlined previously in this section) would not solve all the problems discussed by Goodfriend and Lacker. In particular, if the central bank is still permitted to lend outside the banking sector, there is nothing in the proposal to prevent the Fed from lending excessively to financial institutions deemed to be systemically important. Anticipating that, those systemically important financial institutions will behave in suboptimal ways.

Both proposals in this section have the flavor of the ideas in Good-friend and King (1988), who argue that most of the goals of the central bank can be accomplished through conventional monetary policy. Goodfriend and King (1988), for example, cast doubt on the value of crisis intervention, arguing that such lending is prone to moral hazard problems, and thus dominated by indirect injections of liquidity in a crisis, through open market operations.

Conclusion

Much of Marvin Goodfriend's work was both innovative and prescient. That certainly applies to his work with Jeff Lacker in 1999. Nine years before the financial crisis, Marvin and Jeff grappled with issues of moral hazard and central bank intervention that would be key to how the crisis unfolded in 2008-09. They may not have seen the global financial crisis coming, but their analysis helped provide important background for policymakers during the crisis and afterward.

References

Andolfatto, David, and Stephen Williamson. 2015. "Scarcity of Safe Assets, Inflation, and the Policy Trap." Journal of Monetary Economics 73 (July): 70-92.

Bagehot, Walter. 1873. Lombard Street: A Description of the Money Market. New York: Scribner, Armstrong & Co.

Bernanke, Ben. 2017. The Courage to Act: A Memoir of a Crisis and its Aftermath. New York: W.W. Norton & Co.

Diamond, Douglas W., and Philip H. Dybvig. 1983. "Bank Runs, Deposit Insurance, and Liquidity." Journal of Political Economy 91, no. 3 (June): 401-419.

Ennis, Huberto M. 2019. "Interventions in Markets with Adverse Se-lection: Implications for Discount Window Stigma." Journal of Money, Credit and Banking 51, no. 7 (October): 1737-1764.

Fisher, Irving. 1935. 100% Money. New York: Adelphi.

Friedman, Milton. 1960. A Program for Monetary Stability. New York: Fordham University Press.

Friedman, Milton, and Anna Schwartz. 1963. A Monetary History of the United States, 1867-1960. Princeton: Princeton University Press.

Goodfriend, Marvin. 2011. "Central Banking in the Credit Turmoil: An Assessment of Federal Reserve Practice." Journal of Monetary Economics 58, no. 1 (January): 1-12.

Goodfriend, Marvin, and Robert G. King. 1988. "Financial Deregulation, Monetary Policy, and Central Banking." Federal Reserve Bank of Richmond Economic Review May/June: 3-22.

Goodfriend, Marvin, and Jeffrey Lacker. 1999. "Limited Commitment and Central Bank Lending." Federal Reserve Bank of Richmond Economic Quarterly 85, no. 4 (Fall): 1-27.

Hart, A.G. 1935. "The 'Chicago Plan' of Banking Reform: A Proposal for Making Monetary Management Effective in the United States." Review of Economic Studies 2, no. 2 (February): 104-116.

Pennacchi, George. 2012. "Narrow Banking." Annual Review of Financial Economics 4 (October): 1-34.

Cite as:

Athreya, Kartik, and Stephen D. Williamson. 2022. "Central Bank Learning and Incentives." In Essays in Honor of Marvin Goodfriend: Economist and Central Banker, edited by Robert G. King and Alexander L. Wolman. Richmond, Va.: Federal Reserve Bank of Richmond.