On Lender of Last Resort facilities. By A J R, contributing blogger.
Nov 28, 2009
As is elegantly summarized in Alexander, Dhumale and Eatwell [ADhE 2006], to manage risks, especially of systemic nature ("the awkward tendency for financial crises to cluster over time" [S 2000]), regulators choose from a "crisis management menu" [T 2009] of ex ante and ex post measures. Among the former we can find:
capital adequacy requirements - e.g., the Basel I and II frameworks;
large exposure limits;
limitations on lending
The ex post measures they mention are deposit insurance and the lender of last resort (LOLR) function. Basically, "[t]hese regulatory measures compose the main framework of bank prudential regulation."
A detailed list of steps to be taken after weaknesses are detected after ex ante measures, but before direct support is warranted by the bank being in a distress situation, and of course before closure or liquidation, is printed in Figure 1 of BCBS "Supervisory Guidance on Dealing with Weak Banks," p. 5 [BCBS 2002]. We show several we'll mention later:
Cash (equity) injection by shareholders (*)
Suspension of shareholders' rights, including voting rights
Prohibition on distribution of profits or other withdrawals by shareholders
Removal of directors and managers
Immediate and/or enhanced provisioning of doubtful assets (*)
Stop principal or interest repayments on subordinated debt
Appointment of an administrator
Most of the measures are designed to reduce the hemorrhage of money in case of weakness and to help the bank to strengthen itself. But the two measures highlighted with asterisk are direct ways (if successful) to avoid invoking the LOLR facilities.
How it differs LOLR support from those weak banks measures and other ex post ones? Which is the nature of lender of last resort facilities?
Definition, purpose of LOLR support
Freixas et alii [FGHS 1999] define LOLR as
the discretionary provision of liquidity to a financial institution (or the market [...]) by the central bank in reaction to an adverse shock which causes an abnormal increase in demand for liquidity which cannot be met from an alternative source.
This discretionary element that [FGHS 1999] mentions is important, and is used in the guidance to deal with weak banks [BCBS 2002]. It is obvious the need to keep awake in the night those who may need funds. If the LOLR function were automatic, moral hazard would be even greater than it already is, with the repeated bail-outs of unsecured creditors, the Treasuries support to any amount in individuals' accounts, and the more easy criteria for acceptance to be a member of the too-big-to-fail category.
Humphrey [H 1989], focusing in the system, sees the LOLR role as defined by Thornton and Bagehot [B 1873] as valid back in 1989:
…(1) to protect the money stock, (2) to support the whole financial system rather than individual financial institutions, (3) to behave consistently with the longer-run objective of stable money growth, and (4) to preannounce its policy in advance of crises so as to remove uncertainty. They also advised the LLR to let insolvent institutions fail, to lend to creditworthy institutions only, to charge penalty rates, and to require good collateral. Such rules they thought would minimize problems of moral hazard and remove bankers’ incentives to take undue risks.
These precepts, though honored in the breach as well as in the observance, continue to serve as a benchmark and model for central bank policy today.
So we can say that until 1988 at least, LOLR followed closely the Thornton-Bagehot doctrine. We'll see below today's changing picture.
Aside from these valuable definitions, I'd like to add the following as main differentiating characteristics of this kind of support relative to the measures for weak banks:
LOLR facilities are not designed for weak banks (or financial systems), but for those institutions or systems with liquidity/solvency problems – as such, most of the [BCBS 2002]measures must be tried first.
The many different pathways than can be followed by the regulator are even less clearly determined.
There is a need for high publicity in case of panics, both for just a single institution (to fight "internal descredit" (Bagehot [B 1873]) and in general failure ("to remove instability") – but the [BCBS 2002] measures are recommended to be applied as silently as possible, to prevent doing harm to the bank that is under such intervention.
Transparency of last resort support
Obviously, the last point is not applicable to many of the measures for weak banks. Those printed supra in italics will be highly publicized or at least will be known by rival banks pretty soon. But the intention is to delay as much as possible public knowledge about them. In many cases it is possible to avoid the knowledge the suspension of rights or the prohibition of distribution of profits until a general meeting of shareholders, and that can be almost a year, if you can wait for the next one to pass without intervention.
In the case of other soft ex post measures, it is conceivable to keep them under wraps for some time, supposedly more than a year. In his ‘Emergency Liquidity Facilities’ [DH 2002, p. 124], Dong He details support for a Finnish bank. The last resort help didn't get published until more than a year later. Also, not only the central bank did help, but this wasn't known until it was published with such delay in the central bank bulletin.
This is just a schematic way to explain the differences. In practice, some of the [BCBS 2002]measures that we've referred to as clearly separated in time overlap to some degree. Nothing prevents the regulator from injecting emergency funds, both to the interbank market or to the institutions, and simultaneously adopting many of the [BCBS 2002] measures.
There is few guidance on LOLR facilities if we exclude national laws. Dong He summarizes the relevant information for twenty nations in his paper's appendix, but, as he says, the "study of emergency lending operations is hampered by a general lack of information on country practices in this area".
LOLR support types – by need
We can try to classify LOLR support according to the need of it. By this criterium, it can be invoked mainly by two needs:
because of undercapitalization by day-to-day operations – this is not a contentious point;
because of a panic – when many players "run for the exit at the same time" [W-P 2009] due to the sequential service constraint, which was noted by both Thornton and Bagehot.
[W-P 2009] summarizes many cases of panics in the XIX century and explains how central banks appeared in several countries, and with them the LOLR support in both "emergency lending in normal times" and "in systemic crises", as [DH 2002] calls them. This last support is subject of much discussion today.
Liquidity problems caused by normal, day-to-day operations
If the bank is not suffering "internal discredit", but the regulator, thru routine work, discovers that the financial institution is in distress, which is more common an ocurrence than panics, it may be decided to provide "advances [...] to an immense amount" as Bagehot registered in his book, aside from possible using many of the measures for weak banks we mentioned at the beginning.
Although there is agreement on this, the path is full of perils. Bagehot notes:
But though the rule is clear, the greatest delicacy, the finest and best skilled judgment, are needed to deal [with such affairs].
I've counted twelve times the root "delic-" in his book applied to how fragile and difficult to manage is the system. See References.
Liquidity problems caused by panics – both individual and systemic
Why runs happen even when the bank is not insolvent and can find ways to gather capital at reasonable cost? The work of Diamond and Dybvig [DD 1983] shows that the uncoordinated character of dispersed, uninsured, small depositors can reach what seem several equilibria, among them a bank run, even in such favorable circumstances. Later, Morris & Shin confirmed this [MS 1999]:
Creditors of a distressed borrower face a coordination problem. Even if the fundamentals are sound, fear of premature foreclosure by others may lead to pre-emptive action, undermining the project.
Bagehot offers what it seemed a counterintuitive method to stop a run:
In opposition to what might be at first sight supposed, the best way for the bank or banks who have the custody of the bank reserve to deal with a drain arising from internal discredit, is to lend freely.
He also explains why this solution seems to work:
This discredit means, 'an opinion that you have not got any money,' and to dissipate that opinion, you must, if possible, show that you have money: you must employ it for the public benefit in order that the public may know that you have it. The time for economy and for accumulation is before. A good banker will have accumulated in ordinary times the reserve he is to make use of in extraordinary times.
And later adds, speaking of panics thru an example of successfully solving it (my emphasis):
The holders of the cash reserve must be ready not only to keep it for their own liabilities, but to advance it most freely for the liabilities of others. They must lend to merchants, to minor bankers, to 'this man and that man,' whenever the security is good. In wild periods of alarm, one failure makes many [...]. 'We lent it [...] by every possible means and in modes we had never adopted before; [...] we not only discounted outright, but we made advances on the deposit of bills of exchange to an immense amount, in short, by every possible means consistent with the safety of the Bank [...].' After a day or two of this treatment, the entire panic subsided, and the 'City' was quite calm.
In a way or the other, these considerations keep living in current legislation, at least until Dong He wrote his working paper for the IMF. We can also say that no single country preserves all those prescriptions in writing, there are always subtle changes that well check in the last section, "LOLR support types – by differences in application".
LOLR support types – by differences in application
We can too classify LOLR support according to its differences in application or last resort support.
In some countries there are great limits to the amount of liquidity that can be provided – that is, they do not follow the principle of lending freely. As D. He says, p. 129, Argentina, Hong Kong, Turkey and the Philippines have clear limitations in the size limit of last resort support. Dong He adds that this kind of limits is controversial, as they can encourage runs.
Independence of regulator
In others, there is a need to coordinate the bank with the money reserves with politicians at the executive departments, because of law or because of subtle pressure. In the end, in many cases the much trumpeted independence of the regulator is not fully respected. In the IMF report on Turkey [IMF 2007] we find (details in References):
81. In practice the BRSA lacks full operational independence in regulatory and budgetary matters. Before issuing a regulation [...] must conduct consultations with the related Ministry and the Undersecretariat of the State Planning Organization. While nonbinding consultation is appropriate, the process is perceived as binding, thus raising the possibility of undue government interference. [...] All visits abroad by the BRSA management or staff, such as for on-site inspections or for training, must be approved by the government. [...]
[The situation is not as bad as it sounds in this text selection we made. Turkey made great progress in the years before this report was written, as the report says.]
In some countries, the quality of the collateral doesn't follow Bagehot's criterium (to relax the standard). The Hong Kong Monetary Authority's policy is to accept only high-quality paper. Other countries follow the criterium: e.g., the Bank of Korea may support banks with "the collateral of any assets which are defined temporarily as acceptable security" in some cases. See examples in [DH 2002, p. 126].
Fisher [F 2000] argues that requiring adequate collateral diminishes moral hazard: the prospective borrower would reduce excessive risks because the risky assets won't be accepted as collateral.
Solvent and insolvent institutions
If the firm is just temporarily illiquid but it is solvent, Bagehot recommends to lend freely (although at "very high" interest rates). If the company is insolvent, he recommends to let it fail.
Several issues are raised here:
In the hurry in which these crises are dealt with it is very difficult to know if the institution is solvent or not.
The pressure to help the firm if it is systemically important is enormous, and Bagehot view is increasingly not followed today. See next section.
Impact of recent developments on LOLR principles
Several reasons force our contemporary regulator to provide assistance to institutions that previously would not qualify. Paramount among them are Too Big To Fail (TBTF) issues:
bank failures have direct large costs of liquidation (see abstract of James below [J 1991]), which affect management and shareholders – much more so with big institutions
the payments system is affected – stopping payments by a TBTF bank touches too many economic agents
the interbank market (both as borrower and lender) is greatly affected if such a bank is liquidated
all this helps reduce economic activity – but even worse, the bank knowledge about the customers (how good or bad is someone repaying his debts, his bills, etc.) is lost, further slowing the economy
the Treasury also suffers, since it must support the interbank market and try to soften the situation
compounding this, the size of such a bank can affect the banking system credibility much more than a small bank's failure – system disorders can emerge
Abandoning punitive rates
Rustomjee [R 2009, p. 19] offers two reasons why many "regulators have opted not to provide support at penalty rates":
"charging a higher than market rate could aggravate [...] the bank’s liquidity or solvency challenge"
"public knowledge that a bank is being offered emergency lending at a penalty rate could deepen the loss of confidence in bank management and could itself exacerbate any concerns either by banks that have lent to the weak bank, or depositors, so prompting a bank run."
Market makers of last resort (MMLR)
BoE's Tucker [T 2009], which defines LOLR as putting "a finger in the dyke," comments on the increasing "amount of credit that gets intermediated via markets" rather than via institutions in a speech this month and discusses the need to stand ready to act as MMLRs, which is really complex: When you buy from an individual bank you have some leverage to change things, but once you enter into the market as a buyer (that means potentially a really large number of transactions and counterparties) you cannot later renegotiate the price of the assets if their quality deteriorate.
[ADhE 2006] Alexander K, R Dhumale and J Eatwell /2006): 'Global Governance of the Financial System - The international regulation of systemic risk'. Oxford UK: Oxford University Press.
[B 1873] Bagehot W (1873) Lombard Street: A Description of the Money Market, London: HS King. http://www.econlib.org/library/Bagehot/bagLom.html
But in exact proportion to the power of this system is its delicacy I should har dly say too much if I said its danger.
And this foreign deposit is evidently of a delicate and peculiar nature.
That such an arrangement is strange must be plain; but its strangeness can only be comprehended when we know what the custody of a national banking reserve means, and how delicate and difficult it is.
But though the rule is clear, the greatest delicacy, the finest and best skilled judgment, are needed to deal at once with such great and contrary evils.
And great as is the delicacy of such a problem in all countries, it is far great er in England now than it was or is elsewhere. The strain thrown by a panic on the final bank reserve is proportional to the magnitude of a country's commerce [...].
2ndly. Because, being a one-reserve system, it reduces the spare cash of the Money Market to a smaller amount than any other system, and so makes that market more delicate.
But it is of great importance to point out that our industrial organisation is liable not only to irregnlar external accidents, but likewise to regular internal changes; that these changes make our credit system much more delicate at some times than at others; and that it is the recurrence of these periodical seasons of delicacy which has given rise to the notion that panics come according to a fixed rule, that every ten years or so we must have one of them.
This is the surer to happen that Lombard Street is, as has been shown before, a very delicate market.
In a time when the trading classes were much ruder than they now are, many private bankers possessed variety of knowledge and a delicacy of attainment which would even now be very rare.
So that our one-reserve system of banking combines two evils: first, it makes the demand of the brokers upon the final reserve greater, because under it so many bankers remove so much money from the brokers; and under it also the final reserve is reduced to its minimum point, and the entire system of credit is made more delicate, and more sensitive.
And this is the reason why the Bank of England ought, I think, to deal most cautiously and delicately with their banking deposits.
We must therefore, I think, have recourse to feeble and humble palliatives such as I have suggested. With good sense, good judgment, and good care, I have no doubt that they may be enough. But I have written in vain if I require to say now that the problem is delicate, that the solution is varying and difficult, and that the result is inestimable to us all.
[BCBS 2002] Basel Committee on Banking Supervision (2002): ‘Supervisory Guidance on Dealing with Weak Banks’, Basel: Bank for International Settlements.
[DD 1983] Diamond D and P Dybvig (1983) ‘Bank Runs, Deposit Insurance and Liquidity’, Journal of Political Economy, Vol. 91, pp. 401–19.
[DH 2002] Dong He (2002) ‘Emergency Liquidity Facilities’, Chapter 5 of C Enoch, D Marston and M Taylor, Building Strong Banks through Surveillance and Resolution, Washington DC: International Monetary Fund.
With the same title as a working paper (2000), Washington, DC: IMF.
[F 2000] Fischer S (2000) 'On the Need for an International Lender of Last Resort ' Essays in international economics, no. 220. Princeton University. http://www.princeton.edu/~ies/IES_Essays/E220.pdf
[FGHS 1999] Freixas X, C Giannini, G Hoggarth and F Soussa (1999) ‘Lender of Last Resort: A Review of the Literature’, Bank of England, Financial Stability Review, November. http://www.bankofengland.co.uk/publications/fsr/1999/fsr07art6.pdf
[H 1989] Humphrey T (1989) ‘The Lender of Last Resort: The Concept in History’, Federal Reserve Bank of Richmond Economic Review, March/April, pp. 8-16. http://www.richmondfed.org/publications/research/economic_review/1989/er750202.cfm
[IMF 2007] International Monetary Fund (2007) ‘Turkey: Financial System Stability Assessment’, IMF Country Report Number 07/361, November, Washington DC: IMF.
62. The 2001 amendment of the CBRT Law gave the CBRT [the central bank] substantially greater independence in the implementation of monetary policy and freed it of any obligation to finance the public sector. [...] However, a lacuna in the CBRT Law is lack of specificity concerning the removal from office of members of the governing bodies [...].
81. In practice the BRSA [very much like the FSA] lacks full operational independence in regulatory and budgetary matters. Before issuing a regulation, the BRSA must conduct consultations with the related Ministry and the Undersecretariat of the State Planning Organization. While nonbinding consultation is appropriate, the process is perceived as binding, thus raising the possibility of undue government interference. Moreover, the related Ministry is legally empowered to file a lawsuit for the cancellation of the Board’s regulatory decision. [...] All visits abroad by the BRSA management or staff, such as for on-site inspections or for training, must be approved by the government. [...]
125. The CMB [the Capital Markets Board] has adequate funding [...]. When markets are inactive, the potential shortfall must be covered by a budgetary transfer from the Ministry of Finance, which could have a bearing on its independence. [...]
[J 1991] James C (1991) ‘The Losses Realised in Bank Failures’, Journal of Finance, September, pp 1223–42. James-TheLossesRealisedinBankFailures1991.pdf
This paper examines the losses realized in bank failures. Losses are measured as the difference between the book value of assets and the recovery value net of the direct expenses associated with the failure. I find the loss on assets is substantial, averaging 30 percent of the failed bank's assets. Direct expenses associated with bank closures average 10 percent of assets. An empirical analysis of the determinants of these losses reveals a significant difference in the value of assets retained by the FDIC and similar assets assumed by acquiring banks.
[MS 1999] Stephen Morris and Hyun Song Shin: Coordination Risk and the Price of Debt. Cowles Foundation Discussion Paper no. 1241. December 1999
[R 2009] Rustomjee C (2009) 'Bank Regulation and the Resolution of Banking Crises, Unit 2'. London: School of Oriental and Management Studies-CeFiMS.
[S 2000] Sinclair, Peter JN (2000) ‘Central Banks and Financial Stability’, Bank of England Quarterly Bulletin, November: 377–89. http://www.bankofengland.co.uk/publications/quarterlybulletin/qb000403.pdf
[T 2009] Paul Tucker (2009) 'The crisis management menu'. Speech by Mr Paul Tucker, Deputy Governor for Financial Stability at the Bank of England, at the SUERF, CEPS and Belgian Financial Forum Conference: "Crisis Management at the Cross-Roads", Brussels, November 16, 2009.
[W-P 2009] Wickman-Parak B (2009): Financial stability in focus. Speech by Ms Barbro Wickman-Parak, Deputy Governor of the Sveriges Riksbank, at the Swedish Chambers, Gothenburg, October 29, 2009. Wickman-Parak-FinancialstabilityinfocusOct292009.pdf