Saturday, September 29, 2007

Redesigning the UK Financial Stability System

A pdf file of a Powerpoint presentation titled "What should the authorities have done?", prepared for The London Financial Regulation Seminar, 'The Financial Crisis Conference' on October 1, 2007 at the London School of Economics, can be found here.

The mess surrounding the rescue operation for Northern Rock demonstrated that the UK’s Tripartite arrangement for handling financial crises is not working properly.

There are three distinct sets of problems. First, the UK deposit insurance scheme is both limited in the degree to which it guarantees retail deposits and too slow in paying out on any claims submitted under the scheme. Second, the division of labour among the Treasury, the Bank of England and the Financial Services Authority, as expressed in the Memorandum of Understanding, was disfunctional, mainly because it separated the agency in possession of the relevant information from the financial resources to act effectively upon that information. Third, the Bank of England’s liquidity-oriented open market operations through repos, and its discount window are flawed in three ways: first, the eligible collateral is too restricted; second, the maturity of the operations (loans) is too short; and, third, the list of eligible counterparties is too restricted.

Specifically, five issues can be raised about the current set of arrangements:

(1) The UK deposit insurance arrangements did not work properly.

(2) The lender of last resort (LOLR) mechanism for dealing with individual financial institutions in distress did not work properly.

(3) The Bank of England’s Standing Lending Facility (its discount window) did not work properly.

(4) The Bank of England’s liquidity-enhancing open market operations did not work properly.

(5) The financial stability mess, and the Bank’s about face as regards the collateral requirements and the maturity of its liquidity-enhancing open market operations have created confusion about exactly what it is the Monetary Policy Committee decides on when it sets the official policy rate, or Bank Rate.

All these points will be considered in what follows.

Deposit insurance This needs to be overhauled to provide guaranteed 100 percent cover up to, say, £50,000.00 per person per institution. This would correspond to the level of coverage currently in effect in the USA ($100,000). The £100,000.00 figure that has been bandied about seems excessive. The same person could have accounts in different institutions. Provided these institutions are indeed separate legal entities, the same person could have £50,000.00 insurance cover in each one of number of separate banks.

The insurance should be for retail accounts only. Wholesale deposits would not be covered. A simple rule could be that only deposits owned by natural persons would be covered. Deposits of entities with legal but not natural personality (partnerships, charities, companies etc.) would not be covered.

Apparently, the Chancellor’s deposit guarantee for Northern Rock covers not only all deposits (retail and wholesale), but most other unsecured creditors of Northern Rock as well. Only holders of subordinated debt appear not to be covered. This degree of coverage is ludicrously excessive. Such ex-post insurance and socialisation of investment risk by municipalities, charities and other institutions who were chasing the above-market rates offered by Northern Rock is without justification on equity, efficiency or systemic stability grounds. For these non-widows and non-orphans, the lesson that above-market returns often represent risk premia, and that risk has the unpleasant habit of materialising from time to time, would have been a highly salutary one.

The deposit guarantee scheme should be able to pay out on claims effectively instantaneously, and certainly no longer than 2 working days after a claim has been submitted. The current situation where a bank that goes into administration has its deposits frozen, clearly has to be addressed with legislation.

Deposit insurance, or any form of consumer protection, should not be the responsibility of the monetary authority. The FSA would be a natural body for administering it. The scheme should be self-financing, through levies on the deposit-taking industry. Should there be widespread insolvency in the banking sector, the financial resources to meet the deposit insurance guarantee might exceed the combined resources of the deposit-taking institutions. For such system-wide calamities, a fiscal back-stop would be required. One easy way to do this is to give the deposit insurance agency an overdraft facility with the central bank (the Bank of England), guaranteed by the Treasury.

Liquidity provision

The UK’s arrangements for dealing with illiquid institutions and illiquid markets are a shambles.

Open market operations The Bank has to extend its recently announced policy of providing liquidity to the markets at maturities longer than overnight and against a wider range of collateral. It should effectively adopt the policies of the ECB and the Eurosystem, which accepts as collateral in repos (overnight and at longer maturities), private instruments, including illiquid and non-marketable instruments, as long as they are rated at least in the A category. Clearly, intervening in the markets at the same time in different maturities makes no sense when markets are orderly; when markets are disorderly, however, there may be extraordinary liquidity premia at different maturities (on top of the regular term premia, conventional market risk or default risk premia and expectations of future changes in the policy rate) that can be influenced both by repos at varying maturities and by outright purchases of securities with differing remaining maturities. Such open market purchases or outright purchases should not be at penalty rates. This would blur the distinction between open market operations and discount window operations. The Bank of England’s 3 month repos against illiquid collateral (mortgages and mortgage-backed securities) (the first of which did not attract any takers) was therefore in my view a mistake.

The discount window

For illiquid but solvent institutions, the discount window at the Bank of England (its Standing Lending Facility) has to be the port of call. As currently constituted, the Bank’s Standing Lending Facility is useless: its list of eligible collateral is too restrictive - all it does is trade longer-maturity liquid assets for instantaneous liquidity; it only lends overnight; and it only lends to banks. This discount window should be modified in three ways.

(1) The Bank should create a wider range of eligible collateral. The Bank of England should accept as collateral at its discount window (Standing Lending Facility) private securities, including illiquid and non-marketable private securities. Their market price or fair value would be subject to a ‘haircut’ that would be larger to the more illiquid the collateral that is offered. Where no market price is available, the Bank should ‘make market’ for the illiquid securities, by holding auctions in which it purchases these securities; it should then hold them on its books, taking the credit risk, until they can be sold off again under more orderly market conditions, preferably at a profit for the tax payer. To encourage participation by investors who do not want to mark-to-market their securities, the Bank and FSA could require that all similar securities not priced at the auction be marked-to-market at the price established in the auction.

(2) The Standing lending Facility should offer longer maturity loans than overnight. The Fed already offers up to 1 month maturity loans. I see no reason why both 1 month and 3 month collateralised loans could not be offered at the discount window. The penalty rate (current 100 basis points for overnight loans) could be made to increase with the maturity of the loan, say 150 basis points for 1 month maturity and 200 basis points for 3 month maturity. The Fed’s decision to cut the penalty premium of the discount rate over the policy rate from 100 basis points to 50 basis points was a big mistake – pandering to the profits of those banks willing and able to borrow at the discount window.

(3) There should be a wider range of eligible counterparties. In the UK the Standing Lending Facility of open only to a limited number of banks and other deposit-taking institutions. I would favour widening this to all financial institutions that are subject to and meet the demands of, a regulatory and supervisory regime approved by the Bank. This could include investment banks, hedge funds and private equity funds. The Fed can do much more than that, and can, in principle, open its discount window to individuals, partnerships and corporations (financial and non-financial).

With (1), (2) and (3) in place, it is clear that the Standing Lending Facility, which is open to all eligible institutions on demand, and for any amount of funding for which they can provide eligible collateral, is not so much part of the traditional lender of last resort arsenal, which is targeted at specific institutions that are in trouble, but instead is a form of market support, specifically support for markets trading normally liquid securities that have become illiquid.

The Lender of Last Resort

Institutions that are insolvent as well as illiquid should not be bailed out unless they are deemed to be systemically important. It is hard to think of any bank in the UK that would be systemically signficant, once adequate deposit insurance removes the risk of bank runs. Bail-outs should therefore be for two kinds of institutions: those that are illiquid and about whose solvency there is some uncertainty and those that are insolvent and systemically important.

The decision on whether to bail out the institution should be made by the regulator (the FSA), which has the institution-specific knowledge and information, and the Treasury, which has the resources. The Bank of England’s input will not doubt be required, as it is the systemic significance of an individual institution or set of institutions that is at stake, and the primary responsibility for and understanding of systemic risk is presumably found in the Bank.

Should the Standing Committee on Financial Stability (chaired by the Treasury, with representatives of the Treasury, the Bank and the FSA) decide that a specific institution needs to be bailed out, there are a number of options.

For a bank that is illiquid and perhaps insolvent, the kind of dedicated lending facility created for Northern Rock would be appropriate. It wouldn’t be named a “liquidity support facility”, since more than an injection of liquidity may be required. It is clear that such a LORL facility targeted at an individual institution should not be initiated or managed by the Bank. The Bank does not have the information about individual institutions. It should not be asked to take decisions about individual institutions. Nor should it be required to put its resources at risk.

The institution that should decide who gets a LOLR facility has to be the regulator and supervisor, that is, the FSA, because only the FSA has the necessary information. It does not, however, at the moment have the financial resources to act as LOLR. It should therefore be given the resources to fulfil the LOLR function. These resources can only come from the Treasury. Operationally, this could be done conveniently through a credit line or overdraft facility of the FSA with the Bank (uncapped and open-ended), guaranteed by the Treasury. The Bank’s role in the LOLR function vis-à-vis individual banks is therefore entirely passive. The decision is made by the FSA and the Treasury and the funds are provided through the FSA by the Treasury. There would be a presumption that the existing management of a bank in need of an LOLR facility would be fired, sans golden parachute.

Insolvency and public ownership

The facility would be operated until it is clear to the FSA whether the troubled bank is insolvent or not. If it is solvent, it is weaned off the facility. If it is deemed to be insolvent and systemically insignificant, it goes into the normal (hopefully revised) insolvency procedures for banks. If it is insolvent and deemed to be systemically important, it is taken into public ownership, rather like Railtrack was. Once in public ownership, the bank could continue to operate and meet its existing commitments.

The nationalisation would be temporary. Once orderly conditions have been restored and the value of the bank’s assets and liabilities has been established, the bank can be privatised again as a going concern, sold off in toto to its competitors or liquidated, broken up and sold in parts. The incumbent management would presumably lose their jobs as soon as the bank was taken into public ownership, if they had not already lost it while the bank was using the LOLR facility (if it went through the LOLR process earlier on, because its insolvency was then not yet obvious). After compensating itself for the cost of the LOLR facility, the FSA would pay the creditors of the bank, including those depositors who were not covered by the deposit insurance scheme. The original shareholders of the bank would be last in line and would, if the bank was indeed insolvent, receive nothing.

The schematic below shows how illiquid institutions would be dealt with in a new, improved Tripartite Arrangement. OMOs + and Discount Window + stand for the augmented open market operations and discount window operations I advocated earlier (wider set of eligible collateral, longer maturities, wider set of eligible counterparties at the discount window and in OMOs.)

The overdraft facility or credit line with the Bank of England, guaranteed by the Treasury, which the FSA would have, could serve both to finance its LOLR activities and its deposit insurance activities, were these to exceed the financing capacity of the banking system.

What does the MPC set when it sets Bank Rate?

The monetary policy committee of the Bank of England sets Bank Rate. What is Bank Rate? Under the current set of arrangements for implementing monetary policy, the Official Policy Rate, or Bank Rate, is the target rate for the overnight sterling interbank rate (also called the sterling money market rate). This target is pursued through the sale and purchase of ‘repurchase agreements’ (repos and reverse repos). To the layman these are collateralised loans.

You will be forgiven for wondering why, if the MPC sets the target rate for the overnight rate, the actual overnight rate in the interbank market can ever differ from that rate. Basically, there are two reasons for any discrepancy between Bank Rate and the overnight interbank rate. The first is that the Bank does not rigorously fix the repo rate every minute of very day. They could do this. They could simply stand ready to repo or reverse repo at any time any amount the private sector wants to throw at them. They don’t do that. Instead they inject a certain amount of repos or reverse repos into the market – and amount they expect will meet the normal market demand, and wait to see what happens. Why they do this, I do not know. I think it would be helpful if they simply pegged the repo rate by standing ready to buy or sell in any amount at that rate.

The second reason why the actual overnight interbank rate can differ from Bank rate, is that the overnight repo rate can differ from the overnight interbank rate. This is simply because the overnight interbank rate is a rate on unsecured lending, why the repo rate is a rate on a collateralised, secured loan. When the Bank of England expands liquidity through a repo, the loan to the private sector is almost free of default risk. Both the borrowing bank and the issuer of the collateral would have to default for the Bank to be exposed to counterparty risk through the repo. When a bank lends to another bank overnight in the interbank market, there always is a small probability that the borrowing bank will fail overnight.

Once there is counterparty risk, illiquidity risk becomes a possibility. So there can be a gap between the overnight repo rate and the overnight interbank rate because of market perceptions of default risk and illiquidity risk.

It would clarify the division of labour between the Bank of England’s Monetary Policy Committee and those in the Bank of England responsible for market operations and financial stability, if the Bank were to give the highest priority to its task as agent for the MPC, by keeping the overnight interbank rate as close as possible to Bank Rate. I would start by pegging the overnight repo rate, undertaking repos or reverse repos in any amount required to keep the market repo rate equal to Bank Rate throughout the maintenance period. Market perceptions of overnight default risk is, of course, not something the Bank of England should try to do anything about. Overnight illiquidity can, however, be addressed, by injecting additional liquidity into the repo market, over and above what is required to keep the overnight repo rate at the level of Bank Rate, up to the point that the overnight interbank rate, minus the market’s counterparty risk premium, is at the level of Bank Rate. It is possible that this requires the overnight repo rate to be below Bank Rate. So be it.

The Bank’s operations in the money markets at maturities longer than overnight, and the Bank’s Standing Lending Facility operations are not part of the remit of the MPC.

No comments: