Where the Bank of England went wrong, and how to prevent a recurrence

The joy of pointing fingers and apportioning blame
There is now widespread agreement on how the collective actions and inaction of the three parties to the UK’s Tripartite Arrangement for dealing with financial stability crises contributed to the Northern Rock debacle.  There is also reasonable clarity about the individual mistakes and misjudgements of two of the three parties – the Treasury and the Financial Services Authority (the regulator).  There is less agreement on what errors the Bank of England has made.

Collectively, the Treasury, the Bank and the FSA failed to act, speak and appear jointly, cohesively and harmoniously, to reassure  the markets, depositors and the public at large. The absence of a joint appearance,  when the rescue facility for Northern Rock was announced, by the Prime Minister, who as former Chancellor created the Tripartite arrangement, the current Chancellor, the Governor of the Bank of England, the Chairman and the Chief Executive of the FSA , is incomprehensible. The continued sniping at the Bank since that day by the Treasury and the FSA, and the repeated covert briefing against the Bank by the Treasury further undermined public confidence in the ability of the authorities to manage the nation’s financial stability arrangements.

The FSA was asleep on the job.  The Treasury was responsible for the Memorandum of Understanding between the three parties.  This MoU institutionally separated information about individual banks, including their liquidity positions, from the financial means to address any problems faced by illiquid but solvent banks.  The Treasury was also responsible for the inadequate, indeed bank-run-inviting, state of the UK deposit insurance scheme. 

And the Treasury decided, even after the risk of a run on Northern Rock’s deposits had been eliminated, that the Liquidity Support Facility created at the Bank of England to lend to Northern Rock, at a penalty rate, against its illiquid collateral, would be kept in place until a private buyer for all or part of Northern Rock or its assets could be found.  The alternatives of (1) switching off the life-support system and letting Northern Rock live or die on the strength of its own resources, or (2) taking it into public ownership and re-privatising it or selling its assets when market conditions normalised, were not considered. 

Of these two options, letting Northern Rock sink or swim on its own, with no special access to public resources, would have been my preferred option.  This is because, with the risk of bank run contagion removed, Northern Rock (with between 2 and 3 percent of the assets of UK-registered banks) was small enough to fail; its failure would not have adverse implications for financial stability in the short run and would have a mightily positive impact on financial stability in the long run, because the cautionary tale of Northern Rock this would change the incentives facing banks inclined to excessive risk taking.

The Bank of England as LoLR and MMLR
The Bank of England’s failed as a provider of liquidity.  It failed in two distinct tasks; as a lender of last resort (LoLR) and as a market maker of last resort (MMLR). It is  key to distinguish between these two functions. (see my earlier contributions on the subject, many joint writings with Anne Sibert at (1), (2), (3), (4), (5).) The  lender of last resort provides liquidity to individual banks and other individual financial institutions in distress.  Following Bagehot it does so on demand, at a penalty rate and against collateral that would be good in normal times, when markets are orderly, but that may have become illiquid in abnormal times because of disorderly markets.  The market maker of last resort provides liquidity to markets for financial instruments that are normally liquid but have become disorderly and illiquid.  It does so either by purchasing the potentially illiquid assets outright or by accepting them as collateral in ‘sale and repurchase agreements’ or repos.  Since it corrects a market failure – illiquidity of a market is a public bad -  there is no need to impose a penalty rate on repos used to mitigate such illiquidity.  No adverse incentives (‘moral hazard’) are created as long as the securities that are offered as collateral are priced fairly and subject to appropriate liquidity ‘haircuts’ or discounts on these fair valuations.  I discuss how such fair valuations can be established when the securities offered as collateral do not have a relevant market reference price or benchmark because they are or have become illiquid.  The Bank of England mismanaged both the lender of last resort function and the market maker of last resort function.

The lender of last resort
As regards the LoLR function, the Bank of England should have lent to Northern Rock at a penalty rate, and against normally good collateral, provided the institution was illiquid but solvent.  While I have no first-hand knowledge of the asset side of Northern Rock’s portfolio, those who do (or who assert they do) assure us that most of its assets are sound.  Northern Rock should have been able to access the Bank’s LoLR facilities in that case.  The discount window of a central bank (called the Standing (collateralised) lending facility in the case of the Bank of England, the Marginal Lending Facility in the case of the ECB and the primary discount window in the case of the Federal Reserve System is the place where lender of last resort funds should be provided.

The discount window of the Bank of England is, however,  ineffective as a source of liquidity for two reasons.  First, it only provides overnight loans.  The Fed lends at up to one month maturity at its primary discount window.  Second and most importantly, the Bank of England only accepts as collateral securities that are already liquid: UK government securities, sovereign debt instruments issued by other European Economic Area governments, high-rated debt instruments of a few international organisations (e.g. the World Bank), and, under exceptional circumstances, US Treasury securities.  So the Bank of England’s discount window only does maturity transformation of long-maturity liquid assets into overnight liquidity.  This is useless in a crisis.  The ECB accepts as collateral at its discount window euro-denominated debt instruments (rated at least in the A category) issued by both sovereign and private issuers.  It can accept nonmarketable and illiquid assets, including asset-backed securities.  The Fed can accept as collateral at its discount window anything it deems fit, including cats and dogs. 

Stuck with an underpowered discount facility, the Bank of England and the Treasury had to create the purpose-designed Liquidity Support Facility for Northern Rock to be able to lend to Northern Rock.  In the Euro area and in the USA, Northern Rock would have been able to use its prime mortgages as collateral at the discount window.  Indeed, if Northern Rock had been more on the ball, it could have used its Irish subsidiary (which is in the Euro Area) to borrow at the ECB’s discount window, or to obtain liquidity in the ECB’s liquidity-oriented repos.  The Liquidity Support Facility would have been redundant  had there been a properly designed collateral and maturity policy at the Bank of England’s discount window.  The discount window is, in principle, anonymous.  The Liquidity Support Facility was in the public domain, because of the mistaken belief of the Governor that undercover, secret LoLR support for Northern Rock would have violated the Market Abuse Directive. That, of course, is nonsense, and it was contradicted the very afternoon the MAD assertion was made, by a spokesman for the European Commission.  The ECB is reported to have provided undercover support to troubled Eurozone banks without creating any MAD problems.  Obviously, the MAD was not created to rule out discrete, secret LoLR support where it is needed to limit unnecessary damage.

The market maker of last resort
The Bank’s provision of liquidity to disorderly financial markets was also flawed, and continues to be so.  Again, it accepted (until it changed its mind after Northern Rock hit the rocks and interbank markets remained in turmoil – as they continue to be to some extent especially at 3-month maturities; yesterday (November 16), the spread of 3-month sterling Libor over the market’s estimate of the UK official policy rate (Bank Rate) over the next three months, as measured by the Overnight Indexed Swap (OIS) swap rate one again crawled up to around 75 basis points.  The Bank of England accepts as collateral in repos only the same short list of highly liquid and highly rated securities that it accepts at its discount window.  The menu of securities it is willing to purchase outright in open market purchases is equally restricted.  The Bank only sets the overnight repo rate (using its restricted list of collateral) and used to intervene at longer maturities only ‘at market rates’, not seeking to influence market rates at longer maturities than overnight.  It attributed the gap between, say, 3-month Libor and the expected policy rate over that three month horizon as reflecting market participants perceptions of default risk.  The Bank did not consider it its business to address default risk of UK-registered banks through its repos and open market operations, and I fully concur.

However, to identify the gap between 3-month Libor and the expected policy rate over a three month horizon as reflecting mainly default risk represents a serious error.  That spread is the sum of the default risk premium and the liquidity risk premium.  In my view most of the spread (at least 60 basis points of yesterday’s 75 basis point spread, say), represents the market’s perception of liquidity risk.  There is a term structure of  liquidity risk which the Bank can and should try to address by intervening aggressively at longer maturities in the interbank markets (through repos at the appropriate longer maturities) to drive down the excessive spreads.  It should do so accepting as collateral in its longer-maturity repos a wide range of private sector financial instruments, including illiquid and nonmarketed instruments.  It announced it would do so in September, but the 3 interventions it announced at a 3-month maturity were subject to a penalty floor (the interest rate could not be below the policy rate plus 100 basis points).  Nobody came to the party.

The fact that the Bank made its proposed market support operation subject to a penalty floor shows it still does not get the distinction between assisting an individual institution that is in trouble and supporting an illiquid market or financial instrument instrument.  LoLR operations have to be at a penalty rate to mitigate moral hazard.  MMLR operations correct a market failure.  Illiquidity is a manifestation of a breakdown if trust and confidence.  Liquidity is a public good, subject to ‘intertemporal network externalities’ – even when I have liquid resources today in excess of my own requirements, my willingness to lend these resources to you today at three month maturity, depends on my perception of the risk that both you and I will be illiquid three months from now.  If I expect both of us to be illiquid three months from you, I will not lend to you today.  So fear of future illiquidity supports an inefficient equilibrium with lack of liquidity today.

If the authorities could credibly commit themselves today to provide unlimited liquidity in three months’ time, such liquidity crises would not arise.  But such a commitment is impossible.  The best substitute is for the monetary authority to provide the 3-month liquidity today.  The Bank of England is the only agency whose sterling liabilities have unquestioned and unconditional liquidity.  They can provide this liquidity instantaneously, continuously and at no cost. They should do so; they did not. So the Bank failed in its market maker of last resort role.  It is possible, but socially inefficient for private banks and other financial institutions to carry on their own balance sheets enough liquid assets to weather virtually any market illiquidity storm.  They should have enough liquid assets to manage their affairs when markets are orderly.  It is the job of the monetary authority to prevent or mitigate market illiquidity.  This permits banks and other financial institutions to perform their valuable social functions of borrowing short and lending long, and of borrowing in markets that are normally but not always liquid, and lending in the form of assets that are often illiquid.

To avoid creating adverse incentives for the would-be borrowers in the repos, borrowers who might offer illiquid, nonmarketable instruments as collateral, it is key that the collateral be priced in such a way as not to offer a subsidy to the borrowers.  The Bank need not have superior information about what mortgages are worth, or about the fair value of the underlying assets backing asset-backed securities that might be offered as collateral.  It can organise auctions that serve as reservation-price revelation mechanisms for the parties offering the illiquid securities.  The reverse Dutch auction, with the Bank acting as monopoly buyer, is one example.  To discourage unnecessary complexity in structured products, the Bank should have a positive list of financial instruments that it will make market in.  To avoid a situation where private holders of illiquid instruments prefer not to have them valued at these auctions, there should be a regulation that all securities in a class for which there has been an auction, should be marked-to-market at the prices established in the auction (in the case of the reverse Dutch auction, this could even be at something close to the highest  price in the sequence of prices established in the auction; as the sole buyer in the auction, the central bank should get enough of an advantage from its monopsony power to ensure that the tax payer would not lose on these transactions).

The Bank should carry these illiquid securities on its balance sheet.  If, subsequently, a transparent normal market is re-established, the Bank could sell the securities again.  Otherwise it can hold them till maturity.  It never has to develop an informed view as to how much these illiquid assets are worth.  After all, the Bank is never liquidity-constrained and is uniquely well-placed to take the long view.

So the Bank failed in its lender of last resort function and in its market support or market maker of last resort function, because it operated against too restrictive a list of collateral and at too short a maturity.

Quo Vadis, Northern Rock?
Of course, Northern Rock knew, or should have known, about the collateral policy of the Bank of England, both at the discount window and in its liquidity-oriented market interventions.  Northern Rock had to live in the real world, and its funding policy, and its low liquid asset holdings were therefore reckless. 

The creation of a one-off support facility for a specific financial institution (as opposed to a discount window open to all eligible counterparties on demand, given the right collateral) should only be considered for institutions that are systemically important. For such systemically important institutions, being taken into public ownership is also an option. Northern Rock, which accounts for 2 or 3 percent of the assets of UK-registered banks, is not systemically important–it is small enough to be allowed to fail. It should therefore not have any stronger claim on public financial support that a failing ball-bearings company in Coventry.  It could and should have been allowed to sink or swim on its own.  If the protection of retail depositors, or the prevention of a bank run, are deemed to be important, a decent deposit protection scheme would have been sufficient.  The UK deposit insurance scheme is a shambles.  But once the Chancellor guaranteed all deposits of Northern Rock (retail and wholesale) as well as most other unsecured creditors, the widows and orphans were safe and the risk of a run was ended.  The Liquidity Support Facility for Northern Rock could and should have been ended at that point.

The odds against Northern Rock ever paying a fair price for its access to public funds and for the Treasury’s provision of deposit guarantees are overwhelming. The continued operation of Northern Rock as a publicly subsidised private bank therefore distorts the competitive level playing field, and will lead to justified trouble with Brussels about illegal state aid if it goes on much longer.  The  option value to the rest of the  UK-registered banking system  of the Chancellor’s statements that the Liquidity Support Facility  and the Treasury’s deposit guarantee will also be made available to any other bank that gets into Northern Rock-type dire straits, is worth  less than the  actual subsidy  received by  Northern Rock.  In any case, the socialisation of  bank creditor  risk in the UK is not  a desirable feature of the financial landscape.  The Treasury should cut Northern Rock loose forthwith and not  extend any financial  largesse to potential private sector suitors.

Maverecon: Willem Buiter

Willem Buiter's blog ran until December 2009. This blog is no longer active but it remains open as an archive.

Professor of European Political Economy, London School of Economics and Political Science; former chief economist of the EBRD, former external member of the MPC; adviser to international organisations, governments, central banks and private financial institutions.

Willem Buiter's website