Thursday, October 04, 2007

Murder in the Markets: Whodunnit?

Joint blog by Willem H. Buiter and Anne C. Sibert

An earlier version of this blob appeard in the Financial Adviser, 27 September 2007, under the title: "Walking the line, not chalking it".

A brief history of securitisation and off-balance sheet finance

Once upon a time banks dominated the financial landscape. They raised funds through deposits that could be withdrawn on demand on a first-come-first-served basis. Their assets consisted mainly of loans, both secured and unsecured, to businesses and households. Specialised mortgage lenders (called building societies in the UK) made long-term loans to homeowners secured against property. Northern Rock was one of these. It was a mutual society, not a public limited company. The home loans were non-marketable and therefore illiquid. Its other assets were typically safe and liquid government securities

With liquid liabilities and mostly illiquid assets, banks, including mortgage banks, were vulnerable to ‘runs’. If enough depositors believed their money was safe, they would all keep their money in the bank. If enough depositors believed that enough other depositors wanted to withdraw their money, there would be run and the bank would fail. When there was a run on a solvent, but illiquid, bank, the central bank, acting as lender of last resort (LOLR) stood ready to provide that bank with liquidity by lending to it freely, but at a penalty rate of interest and against collateral that would be good in normal times. The government provided or mandated deposit insurance, at some cost to the banks; government-imposed capital requirements forced banks to hold more capital than they liked. There were restrictions on what banks could do with their assets; serious reporting obligations and thorough independent audits were the rule.

Banks, however, did not like holding illiquid assets. Thus, in the 1970s securitisation was invented and pools of previously illiquid asset were sold to ‘off-balance-sheet’ special purpose vehicles (SPVs) that sliced and diced them, mixed them with other assets, enhanced them in various ways and issued tranched securities backed by the entire asset pool. We got ‘multi-layered securitisation’ as securitised assets themselves became the underlying assets for the next level of securitisation.

There are good aspects to this, pooling assets reduces risk and making non-traded assets tradable enhances opportunities for risk trading. But, securitisation destroys information: the orginator of the loan (often the party that continued to monitor the original mortgage borrower on behalf of the SPV) was now the agent rather than the principal in the investment relationship. Incentives for acquiring information and for monitoring inevitably decline when these activities are delegated. The information that was acquired ended up in the wrong place because it remained with the originators of the loans. By the time a conduit of a German Landesbank bought some tranche of the securitized home loans from a Paris-based hedge fund, neither the buyer nor the seller knew much about the risk associated with the underlying assets.

In came the rating agencies to solve this information problem. How they would acquire the information dispersed among myriad individual originators of the underlying loans was a question no one asked. Moreover, there were conflicts of interest as the rating agencies were paid by the issuers of the products they were rating. They often advised those whose structured investment products they would rate on how to engineer the product to obtain the best rating! The result was a ratings inflation for structured instruments. In the US, triple A silk purses were made out of the pigs ears of subprime-mortgage-backed securities. Regulators could not keep up and central bank warnings about excesses in credit markets were ignored.

Banks also did not like the restrictions, such as capital requirements and reporting obligations, that were the quid pro quo for access to the lender of last resort. Many of them (although not Northern Rock) outsourced their riskier activities to off-balance sheet vehicles and other less regulated or unregulated entities, including conduits or other structured investment vehicles (SIVs), hedge funds and private equity funds. The banks often remained exposed to these entities through credit lines or reputation considerations, but did not pay much attention to this. More information vanished; nobody knew any longer who owned what and who owed what and to whom.

Banks, including Northern Rock, disliked having to raise funds by attracting depositors and found capital markets to be an obvious alternative. Northern Rock de-mutualised and became a regular public limited company. Asset-backed securities (including covered bonds) and unsecured corporate loans permitted a faster expansion of business than the pedestrian process of attracting depositors.

The securitisation and disintermediation boom caused many market players to believe that risk had not only been repackaged, but that it had disappeared. After 2003, credit risk spreads of all kinds shrunk to miniscule proportions, assisted by unduly expansionary monetary policies in the US, Japan and the Eurozone. Low long-term real interest rates further boosted the leverage frenzy in the new financial sector.

Then it happened. In 2006 default rates on US subprime mortgages rose to levels inconsistent with the ratings of the securities they backed. By July 2007, many asset-backed securities markets were becoming illiquid, and de-securitisation was beginning. In August, interbank markets and asset-backed corporate paper markets began to seize up in the US, the Eurozone and the UK. The first victims were hedge funds (in the US and in France) and small banks in Germany that had been directly exposed to the US subprime and other risky mortgage markets. On September 13, Northern Rock had to seek emergency funding from the Bank of England, not because it had any significant exposure to the US subprime sector, but because it was unable to fund itself in the wholesale markets.

Although the bail out of Northern Rock was a joint decision of the UK Treasury, the Financial Services Authority and the Bank of England, these parties’ Memorandum of Understanding (MOU) makes it clear that ultimately the decision belonged to the Treasury: Ultimate responsibility for authorisation of support operations in exceptional circumstances rests with the Chancellor”. This makes sense, because ultimately the tax payer funds the losses when public resources are put at risk. The reputational damage of this debacle, however, affects mainly the Bank: it had to provide the line of credit within days of taking a strong public line against bail outs.

Northern Rock engaged in reckless borrowing. It could not survive without external assistance not because its assets were bad (its exposure to the US subprime market is tiny) but because it used a high-risk funding policy to finance its breakneck expansion, with three quarters of its funds coming from the wholesale markets rather than from depositors. It gambled and lost and it urgently needs a buyer with deeper pockets and a more sensible funding strategy.

Northern Rock’s bail-out cannot be justified on systemic financial stability grounds: as the UK’s fifth largest mortgage lender it is just too small to be a threat to financial and economic stability. A bail out should only be undertaken if there is, in the words of the MOU, “… a genuine threat to the stability of the financial system to avoid a serious disturbance in the UK economy.”

Ironically, the Liquidity Support Facility failed to restore confidence and a run on the deposits of Northern Rock developed. Calm was restored only when the Chancellor guaranteed in full all deposits of Northern Rock and of any other bank that might be granted recourse to a Liquidity Support Facility in the future. Instead of this socialisation of UK-wide depositor risk, it would have been preferable to take Northern Rock into public ownership. It could have resumed operations immediately in support of its existing commitments and could have been re-privatized at some later date.

Getting the monetary authority out of the Lender of Last Resort business and into the Market Maker of Last Resort business

We believe that the Bank’s understanding of the distinction between its Lender of Last Resort (LOLR) role for individual banks and its responsibility for providing broad liquidity support for financial markets, and specifically for the longer-term money markets (see e.g. the Governor’s Paper submitted to the Treasury Committee on12 September 2007), is flawed. The Bank should support key financial markets and other institutions such as the payments system and the clearing and settlement systems. The Bank should leave bailing out individual banks to the FSA, which has the institution-specific knowledge, and the Treasury, which can call on tax payers for funding. Ending the active role of the monetary authority as LOLR would require the FSA to have a credit line or overdraft facility with the Bank, guaranteed by the Treasury and would require a change in the MOU. The Bank would take no part in the decision as to whether some bank should be bailed out, and the Bank’s role in funding any bail-out decided by the Treasury and the FSA would be entirely passive.

Bank intervention in longer-term money markets

The Bank has made a mistake in its unwillingness to intervene at longer maturities than the overnight market. The Bank’s own primary money market objective is for “Overnight market interest rates to be in line with the official Bank Rate, so that there is a flat money market yield curve, consistent with the official Bank Rate, out to the next MPC decision date ….”. This means that, following the last MPC meeting on 6th September 2007, when there was a month to go till the next scheduled MPC meeting, the one-month interbank rate on unsecured lending (LIBOR) should have been close to the Bank’s policy rate of 5.75%. In fact it was only just below the three-month LIBOR rate of 6.68% (see Chart 1).

There are four explanations for the sizable spread of three-month LIBOR over the Bank Rate. The first is an expectation that the Bank Rate will rise over the next three months, but this highly unlikely. Second, there could be a pure term premium, but this must be tiny over such a short horizon. Third, there is a risk that borrowers will default. This is clearly not zero, but it is difficult to believe that there is a one percent probability that a typical UK money centre bank will default with a zero recovery rate during the next three months. Finally, there is a liquidity risk and we attribute the lion’s share of the recent spread of three-month LIBOR over Bank Rate to liquidity factors.

Currently liquid banks may be reluctant to make three month loans, not because they are afraid that their borrowers will be insolvent in three months, but because they are afraid that both they and their borrowers will be illiquid in three months. If enough banks have these fears, an interbank ‘lending strike’ results.

Banks everywhere are gearing up to take on their balance sheets the illiquid assets of conduits, other SIVs and other off-balance sheet SPVs that they are exposed to through credit lines or reputational considerations. Fear of future illiquidity is widespread and banks are hoarding excess liquidity rather than lending it out in the interbank market, even at nearly seven percent. The Bank could address this unfortunate situation by injecting liquidity, through repos with, say, a three-month maturity to eliminate the liquidity premium. Such repos are likely to be more effective if they are against a wider range of eligible collateral that what the Bank currently accepts, including illiquid assets.

It is true that the ECB’s massive injections of three-month have not prevented significant spreads (albeit lower than in the UK) of Euribor over the policy rate in the Eurozone (see Chart 2). But, in the United States, the spread on three-month LIBOR has fallen to less than 50 basis points since the Fed lowered its discount rate by 50 basis points on 17 Aug 2007, in addition to more modest open market and discount window operations (see Chart 3). However, what is most important is not the spread, but the amount of lending taking place. In the UK 3-month LIBOR has become the rate at which banks will not lend to each other.

The Bank as Market Maker of Last Resort

We have proposed (Willem H. Buiter and Anne C. Sibert “Three Steps to Calm the Storm”, Financial Times Comment, 5 Sept 2007) that the Bank should accept a wider range of collateral, including lower-rated illiquid private assets, as long as it is punitively priced, and subject to a suitable ‘haircut’ (discount) as well. The Bank should ‘make market’ for such 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.

There would probably have been no need for a bail-out of Northern Rock if the Bank of England had had a sensible collateral policy for its regular open market operations. Unlike the ECB and the Fed, the Bank only accepts European Economic Area sovereign debt instruments, high quality debt issued by a few international organisations and, exceptionally, US Treasury debt. More sensibly, the ECB also accepts private securities rated at least A-, including asset-backed securities. If Northern Rock had a Eurozone subsidiary, it could have funded itself through the eurozone three-month repos conducted by the ECB last week, using its high-grade mortgages (or securities backed by them) as collateral. The Fed can, in an emergency, accept anything it deems appropriate as collateral at its discount window, not only from banks, but from individuals, corporations and non-bank financial institutions as well.

On September 19, the Bank suddenly had a double change of heart: it announced it would initiate repurchase operations at 3-month maturity and would accept as collateral private assets, including mortgages.


We know that the Chancellor authorised the Bank to “provide a liquidity support facility to Northern Rock against appropriate collateral and at an interest rate premium.” But, this is not sufficient information. Is the support uncapped and open-ended, as Northern Rock informs us? What is the premium charged for the use of the facility? What is the arrangement fee for the facility? Exactly what collateral will be offered, how will it be valued and what ‘haircut’ will it be subject to? The public are funding this risky venture and they are entitled to know. There is no commercial confidentiality argument for keeping any of this information secret. The accountability of the Bank and the Treasury are at stake. The Chancellor should provide this information forthwith, and if he does not, Parliament should insist.


The Chancellor, and possibly also the FSA and the Bank, do not want even a systemically insignificant mortgage lender to fail on their watch, regardless of the moral hazard created by the bail-out. The Chancellor is willing to risk tax payer money to prevent it. The Bank, however, should not be directly involved in the decision making; nor should it play an active role in the funding of the liquidity support facility. Depositor protection is the job of the FSA and the Financial Services Compensation Scheme. Redistribution of income belongs to the Treasury. If the Bank remains an active participant in an inherently political bailout process for individual banks, the Bank’s independence in the realm of monetary policy could be compromised.

Charts Source: Bank of England

Source: European Central Bank

Source: Federal Reserve Board


Anonymous said...

From BOE

"(i) The Bank conducts long-term repos (LTRs) of three-, six-, nine- and twelve-month maturities.
LTR OMOs are conducted on a monthly basis, according to a timetable published on the Bank of
England wire service pages (listed in Appendix IV(a)) and available on its website at
(ii) All maturities are offered at each tender. Maturities are normally 91 days, 182 days, 273 days and
364 days, but are occasionally slightly longer to ensure that they coincide with the settlement date
of the monthly LTR OMO. Twelve-month OMOs are never longer than 364 days, so may on
occasion mature before the date of the relevant LTR.

(i) In its routine short-term Open Market Operations, long-term Open Market Operations, liquidityproviding
fine-tuning operations and in the standing lending facility, the Bank normally lends via
reverse repo of the following types of eligible securities:
• Gilts, including gilt strips
• Sterling Treasury bills
• HM Government non-sterling marketable debt
• Bank of England Euro bills and notes
• Sterling-denominated securities issued by European Economic Area (EEA) governments and
major international institutions (see Appendix VI).
• Euro-denominated securities, including strips, issued by EEA central governments and central
banks and major international institutions (see Appendix VI) where they are eligible for use in
Eurosystem credit operations. These securities can either be those issued directly into
Euroclear and Clearstream or can be ‘CCBM securities’ where the central bank in the country
in which the relevant securities were issued has agreed to act as the Bank’s custodian under
the Correspondent Central Banking Model (CCBM)."

From ECB.
"The CCBM is also available to counterparties of the Bank of England, Danmarks Nationalbank and Sveriges Riksbank."

Anonymous said...

It apperas there is no shortage of
eligible collateral.

Tim Young said...

I am afraid the three month LIBOR spread to repo is still a mystery to me. I can understand that banks were reluctant to lend their existing funds out to another bank for three months, but they could always access more overnight money at close to the repo rate. Why did some banks not expand their balance sheets by lending to other banks for three months at 6.9% and take a punt on being able to fund that lending from overnight money for three months? If my bank had offered me an extra 1% to convert my instant access deposit to a three month time deposit, I would have readily agreed!