Collateral damage limitation

No, this blog is not about ways of locking Dick Cheney in the broom closet.  It is about the collateral policy of the ECB at its discount window and in in its repos.

In the recently published Biennial review of the risk control measures in Eurosystem credit operations, the ECB announced some measures to address the problems of excessive credit risk exposure for the ECB and the implicit subsidisation of banks borrowing from the Eurosystem using dodgy ABS as collateral.

The reason behind this change is the fact that the ECB/Eurosystem has both the largest set of eligible counterparties at the discount window and in its open market operations (including repos) and the widest collection of eligible collateral.  This has made the ECB the most effective market maker of last resort of all central banks in advanced industrial market economies during the current crisis.  It has also led, not surprisingly, to abuses.

There is a widespread suspicion that troubled banks that were locked out of the private securitisation markets have been packaging substandard asset-backed securities (especially RMBS – residential mortgage backed securities) in ways that would make them eligible collateral at the Eurosystem.  Discount window operations (collateralised borrowing at the marginal lending facility) and repos are implemented not by the ECB in Frankfurt, but by the 15 national central banks of the euro area. It is very difficult to ensure uniform standards and criteria in such a decentralised system with very different national traditions, cultures and financial structures.

There are two potential problems if the Eurosystem accepts collateral with a significant risk of default on its balance sheet.  The first problem  is that there may be a subsidy to the borrower, that is, the risk-adjusted rate of return on the collateralised loan is less than the risk-free interest rate.

The second is that,  even if the risky collateralised loan is priced properly ex-ante, the Eurosystem may, if ex-post default occurs,  suffer a capital loss so large that it can only restore its solvency itself by creating money on a scale that threatens the price stability mandate.  The alternative of being recapitalised by some convex combination of the national fiscal authorities of the euro area, would threaten the independence of the ECB.

Has there been abuse of the Eurosystem’s generous collateral facilities?  We cannot be sure because, as is always the case with central banks, the Eurosystem does not make public the information required to judge whether it is doing its financial stability job properly.  It does not give us the models it uses for deriving the ‘theoretical prices’ it assings to illiquid collateral for which there is no market benchmark.  It does not provide us with details about the actual prices assigned to individual illiquid collateral items, to allow us to determine, given enough time, whether there has been an oppropriate use of public money.

But look at the following straws in the wind.

Between August 2007 and July 2008, the share of Spanish banks in the Eurosystem’s allocation of main refinancing operations and longer-term refinancing operations went up from about 4 percent to over 10.5 percent.  The share of Irish banks went up from around 4.5 percent to 9.5 percent.  It can of course be a coincidence that Spain and Ireland are the euro area member states with the most vulnerable construction and real estate sectors.  Another measure of the increase in the scale of the Eurosystem’s lending to the Spanish banks since the beginning of the crisis in August 2007, is the value of the monthly loans extended to Spanish banks by the Banco de España.  This went from a low of about €23 billion in August 2007 to a high of more than €75bn in December 2007 (for those worried about seasonality, the December 2006 figure was just under €30 billion).

It is certainly possible that all these securities are either safe or priced appropriate ex-ante to reflect their impaired creditworthiness, but without more information we cannot be sure.  Repeated worried public statements by ECB Governing Council members, including notably Governor Ives Mersch of the Luxembourg central bank suggested that even among those in the know there was concern.  And the three and a half measures announced on September 4, 2008 also support the view that some of the collateral was pigs’ earns masquerading as silk purses, and valued as such.

Lets look at the annoucement in greater detail.

First, all asset-backed securities (ABS), now put into a new liquidity category V,  will be subject to a haircut of 12% regardless of their residual maturity and coupon structure. Previously only ABS with a fixed coupon and a residual maturity of over ten years were subject to this large a haircut. Furthermore, ABS that are not valued with reference to a market price, but instead are marked-to-model (given a ‘theoretical value’, in ECB speak) will be subject to an addition 5% valuation markdown on that theoretical price (which is equivalent to an additional haircut of 4.4% – don’t ask).

I still consider these valuation haircuts to be rather generous to the borrowing banks.  At the Bank of England’s Special Liquidity Scheme, for instance, wich also accepts RMBS, covered bonds and credit card ABS (if AAA-rated), the haircuts range from 12% to 22%.

Second, if a bank offers an ABS as collateral when it (or any third party that has close links to it) provides support to that ABS by entering into a currency hedge with the issuer or guarantor of the ABS or by providing liquidity support of more than 20% of the nominal value of the ABS, then that collateral will be ineligible.  The logic behind this is that the probability of a default on the ABS is the joint probability of the bank offering the ABS as collateral going broke and the issuer or guarantor of the ABS defaulting.  If the bank offering the ABS as collateral is also the issuer or guarantor of the  ABS, the credit risk faced by the Eurosystem will be higher than if the bank offering the ABS as collateral and the issuer or guarantor of the ABS are independent parties.  Under exising ECB rules, ‘close links’ between the bank offering the ABS as collateral and the bank issuing or guaranteeing the collateral were not allowed.

The practical issue then became: how close is close?  Even under the old rules, a bank offering an ABS as collateral is not supposed to have orginated itself the assets that back the securities.  If, for instance, the Celtic-Castillian bank were to offer RMBS (residential mortgage-backed securities) as collateral to the Eurosystem, the mortgages could not have been orginated by the Celtic-Castillian bank itself.  Now, even much weaker links between the bank offering the collateral and the issuer of the collateral are sufficient to make the collateral ineligible for repo with the Eurosystem.  This modification is sensible, because it closes a loophole.  However, new loopholes will continue to be found.  Instead of trying to restrict ‘closeness’, the ECB could instead penalise ‘closeness’ through more aggressive model-based pricing of the ABS and/or through larger haircuts.

Third, the rating requirements for ABS eligibility are tightened.  Again this is good in principle, but unless you attach material value to the ratings agencies’ ability to evaluate the creditworthiness of ABS, this is unlikely to offer much additional protection to the Eurosystem.  Given the rating agencies’ dismal track record in rating complex structures, I would not attach much significance or value to even the most recent ABS ratings coming from the same old rating agencies

Finally, there is a vague, generic reminder/warning “…that the Eurosystem has the possibility to limit or exclude the use of certain assets as collateral for its credit operations, also at the level of individual counterparties, if required, to ensure adequate risk protection of the Eurosystem”. This could, in principle, be quite powerful, because it asserts that the Eurosystem has complete discretion in accepting or not accepting any specific collateral item.  The downside of this is that discretion can mean arbitrariness and the absence of a level playing field.  This clause appears to target mainly those ABS where the underlying assets are not denominated in euro.

Two things are clear.  First, much more will have to be done to clean out the Augean stables of the EU collateral universe.  Second, nothing serious will happen until the current financial crisis is history.  Stable-cleaning measures will tighten credit and worsen liquidity conditions in unpredictable ways.   Now may not be the right time for a great leap forward towards the first-best.  Even the minor tinkering measures announced on September 4 won’t become effective until February 1, 2009.  So go for it, boys and girls.  Repo your rubbish while you may!

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

Maverecon: a guide

Comment: To comment, please register with, which you can do for free here. Please also read our comments policy here.
Contact: You can write to Willem by using the email addresses shown on his website.
Time: UK time is shown on posts.
Follow: Links to the blog's Twitter and RSS feeds are at the top of the page. You can also read Maverecon on your mobile device, by going to