April 29, 2008
Some further thoughts on the Bank of England’s Special Liquidity Scheme
Tim Young makes three interesting comments on my blog on the Bank of England’s Special Liquidity Scheme:
(1) The Treasury bills involved are of nine months original maturity, not one year.
(2) In the event that the borrower defaults, the public sector gets stuck with a loss if the value of the collateral is less than the value of the t-bill loan, even if the issuer of the securities posted as collateral does not default. Presumably this is much more probable than a simultaneous default, especially if the borrower is widely known to be a holder of such securities.
(3) US mortgages are not in general non-recourse.
As regards Tim’s first point, the Bank of England Information sheet states: “To provide banks with the certainty about liquidity that is needed to boost confidence, assets will, unless they mature within one year, be swapped for one year and banks will have the opportunity, at the discretion of the Bank of England, to renew these transactions for a total of up to three years.” In the Market notice for the SLS, it says that “The Bills lent under the Scheme will be for an original maturity of nine months and will have been created within the month preceding the drawdown. Bills must be delivered back to the Bank 10 days prior to their maturity and will be exchanged for a further 9-month Bill.” So Tim is correct. Although the banks will be provided with liquidity through Tbills for a year, the Tbills that are issued as part of the initial swap of Tbils for MBS will have an orginal maturity of nine months. These Tbills will be swapped 10 days before maturity for further 9-month bills, which will be held by the banks for no more than 4 months before the MBS swap reverses (unless the scheme is extended, as is likely).
As regards your second point, the Bank of England Information sheet states, in a section headed (ii) Credit risk stays with the banks: “….the Scheme is indemnified by the Treasury but is designed to avoid the public sector taking on the risk of potential losses. That risk will remain with the banks and their shareholders. The assets are pledged by banks as security against which they will borrow the Treasury bills. When a swap transaction expires, the assets are returned to the banks in exchange for return of the Treasury Bills.”
I interpreted the statement that “The assets are pledged by banks as security against which they will borrow the Treasury bills”, as meaning that the transaction was a collateralised loan of Treasury bills, where the lender of the Treasury bills (the Bank of England) has recourse to the borrower (the bank) and to the collateral. In that case the risk to the Bank of England would indeed be the risk of the joint default of the borrowing bank and of the issuer of the collateral offered by the bank. This interpretation was supported by the Bank of England’s News Release on the Special Liquidity Scheme . The News Release and the Information sheet also makes the separate point that “If the value of those assets were to fall, the banks would need to provide more assets, or return some of the Treasury Bills. And if their assets pledged as security were to be down-rated, the banks would need to replace them with alternative highly-rated assets.”
However, some recent discussions with Helen Parry of Complinet.com have convinced me that Tim may well be right, and that the so-called swap does in fact shift the default risk on the illiquid bank assets to the Bank of England, without recourse to the borrowing bank. The swap is technically/legally a customised stock lending agreement. In a note on the issue, Helen writes: “Officials at the Bank of England have confirmed that this is the case. Furthermore, while officials at the International Stock Lending Association have indicated that the contractual terms differ in some respects from the standard terms to be found in the ISLA Global Master Securities Lending Agreement, the agreement does provide for the fact that title to the borrowed securities passes to the borrower. This is the normal provision in a stock lending agreement. This means that the title to the illiquid securities will pass to the Bank of England.”
If Tim’s and Helen’s interpretations are correct, the language used by the Bank (and also by the Treasury, as is evident from the Chancellor’s statement to the House of Commons on 21 April 2008) are misleading, to say the least. The assertion by the Chancellor that “This means that the banks will continue to hold the risk on the securities they provide, so it is them rather than the Bank of England that will be exposed to any fall in value. ” would be the exact opposite of the truth.
I look forward to further enlightenment on the issue.
As regards the third point, Tim is right again. While most first mortgages in the USA are non-recourse, this is not true for all states, and it is also generally not true for second mortgages.[1]
[1] From Wikipedia, “nonrecourse debt or non-recourse debt or nonrecourse loan is a secured loan (debt) that is secured by a pledge of collateral, typically real property, but for which the borrower is not personally liable. If the borrower defaults, the lender/issuer can seize the collateral, but the lender’s recovery is limited to the collateral”.
I do not know the terms of the legal agreement that the Bank of England is using. But the interpretation of a standard securities lending transaction (eg under the Global Master Securities Lending Agreement) above is incorrect. In such a transaction, legal title over the mortgage-backed securities (the Collateral) would indeed pass to the lender (in this case the Bank). But the lender has a contractual claim on the borrower (in this case the commercial bank) to deliver Equivalent Collateral if the value of the Collateral falls below that of the Loaned Securities (ie it does have recourse to the Borrower). If the Borrower defaults, obligations are accelerated and the current market value of the Collateral is set off against the current market value of the Loaned Securities so that one party has a net claim on the other. In this way the lender will only lose money if the borrower defaults and the value of the Collateral falls below that of the Loaned Securities.
Posted by: David Rule ISLA | April 29th, 2008 at 12:21 pm | Report this commentWhether or not the BOE takes a direct hit on losses — as is under debate — it seems unlikely that the Bank will go after other collateral if it will force the borrowing institutions into insolvency. It’s quite conceivable that rigorously marking to market may fall by the wayside as well if it would have the same consequence. Simple public disclosure of the underlying pricing, without naming institutions, would avoid these “bailout” issues.
On a related note, S&P (always first out of the gate as we now know) has revised its expected recovery rates on AAA CDO’s (MBS and ABS) to 60% with no recovery expected below A and 5% for AA rated securities. The type of haircuts on collateral value that BOE is proclaiming hardly seem reassuring under the circumstances.
Posted by: SS | April 29th, 2008 at 12:53 pm | Report this commentThe BoE is taking risk on the borrowing bank not on the pledged assets, unless the borrowing bank defaults. In that case, the BoE is a super senior lender to the bank and as it holds title to the pledged collateral can do as it pleases to recover what is owned without interference from administrators/bankruptcy courts. This is standard for repo agreement, the title transfer is “just in case” and the pledged assets remain on the borrowing bank as does both credit and price risk.
Posted by: jck | April 29th, 2008 at 1:16 pm | Report this commentOne other question: Where will the Old Lady get the T-Bills?
Posted by: SC | April 29th, 2008 at 1:47 pm | Report this commentSince she patently does not have enough on her balance sheet (unlike the Fed), she will have to buy them from HMT, one supposes.
But, if so, is this not highly inflationary - a veritable Reichsbank of a policy?
Re: David Rule’s comment. So the correct understanding is that, although legal title over the MBS passes to the Lender of the TBills (the Bank of England), the Bank has recourse to the Borrower because, in the words of the Bank’s statement: “If the value of those assets were to fall, the banks would need to provide more assets, or return some of the Treasury Bills. And if their assets pledged as security were to be down-rated, the banks would need to replace them with alternative highly-rated assets.”
It is very helpful to have that clarified. It could mean, I suppose, that the only way the Bank could perfect its security might be by pushing the Borrower into default & insolvency. The Bank might prefer to swallow a loss on its collateral instead. But this does not affect the validity of the legal/contractual point you have made.
Posted by: Willem Buiter | April 29th, 2008 at 2:26 pm | Report this commentRe all the above comments: does this not suggest the SLS is patently a time-inconsistent policy? As SS suggests, it’s not clear that they can credibly commit to maintaining the value of the collateral, particularly if doing so increases systemic risk - presumably what the policy was designed to reduce in the first place.
Posted by: Robert Hillman | April 29th, 2008 at 4:59 pm | Report this commentThanks for highlighting my comments Willem.
Discussion of the SLS can get confusing, because it involves a loan of treasury bills by the authorities to a bank, against collateral which comprises securitised mortgage loans made by a bank (which may or may not be the same as the first bank) to homebuyers.
When the BoE press release says “responsibility for their loans, however, remains with the banks”, it is referring to the securitised mortgage loans. Since these securities return to the banks at the end of the t-bill loan, the banks remain fully exposed to decreases (or increases) in their value for any reason, including defaults by mortgage borrowers.
The authorities are, however, exposed to loss on the t-bill loans, which was my concern. For the SLS to be subsidy-free, this exposure should be fully reflected in the price of the t-bill loans. For the authorities to incur a loss, the bank has to default on the loan (ie fail to return the t-bills) AND the collateral to which the authorities have legal title has to be worth less than the value of the t-bills.
The terms of the t-bill loans give the BoE the right to adjust the amount of collateral daily so that the value of the t-bills in each case is more than fully covered (the excess being the “haircut”). If the collateral falls in value because of mortgage defaults, the BoE would require the t-bill borrower to replace the impaired securities to maintain the value of the collateral, but provided that the t-bill borrower remains solvent despite their loss and is able to pledge new collateral, the t-bill loan can continue.
In theory, the haircut and daily adjustment make it unlikely that the t-bill loan is ever less than fully secured. Even if the t-bill borrower is driven to bankruptcy by the declining value of its holdings of mortgage backed securities, it should have pledged a growing amount of these to the BoE in the preceding days. In practice, I can imagine two problems with this mechanism though. One is that news of the bankruptcy of a large holder of mortgage backed securities would probably generate an intraday drop in their value, in which case the proceeds of selling the collateral might still not cover the value of the t-bills. The other problem is that, at the end of the process, the BoE might be seen as having sucked the lifeblood from the ailing bank, leaving nothing for the little old lady with her savings on deposit there.
Posted by: Tim Young | April 29th, 2008 at 8:36 pm | Report this commentSC
You raise a good point. According to the BoE briefing note, the t-bills are being issued by the government and lent to the BoE. It does not say whether the BoE credits a government account in return, but if this is done, I would doubt that it was more than a formality. While it is true that if a government account was credited and drawn on the supply of base money would increase, since the t-bills are only being lent to the BoE any such increase would unwind when the scheme ended. I do not think that the SLS is highly inflationary.
Posted by: Tim Young | April 29th, 2008 at 9:07 pm | Report this commentRe: Tim’s comments
It’s the “in theory” aspect of the daily MTM on the illiquid collateral that has me stumped. As a practical matter, it’s unlikely the Bank is going to be coming up with a slew of margin calls on a broad range of complex bundled securities on a daily, weekly, or perhaps even monthly basis. Numerous questions occur with respect to what information is going to be used to revalue these securities, where the Bank will get this information from (the borrowers?!), and the appropriateness of its methods.
Potentially, this leaves tremendous exposure on the table in the event of a discontinuity such as a bankruptcy or a run on the bank. Under these scenarios, it wouldn’t be hard to imagine that the T-bills loaned have been swapped for cash and spent. The probability of the Bank being made whole may more likely resemble a binary rather than a continuous MTM outcome.
The above questions form some of my rationale for advocating ongoing public disclosure of the marks. That way there can be no question that a) the Bank is doing its job and b) the taxpayer is being protected. Ultimately, there may be better ways to risk 50B sterling.
Posted by: SS | April 29th, 2008 at 10:34 pm | Report this commentRe where the T bills came from..The Bank borrowed them under a stock lending agreement..They were created on the National Loans Fund and sold to the DMA Debt Management Account. The bills were then lent by the DMO (from the DMA) to the Bank via a stock lending agreement. This was done because under schedule 5A of the 1998 Finance Act, the DMA has vires to provide facilities to other public bodies (like the BoE) to acquire Govt securities (para 1d).
Posted by: helen parry | April 30th, 2008 at 9:45 am | Report this commentRe: SS’s Comment (unfortunate initials, incidentally…). I agree that the Bank’s valuation of the collateral should be in the public domain. We already know the haircuts applied to these valuations, but unless we know the valuations themselves, we cannot judge whether there is a subsidy from the Bank to the banks.
Posted by: Willem Buiter | April 30th, 2008 at 10:00 am | Report this comment