Tuesday, April 22, 2008

Willem Buiter on the Bank of England SLS

The actual baby turns out not to be too different from the one expected. The Bank of England is now wholeheartedly committed to acting as market marker of last resort for systemically important securities for which the markets have become illiquid, not to say defunct, since the start of the crisis in August 2007.

The market maker of last resort provides market liquidity in the transactions-based model of financial capitalism the same way the lender of last resort provides funding liquidity to banks in the relationships-based model of financial capitalism. The same institution, the central bank, can play both roles. All real-world versions of financial capitalism are convex combinations of the transactions-based model and the relationship-based model. The relative weight on the transactions-based model is highest in the US and the UK. Anne Sibert and I called for the Bank of England and other central banks to act as market makers of last resort on August 12, 2007, when the crisis had barely hatched from the wind egg of the financial bubble of 2003-2006.

So I am reasonably pleased that the Bank’s early opposition to this notion has been overcome. It would have been better to have had this facility in place in September 2007, but it is not too late to prevent unnecessary systemic damage to the British banking sector and the negative fallout this would bring for the rest of the UK economy.

The swap of one-year maturity Treasury bills for mainly residential mortgage backed securities (RMBS) - although covered bonds and asset backed securities based on credit card receivables can also be offered by the banks - amounts to a collateralised loan of Treasury bills of one-year maturity to the banks. Another way of looking at it is that the Bank of England lends Treasury bills to the banks, in exchange for the banks lending a commensurate amount of RMBS to the Bank of England, with the banks paying the Bank of England the difference between the unsecured borrowing rates of the two institutions (for the Bank of England, the secured borrowing rate is the same as the unsecured one).

The facility can, at the discretion of the Bank of England, be extended for another two years. If we turn out to need the full three years of the facility, I will emigrate to the Isle of Yap, where there are no liquidity crises and the currency is solid.

What credit risk are the Bank and the government exposed to?

Under the swap arrangement, the banks retain ownership of the RMBS they offer to the Bank of England. It is therefore not quite correct to say, as both the Bank and the Treasury have said, that the Bank of England and the tax payer are not exposed to the default risk on the RMBS. The Bank of England is not exposed, because the Treasury bills to be offered in the swap don’t come from its holdings of Treasury bills. As I see it, the Bank of England is just acting as an agent for the Treasury in administering the Special Liquidity Scheme.

The Treasury and the tax payer are exposed to the risk of both the bank offering the RMBS failing and the issuer of the RMBS defaulting at the same time. RMBS backed by mortgages originated by the institution offering the RMBS to the Bank of England, or by entities in the same group as that institution, are permitted. This means that the probability of the RMBS defaulting and the bank lending the RMBS defaulting are not independent events. Greater care in characterising the default risk exposure of the Treasury and the tax payer is necessary.

Repos are better than Treasury bill - RMBS swaps

It is good news that the additional liquidity is provided for a year. It would have been even better if, instead of swapping Treasury bills for RMBS, the bank had agreed to swap the same aggregate amount of RMBS (£50 billion, say), not for Treasuries, but for perfect liquidity (cash) at a range of maturities, (say 1, 3, 6, 9 and 12 months) through (reverse) sale and repurchase operations (repos). Now the banks still have to go through the process of turning one-year Treasury bills into pure liquidity of the maturities that would be most useful to them.

Too many A’s

The securities accepted in the swap have to be triple A rated by at least two of the not-any-longer-quite-so-holy trinity of rating agencies, Fitch, Moody’s and Standard and Poors. I believe that’s too restrictive, and that in future liquidity support operations the standard will have to be relaxed to single A or better. Provided the Bank of England prices these securities sufficiently punitively, and applies a hefty discount to its valuation of these securities, the Bank and the tax payer will be properly rewarded (ex-ante) for the higher default risk attached to these securities. There need be no element of subsidy involved in accepting lower-grade collateral.

It’s no bail-out

The “£50 billion bail-out for banks” headline of that fount of orginal analysis, the (free) London paper, is rubbish. The Special Liquidity Scheme is priced quite unpleasantly for the banks. The so-called fee (the difference between 3-month libor rate and the 3-month General collateral gilt repo rate) will be quite hefty, because it reflects not only bank default risk (libor is the rate for unsecured bank lending) but also liquidity risk. In addition, the haircuts (discounts) applied to the Bank of England’s valuation of the RMBS, covered bonds and credit card ABS ranges from 12 percent, for the shortest maturity to 22 percent for the longest maturity. That valuation will either be based on observed market prices that are independent and routinely publicly available or on the Bank of England’s own calculated prices. As most of the assets are illiquid, the prices of these assets will for the foreseeable future be set by the Bank of England. In that case it will also apply a higher haircut to the valuation. In addition, Bank officials have told me that should any of the assets lent by the banks to the Bank fall, during the life of the swap, below the required AAA rating, these assets will have to be replaced by AAA-rated securities.

All this is a lot more sensible and likely to minimise adverse selection than the insane valuation of the collateral offered by US Primary Dealers to the Fed at the Term Securities Lending Facility and the Primary Dealer Credit Facility. The collateral is priced by the clearing bank acting as agent for the Primary Dealer! If ever a scheme was designed to encourage collusion between Primary Dealer and Clearer to dump useless securities at inflated prices on the Fed, this is it.

The UK scheme prices what amounts to a collateralised loan of Treasury bills by the Bank of England quite harshly. And appropriately so. There is no eartly reason for giving the banks a tax payer handout. The banks and their borrowers made the mistakes. They and their borrowers should pay the price.

The unavoidable accounting shenanigans when the Treasury is involved

The Bank will not take the Treasury bills it will offer in the swap out of its own Treasury bill holdings. It has less than £14 billion of government securities on its balance sheet. The £50 billion worth of Treasury bills will be additional Treasury bills issued by the Treasury and parked in some ring-fence facility.Why Treasury bills, you may ask. Why not a representative basket of Treasury bills (maturity a year or less) and gilts (more than one year maturity). The answer appears to be that for some incomprehensible accounting reason, Treasury bills don’t count towards the public debt for certain purposes, but gilts do. Just remember this next time you talk to your banker and he asks you whether you have any debt outstanding. If it’s debt you have to repay within a year, it isn’t debt.

Accounting truly is confusing. Swaps are supposed to be off-balance sheet instruments. After a year, the Treasury bills revert to the Bank of England and the RMBS revert to the banks (the Treasuty bills will of course expire after a year). Therefore, looking at the year as a whole, the risk faced by the government, its only exposure is the risk that the banks will not be able to pay the agreed fees over they year (this could have been avoided if all fees had to be paid up front) and that the value of the RMBs is less than the value of the unpaid fees.

What’s with the ABS backed by credit card receivables?

I am somewhat confused by credit card ABS being included in the illiquid asset pool that can be swapped for Treasury bills. The reason RMBS are included is that the normal market for RMBS is in a state of suspended animation, the banks don’t have alternative sources of housing finance and the government fears the flow of new home lending will dry up. Is the flow of credit card lending also threatening to dry up? Are first-time credit card users at a particular disadvantage and are they a social group deserving special attention?

What good will it do?

For at least a year and possibly for up to three years, £50 billion of illiquid bank assets will be swapped for liquid Treasury bills. The banks’ balance sheets will be more liquid. That should help if illiquidity is the binding constraint on bank lending. If the take-up is huge, the scheme can and will be expanded in size. I expect it will be.

If the Special Liquidity Scheme is not used by the banks, this means that liquidity is not the binding constraint on bank lending and may not be much of a problem of any kind. Bank solvency instead would be the issue. Dealing with that is not the province of the bank. If the tax payer has to recapitalise insolvent systemically important or too-big-to-fail banks, the Treasury should do so directly and on-balance-sheet.

It is certainly possible that if the 20-30 percent declines in house prices many observers are expecting materialise, there will be massive negative equity (given the amount of 95% and 100% LTV mortgages that were issued in recent years). If homeowners in negative equity walk away from their debt, the banks could suffer very large losses. Unlike the US, where mortgages are non-recourse debt, the UK does not have mortgages that are non-recourse. The lender has a claim on the defaulting borrower over and above the value of the collateral, and this claim can be enforced for years. Still, a large increase in the incidence and magnitude of negative equity will be bad for banks’ financial health.

If the economy weakens significantly and unemployment starts rising, arrears and repossessions will start to rise. Credit card debt and other consumer loans could also become impaired assets of the banks on a much larger scale than at any time since 1992. Deleveraging households and consumers will join deleveraging financial institutions in the economic-financial emergency room. Given the amount of careless lending to households and careless borrowing by households there has been during the seven fat years, the notion that it would not take much to put a large question mark over the solvency of a number of banks does not see far-fetched. So if the banks don’t rush in and take advantage of the new liquidity facility, I will get seriously worried.

No comments: