Insuring toxic assets: throwing good tax payers’ money after bad private money

The UK government has offered, under its asset protection scheme (APS), to guarantee (or insure) up to £600 bn worth of toxic assets held by British banks- up to £300 bn for RBS and up to £250 -£300 bn for the Lloyds Banking Group.  Barclays may be waiting in the wings.  The APS insures the banks (that is, their CEOs, shareholders, junior and senior unsecured creditors other than retail depositors – already covered by deposit guarantees up to £50.000 – and staff) against losses on these toxic assets over and above a certain deductible or ‘first loss’ for the bank.

There is ample precedent for this kind of guarantee scheme.  In the US, the Fed, the FDIC and the US Treasury have guaranteed a large chunk ($300 bn) of toxic assets of CItigroup.  In the Netherlands, the Dutch state insured a portfolio of $39 bn (face value) worth of securitised US Alt-A mortgages held by ING.

Like its American and Dutch counterparts, this toxic asset insurance scheme is without redeeming social value: it is inefficient, unfair and expensive to the tax payer.  Apart from that it is great.  There also are superior alternatives available: full nationalisation and, best of breed, the ‘good bank’ solution.

The proposed UK insurance scheme is in design rather like the Dutch guarantee for ING’s toxic assets.   The Dutch assistance took the form of a back-up guarantee facility for ING, under which the state shares with ING any gains and losses on this portfolio relative to a benchmark value for the portfolio of $35.1 bn.  The shares of the state and ING in any gains/losses relative to this benchmark are 80% and 20% respectively.  Ignoring the sharing of the gains (the possibility that the portfolio will turn out to be worth more than $35.1 bn is remote), this amounts to a 10 percent or $3.9 bn first loss tranche for ING, with the state taking 80 percent of the losses in excess of $3.9 bn, and ING 20 percent. There is therefore 20 percent private co-insurance. People familiar with the market believe the portfolio to be worth no more than 65 cents on the dollar, so a ten percent first loss seems generous to ING, and not really balanced by the 80 percent share of the state in the upside.

So to determine how good a deal for the tax payer the UK government’s APS is, the magnitude of the first loss and of the subsequent co-insurance shares of the banks are key.

In the Dutch case, the bank pays a guarantee fee to the state and the state pays ING a management and funding fee.  Incredibly, with ING the combination management and funding fee exceed the guarantee fee, so the state is paying ING for the privilige of providing the bank with a guarantee.  Since ING continues to own the assets (the same holds for the APS) and would have had to manage and fund them in any case, the payment by the state of a management and funding fee to ING is ludicrous.  Again, Parliament in the UK has to scrutinise very carefully the magnitude of the guarantee fee paid by the British banks to the Treasury and the size of any other fee income (if any) going the other way.  The 3-6 percent fee size rumoured by the Press would, cet. par. be a reasonable start, as long as there or no offsetting management and funding fees going towards the banks.

The upside for the tax payer in the UK comes through the payment of the guarantee fee in the form of convertible preference shares or warrants.  This does indeed provided ex-ante upside, but also ex-ante downside.  Only if the common stock of the banks is fundamentally undervalued when the preference shares are converted or the warrants exercised, will there be any real shareholder upside.

ING is also to provide 25 bn euro of additional credit to Dutch businesses and Dutch households; there will be no bonuses for 2009 and until a new remuneration policy is adopted, and the CEO was told to leave.

It is unclear whether the British scheme will have further bonus-bashing features, human sacrifices and forced lending attached to it.  The new economic nationalism and financial protectionism that are inevitably associated with the bailing out of banks with tax payers’ money will no doubt be part of the UK plan for the banks as well: when the British tax payer bails you out, it isn’t politically acceptable to increase lending to households and enterprises in Zanzibar.  Of course, the enforceability of even the most solemn pledge to increase lending is dubious.

Why guarantees rather than outright purchases by the state of toxic assets?

A toxic asset is an asset whose fair value is highly uncertain.  Its current market value, if a market price can be found at all, is likely to understate its fair value because the markets these assets used to be traded in have become illiquid and may have dried up completely.  The hold-to-maturity value of a toxic asset, obtained by discounting its (risky) expected future cash flow using a discount rate that abstracts from illiquidity premia, is a highly subjective concept.  The toxic assets may eventually turn out to be good assets (future sale price close to fair value or held-to-maturity value close to notional or face value) or bad assets (future sale price  or held-to-maturity value well below face value).

The banks argue that their toxic assets are bound to turn out to be good assets.  “Well they would, wouldn’t they”, if I may paraphrase that well-known appraiser of illiquid complex structured products, Mandy Rice-Davies.  With the earlier destruction of asset values through the implosion of a brace of asset price bubbles and credit bubbles followed by a vicious downturn in the real economy that will feed back negatively on what’s left of the value of mortgages, credit card receivables, student loans, car loans and regular corporate lending, there is a strong likelihood that most toxic assets will turn out to be bad assets once the veil of uncertainty is lifted.

Instead of RBS and Lloyds retaining ownership of the toxic assets and the government providing insurance against losses, the state could have purchased the toxic assets outright from the banks.  From an economic point of view, the two approaches are equivalent: there exists a price at which the government acquires a toxic asset outright that is equivalent to insuring the asset againts losses for a given first loss tranche for the banks (before the insurance would kick in)  – likely to be less than 10 percent -, a given co-insurance share for the banks (the share of the additional losses beyond the first-loss tranche borne by the banks) – likely to be 10 percent -, a given net insurance premium, somewhere between three and six percent, and a given duration of the insurance.  You need to be able to value and price the asset if you are willing to insure it.

But from a public sector accounting and government accountability point of view, transparency-dodging and accountability-avoiding governments much prefer insurance or guarantees to outright purchases.  The reason is that insurance is off-balance sheet.  It is a contingent liability, which is not entered at its fair value on the liability side of the government balance sheet under the uninformative government accounting practices followed in the UK and most other countries.  It ought to be, if we take an informed, comprehensive approach to valuing all the government’s assets and liabilities, even the contingent one.  I have been advocating such a comprehensive balance sheet approach for more than 25 years.  Only in a few enlightened countries like New Zealand do we find governments that subject themselves to something approximating comprehensive asset, liability and net worth accounting.  But in most of the world, including the US, the UK and the Netherlands, contingent liabilities don’t count towards public sector debt.

In a rational world (or even in a world that is merely not terminally stupid and/or bent on escaping accountability), the fair value of the contingent liability represented by the government’s insurance scheme would be entered as a liability on the comprehensive public sector balance sheet.  The fair value of the future cash flow of the insurance or guarantee premia would be entered on the asset side.

Should the government instead purchase the toxic assets outright and borrow funds to make the purchase, the government debt thus incurred would, both under real-world UK and US-style accounting and under economically sensible, comprehensive accounting methods à la New Zealand, add to the liabilities of the government.  The fair value of the toxic assets would be on the asset side of the government’s comprehensive balance sheet.  If reason prevailed, the excess of the fair value of the contingent insurance liability over the fair value of the future insurance premia under the insurance scheme would be the same as the excess of the value of the public debt issued to finance the outright purchase of the toxic assets over the fair value of the toxic assets acquired by the government.

Under current UK accounting conventions, the insurance scheme does not show any immediate increase in government liabilities, while the outright purchase scheme does.  Under neither scheme would the assets be recorded as such in the government balance sheet.  Guess which of the two schemes the UK, the US and the Dutch governments have chosen?  You guessed right.  With nothing showing up on the government balance sheet or in the government budget unless and until a loss materialises, the government does not, given the general economic illiteracy of its interlocutors in Parliament, the media and the public at large, have to subject itself to up-front accountability for the contingent exposure it has incurred.

Why should the government guarantee/insure or purchase the toxic assets at all?

All the toxic assets insured by the government are already existing assets. They are the result of investment decisions made in the past.  One of the oldest insights in economics is that bygones are bygones.  You cannot undo past mistakes.  We may not know yet the magnitude of the losses that have been incurred on the underlying assets (how bad the toxic assets will turn out to be), but there is little if anything the banks or anyone else can do, other than sensible macroeconomic management (monetary, credit and fiscal policy) to affect the eventual magnitude of these losses.

By insuring losses that have already been incurred (although we may not know their magnitude as yet), the government simply redistributes these existing losses from the shareholders, creditors, management and employees of the banks to the tax payers.  This is both unfair – those who break something should own it – and inefficient: it encourages future reckless lending and investment by the banks, that is, it creates serious moral hazard.  As I will argue below, it creates this moral hazard quite unnecesarily.

There are two efficiency arguments for this ex-post insurance of losses that have already been incurred – losses on stocks of existing assets.  The first argument is that the overhang of a stock of toxic assets on the balance sheet of a bank discourages new lending by that bank.  The toxic assets require considerable capital to be held against them, and perhaps additional liquidity as well; they increase the funding costs of the bank.  Toxic assets on a bank’s balance sheet act as a tax on new lending.

It is true that toxic assets will act as a tax on new lending by the bank that holds them.  Insuring the toxic assets (or purchasing them from the bank using public money (or a combination of public and private money as proposed in the US) is, however, likely to represent an inefficient use of public money, because the terms of the insurance offered to the banks or the price offered to purchase their toxic assets outright, are likely to be far too generous.  Knowing that the government is desperate for a deal, because the government knows the banks will fail without either the insurance scheme or the removal of the toxic assets from their balance sheets, the government invariable gets pushed into terms that imply a massive subsidy to the shareholders and unsecured creditors of the troubled banks.

The second potentially sound economic reason for the government intervening in the market for toxic assets is that these assets are systemically important, liquid under normal circumstances but temporarily illiquid as a result of pervasive uncertainty, fear and loathing.  In that case the state can act as a market maker of last resort, engaging in a price discovery process, through reverse auctions and similar mechanisms, to restore the market for toxic assets more quickly to its desirable liquid state.  This argument for government intervention applies, at most, only to outright purchases of toxic assets by the government or the government in combination with the private sector.  It does not apply to the UK insurance scheme, even though it is equivalent from an economic perspective, because most of the market cannot back out an implied market price from an insurance contract.

Even if the government were to purchase the toxic assets outright, I believe that this argument is not relevant for most existing toxic assets.  Far from being normally useful and systemically important securities, they tend to be complex, opaque and indeed incomprehesible structured products that should never have been created.  No wider social purpose is served by the state making market in these assets.

The good bank solution: more efficient, cheaper and fairer

There is a much simpler solution.  Leave the bad assets with the banks that own them now (Old RBS, say).  Create a new bank (New RBS, say), capitalised with government money or, if possible, with a combination of government money and new private money.  Take away its banking license from Old RBS.  Transfer the deposits of Old RBS to New RBS.  Let new RBS purchase any of the good assets of Old RBS it is interested in at market value or fair value.

Apart from some financial assets, the good assets purchased by New RBS from Old RBS will include most of the UK commercial banking franchise (the physical and organisational infrastructure of UK high-street banking and corporate lending).  New RBS would also hire most of the staff of the UK retail banking and corporate lending franchises of Old RBS.  The good assets are, of course, by definition, easy to value.  Old RBS would not be allowed to make new loans or investments.  It would fund its existing asset portfolio of toxic and bad assets, plus the money received from the sale of the good assets to New RBS, plus the excess of the value of the sale of the good assets over the value of deposit liabilities transferred out of Old RBS.  This excess could take the form of government securities or (if negative) a debt to the government.

All present and future government financial support would go towards the capitalisation of the new good banks and towardx guaranteeing new lending and new borrowing by these new good banks.  The old legacy banks would not get another penny of new government support.  Should they go bust, as would be quite likely, their existing shareholders would lose what is left of their investment (not much in any case).  The unsecured creditors, junior and senior, might also lose some or all of their investment.  Partial or complete conversion of their debt into equity would be a possible next step in the after-life of Old RBS.  Old RBS would manage the existing assets either by selling them or by holding them to maturity.  When the last asset is sold or matures, Old RBS would effectively cease to exist as an economic actor.

It may be possible to use the Special Resolution Regime of the new Banking Act of 2009 to implement the ‘good bank’ solution.  Under the SRR, the Triparite authorities can (according to the Bank of England’s website):

  • transfer all or part of a bank to a private sector purchaser
  • transfer all or part of a bank to a bridge bank – a subsidiary of the Bank of England – pending a future sale
  • place a bank into temporary public ownership (the Treasury’s decision)
  • apply to put a bank into the Bank Insolvency Procedure (BIP) which is designed to allow for rapid payments to Financial Services Compensation Scheme (FSCS) insured depositors
  • apply for the use of the Bank Administration Procedure (BAP) to deal with a part of a bank that is not transferred and is instead put into administration

Only the first of these powers would be necessary to implement the ‘good bank’ solution.

Why it does not pay to be a little bit pregnant

The Treasury is surprisingly reluctant to bite the bullet and nationalise the dead banks walking in the UK.  This includes at least RBS (already 70 percent publicly owned) and Lloyds (43 percent publicly owned).  As regards the other UK high street banks, we are likely to find out before the end of the year how they stand up under the battering the are taking because of their emerging market exposures in South America and the Far East, and under the common strain of a deep and long slowdown in their markets in the overdeveloped world.

Partial public ownership (or the threat of future partial or complete public ownership) provides the banks with powerful incentives to do everything except lending to the (risky) real economy.  First and foremost,  the leadership (CEO plus minions and non-government-appointed Board members) of these partially state-owned banks and banks under threat of state ownership and control want to get rid of the state ownership stake or prevent the state from acquiring an ownership state.  They have this incentive whenever state ownership is financially costly or imposes constraints on the management to act in its best interest: constraints on dividend pay-outs and share repurchases; caps on bonuses and on executive and board remuneration in general; pressures to lend beyond what the banks consider commercially desirable, which even if they are resisted, are a pain in the neck for bank executives who have to come up with new excuses in front of the Treasury Committee (“we are very sorry”), and who have to put up with public chastisement by the Chancellor and the Prime Minister .

These state-ejecting or state-evading banks pursue their strategies by building up capital and boosting liquidity rather than lending to the real economy.  This hoarding of capital and liquidity can be observed in the US and in the UK and whereever else the state has partial ownership of banks or threatens to achieve (partial ownership), and this partial ownership cramps the style of the banks’ management.

The halfway-house of partial state ownership with costly conditionality is the worst of all possible worlds.  There are two alternatives.  The first is the solution preferred by the US regulators and the US Treasury.  It is to make government support for the banks as cheap as possible – financially and in all other ways.  Populist outrage and blood-lust in the Congress have compelled the imposition of banker-bashing measures like bonus caps and other ceilings on bankers’ remuneration, but it is clear that no-one in the US administration or regulatory agencies cares about moral hazard – the incentives created by current actions for future excessive risk taking – and that massive redistribution is taking place from the tax payer to the banks’ executives, shareholders and unsecured creditors.

The second alternative to partial nationalisation is full nationalisation.  It would reduce the incentives not to engage in new lending. It is not as cheap (in terms public funds required) as creating good banks – i.e. the (partial) nationalisation of new lending.  That is because full nationalisation would involve the purchase of the toxic assets as well as the good assets.  But is is better than the limbo of partial public ownership, which is the worst of all possible worlds from the perspective of incentivising new lending.

Why does the inefficient and inequitable guarantee scheme prevail over the superior ‘good bank’ solution?

Why do governments invariably prefer insuring toxic assets or even the ‘bad bank’ solution, where the government (possibly jointly with the private sector) purchases the toxic assets outright, to the ‘good bank’ solution? The reason is explained in a famous book straddling economics and political science, by Mancur Olson, The Logic of Collective Action: Public Goods and the Theory of Groups, first published in 1965. In the book, Olson develops a theory of concentrated benefits versus diffuse costs.  In a democratic society, a limited number of economic agents, each with much to gain from a policy or regulation that is socially inefficient, will generally prevail over a much large number of opponents, each of whom may not stand to lose very much individually, even though the aggregate loss of the losers exceeds the aggregate gain of the winners.

The reputation of the banks and of some of the other highly leveraged financial institutions has taken quite a battering as a result of the financial crisis and economic slump.  Despite this, the powers of persuasion, lobbying prowess and influence of the established banks and other financial institutions and of their representatives exceeds that of the millions of tax payers who stand to lose as a result of this bail-out, compared to the economically more efficient and fairer alternative offered by the ‘good bank’ solution.  No-one speaks for the legions of tax payers with the same ‘voice’, eloquence and powers of persuasion that the City establishment can turn on when it makes its case in the corridors of power.

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