Good Bank vs Bad Bank
The Good Bank solution differs significantly from the Bad Bank solution as regards its distributional implications, its medium-term and long-term incentive effects and its immediate financial stability impact.
Under the Bad Bank approach, the authorities either purchase toxic assets from the banks that made the toxic investments/loans, or they guarantee (insure) these toxic assets. Toxic assets are assets whose fair value cannot be determined with any degree of accuracy. Clean assets are assets whose fair value can easily be determined. Clean assets can be good assets (assets whose fair value equals their notional or face value) or bad assets (assets whose fair value is below their notional or face value). When the authorities acquire the toxic assets outright, they establish a legal entity to manage these assets – the Bad Bank. The publicly-owned Bad Bank either sells these toxic assets as and when they cease to be toxic and a liquid market for them re-emerges, or holds them to maturity.
Under the Bad Bank approach, the legacy banks, either sans the toxic assets or with the toxic assets guaranteed by the state, live to fight another day. The presumption is that the state overpays for the toxic assets. The price it pays is certainly greater than the immediate liquidation value of the assets by their owners. It is also likely to exceed the present discounted value of the future cash flows of the assets, or their hold-to-maturity value. Similarly, the cost of any guarantees provided by the state in the case where the toxic assets continue to be held by the banks, is likely to be less than the fair value of these guarantees.
The rationalisation for the creation of Bad Banks and for toxic asset purchases by the state that was part of the original TARP proposal - it would serve as a price discovery mechanism for potentially socially useful financial instruments that had temporarily become illiquid – is no longer credible. Most of the toxic assets ought never to have been created and, with a bit of luck, will never be seen again. So the fundamental rationale for the creation of Bad Banks and for toxic asset purchases by the state is the provision of a subsidy to the banks that made the toxic loans and investments. These beneficiaries include the top management and board of these banks, the shareholders and the unsecured and non-guaranteed creditors.
The subsidies for the legacy banks inherent in the purchase by the state of the toxic assets and/or in the guarantees provided by the state for these toxic assets are further boosted by the myriad modalities of further official financial support for these banks. These can be additional capital injections, guarantees for new borrowing or guarantees for new loans and investments by the banks.
Under the Good Bank approach, the state creates a new bank, the Good Bank, which gets the deposits and the clean assets of the old banks. The old bank gets compensation equal to the difference between the (known) value of the clean assets it loses and the value of the deposits it gives up. The state may also inject additional public capital into the Good Bank, or it may invite in additional private capital. Government financial support is given only to new lending, new investment and new funding by the Good Bank. The legacy (ex-)bank has its banking license taken away and simply manages the existing stock of toxic assets. The legacy (ex-)bank does not get any further government support.
The Hall-Woodward-Bulow good bank solution
A particularly neat example of the Good Bank solution has been proposed by Robert E. Hall of Standford University and Susan Woodward of Sand Hill Econometrics. It can be found on the Vox website. They attribute the key idea to Jeremy Bulow. In what follows I merely adapt their numerical Citicorp example to the RBS Group.
The data for the Table below come from the Annual Report & Accounts 2008 of RBS. I am doing the exercise for the whole RBS Group. As it is unlikely that home country governments would be willing (or even able) to support the foreign subsidiaries of their banks, it might have been more appropriate just to consider the UK high-street banking units of the RBS Group. I leave that as an exercise for the reader.
Total equity of the RBS Group at the end of 2008 is reported on the balance sheet as just over £80bn. Market capitalisation is around £ 8bn. I therefore subtract £72 bn from the £2,402 total assets reported for the end of 2008, which leaves adjusted total assets at £2,330 bn. The tax payer has already put £45 bn into RBS. In addition, RBS has placed £325 bn of toxic assets in the government’s Asset Protection Scheme.
This means that RBS is a dead bank walking, a zombie bank, with its market capitalisation much less than past and present government financial support, let alone past present and anticipated future government financial support, which would also be reflected in today’s market capitalisation. I could have done the same type of exercise for Lloyds Banking Group, for Citicorp, for Bank of America or for UBS and many other zombie border-crossing banks.
|Good Bank -Bad Bank
Deconstruction of RBS Group end-2008 Balance Sheet
Hall-Woodward-Bulow approach) (£ bn)
|RBS||Good Bank||Bad Bank|
|Clean assets (good & bad)||1,012||1,012||-|
|Equity in other bank||-||-||114|
|Debt securities & other
|Total liabilities & equity||2,330||1,013||1,432|
On the asset side of RBS group are clean assets (good and bad, but with known fair values) and toxic assets (assets with unknown fair values and derivatives. On the liability side, I distinguish deposits, debt securities and other non-deposit liabilities, and derivatives. In the US, the derivatives on both the asset and liability sides of the balance sheet would have been netted, which would have reduced the size of the balance sheet by almost one trillion pounds.
I assume that the £325 bn worth of toxic asset insurance offered by the authorities to RBS equals the stock of toxic assets on its balance sheet. This leaves RBS with just over £ 1 trillion worth of clean assets (and the derivatives, just under £1 trillion). ‘Deposits’ is shorthand for guaranteed or secured creditors. The £899 bn worth of deposits on the RBS balance sheet is, however, larger than what is formally covered by UK deposit insurance (or by the applicable deposit insurance schemes of the foreign subsidiaries). Debt securities and other non-deposit liabilities are claims on RBS by unsecured and non-guaranteed creditors. They include all unsecured debt, including subordinated debt, other junior debt and senior debt. RBS had £452 bn of this unsecured and non-guaranteed debt (plus of course some non-guaranteed and unsecured liabilities included in ‘deposits’). Then there is just under £1 trillion worth of derivatives on the liability side of the balance sheet.
Equity – market capitalisation – is a mere £ 8 billion, giving a capital ratio (equity as a percentage of assets) of 0.34%. Even with all the government support it has received, RBS group is effectively worth nothing.
The Hall-Woodward-Bulow Good Bank – Bad Bank deconstruction of the RBS balance sheet requires one key condition to hold: the value of the clean assets of RBS has to exceed that of its deposit liabilities. This will be more likely the larger the amount of non-deposit funding RBS engages in.
We split RBS into a Good Bank and a Bad Bank by giving the deposits and the clean assets of RBS to the Good Bank, leaving everything else with the Bad Bank, and giving the Bad Bank all the equity in the Good Bank. (The derivatives on both sides of the balance sheet could be given to the Good Bank instead of to the Bad Bank, assuming they are clean). Since the value of the clean assets (£1,012 bn) exceeds that of the deposits (£ 899 bn), the good bank has equity of £114 bn (mind the rounding errors!). Its capital ratio is a healthy 11.25%. If I had used a more restrictive definition of ‘deposits’, the capital ratio could easily have been over 20% or even 30%.
The Bad Bank keeps the toxic assets and derivatives of RBS. It also has the equity in the Good Bank as an asset on its balance sheet. On the liability side it has just the unsecured and non-guaranteed debt securities and other non-deposit liabilities. Its equity is, of course, the same as that of RBS: £8 bn. Its capital ratio will therefore be higher than that of RBS, because the balance sheet of the Bad Bank is smaller. Neither the equity owners of the Bad Bank nor the unsecured and non-guaranteed creditors of the Bad Bank are worse off than, respectively, the equity owners of RBS and the unsecured and non-guaranteed creditors of RBS.
To achieve the deconstruction/decomposition of RBS into a Good Bank and a Bad Bank according to the Hall-Woodward-Bulow principles would require that RBS be put into temporary administration. The new Special Resolution Regime (SRR) introduced for the UK in February 2009 provides the ideal legal setting for this. It should not take long, a weekend at most. Basically, the Bad Bank just becomes a financial portfolio of toxic assets and derivatives, plus its stake in the Good Bank. It would not be allowed to invest in any new assets or to engage in any banking activities. It would manage the existing asset portfolio down and would cease to exist once the last asset has been sold or has matured. Among the clean assets the Good Bank buys would be the buildings, equipment etc. necessary for conducting the banking operations of the Good Bank.
If the UK government had not been daft enough to guarantee the £325bn worth of toxic assets on the balance sheet of the Bad Bank, there can be little doubt that the Bad Bank I have just constructed would have failed soon after coming out of the SRR. The Bad Bank, which is just a fund restricted not to invest in new assets, would be put into administration. The shareholders would be wiped out (more than 70 percent of RBS is now government-owned), and the unsecured and non-guaranteed creditors would determine what to do with the Bad Bank and its assets. Most likely there would be a significant debt-to-equity conversion and/or a large write-down of the debt.
The government would focus its financial support on the Good Bank, either by providing it with additional capital or by guaranteeing new lending and/or new borrowing by the Good Bank. Private capital could be attracted into the Good Bank too.
Distributional differences between the Good Bank and the Bad Bank solution
The Good Bank solution favours the tax payer. The Bad Bank solution favours the unsecured and non-guaranteed creditors of the zombie banks. ‘Tax payer’ includes those beneficiaries of public spending programmes that may have to be cut to meet the fiscal cost of purchasing or guaranteeing the toxic assets under the Bad Bank solution. It also includes those who lose as a result of future inflation or sovereign default, should either of these two solutions to dealing with the public debt created as a result of the Bad Bank solution eventually be adopted.
The Bad Bank solution also favours the shareholders of the zombie banks, but in the case of RBS, this is mainly the government and therefore the taxpayers. The amount of shareholder equity involved in the zombie banks is, in any case, negligible compared to the exposure of the unsecured and non-guaranteed creditors. The Bad Bank solution also saves the jobs and perks of the top management and the boards of the zombie banks - often the very people responsible for turning a once-healthy bank into a zombie bank.
There can be no doubt that, from a distributional fairness perspective, the Good Bank solution beats the Bad Bank solution hands down.
Incentive effects of the Good Bank and the Bad Bank Solution.
The Bad Bank solution creates moral hazard, because it rewards past reckless investment and lending. It also represents an inefficient use of public funds in stimulating new lending by the banks. To stimulate new lending, a subsidy to or guarantee of new lending is more cost efficient than the ex-post insurance of losses that have already been made on old lending, even though their true magnitude is not yet known. The Good Bank solution leaves the toxic waste with those who invested in it and with those who funded these activities, freeing government funds for reducing the marginal cost of new lending or increasing the expected return to new lending.
Both as regards moral hazard (incentives for excessive future risk taking) and as regards the efficient use of government funds (‘new lending bang per buck’), the Good Bank solution beats the Bad Bank solution hands down.
Financial stability implications of the Good Bank and Bad Bank solutions: saving banking, without saving bankers or the existing banks
The holders of bank debt, with the possible exception of perpetual subordinated debt (which counts as tier one capital in some countries), have become the sacred cows of this financial crisis. Regulators, central bankers and Treasury ministers are quite willing to see shareholders wiped out. After the demise of WAMU and Lehman Brothers, however, the unsecured creditors have become inviolable. Somehow, those in charge of macro-prudential stability, notably the Fed, have bought into the notion that if there is either a further default on bank debt, or a restructuring involving a significant debt-to-equity conversion, or a signficant write-down of the claims of bank bond holders, this will be the end of the world.
I just don’t buy it. Fortunately, I am not the only one. Luigi Zingales, the Robert C. McCormack Professor of Entrepreneurship and Finance at the Chicago Business School, has been advocating the case for mandatory debt-into-equity conversions, debt forgiveness and other up-tempo Chapter 11- style financialestructuring of banks and other defunct behemoths like GM, since the first days of the crisis (see e.g. (1) and his book Saving Capitalism from the Capitalists, co-authored with Raghuram Rajan). Robert Hall and Susan Woodward also feel no need to pay any special attention to or lavish any public funds on the toxic assets, their owners and those who funded them (the unsecured and non-guaranteed creditors) once the Good Bank has been established and sent on its way.
Part of the reason there appears to be this widespread belief that you have to guarantee all bank liabilities is that this is what the Swedish authorities did during their 1991-1993 banking crisis (see e.g. Lans Jonung’s paper “The Swedish model for resolving the banking crisis of 1991-93. Seven reasons why it was successful” . First, Jonung lists seven criteria for ‘successful’ resolution of a banking crisis. The paper does not demonstrate that this septet constitutes a set of necessary and sufficient conditions for success – if indeed the Swedish approach is deemed to have been a success. Second, success is in the eyes of the beholder. The Swedish banking system has been hard hit again in the current crisis by its overexposure (30 percent of annual GDP) to risky investments in Central and Eastern Europe, including the Baltics and Ukraine.
Every financial boom/bubble has been characterised by rising and ultimately excessive banking sector leverage, that is, by excessive lending to banks. If all the unsecured creditors of the banking system were made whole in the previous systemic crisis, it is not really surprising that the banks, and their creditors, are back for more.
In a more systematic study of the use of blanket guarantees of bank liabilitiesLuc Laeven and Fabian Valencia find the following:
The proposition that the consequences of inflicting losses on holders of bank debt are awful beyond our wildest imagining is voiced incessantly and loudly by bankers and by those long bank debt, especially insurance companies and pension funds. And a vigorous campaign is underway to extend the no-default presumption to the debt of pseudo-banks like AIG.
The most over-the-top, ludicrous piece of attempted scare mongering about the systemic risk implications of default by any institution I have ever read is the internal memorandum “AIG: Is the Risk Systemic”, of 26 February 2009, which is now all over the internet. Just one small sample: “The failure of AIG would cause turmoil in the U.S. economy and global markets , and have multiple and potentially catastrophic unforeseen consequences”.
I would have thought that, on the contrary, markets have discounted the likelihood of default by many of the major border-crossing banks, and by AIG, pretty comprehensively by now. After the US authorities bailed out Bear Stearns in March 2008, letting Lehman go belly-up in September 2008 was a bad surprise. Even then, I don’t accept the interpretation that it was Lehman’s filing for bankruptcy protection that triggered the cardiac arrest in global financial markets in the second half of September 2008. Instead the financial sector convulsions of the last quarter of 2008 were caused by the realisation that (1) most of the US and European border-crossing banks were insolvent without government financial support, that (2) the rot extended to the shadow banking sector (AIG), and that (3) the US authorities (Treasury, Fed, SEC) were not on top of the issue and that Congress was bound to act irresponsibly.
But even if it had been Lehman that triggered the financial upheaval, that was then. This is now. Banks, counterparties, investors and policy makers have had 6 months to adjust to the new reality and prepare for the eventuality of default on zombie bank debt and even on AIG debt. The bonds of large zombie banks trade at spreads over government yields comparable to those of automobile manufacturers (600 – 650 basis points). Their CDS spreads put many of these banks well into the default danger zone. Their stock market valuations are consistent with those of institutions not a mile away from insolvency and default.
The fact that zombie banks or AIG are self-serving when they plead systemic risk as an argument for further government hand-outs does not mean that they are wrong. It does lead one to verify more carefully the logic of their arguments and the quality of the empirical evidence offered in support. Let me just consider the argument that the main investors in bank debt (and AIG debt) - pension funds and insurance companies – are too vulnerable and too systemically important to permit the banks (or AIG) to fail.
Pension funds don’t go broke with adverse effects on systemic stability. If they are funded, defined-contribution funds, a reduction in their asset value means that pensioners will get lower pensions. If they are defined-benefit schemes (including ‘final salary schemes’), the risk of investment surprises is shared by the sponsors and the beneficiaries. When the Dutch pension fund ABP took a big hit last year, my parents did not get any indexation of their pension benefit. In past years, pension benefits had tracked earnings inflation, and occasionally price inflation. Should coverage ratios decline enough, even nominal cuts in pension benefits can be implemented. This may cause hardship, but not financial instability. No reason to favour the pensioners over the tax payers.
As regards insurance companies, I doubt whether “Insurance is the oxygen of the free enteprise system”, as AIG would like us to believe. Certainly insurance companies like AIG are not the only suppliers of oxygen. Insurance is a regulated industry. Orderly restructuring following administration and insolvency need not interfere with the provision of any of the essential infrastructure services required for the proper functioning of a market economy.
Regulators, especially but not just in the US, have bought the ‘don’t touch the unsecured creditors’ argument. The Fed especially appears to have swallowed it hook, line and sinker. This cognitive regulatory capture has turned the Fed, with the enthusiastic support of the FDIC and the US Treasury, into the most powerful moral hazard propagation machine ever.
If a bank or an insurance company like AIG is at risk of failing but is truly too big or too interconnected to fail (rather than merely too politically connected to fail), and if a Good Bank solution is not feasible, then the institution in question should immediately be taken into public ownership or put into a special resolution regime, if one is available. From public ownership it can be put into administration. Once in administration or under a Special Resolution Regime, it can be restructured decisively through a mandatory debt-to-equity conversion or debt write-down. There is no case for sparing the unsecured creditors.