Save banking, not the bankers or the banks; the case of ING

I had been planning to blog today on US Treasury Secretary Timothy Geithner’s proposals for saving/reviving financial intermediation in the USA.  However, picking through the entrails of this multi-faceted, surprisingly incomplete, seriously underfunded, occasionally well-designed but mostly inadequate, counterproductive and unnecessarily moral-hazard-creating set of proposals was just too depressing.  I will wait till I am at my parents’ home this weekend, mollified and mellowed by my father’s good claret, before I review the Geithnerbharata.  But as a four-finger exercise before the main concert, I shall discuss here the second Dutch government bail-out of ING.  Many of the issues involved in and principles raised by this deeply unfortunate exercise also are central to the Geithnerbharata.

On 26 January 2009 (my sister’s birthday – happy birthday, Hens!) the Dutch government mounted a second financial support operation for ING Group.  The state had earlier injected 10bn euro worth of capital into ING.

The assistance takes the form of a back-up guarantee facility for a portfolio of $39bn (face value) worth of securitised US Alt-A mortgages.  Under the deal, 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 are 80% and 20% respectively.

The bank pays a guarantee fee to the state.  The state document I saw did not specify the magnitude of the guarantee fee, or how it was arrived at.

The state pays ING a management and funding fee.  Again, I don’t know the amount or how it was arrived at (it would be cute, however, if the guarantee fee and the management and funding fee just happened to cancel each other out!).

The other relevant conditionality is that ING is to provide 25 bn euro of additional credit to businesses and households and that there will be no bonuses for 2009 and until a new remuneration policy is adopted.  The CEO was told to fall on his sword.

I am sure the stock market loved this deal. And I am sure the politicians and civil servants who put it together loved this deal.  Politicians and their advisers love off-balance-sheet and off-budget financing.  Guarantees are great from this perspective.  Because they represent a contingent liability, they are off-balance sheet.  Only if and when the guarantee is called, will the payments under the guarantee show up in the budget.  The tax payer, however, should hate it.  And so should all those concerned about moral hazard: the effect on future incentives for excessive risk taking by ING and other Dutch banks.  It is also unfair: it bails out the shareholders and unsecured creditors of the bank, who made lousy investment decisions and should pay for that.  Except for the departure of the CEO (which was appropriate, and indeed overdue), the rest of the executive and supervisory boards are still in place.

I am not a Dutch taxpayer, but I too think this is a dreadful deal.  The objective of strengthening financial stability and fighting the economic downturn by encouraging new lending could have been achieved at much lower cost to the tax payer, without distorting incentives for future risk taking and without doing violence to fundamental notions of fairness.

Why is the guarantee a bad deal for the tax payer and an unnecessary source of moral hazard?

The guarantee is a good deal for ING and a bad deal for the tax payer because the market valuation of the Alt-A portfolio did not imply the 10% discount (from $39 bn to $ 35.1 bn) that was used to define the reference value for the guarantee, but a 35% discount (from $39 bn to $25.4bn).  It is possible that the hold-to-maturity value of the portfolio (the present discounted value of its current and future cash flows, discounted at an interest rate that is not distorted by illiquidity premia, is $35.1 bn or more.  Possible, but not likely.

It is possible that the guarantee fee appropriately prices the risk assumed by the state.  Until I see the numbers and can verify the assumptions on which they are based, I consider it possible but not likely.  It is possible that the management and funding fees reflect the true incremental cost to ING of the deal with the government.  Possible, but extremely unlikely.  ING was and remains the owner of the portfolio.  They would have had to manage and fund the portfolio even if the state had not guaranteed it.  For the state to pay ING a management and funding fee is ludicrous.

The Dutch state is heavily exposed to its oversized banking and insurance sectors.  It is possible that those Dutch banks and insurance companies that are technically solvent today only because of past, present and anticipated future state financial support collectively have bad assets on their balance sheets that makes them to large to save.  The Netherlands is not as exposed as Iceland (where the banking sector was to large to save), Belgium (where the fate of Fortis can be viewed as evidence that the Group was too large to save), or Ireland (which will in all likelihood provide the next test of the too large to save theory). At least one Dutch bank, Rabobank ( a mutual bank!) is widely help up as an example of viable cross-border banking.  But there is at least a material risk that the fiscal spare capacity of the Dutch authorities would be insufficient to fill the solvency gap in the balance sheets of the Dutch banking and insurance sectors.

Let’s ask ourselves how the Dutch authorities could stimulate lending to the real economy by the two largest banks, ING and the Dutch rump of ABN-AMRO/Fortis.  ABN-AMRO and Fortis Nederland (which contains insurance as well as banking) is 100 percent state-owned.  State ownership does not, of course, imply that there is a sovereign guarantee for the unsecured creditors of the bank, including its bond holders.  Only the retail depositors are insured and guaranteed by the state, up to a limit of €100,000.  But the state can and should make it clear that it will not guarantee the other creditors of the bank.  It can do this easily, because it is the sole shareholder, by splitting ABN-AMRO and the banking parts of Fortis it owns into a good bank and a bad bank.  The good bank would get the deposits and the good assets of ABN-AMRO and Fortis Nederland.  It would get adequate new capital, probably from the state, although private participation could be invited also.  Government guarantees would be provided only for new lending and/or new borrowing by the good bank.

The bad bank would be left with the toxic, bad and dodgy assets of ABN-AMRO and Fortis Nederland and the capital of the original banks. No new capital would be injected by the government into the bad bank.  The bad bank would lose its banking license and would simply manage the inherited portfolio of bad assets so as to maximise the discounted value of its future cash flows.  It would not be allowed to invest in any new assets or to engage in any other activities.  It would not receive any further guarantees for its assets or its funding. The assets could be held to maturity or sold if and when a liquid market for securitised US Alt-A mortgages ever revived.

It is quite possible (likely?) that the bad bank would go into administration and would be declared insolvent.  In that case the existing shareholder (the state) would lose its investment and the other creditors would effectively end up owning the assets of the bad bank.  The state would of course have to make good on any guarantees it has already, unwisely, provided.  These unsecured creditors (mainly pension funds, insurance companies and other institutional investors) would be quite likely to suffer large losses.  That is too bad, but it is better than shifting the burden of these losses onto the tax payer.  The losses have been made.  It minimises moral hazard to have these losses born by those who made the bad lending and investment decisions.  When this can be done without aggravating the financial crisis, it certainly ought to be done. The bad bank is not systemically important. It should be left to sink or swim on its own.  The same approach can be applied to the fully state-owned insurance company.

The simplicity of this solution is due to the fact that the state already owns all of ABN-AMRO and Fortis Nederland.  What should be done about ING, where the state does not own any common equity?  The easiest solution is to create a new ‘good bank’, New ING, say, with capital provided by the state and possibly also by the private sector.  The good bank would take the deposits of ING and purchase any of the good assets of ING it is interested in.

The valuation of these good assets would not represent a problem, because part of the definition of ‘good asset’ is that there either is a liquid market price for it or, in the case of non-traded assets, that the buyer can determine their value in a straightforward and transparent manner.  It is possible that none of the existing assets of ING would be bought by New ING.  In that case, the assumption of ING’s deposit liabilities by New ING would be effected by a loan from the state to ING, and the asset-side counterpart on New ING’s balance sheet to the deposits acquired from ING could be a matching amount of government debt.  ING (now Old ING) would also have its banking license withdrawn and would not be able to engage in any new lending or investment activities.  It would not receive any further government guarantees, nor would the government purchases any of its assets.  It would be left to sink or swim on its own.

Scarce public funds should not be used to purchase or to guarantee stocks of existing assets.  They should instead be used to guarantee or fund new lending and borrowing by the new good banks (like New ING) or directly by non-financial enterprises.  Again, this preserves systemically important banking activities (new lending and borrowing to enterprises and households), without saving either the bankers or the banks.  This minimizes moral hazard and creates the right incentives for the future exercise of prudence in investment and lending decisions.

To be able to achieve the economic efficiency objective of stimulating flows of new lending and borrowing without subsidising existing stocks of assets and liabilities, it is necessary to achieve a legal and institutional separation between the owners of the existing stocks and the owners of the entity that will be engaging in the new lending.  The Good Bank model is one way to achieve this.  Soros’s ‘Side Pocket’ solution is economically equivalent to the Good Bank solution.

The difficulties associated with trying to be a little bit pregnant

Often government financial assistance to banks imposes conditionality, costs and constraints on the bank’s management and existing shareholders without taking full ownership and control of the bank.  Examples are; onerous financial terms; constraints on bonuses and other aspects of executive and board remuneration; constraints on dividend pay-outs and share repurchases; constraints on new acquisitions and on foreign activities; guidance and direction on how much to lend and to whom.  All these encumbrances last until the state has had its stake repaid.

This creates terrible incentives encouraging banks that are already in hock to the government to hoard liquidity and hold back on new lending activities to get rid of the government’s interference.  Banks that are not yet financially dependent on the government will likewise pursue extremely cautious and conservative strategies, hoarding liquidity and capital to stay out of the clutches of the state.  Such half-way houses are inherently dysfunctional.  To get banks to lend again, you have to move towards one of the two extremes: either give the banks state money without any strings attached and at a very low financial cost or take full control (possibly but not necessarily through full state ownership) of the banks that are the beneficiaries of state money.  Towards the end of the Bush administration, US policy was evolving towards throwing money at the banks with little or no conditionality or cost.  Under the Obama administration, much more onerous conditionality is likely to be imposed.  The willingness to move towards full public control and ownership does not yet exist in the USA, however.

Throwing unconditional money at wonky banks creates terrible moral hazard.  It is also very wasteful of public resources, as it often subsidizes and makes up for past losses on existing assets rather than focusing on stimulating new flows of lending and borrowing.  The Good Bank solution saves resources, because it leaves the existing bad assets to the old bank’s owners and creditors, and minimizes moral hazard.

The Good Bank solution has two problems associated with it.  First, the state is likely to be the main, perhaps for a while the only, owner of the new good bank.  The state is a lousy banker.  So of course, is the private sector, at least for most border-crossing universal banks in the North Atlantic Region for the past ten years.  The question of how to get the state out of the banking business again has to be addressed as soon as the state goes in.

Second, the Good Bank solution has to be imposed swiftly and preferably as a surprise. If the public anticipates that a wonky existing bank is likely to be transformed into a bad bank, there will be a flight of depositors and other creditors, and the existing bank might collapse before the new Good Bank has been put in place.  As the new Good bank would use most of the real assets and personnel of the old bank (the bad bank would really be just a balance sheet with the bad assets and a few fund managers), it could be set up and functioning almost instantaneously.

Financial nationalism

Inevitably but regrettably, the ING guarantee package involves a form of economic protectionism – a repatriation of cross-border banking.  The requirement that ING is to provide 25 bn euro of additional credit to businesses and households is unlikely to be satisfied by ING lending and additional 25bn euro to businesses and households in Turkey, South America or even Germany.  This is meant to be lending to Dutch businesses and households.  As it is Dutch tax payer money that is funding the bail-out, such financial protectionism is hardly surprising.  But it will limit the scope for global diversification of assets and of funding compared to what would be possible in a world with a supranational fiscal authority and compared to a world in which the importance of the fiscal authority as the recapitaliser of last resort and the ultimate source of bail-out support had been forgotten, as it had been in the North Atlantic region since the Nordic banking crises of 1992 and 1993.


So the Dutch state’s ING guarantee got it 100 percent wrong.  It does very little to stimulate new lending to the real economy.  Instead it subsidises/bails out the owners and unsecured creditors of a bank that holds large stocks of (bad) assets,  that is, those who in the past have (carelessly/unwisely) extended credit to the bank.  This maximises moral hazard for very little gain in systemic financial stability and for very little direct stimulus to new lending.  It also incentivises the bank to hoard capital and liquidity to enable it to pay off its obligations to the state and regain its operational independence as soon as possible.  Finally, it exacerbates the tendency to restrict cross-border banking, quite possibly to a degree that could harm the efficient diversification of asset and funding risk.  This deeply defective bail-out should not be repeated, in the Netherlands or elsewhere.

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