Tuesday, April 14, 2009

Nationalize or else?..

I just received a good comment to an earlier post (here) that warrants a separate attention.

"Regarding NAMA, it seems to me that the one big advantage to this scheme is that it means someone will lend us enough money to cover the bank's bad debts, via the sleight of hand of issuing government bonds to the banks and then them redeeming this in hard cash from the ECB. I strongly suspect the Irish government would be hard pressed to borrow this kind of money from anywhere else.

What I don't understand is why we don't first just nationalize the banks. The question of proper pricing then becomes less of an issue. We'd be just moving money between different arms of the state.

One thing I've wondered about: can this device for swapping government bonds for euros only be done by a commercial entity? If we first nationalized the banks would such a move then be precluded? If so, maybe the government do secretly intend to largely nationalize them at a later stage after the cash has already been received from the ECB. I do hope there's some technical reason like this for not first nationalizing the banks, that the reasons are not purely political, because I've no confidence that the taxpayers will end up paying a fair price for these assets. Finbar."

There are several arguments in favour of nationalizing first, then deleveraging bad assets, recapitalizing and re-floating the banks. And there are several arguments against such an approach. I will first deal with arguments in favor of nationalization...

Pro-nationalization arguments:
  1. Clarity of valuations: banks are not going to willingly reveal all pertinent information concerning loans quality to NAMA, so nationalizing them and then opening their books will provide much needed clarity concerning fundamentals relevant to valuations and pricing;
  2. One-shot recapitalization: whatever price NAMA sets for impaired and stressed assets, such a price will either be too low to allow the banks to continue operating without further recapitalization injections, or too high to allow the Exchequer to recoup significant share of losses. Nationalizing the banks will resolve the problem, as capital requirements can be dropped significantly under a public guarantee on publicly-owned banks. The upside here is significant (see below);
  3. Ownership-liability symmetry: under nationalization, ownership of banks assets will be fully coincident with the holder of liabilities - the State. This prevents a situation where taxpayers money is being used to underwrite private shareholders and bondsholders objectives;
  4. Bond holders can get a haircut: under nationalization scheme, the Government can impose a stamp duty on bondholders in Irish banks, allowing for a partial recovery of funding and imposing a haircut on banks bondholders (currently covered by a blanket taxpayers'-financed guarantee);
  5. Maximizing recovery for the taxpayers: if the objective of NAMA is to deliver value to the taxpayers, while deleveraging the banks balance sheets, nationalization, with a clear pre-commitment by the state to disburse banks equity via a voucher-based privatization within say 3-5 years will deliver both (see below for an outline of the scheme);
  6. Avoiding discriminatory treatment of individual loans: Under NAMA arrangement, some developers / business owners that have performing loans against them might not want to face an arbitrary transfer of their loans to NAMA. This might be a litigious issue that can be fully resolved by an outright nationalization of the banks;
  7. Change of the guard: Under nationalization, the Government will have a full right to change the executive structure of the banks and their boards to bring in new blood to run these institutions, breaking away with legacy issues in management.

Voucher scheme

To pre-commit to such a scheme, the Government can issue 3 or 5 year options on shares of the banks. For example, a part of existent equity in AIB can be converted into options at a price on the day of nationalization. Suppose, for the sake of illustration, that nationalization takes place on May 4, 2009.

Suppose that the Government commits to voucher-privatizzing 50% of the value of shares, retaining 50% shares in own account. April 30 closing price for AIB is €X. The European-style call options are issued on May 4, 2009 at an exercise price of €X with maturity date of, say, May 4, 2012.

The Government re-floats a part of its share holding in AIB on May 4, 2012 (Swedish Government retained ca25% of the banks shares on own account after re-privatization, so Irish Exchequer might want to do something similar). This sets the expiration price on AIB shares at S. If S>X, households holding options will exercise them, collecting S-X in profit. If not, they will forefeit any gains with no loss.

Two questions arise concerning such transaction:
  • How the vouchers should be disbursed? My preference is to issue vouchers on a flat-rate basis to all households in Ireland in order to achieve a voucher-distribution that is reflective of the economic stimulus in line with an across-the-board tax cut;
  • What will happen to AIB shares when vouchers are exercised? Nothing: markets at IPO will be pricing in an inflow of shares from the households as it will be pre-announced in advance.
The Government can collect a special rate CGT on such profit realization at, say, 30%, so that in effect there will be a 0.3*(S-X) payout to the Exchequer in addition to the retained shares value.

The upside to capitalization savings

Banks equity capital (BEC) = assets net of liabilities must legally not fall below 8% of the Risk-Weighted Assets (crudely for any given asset - e.g a loan or a bond - held by the bank, RWA =risk weight*asset value=RW*AV).

At the end of 2008 both banks hold ca €80bn in property loans of various quality. Not all of these loans will be earmarked for NAMA, so, having no better guidance from the NAMA itself, assume that the banks would want to off-load ca 3/5ths of this amount or €48bn.

(How do I get to this number? AIB has total assets of €182bn, RWA of €116bn, RW of 116/182= 63.7%, BEC €9.28bn and the actual Tier 1 capital of €9.9bn. BofI has assets of €204bn, RWA of €134bn, RW of 65.7%, BEC requirement of €10.72bn against the actual T1 capital of €10.1bn. Note that RW(BofI)>RW(AIB) implies lower quality of the BofI book. Prior to the first round of recapitalization, combined RWA €250bn, BEC Tiers 1&2 requirement of €20bn (0.08*250bn) just covered by the actual Teir 1 held. Any change in the NAV of underlying assets would have triggered a rise in RW thus driving the banking system below the 8% requirement, so the Government injected €7bn, thereby providing for the €87bn RWA cushion and raising Tier 1 capital to 10.8%. While sounding like a high number, this is pittance compared to the US and UK trend toward raising T1 ratios to 12-14% that would require a further injection of €3-8bn in cash, assuming there has been no deterioration in the assets quality since the end of 2008. Further note that total 6-banks property exposure ex Poland for AIB is €165bn. So far, we do not know how much NAMA will take on, but in the case of Securum - Sweden's bad bank - only took on non-performing loans. Now, AIB assumes max 25% non-performing loans on total development & property investment loan book, with current running non-performing loans at 3.5%, so our €48bn assumption is about coincident with the ca 25% non-performing loans assumption on property exposure across the 6 banks).

As Government bonds carry a RW=0, the value of NAMA bonds replacing specific assets will be excluded from RWA calcualtions. If NAMA buys €Xbn in loans at discount
d%, then banks will get to write off €Xbn of assets, get €(1-d)*Xbn in state bonds in return and face a net cost of €dXbn to their capital, so that the combined banks RWA becomes €(250-(1-d)X)bn against Tier 1 capital of €27bn post re-capitalization. Writing off €dXbn of the value of the loans will hit the banks straight into their book value, thus cutting their equity capital - and directly hit their Tier 1 capital as well.

So Tier 1=27bn-dX=8% of RWA=250-(1-d)X. In other words, 0.08*[250-(1-d)X]=27-dX. Now, solving for discount factor:
d=[7+0.08X]/(1.08X).

If the Government wants to buy 3/5ths of the property-related loans, X=€48bn and d=20.9% - a scenario that would see the state issuing €38bn in new bonds - over 1/2 of the entire current Government debt.

Analysts estimate that the total loans impairments across BofI and AIB can run between €19-25bn. Adverse selection under the voluntary NAMA scheme imply that the banks will dump the lowest quality assets first. This means that under the scheme of 60% of loans being purchased by NAMA, the cost of the scheme - €38bn will be underwriting the asset base with expected recovery of just €48bn-19 or 25bn = 23-29bn, making an immediate loss to the taxpayers of €9-15bn.

Under nationalization scheme, the Government can blend assets at its own choosing, spreading the loss-implying assets across the books and it can drive T1 capital to 6% if it wants to. This would imply that, under an unbiased weighting scheme, NAMA will get €11.4-15bn in loss-inducing assets against the book that has
d=[14.9+0.06X]/(1.06X)=[14.9+0.06*48]/(1.06*48)=35%
costing the Exchequer €31.2bn in new bonds for an asset base with underlying recovery of €33-36.4bn - a nice expected profit of €1.8-5.2bn.

And this is the exact value of nationalization...

Arguments against nationalization will be dealt with in the follow up... (I need a smoke break!)

4 comments:

yoganmahew said...

I'll take you maths and argument at face value, but I tend to agree with Finbar's comment that the reason the banks will be only RBS nationalised (80-90% by equity recapitalisations) is that the whole scheme is posited on the EU/ECB allowing the government to quantatively ease by issuing bonds directly to NAMA and the banks (for recapitalisation of their losses on transfer of assets).

As these bonds won't have gone through the bond market, they are not 'priced', therefore there is no real limit to the issuance? Would they possibly be zero coupon bonds? Just for repo purposes? We'll have to wait and see, I suppose.

TrueEconomics said...

Yoganmahew, I don't see a barrier to nationalized banks accessing ECBs discount window - i.e monetizing government bonds. In terms of pricing - you are correct - they will have to set some 'arbitrary' price with some linear relationship to the past yields, but this price will not be tested in the market directly. Indirectly, however, I certainly hope the markets will not be stupid enough to take other issues (and we will have to issue other bonds to finance current spending habits of this Government) as being in addition to the NAMA issues, so overall, cost of public financing will rise in line with NAMA-led deterioration of debt/GDP ratio. This will constitute another layer of subsidies to the banks, as under NAMA their borrowing will be financed at a preferential rate to the direct public borrowing, i.e we will be paying higher price on public debt because the banks will need to be recapitalized and cleaned up. Then again, this is the same as with direct recapitalization...

yoganmahew said...

I don't disagree with you at all Dr. G. in terms of nationalising the banks - the case would seem economically apparent, indeed, it would also be politically popular. But it is not being done, so what could be stopping this?

It must be external political pressure. BBM would sell their grandmothers to get re-elected (as would most politicians!), so why have FF turned from populism to prudence?

My further guess is that government bond auctions will continue to be sold to domestic buyers, at rates at or about the current level. These will be paid for by, er, repo money from recapitalisation bonds issued directly to the banks. The market in Irish government treasuries is small, illiquid and so price is easily, eh, 'adjusted'.

I suspect that it would stick in the craw of the straight-backs in Europe if it was state-owned banks buying government debt, so the conscience fig-leaf of at least partial ownership and partial management is maintained.

Or should I take the tinfoil hat off now?

Anonymous said...

Hi Constantin. Thanks for perhaps the most comprehensive info on nama/nationalization I've seen! I'm no economist but your numbers will make interesting food for thought. I like your idea of a voucher scheme. My thoughts do tend along similar directions to yoganmahew that there may be other external reasons for the government going with the current scheme. As he says, a good portion of recently issued government bonds seem to have been purchased domestically. It might be a step too far for a major purchaser of Irish debt to be also wholly state-owned. Finbar