Here we opine on NAMA when it was but a twinkle in our eyes…and a tingle in our wallet
To date, the international discussion on how to rescue failed banks has focused on repairing their balance sheets. This ignores the underlying cause of the problem – the deterioration of their asset base. In fact, in the case of the National Asset management Agency (NAMA) “bad” banks arrangements, the cure compounds the asset-base problems.
Two questions regarding NAMA:
First, we do not know how the assets can be priced to align the NAMA objectives of repairing banks’ balance sheets (with an incentive to pay high prices for transferred assets) and its duty to safeguard taxpayers interest (requiring the price to be set below the expected risk-adjusted value of the loans total).
Second, we do not know how the impaired assets will be treated under NAMA. One option is to keep them alive as zombie-development projects awaiting realisation decades from today. Another is to shut them down. Which option will be pursued will, in the end, seal the fate of largescale development land banks and half-baked development schemes.
And this is before we consider the fallout from a virtually inevitable creep of the NAMA remit to cover defaulting mortgages on principal residencies, credit-card debts and bad car loans.
Extent of the NAMA liability With respect to the first question, the US Treasury Department identifies the bad assets before they are fully impaired, using financial models that estimate future loan values under different economic scenarios. Ireland is yet to make even this first step but currently neither the Central Bank and Financial Services Authority of Ireland (CBFSAI) nor the Department of Finance DoF) and, least of all, NAMA have the capacity to develop and administer such model-based testing procedures. The CBFSAI has virtually no real expertise in risk management and pricing, while the DoF hasn’t enough real economic, financial and analytical capabilities to oversee a minor credit union, let alone control the NAMA.
Thus, ex-ante pricing transparency is the only guard the taxpayers have to limit NAMA’s monopoly powers.
So let us consider the loans that are nonperforming, stressed or rolled over with little chance of repayments any time soon. Banks’ provisions for future impairment charges are currently running at 4-5%. Independent and In house analysts are forecasting that some 12-15% of the entire asset pool of the Irish banks will be under stress by 2010.
In our view, this is a lower bound of the true state as:
(1) loans under threat to date will almost certainly remain under threat in 2010;
(2) the first quarter of 2008 saw a relatively benign trading environment, so 2009 is going to see even greater rates of impairment; and (3) the economic troubles underlying the rapid asset-quality deterioration are set to deepen in 2009.
We know nothing about the recovery rate on these risky assets. But, globally, AAA- rated collateralized debt obligations (CDOs) carry the recovery rate of only 32% on face value while for mezzanine vehicles, the recovery rate is only 5%. The default rates on the US corporate junk bonds (which are less risky than Irish development-linked loans due to their higher diversification, liquidity and transparency) is estimated to reach a whopping 53%, with a recovery rate of zero.
Given the perilous state of the Irish economy and the extent of the property related exposure for Irish banks, we see as reasonable (or potentially even generous) a 45-50% average recovery rate on the stressed loans. This implies that the expected final losses on the entire six banks pool of EUR165bn in property exposure (ex-Poland) will be closer to 25-30%, or EUR50bn.
For anyone who thinks that this figure is unrealistic, a recent McKinsey study showed that out of $2 trillion of impaired US assets, the eventual writedowns may total $1.5 trillion or 67%.
The above loss rate implies that NAMA will be purchasing the impaired assets at less than 28% discount to their face value, should the Government set the price to keep the six banks’ capital ratios at 8% minimum required levels.
Such a discount will imply an issuance of EUR36bn in fresh bonds to the banks, underwriting only EUR25bn in risk-adjusted assets on the NAMA-held EUR50bn book of loans.
The implied expected loss to the taxpayers from such an operation is EUR11bn in capital cost, plus ca EUR11.5bn in interest costs for a five-year bond to be covered out of tax revenue and a higher cost of banking.
It is worth noting that these costs of over EUR22bn for NAMA operations assume that Irish banks keep capital ratios at the required legal minimum after deleveraging their balance sheets.
In other words, these losses do not fully insure the banking system against future capital demands.
But an 8% capital requirement is now considered to be insufficient for operating a private bank. Instead, markets are demanding a minimum 10% capital ratio, with 12-14% being a golden target. If NAMA was to keep Irish banks private, the recapitalisation demand for the six-banks system, due to the NAMA-assets transfers, will add another EUR4-Bbn in costs to the exchequer bill.
Note: should NAMA buy EUR80bn in loans as discussed in recent reports, the associated required maximum discount rate will be 23% and the total losses to the taxpayers will be EUR43-51bn.
How can toxic assets be priced? Generally, assets on bank-balance sheets are valued either at a hold-to-maturity value or at fair value. Both frameworks fail in the current environment.
An alternative solution is that the Government can set up a two-stage process of buying stressed assets into NAMA. The first stage will involve a quasi-voluntary scheme that would establish a functional resale market for the stressed loans to be used in the second stage of purchasing.
To do so, the Government should set a basic level of discount on the assets, based on the publicly verifiable valuation model.
The discount should be fixed on the date of the scheme announcement to prevent future manipulation of the fair value by the banks. It should apply to all systemically important banks regardless of who holds the specific loan or what project it is written against. This will avoid political interference in the pricing of stressed assets.
Loans with interest and principal non-payment of less than three months can be sold at a fixed discount of, say 15% (reflective of the current expected default rates). Loans with non-payment of three to six months can be sold at a fixed discount of 25% (a rate that is more consistent with the US experience and the ECB discount lending criteria). Non-performing loans in excess of six months and repeated roll-up loans can be traded at a 50% discount equal to their estimated default risk. This first- stage transfer will remove the most toxic paper off the balance sheets of the banks.
After the first stage establishes a quasi-market pricing of the assets transferred to NAMA, the Government can retain the face-value discount on other stressed assets while allowing for some recapitalisation support to be given to the banks that need it.
The second stage involves using the same discounts on loans as in the first stage with the Government using bonds to swap for banks shares to cover some fixed proportion of the discount. In other words, the banks will still sell most impaired assets at a 50% discount but they will have an option to receive a roll-back of say 10-15% of the discounted value in the form of the NAMA taking new shares in the banks.
For example, a loan package of EUR10bn with an average non-payment of more than six months will be sold to NAMA for EUR5bn, but the bank will have an option to sell EUR500-EUR7,500m worth of new equity to NAMA at the same discount on the share price as on assets sold to NAMA.
The advantage of this scheme is that the clean-up of banks balance sheets will be systematic and non-distortionary. The disadvantage is that it still saddles the taxpayer with the task of recapitalising the banks after they take a hit on their capital base under NAMA. This, however, is inevitable under all possible scenarios for toxic-assets removal.
In our view, the only real option to avoid the need for endless rounds of recapitalisations is to nationalise systemically important banks outright. A nationalisation option will allow the Government to keep the capital base of the banks at an 8% limit, outside the markets’ demand for higher capital reserves. In addition, under nationalisation NAMA can choose specific assets off banks’ balance sheets to create a blended portfolio of loans with lower expected default rates.
Avoiding zombie land banks The second problem with the Government proposal is that we do not have any idea as to how the impaired assets will be treated under NAMA. Upon purchasing the loans, the Government will have an incentive to keep the underlying assets alive as a zombie development projects. This is because as long as the development-zoned land remains “active” as an investment project, it will retain some notional value on the NAMA balance sheet, creating an illusion of value to the taxpayer.
Of course, much of the existent recent vintage land banks that NAMA will end up holding will cover speculatively purchased agricultural and industrial land with virtually no hopes of being developed in the next 15-20 years.
Another option is to shut these projects down, dezone the land and either release it into the market as agricultural land or retain it as public-use land. This option implies NAMA writing down the asset values of such land.
In our view, the Government will be wise to opt for the second option, converting improperly zoned development land into a mixture of leasable publicly-owned land (useable for sustainable developments) and commons (for public amenities, such as parklands). Incidentally, our pricing scheme described above incentivises such conversion as most of speculative land banks will fall under the heaviest discounted-price category, minimising the value of the writedown and maximising land rents to be collected on leasable lands.
This process will only be further enhanced by imposing a direct land-value tax on development sites, mentioned in this column in the previous issue.