Friday, August 6, 2010

Economics 6/8/10: Anglo's plans & systemic risks


Here are some interesting questions (note - just questions for now, on the foot of comments made by Prof Brian Lucey earlier today) regarding the 'Good' v 'Bad' Anglo plans.

Take it from the top: we started with a bank with €71bn on the books valued at valuations of the peak markets. This is now allegedly going to:
  • €25bn of the face value of loans pre-writedowns - to the 'Bad' bank, implying that these loans are so poor in quality, even Nama, with an average 50%+ (LTEV-inclusive) haircut is not touching it. This implies that even in the long run, these loans are not going to generate more than, say, 30% recovery rate (a generous 30% that is, but let's take it as such. Note: that is across the entire loan book of the 'Bad' bank);
  • €10bn of the face value of loans pre-writedowns - to the 'Good' bank, implying that these loans are better than Nama average, so the LTEV on these loans is above 50%. Assume that the LTEV on them is 60% (which makes them better than Ulster Bank's Irish book, per today's results for Ulster - again a generous allowance, but let's entertain it);
  • the remainder is going to Nama.
Now, another little factoid: Central Bank of Ireland has lent Anglo €11.5bn under a MLRA repo agreement secured against the non-Nama loans.

Per Anglo last published results: "Sale and repurchase agreements with central banks include €12.2bn (30 September 2008: €7.6bn) borrowed under open market operations from central banks and €11.5bn (30 September 2008: €nil) borrowed under a Master Loan Repurchase Agreement (’MLRA’) with the Central Bank and Financial Services Authority of Ireland."

Let's do some simple math.

Value (recall - I am factoring referencing to Long Term Economic Value, not the current mark-to-market value, which is even lower):
  • €10bn in 'Good' Anglo can be (optimistically) expected to yield €6bn valuation using LTEV;
  • €25bn in 'Bad' Anglo can be (again, optimistically) expected to yield €7.5bn valuation using LTEV.
  • Allow for 1.5% margin of costs on both sides, to €525mln pa, or over 5 years - i.e. much shorter than Nama horizon - €2.6bn (note: current bank cost structure, which one can expect to be preserved as both banks go about conducting impaired assets recovery - a higher cost activity)
  • Total non-Nama book value is, inclusive of LTEV net of expected management costs, therefore could be already around €10.9bn.
Against this value of €10.9-13.5bn, CB of Ireland holds €11.5bn worth of loans to the Anglo.

Now to the question: Does this mean that CB might be facing a potentially significant loss on the repos?

This possibility raises two issues:
  1. If the repos are spread across 'Good' and 'Bad' bank, then the 'Good' bank is hardly a feasible undertaking, as repos alone already exceed the value of the 'Good' bank even absent impairment charges, while 'Bad' bank has clearly no ability to repay any fraction of these;
  2. If the repos are inherited by the 'Bad' bank, then, either CB has to declare a loss (and I am not sure how it can do this), or the taxpayer is on the hook for the repos by having to pay them down through the 'Bad' Anglo.
Now, alternatively, let's ask the following question: to recover CBs repos from non-Nama assets, we need to have a combined 'Good' & 'Bad' banks recovery rate of 33% (not covering the costs of operating both banks, funding, bond holders etc). And this, once again, refers to the valuations done on the face value of the loans before any recent impairments - before the bubble burst. To recover all other CBs' funds, plus our own - we need a recovery rate of ca 68% (again ex all costs etc). That's really highly unlikely, folks.

Quite a dilemma, then, especially since ECB (see here) didn't approved the repos in the first place... and since Anglo also owes the other Eurozone Central Banks some €12.2bn more.

What could Mr Trichet say about Anglo's priorities in repaying the loans? Would he be (a) so kind as allow CB of Ireland to recoup its repos, which ECB thought were dodgy enough to refuse to take them itself? or (b) insist that other CBs be repaid first before our own repos are covered? If (a) - I'd say Mrs Merkel would have a few kind words to say to Mr Trichet, given her electorate's feelings about having to bailout Greece. If (b) - the above potential negative valuation of the repos will have to be multiples larger...

Just asking some questions for now... Wonder if there are any answers out there...


Anonymous said...

30-40% of 25bn is 7.5-10 and not 3-4bn as mentioned above in the 'bad bank' heading.

Scarab said...

I'm confused by "implying that these loans are so poor in quality, even Nama, with an average 50%+ (LTEV-inclusive) haircut is not touching it."

NAMA acquired land and development and associated loans, regardless of their credit quality. The fact that loans were not purchased by NAMA only indicates that the loans do not relate to land or development or to a developer with outstanding land or development loans.

Hence I don't understand your basis for the value of impairment provisions against good bank/bad bank assets. Indeed per Anglo's accounts 70% of NAMA loans are impaired whereas 26% of the remaining loan book is impaired.

What am i missing?