Note: Karl Whelan's post on Nama Business Plan is available here.
So let us start with Nama Business Plan published tonight: the main claim is that Nama is expected to generate a net present value return of €4.8bn by 2020.
I beg to differ. Here is why in two steps:
“This €77 billion is made up of approximately €49 billion land and development loans (€28 billion and €21 billion respectively) and approximately €28 billion in associated loans.” Of the latter, €14.2bn is in derivative instruments.
Now, land values have fallen by some 70% plus, with some land now valued at a 90% discount. What the recovery rate on these loans? Assume 30-35%, to the total loss of €18-20bn.
Development loans currently carry default and roll-up rates of well in excess of 40%. Suppose Nama buys an average portfolio of these and that the default rate rises to 1/3 of all stressed development loans. Expected loss here is therefore around €7-9bn.
Associated loans include second recourse and non-recourse loans and cross-collateralized loans. They have lower seniority on underlying assets. And this includes (50%) derivatives – instruments that actually cannot be priced directly without requisite information that has not been supplied by DofF. So suppose the default rate here is the average of the above two rates, or ca 50%, to the total loss on this part of the book of €14bn.
Add this up: total expected loss on Nama loans book value is €39-43bn before we factor in roll ups of interest. Day one of operation, Nama will be holding the portfolio of loans with expected value of €77-€41=€36bn against the liability of €54bn, which implies it will be in the red to the tune of €18bn.
Make another clarifying assumption. Assume that for the last segment of the book – the associated loans – derivative instruments are similar to the average market derivative contracts as stipulated in Table 3 of the BP. This pushes losses on this part of the book up by additional 25-35% of the derivatives segment value. The total loss Nama will incur on day one of its operations will then be a staggering €21.7-23.1bn.
“The estimated aggregate average loan to value (LTV) rate for these loans is approximately 77% i.e. the value of the real estate collateral at the time the loans were originated was €88 billion. The loans were made over a number of years and not all were made at the peak of the market.”
Suppose this is true, although I have no confidence that this number is real. Suppose average vintage of the loans is 2005. Land is currently at below 1999 levels in pricing. Development projects are around 2001-2002 pricing and completed property is around 2004. Assuming we are at the bottom, average LTV on these loans today is around:
This is LTV ratio on Nama purchase as of today of 98.2%. Not 77%, but 98.2%.
If average vintage of Nama loans shifts to 2006 (a more likely scenario, as Nama will not be buying an ‘average bank loan, but a non-performing loans portfolio with so-called ‘performing’ loans to be mixed in coming from stressed loans side of the balance sheet), then the actual today’s LTV shifts to:
This is LTV ratio on Nama purchase as of today of 112%. Not 77%, but 112%.
Incidentally, Nama ‘Business Plan’ contains no sensitivity analysis of this sort or of any sorts – neither for expected inflation, nor for spreads on bonds, nor for cost of administration, or for any other assumptions.
Step 2: redoing Nama balancesheet:
Table 5 clearly states that Nama expects life-time default rates for all loans and derivative instruments transferred to be 19.35% of the book value of loans at origination! Business Plan admits (page 9) that in the last year alone the banks took a charge of €7.3bn on the book – just under 10%. Thus, DofF expects 2009-2011 default rate to be only 10% more. This is for a book that overall contains 40% non-performing loans already! It is simply a case of amazing degree of optimism.
Let us do the math for alternative scenario. Suppose the default rate overall will be 33%, in which case without challenging any other DofF assumptions in Table 5, the net gain of €4.8bn turns into a loss in present value terms of €10.2bn. Just like that!
Now, let us challenge the assumption on Nama yields. DoF data is shown in the Table below. The second table changes yield assumptions and retains my default assumption above:
Now, per table 2 above, combined assumptions of more realistic default rate and more realistic yields (consistent with current yield, adjusting for default rate expected through 2011), and recognizing that derivative instruments yields are unlikely to be achieved at all, bottom line Nama is now expected to yield an €8.6bn loss in present value terms.
Shall we move on? Assuming slightly steeper curve on the cost of bonds financing, table below shows that expected Nama losses can reach €11.5bn in present value terms (Table 3).
One last thing left to do. Recall that per Nama own Business Plan admission, 40% of loans are currently producing a yield. This implies that 60% are non-performing. If yield curve were to rise over time as Nama assumes, these loans are not going to start repayments at any time in the future. So suppose the default rate assumption goes to 45%. Table below shows the end game:
A loss of €19.1bn in real terms!
And this is before we compute the opportunity costs of this money.
Conclusion: DoF estimates for Nama make absolutely no sense. The best scenario I get is a loss of €10.2bn. The worst one yield losses of €19.1bn.
Note: the above do not include the cost of managing the Nama loans by the banks. These ordinarily range around 0.5% of the total loan value per annum. Suppose the banks will be able to pass these costs on their paying customers (you and me). The net effect will be an annual added cost to businesses and paying customers of €270mln.
Note: All Nama flows are targeted for 2013, which in effect saddles future Government with the entire obligation under Nama. A rescue package, then, for banks, developers (with a repayment holiday until 2013) and... FF...