Banking Expert Peter Mathews on fixing Ireland's Banking Crisis

Writing my last post and thinking about Deputy McGrath’s political speak raised a couple of questions in my mind about the NAMA business plan and so I’ve gone back, yet again, to look at the numbers. I wanted to study closer the argument about the value of the underlying assets and their ability to cover the NAMA loans.

Peter regularly makes the point that you can’t collect more than 100% of a loan. A common response to that point is that you may not but you can recover the value of the underlying asset. And on that basis, the borrower may default on the loan, the lender takes ownership of the underlying asset, and could, theoretically, sell the asset for more than the amount owned, and so collect more than 100% of the loan. Particularly if the lender can afford to wait ten years before selling the asset, suspending for a moment a critical look at who could afford to do this and at what cost.

Before we get into the detail of the Business Plan, let’s refresh in our minds the general state of our main lending insitutions (all numbers are €bn):

We’re out on a €190bn credit limb beyond where we prudently should be. Hopefully people will end up behind bars for the reckless negligence that allowed such a situation to arise.

But back to the task at hand, understanding what can be salvaged from the underlying assets. Let’s start with the rationale as it is explained in the NAMA Business Plan (13 Oct 2009):

On page 7, “NAMA Portfolio – Key Parameters”:

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.

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.

It is estimated that, of the total portfolio identified for potential transfer to NAMA, approximately €9 billion consists of interest rollup.

NAMA estimates an approximate 47% average decline on an aggregate portfolio basis including the impact of declines in overseas markets.

It is estimated, based on the application of statutory adjustment factors, that the consideration to be paid to participating institutions will be €54 billion, the estimated long-term economic value of the eligible assets. This will result in an uplift of approximately 15% on the current market value of the collateral of loans identified for transfer, equivalent to a discount of 30% on their loan book value.

The table below appears on page 8:

And on page 10, one of my favourites:

The projections assume that, of the €77 billion nominal value of loans acquired, €62 billion will be repaid by borrowers and that loan defaults or debt restructuring will occur on €15 billion (a rate of 20%). Over a five year period in the early 1990s, one UK bank experienced a default rate of less than 10% on its whole book. Given the concentrated nature of the prospective NAMA portfolio and the risk of a prolonged recession, a 20% default rate assumption has been made. It is also assumed that €4 billion will be realised from the sale of underlying assets secured by the defaulting loans of €15 billion. These are conservative and prudent assumptions.

Another chestnut on page 30, “NAMA Proposed Investment Strategy”:

Of the €77 billion to be acquired by NAMA, it is estimated that €31 billion is cashflow-generating (€28 billion in commercial loans and €3 billion in land and development loans). This means that loans totalling €46 billion will not be producing cash flows but may be regarded as performing as they are on interest rollup (as per their contractual terms).

So now we’ve got the relevant information. All of the associated loans are performing (assuming that’s what ‘commercial loans’ is referring to) and €3bn of the land and development loans. We know that the book value of the underlying assets is estimated at €88bn based on a 77% loan to value (LTV) ratio when you strip out the €9bn in rolled up interest. €62bn of loans will be repaid and we’ll collect another €4bn on sale of assets from the €15bn that default. Ignoring interest and time value of money, we’re going to make €66bn. We’re paying out €54bn. That makes sense.

However, let’s look at the value of the underlying asset argument. I’m going to assume we’re only talking about loans that are currently not generating a cash-flow, which is €46.4bn of the land and development loans. Strip out the €9bn in rolled-up interest, which again I’m assuming is only on those loans, €37.4bn, and this is the amount originally loaned which is not generating a cash-flow. Applying a LTV of 77% means that the value of the underlying assets is €48.6bn. Discount that by 47% to notional ‘current’ market value, €25.7bn. Add 15% to get to Long Term Economic Value (LTEV) at some point approximately ten years from now, €29.6bn. Add that to the €31bn in loans that are generating cashflow, €60.3bn. So in a worst case scenario where all of those land and development loans that are not generating cash-flow default, we’re still making money (ignoring again the cost of borrowing and the time value of money, opportunity cost, etc.).

In a slightly different scenario, in which you take the business plan assumption that €15.4bn will default (20% of the total). That would be 33% of the non-cashflow generating land and development loans. But you collect €4bn on the sale of the underlying assets, a recovery rate of 26% and a net loss of €11.4bn. In that case we’ve lost €11.4bn of our €46.4bn land and development loans, making a clean €35bn, plus all of associated loans which are generating cash, €31bn, and we’re still in the black to the tune of €66bn.

So what’s the problem with a bonanza like this?

Here’s where the assumptions start to unravel, and interestingly, let’s continue to use those from the business plan but apply them differently, what we would argue is ‘correctly’.

The error in the application of the 20% default rate to only the NAMA loans has already been pointed out. If you’re going to use the Barclays example as the reference point then you have to apply it to the entire loan book of €400bn odd to compare apples with apples. In which case you’re total defaults are ~€80bn, or all of NAMA loans. Very possible given the extent of bubble lending and how rapidly our economy is deteriorating.

If you apply a recovery rate of 26% on defaulting loans, as the business plan does, but apply it to all non-cashflow generating loans, €46.4bn, then only €12bn is recovered, which when added to the other €31bn collected, results in a loss of €11bn.

Now let’s look at the cashflow generating loans. We have from Peter what he describes as a Rolls-Royce of a performing loan to use as a benchmark. It’s a commercial property that has first class tenants all paying long term leases, it was valued three years ago on the basis of a yield of just less than 5% which valued the property at €72m. Today the bank has applied a 7.5% yield, standard in the pre-bubble days, which values the property at €43m. The loan was made for €52m, a 72% LTV, but now the property’s value has fallen below this amount and so the borrower is in breach of their loan agreement. The bank is therefore demanding payment within 18 months at a much higher loan interest rate. The borrower in this case is more likely to hand over the keys, and now the lender has the problem of realising the value of the asset to collect their loan – a certain loss of €10m based on its current value. Because this is a bank and a commercial entity, it can’t afford to wait ten years in the hope of a 15% appreciation in the value of the property so that it can hopefully make a smaller loss.

Lets assume that all the associated loans which are performing according to the business plan are in a similar situation (if they weren’t why would the bank be looking to put them into NAMA?). By increasing the yield to 7.5% from just under 5% on these loans, their value drops by 42%. Assuming that €27.7bn represents 72% of the asset value, the asset value is €38.5bn. In this scenario it has lost 42% of its value making it now worth €22.3bn. Add this to the 26% collected from the non-performing loans, €12bn, and total recoveries amount to €34.3bn.

The result, a loss of €20bn.

How many schools, hospitals, public infrastructure projects, energy reduction and generation programmes can you buy with €20bn? A rough estimate would be a lot. Here’s an interesting way to spend half of it: NewERA (the author, yours truly, is in no way associated with the publisher of this proposal).