If you say that econ value is interest on 100 + principal repayment of 70 – well, I get principal repayment of 70 only if the borrower defaulted somewhere down the road (otherwise I get 100).

If that’s the case, I will be getting the interest for less than the life of the loan, until t_default, time of the default. I could get the same E[NPV] if I drop the notional – that should mean that new t_modified_default>t_default (ceteris paribus), so I get less interest, but for longer time. Depending on the initial values and the default probabilities, I may well be able to find a new notional that would E[NPV] the same (it may be higher than your estimated recovery, say 80). So, the value for me is the same – but it should be a bit more “certain” (as each repayment payment is a bit more certain, assuming finite budget of the borrower).

The original owner of the loan has no such freedom to modify it w/o NPV impact – more parameters need to be moved for it to keep the same NPV (say extend maturity but increase rate, drop notional but decrease maturity and increase rate etc..) – and none of those are likely to be acceptable to the borrower.

But maybe we’re talking cross purpose.

]]>I said:

“The example depicts a cash flow expectation that is equivalent to an original loan of 100, plus …”

“Equivalent to” means it is as if the buyer takes over the loan as originally booked, with further book adjustments as indicated. The book value of the loan is 50, but the economic value may be higher (by 20). That’s why the buyer won’t do the mod down to the book value.

]]>Disclaimer: I have very little knowledge how it works for individual loans, I know how it works for derivatives and suchlike.

You say “original loan of 100, plus contractual interest payments, with provision for a loan loss of 30”.

Why, when I bought the loan for 50 – which would go on the books as the original loan cost – would I take another impairment of 30? Wouldn’t that immediately create an accounting loss to me? If I bought the loan for 10, would it mean I’d have to take a provision of 90?

I may believe that the loan’s value is 70, and I can then provision for the 20 difference, and say enter the loan in as cost 50, value 70, but take provision of 20 released as appropriate. But that still leaves me with space to do the mod so that the expected cashflows add up to the total value of 70 w/o taking any other provision. If I was the originating bank, I’d not do it w/o taking a loss straight away (I did anyway, as whoever sold it for 50 must’ve taken loss, but that’s immaterial to the seller).

I don’t have to do the mod – it could well be that I’d be willing to take a punt on the real estate markets recovering or whatever.

But my ability to do the mod is less constrained than the originator’s one, as I carry the asset at lower value (regardless of whether it’s 50 or 70). The economic loss on the loan was already realised by the seller.

In the above example, it’s also possible that the recognition of the income corresponding to the 20 component may be deferred until the point where the cash is actually received, but that also doesn’t change the economic rationale for the mod decision.

]]>The bank buys the loan from a distressed bank @ 50 cents on the dollar.

It makes an assessment as to what cash flow it expects from the loan over its lifetime.

E.g. the cash flow it expects may be equivalent to the contractual interest payments, plus repayment of principal @ 70 cents on the dollar.

The point being that, particularly in a distress sale, there’s no reason why the bank’s expectation for principal repayment must be the same as the discounted price it paid for the loan.

The example depicts a cash flow expectation that is equivalent to an original loan of 100, plus contractual interest payments, with provision for a loan loss of 30, with an additional discount of 20 attributable to distress sale pricing. This loan would be booked at a current value of 50. Subsequently, over the life of the loan, income would be booked according to the contractual interest payment, plus accretion of the distress sale discount component of 20. Income would be booked roughly like that of a bond under accrual accounting, with a principal repayment of 70, initially marked at 50.

Anyway, the book value of the loan would start at 50 and rise to 70, assuming eventual cash flows matched expectations. The accretion of the 20 would be part of periodic income and corresponding equity accumulation.

The loss provision I suggested might be more conceptual than formal financial accounting, but it falls out in accounting logic from a comparison of expected cash flow against loan book value first originated by the selling bank. It’s effectively an allowance for eventual expected losses from this original book value. The degree to which the purchasing bank would be “comfortable” with principal mod depends on expected cash flow, not on accounting. Expected cash flow, as in this case, may not correspond simply to the price discount.

If, in the example, the bank accepted a principal mod of 50 (the new net book value of the loan), this would be inconsistent with its expectations for cash flow. It would be undoing that component of income and equity appreciation it expects due to the accretion of 20 of the discount into income over the remaining life of the loan. With such a mod, the bank would have to change its expectation to conform to the lower expected principal repayment (or total cash flow) inherent in the mod. So in that sense, it can’t afford to do such a mod, because it simply doesn’t make sense according to its actual cash flow expectation. The economics of cash flow expectation drive the mod decision; not the accounting.

]]>