In a recent (July 30) posting, Dean wrote that
Since Just Group’s trading update last week, there has been much speculation about the ‘pioneering no negative equity guarantee (NNEG) hedging transaction’ announced by the firm. It is not clear whether the firm has put the hedge in place, or whether they are still waiting to establish ‘appropriate regulatory treatment’ with PRA. The current thinking is that the hedge will be transacted through a major reinsurer, and that it will be a purchase of some form of long dated put option on the housing index.
A related possibility is that the hedge, if there is one, is some kind of NNEG swap, which then raises the question: what is Just’s NNEG valuation?
Short answer: we don’t know and it would be naughty of us to speculate. We don’t have enough information about the firm’s ERM portfolio or enough information about the firm’s valuation approach.
But this got us thinking …
… about how would one go about valuing a firm’s NNEG if one had the information. (We use the term “firm’s NNEG” as a shorthand to refer to the sum of the NNEGs on all the ERMs in a firm’s ERM portfolio, which is a bit of a mouthful.) Remember that we have hitherto focussed on the valuation of the NNEGs associated with individual ERM loans. Going from the one to the other is a nontrivial task.
It is therefore interesting to consider a couple of hypothetical firmwide NNEG valuation examples, on the understanding that, as the movies say, the story is fictional and any resemblance to any real world situation is purely coincidental.
Such an exercise is useful in that it illustrates in a very oversimplified way the steps involved to go from the individual loan NNEG to the firmwide one.
We have to make assumptions about, at a minimum: (a) the valuation approach that a firm uses, (b) the size of its ERM book; (c) the breakdown of its ERM book across borrower ages; and (d) the LTVs used. And note that we have glossed over other important issues, e.g., such as the breakdown by product type, new versus historical loans, differences in loan rates and other product features, etc.
On (a), let’s consider 3 cases: (i) deferment rate = 4.2%, as recommended in our Eumaeus Report; (ii) deferment rate = 0.5%, and (iii) deferment rate = minus 2.75%.
On (b) let’s consider a hypothetical firm with an ERM book of £300 million. As far as we know, there is no UK equity release firm with such an ERM book, so there should be no confusing it with any actual firm.
On (c) and (d), let’s first consider a case of a firm that has just made a set of ERM loans to identical 70 year old male borrowers at an LTV of 30%.
Given these and other assumptions, we obtain NNEG valuations of £188 million, £72.8 million and £15.7 million for each of cases (i), (ii) and (iii) above. 1
Note too that these results scale with the value of the ERM book. So, for example, if the ERM book were £3 billion, then the NNEG valuations reported above would increase by a factor of 10, and so on.
As an alternative, consider an equity release lender that has the same ERM book in total, but this time assume that the firm has made 13.6% of its ERM loans to borrowers aged 60, 57.7% of its loans to borrowers aged 70, 25.3% of its loans to borrowers aged 80, and the rest (3.4%) of its loans to borrowers aged 90. Assume also that the LTVs were 24.3%, 35.1%, 45.4% and 47.1% respectively. These figures are not too loosely based on those reported in recent Equity Release Council’s Market Reports and are roughly reflective of the industry as a whole. 2
Making these assumptions, we then obtain NNEG valuations of £227 million, £101 million and £27.4 million for each of cases (i), (ii) and (iii) above, which are somewhat higher than those we obtained earlier. Geeks will recognise these higher valuations as reflecting a concavity effect, by which the average of the NNEGs for individual ages exceeds the NNEG for the average age, and the average age is in fact around 70.
What is also interesting is if the two parties used the same or a different approach to the valuation of the NNEG securitisation. We then need to consider the approaches used by each firm and compare them against the ‘correct’ approach.
Suppose that a firm laying off the NNEG risk via a securitisation uses a deferment rate of minus 2.75%. Now if the counterparty uses the same approach, and if we are correct that that approach will lead to a NNEG undervaluation, then the NNEG undervaluation problem that we have complained about many times elsewhere is transferred to the counterparty. The terms ‘face’, ‘ripped’ and ‘off’ come to mind. On the other hand, if both firms use the correct approach as we see it, then the first firm will have crystallised its NNEG losses. Then again, if a deferment rate of minus 2.75% is reasonable and both firms use it, then it would follow that the terms of the securitisation were reasonable and there is nothing for anyone to complain about, and that our earlier concerns were misplaced.
It also gets interesting if the firms use different valuation approaches from each other. In that case it would be theoretically possible for both parties to post a profit on the transaction or for both parties to post a loss on it. We can probably discount the latter case because why would both parties enter into a transaction on which they both lost money? In which case, we would presumably be dealing with a case where both parties reported profits on the same transaction, which is curious from a fair value perspective.
In any case, it’s all very interesting and any resemblance to any real world situation is ‘just’ a coincidence.
- Based on the following calibrations: borrower age=70, risk-free rate=1.5%, lending rate=5.21%, ERM loan book=£300m; ltv=30% and vol=14.8%. Results are based on the M5-CBD mortality model calibrated using England & Wales male death rates over sample years 1971:2017 and sample ages 55:89.
- Based on the following calibrations, in addition to those mentioned in the text: risk-free rate=1.5%, lending rate=5.21%, vols are 19.1% for age 60, 14.8% for age 70, 12% for age 80 and 10.7% for age 90. Results are based on the M5-CBD mortality model calibrated using England & Wales male death rates over sample years 1971:2017 and sample ages 55:89.