Kevin Dowd 3 December 2018
Chris Cundy at InsuranceERM has just written a nice piece on Dean’s and my new report, Equity Release: A New Equitable in the Making published by the Institute for Applied Economics, Global Health, and the Study of Business Enterprise in their Studies in Applied Economics series. This new report is a follow-up to the Adam Smith Institute report Asleep at the Wheel: The Prudential Regulation Authority and the Equity Release Sector published on August 6th. To quote Chris’s article:
More criticism has emerged from Dean Buckner and Kevin Dowd on the valuation approach taken by some UK insurers investing in equity-release mortgages (ERMs) [who] earlier this year published a critique [Asleep at the Wheel] of how the no-negative equity guarantee (NNEG) element of ERMs was being under-valued, which results in ERM portfolios being over-valued.
To continue:
The pair reminded the IFoA of its duty to promote the public interest, remain impartial and “avoid even the appearance of conflicts of interest” with commercial interests.
The IFoA, together with the Association of British Insurers, have recently commissioned research on appropriate methods for valuing the NNEG.
A slightly awkward question is whether research commissioned by the ABI and the IFoA could reasonably be independent and avoid even the appearance of conflicts of interest. Both bodies have lobbied extensively on behalf of equity release firms and these firms clearly have a stake in the results of the research being commissioned. A possible response is that the independence or otherwise of the research will depend on the independence of the steering committee that awarded the contract and manages the project, and there are issues there too (see here and here). However, we would certainly agree that the chair of that committee, Professor Johnny Li, is independent.
But let’s leave these quibbles aside and accept the ABI-IFoA project as independent research commissioned in the public interest.
The project, titled Equity Release Mortgages: No Negative Equity Guarantee, will be delivered by the University of Kent and led by Professor Radu Tunaru. Results are due to be published in February 2019.
Radu is a good choice and I am sure that the research will be will as good as can be.
Going back to our report, our work is pro bono publico and independent in the sense that we received no payment for writing it and we have no direct financial interests in the equity release sector. (I say ‘direct’ because I can’t be sure whether my own pension fund, USS, is invested there.) So we are not being paid to promote bullish or bearish views on the stock of any companies.
We did consider applying for the ABI-IFoA research project ourselves but we decided against on the grounds that we could not meet some of the requirements set out in the project specification requirements. There were various issues, but two particular points in the tender document stood out:
To consider the relative merits of ‘real world’ and arbitrage free (risk neutral) methods [KD, i.e., market-consistent methods] and how the assumptions should be set on both bases.
The problem is that we couldn’t find any merits in the ‘real world’ approach and we couldn’t think of any assumptions that would make it work.
To consider whether there are any “halfway house” solutions between real world and risk-neutral approaches given, in relation to the latter, the absence of a deep and liquid market.
The problem here is that they wanted some kind of weighted average between an approach that makes sense and one that does not, and the only weights we could come up with were 1 and 0.
So it seemed to us that these project requirements amounted to a Holy Grail that we knew we could not deliver. We knew that because we had searched for it ourselves.
To quote an IFoA spokesman:
We believe more research needs to be completed to understand if the NNEG is understated currently.
That is why we wrote our report.
We spent maybe six months working on this issue but try as we did, we couldn’t find any way to get credible and low NNEG valuations based on (a) a plausible model and (b) plausible calibrations of key parameters. Moreover, (c) whatever way we tweaked the option pricing calibrations we soon ran into the PRA NNEG lower bounds that were not dependent on any option pricing model at all, so any arguments against our analysis based on criticisms of, say, the Black ’76 option pricing model are moot. These lower bounds also meant that there was no point coming up with alternative option pricing models either, because they would hit the lower bound too. And (d) the equity release industry’s preferred approach, the ‘real world’ or ‘discounted projection’ approach, couldn’t be squared with any respectable (refereed, externally validated) approach to option pricing that we knew of.
Indeed, the only arguments in favour of this approach even put forward by its adherents were that it produced low NNEG valuations. However, and I really shouldn’t have to spell this out, the issue is not whether these valuations should be low because firms have an interest in low valuations but whether they should be reliable. Getting this right is not an option. Actuarial standards mandate that valuations should be reliable, “fit for purpose both in theory and in practice,” neutral, unbiased, and so forth.
The key parameter turns out to be the net rental rate or deferment rate q. Consider this plot of NNEG valuations against q for a set of illustrative calibrations:1
NNEG Valuation Against Net Rental Rate
We see that the q rate does most of the heavy lifting. So if we use the ‘real world’ approach calibrated with an HPI rate of 4.25% we get an implied q rate of about minus 2.75%, assuming a risk-free rate of about 1.5%. But a negative rental rate makes no sense. Our best estimate is a q rate of at least positive 3% and this q rate produces a much higher NNEG valuation.
In the meantime we wish the Kent team the best of luck.