There is a very fine article in InsuranceERM just published. Behind a paywall I am afraid, but it explains the idea of Eumaeus very well, although I would say that. For those without a subscription, an interesting part is here
As InsuranceERM went to press in December, the PRA published its policy statement (PS31/18), which Buckner largely welcomed. “For the first time it settles, with great authority and a wealth of cogent reasoning, how life insurers should correctly value a portfolio of simple European put options,” he says.
But he is scathing of the PRA’s change of heart on applying the rules for valuing guarantees to business written before Solvency II came into effect. Insurers can now apply different treatment to the same type of loans, depending on whether they were written before or after 1 January 2016. “It makes no sense to me,” he says.
Indeed it made so little sense to me that I queried it with the PRA after the interview.
The PRA came back to me and confirmed (18 December) that all ERMs will have to be valued “in accordance with the Solvency II directive regardless of when they were written or whether they are backing liabilities written before or after 2016”. That means passing the effective value test which in turn means the embedded NNEG guarantee must have a minimum value consistent with the minimum deferment rate of 1% specified in PS 31/18.
In plain English (and note, this is my interpretation of the PRA’s words) firms must use the correct methodology for valuing all of their ERM books immediately, rather than waiting 13 years. That is, they must use the net rental yield methodology and not some forecast of house prices. House price growth is irrelevant. Furthermore, they must use a deferment rate consistent with the provisions of PS 31/18.
However, and this is the weird part, while they must value their book correctly, or at least nearly correctly, for the back book they are also allowed to park the misvaluation in the ‘transitionals’. The transitional, as I explained a while ago, is a weird regulatory asset corresponding to no actual asset. It’s just a (large) number the regulator allows you to add to the asset side of your balance sheet. And as I explained in that post, if you increase the asset side, you automatically increase the liability side, in virtue of it being a ‘balance’ sheet, and this number goes into tier 1 capital. Then the value of this invented asset gradually decreases over the 13 years until year end 2031.
To appreciate the full strangeness of this, suppose there is a firm whose entire assets consist of transitionals. On the asset side we have the transitional, on the liability side we have the share capital owed to shareholders. Would you buy a share of that capital? Surely not. The asset doesn’t exist. It produces no income and moreover its value decreases every year in accordance with regulatory fiat until it is zero in 2032. How would you value the company using present value techniques? You would have to include the present value of the imaginary cashflows that you pay back to the regulator over 13 years, to balance the present value of the transitional. To my mind, the number zero fairly reflects the reality.
I tried to explain the idea to a journalist earlier this week and failed. He was incredulous. How can a regulator create an asset that isn’t there? Strange but true, I said.
I didn’t hear back and I suspect he thought I had got something wrong. Possibly, but we should be told.