Another day, and another challenge to the method we have proposed for the valuation of the no negative equity guarantee. In our presentation here (slide 9, equation 8) we use the term r-q in the calculation of the forward house price at time t, so aren’t we incorporating both a house price growth assumption (r = interest rate) and a net rental rate q? Thus aren’t we – implausibly – assuming that future house price growth will be equal to the interest rate?
A spokesman for UKAR got back to me saying I got the numbers wrong in this post. I said that the total loan value of the book bought by Rothesay Life, i.e. the original amount lent accrued at the loan rate is somewhat north of £1bn. Apparently not, for that value, which they call ‘unpaid balances on portfolios sold’, is £860m, hence somewhat south of £1bn. So there are five different values to choose from:
The loan value (i.e. original amount lent accrued at the loan rate) £860m
The book value of c. £750m
The amount that would have been paid if not for the NNEG
The amount paid by Rothesay, which UKAR cannot disclose, but which was greater than book value, and £200m less than the amount that would have been paid if not for the NNEG
Government loan repayment: over £1bn
Fans of linear algebra will spot that there are still too many unknown quantities to make any sense of this.
I was at a discussion today where we were asked whether corporate reporting faced any problems. Most hands went up. Then we were asked whether reporting faced any significant problems. Only two hands went up, including mine. I briefly skirted around the problems with life company reporting.
My point was dismissed. Everyone agrees, said the room, that life company accounting is fundamentally broken, so that objection didn’t count. What else haven’t the Romans done for us, sort of thing.
To be sure, I have heard this many times, including in the corridors of power, even from life insurance accountants. But if everyone thinks that, why has no one done anything? Life insurance is about, well, our pensions and stuff. Could it really be too difficult to fix?
Our contact form is on the menu above, all suggestions welcome. Is there a problem? How do we fix it?
There certainly was a ‘lively debate’ at the LSE on Monday evening. The house was packed, with guests including our friends at the Treasury, the PRA, and a number of analysts. Kevin will be writing some more about this, meanwhile here are the slides. Note that we covered them in a slightly different order. I discussed slides 16-19 (and mostly slide 16) on the ‘upper bound principle’ after Kevin’s main presentation.
The upper bound continues to be misunderstood, as I commented in our reply to the Institute yesterday. It does not depend in any way on arbitrage arguments, complete markets, geometric Brownian motion or any of that stuff.
Simply, jam today worth more than jam tomorrow, and should be valued as such by accountants such as KPMG. Yes?
Some extraordinary developments this week including the resignation of Simon Thomas as Chief Financial Officer of Just Group, and the somewhat stormy presentation by Kevin Dowd and myself at the LSE. I shall comment on the latter tomorrow, today I shall discuss the Institute of Actuaries’ response to CP 13/18.
We replied at the last minute to the CP 13/18, our letter is here. For those who don’t have time to read it, the main points are:
A jolly good set of proposals, but why did the PRA take nearly 4 years to decide on the pricing of a simple European put option? I shall be commenting upon this enigma at the discussion at the LSE this (Monday) afternoon.
The CP does not consider the capital treatment of ERMs, yet an autocorrelated market such as residential property poses considerable problems for Var -type capital treatment.
The Matching adjustment regime is completely impenetrable. “We believe the PRA should make it a priority to work on possible reforms to Solvency II or on a UK successor to Solvency II to bring it into line with accounting standards such as IFRS”.
IFRS 17 is not consistent with the regulatory accounting treatment of Solvency II
I look forward to seeing our readers at the LSE tomorrow. There will be drinks.
[Update: The Institute of Actuaries has just published its response to CP 13/18. We will be commenting on this tomorrow, but note they also bring up the autocorrelation point, although, like many others, they confuse a valuation question with a risk management one.]
There was much talk about the UKAR ‘bad bank’ selling a portfolio of ERM loans to Rothesay Life, an insurance company once owned by the good bank. See e.g. this report by Ralph and Pooley.
The price of a property is overwhelmingly determined by the value of the land it sits on, which makes it inherently positional, and not something you can cut to zero. For example, you can’t reduce the value of the land in the City to zero (or even make a material dent in it) by liberalising planning in Surrey. The reason is that the value of the land is determined by the discounted stream of rent you can get from it. This is unrelated to the cost of building. If you build somewhere where nobody wants to live, the discounted stream of rent, and hence the property value, will be zero.
Kevin writes: ‘I must have been mistaken: the true purpose of a pension fund must be to finance a flutter on the housing market.’
When I first joined the world of insurance, it was explained to me somewhat cynically that banks borrow short term from depositors, and use the money to speculate with. If the bet turns sour and depositors want their money back, it goes horribly wrong very quickly.
Insurance companies, by contrast, borrow long term from future pensioners and use the money to speculate with. If the bet turns sour, it takes about 20 years for things to go horribly wrong, after the prime movers are safely retired (preferably not with a pension for the firm they worked for).
The chart shows the Japan housing index 1975-2018 Q1. There is a fascinating history behind this but I won’t go into that now. The point is about the risk of something similar happening here. Banking and insurance capital models work on the principle of Value at Risk, i.e. the amount required to sustain a loss over a specific time horizon, to a specified probability. For example, the advanced Basel IRB model has a time horizon of 1 year, with a probability of 1 in 1,000. The Solvency II model has the same time horizon, but a probability of 1 in 200. So the first would last us from the succession of Cnut the Great in Denmark in 1018 until now, the second from the birth of Karl Marx in 1818 until now.
But I wonder. Could the probability of a Japanese-style collapse here in the UK (or anywhere else in the West) be only 1 in 200? Ask most professionals in the business of capital management and they will say so. The fact that it happened in Japan doesn’t mean it could possibly happen here. Japan is such a terribly different, utterly different place from Britain, that no connection can be drawn between the two scenarios.
But I still wonder. There is a good FT article (31 August 2018) by Gillian Tett about a Japanese central banker warning that a financial crisis was about to explode. Déjà vu, he was saying, presumably not in Japanese, referring to the banking crisis sparked off by the collapse of the baburu keiki, or bubble of the 1980s, leaving about $1tn of bad loans.
