Float Like a Butterfly

Brexit superhero Steve Baker MP was in fine form on the last (23 June 2021) Treascom meeting. The witness before the Committee was PRA chief executive Sam Woods. Steve began by asking (transcript here, Q69 ff)

‘Q69 Mr Baker: …can I go back to Harriett Baldwin’s question about matching adjustment? If I understood correctly … you said it is about capital being set against expected cash flows. Could you just tell me something about the risk profile of those expected cash flows?

Sam responded:

The way the matching adjustment works, crudely, is that if insurance companies can prove to us, to a very high standard, that they have achieved a fixity of cash flows coming in from the assets, which maps on to what they have to pay out on the other side on liabilities, we then allow them to use a higher discount rate for their liabilities. What we do not allow them to take over into that higher discount rate is something called the fundamental spread, because the fundamental spread is meant to capture the risk that those bonds go wrong. A lot of work went into calibrating the fundamental spread when Solvency II was set up. The point I was making was that, under the construct that we currently have, there is not an allowance for uncertainty around that going forward. If we were looking at broadening eligibility somewhat, if we were looking at de-bureaucratising somewhat, if we were reducing the capital strain from the risk margin, in the mix of all that it would make sense to look at that question. There is probably a way of making that work better, which leaves the whole thing square.

Sam’s explanation is a masterpiece of regulatory transparency.

First point, note Sam’s use of the term, ‘a fixity of cash flows’. You might think that when Sam said ‘fixity’ he really meant ‘fixed’, but he didn’t. There is a deep problem here, but we will dig it up another time.

Now this ‘fixity but not fixedness’ property of the cashflows is then deemed to allow firms to use a higher discount rate on their liabilities. Our response is that even if one grants the principle of the MA (and one should not), then the MA logic [sic] only applies if those cashflows have the ‘fixedness’ property, which they do not. They merely have this ineffable but different property of ‘fixity’.

‘[T]he fundamental spread is meant to capture the risk that those bonds go wrong’ – maybe, but if we were dealing with any firms other than life firms, the discount rate (i.e. the rate that captures the risk that those bonds could go wrong) would be a market discount rate not some ‘fundamental spread’ dreamt up by actuaries with no sense of market reality and calibrated by some regulatory sausage machine that in any case can only respond after the horse that bolted has long since died of old age.

The exam question is this: if the liabilities of life companies have this ineffable quality of ‘fixity’, then maybe other firms like banks, hedge funds and private equity firms, have liabilities that also have ‘fixity’. So, presumably, the PRA should either extend MA benefit so everyone gets it or roll it back so no one gets it.

All of which catapults the issue into the limelight: is the principle of MA sound, e.g. because it is ‘a fundamental principle of actuarial science,’ to quote former IFoA president John Taylor, who put the credibility of actuarial science on the line on the issue (not a clever move, John, if I may say so) or is it hogwash, because financial economists who understand the principles of valuation (including experts within the BoE) have all agreed that it is hogwash.

But then the answer is revealed:

The point I was making was that, under the construct that we currently have, there is not an allowance for uncertainty around that [the fundamental spread] going forward.

What Sam is saying here is that even though the FS ‘capture[s] the risk that those bonds could go wrong’, the system that they put so much work into has still not got around to providing ‘an allowance for uncertainty around [the FS] going forward’.

In other words, the system Sam is depending upon presupposes that the future is certain. Forsooth, the regulators have a crystal ball! As it happens, I bought one of those myself at a yard sale just last week.

‘There is probably a way of making that work better, which leaves the whole thing square.’ Yes, there is. Stop believing that the regulators can ‘improve’ upon the market discount rate and scrap this MA/FS nonsense that follows from this hubris.

‘Q72 Mr Baker: I find myself listening to various figures, and they seem to be quite alarmed about the capital position that some firms might be in because of the matching adjustment. What would your advice be to them if they wish to bottom out this argument? What argument would you want to listen to in order to try to bottom this out?’

Good question. One wonders who these mysterious ‘various figures’ might be.

Sam Woods: I am glad people make that side of the argument, because the insurers make the opposite, and it is good to be attacked on both sides. Honestly, I think it is a good thing, because there should be a debate around how we do this, given how important it is. The argument I would make to them is this. Their contention is that basically you should only ever discount liabilities at a risk-free rate. That is a perfectly respectable worldview. It is a worldview that, in my opinion, would have two negative aspects. First, it would very significantly increase the pricing of annuities. I do not know if you have looked around lately, but the annuity you get for putting away 100 grand is not particularly appetising. There is just a trade-off in there between the degree of risk that we choose to run and the amount of benefit that consumers can get. That is inevitable. They have a corner solution view on that. That is one obvious downside of what it suggests. The other big downside is that the thing the matching adjustment does, which is genuinely very helpful and risk-reducing, is the incentive it provides for them to match. It is a very good thing that, as we go through something like last year with the markets going all over the place, insurers can actually ride that out, because they have achieved that matching. That is a good thing. Those are the counterpoints to them, but it is a good thing that they push us around on it.

Sam’s response is a gracious one, if a little on the wordy side.

Sam states that removing MA ‘would very significantly increase the pricing of annuities’ and goes on to mention the ‘benefit that consumers can get’ from the MA. But how does MA benefit customers who already own annuities they have purchased perhaps many years ago. In fact, those customers demonstrably lose out. The annuity is an asset to the pensioners and a liability to the buyout firm, so if the firm significantly decreases the value of its liability, by the same token it must decrease the value of the asset held by the pensioners.

The other big downside is that the thing the matching adjustment does, which is genuinely very helpful and risk-reducing, is the incentive it provides for them to match. It is a very good thing that, as we go through something like last year with the markets going all over the place, insurers can actually ride that out, because they have achieved that matching. That is a good thing.

But how is the MA risk-reducing, except on paper? The MA allows firms to create phantom capital and obtain associated reg cap relief. All that does is make firms appear to be stronger than they really are: MA is a mask not a cushion.

Regarding the benefit of ‘riding out the storm’, how exactly did masking the true situation help firms to ride the storm out? The markets were saved last year by the successful development of vaccines, but no-one could know in advance that successful vaccines would be developed so quickly. Thus the ship was nearly ran onto the rocks for a while, but the storm just happened to die away before any serious damage was done. But what would have happened in the event that the storm had not died down, spreads continued to blow out, and there were significant defaults like we saw in the 1930s?

As Dean wrote in his posting on the Steve and Sam show, Sam plays it again, ‘This one will run for a long time.’ Indeed it will.