‘Concerned Observer’, writes in to query my recent post on the buyout model. He or she makes two points. First:
… the buy-in premium tends to be lower than the present value of the liabilities using a gilts-based discount rate. So there will be some schemes that are not sufficiently well funded to be able to invest in gilts and still meet the liabilities as they fall due, but which can afford a buy-in.
Well, it is true that the graph above confirms that in recent months (specifically since June 2017) the buyout firm is ‘borrowing’ from pensioners at a rate somewhat higher than risk free. But before that date it was negative, so I question whether the premium ‘tends’ to be lower than the gilts rate. Note that the bold horizontal line on the y axis is set at minus 20bp, not zero.
In any case, a careful investigation of Solvency and financial reports shows that the spread on the risky assets used by the buyout firm is much higher than 40bp. More on this when our second report (‘Asleep at the Wheel II’) comes out.
Concerned Observer’s second point is that there is longevity risk in any pension scheme, so any scheme holding gilts will be in cash deficit if the pensioners live longer than expected. By contrast, a buyout scheme is protected by capital.
You might argue that a scenario where pensioners live for longer than expected might bankrupt the insurance company. However, the insurer needs to hold capital to guard against this, whereas – in our example – the pension scheme merely had assets equal to its liabilities.
But our claim is precisely that any buyout arrangement under Matching Adjustment, which is pretty much all of them, is not supported by capital at all, because Matching Adjustment is a way of creating capital out of thin air. We borrow at 40bp over gilts (to concede Observer’s first point), then invest in risk assets yielding 240bp over gilts. Under the PRA rules, if you can persuade the PRA that the risky assets are risk free, you can discount with up to 240bp spread, creating 200bp of fake capital.
But this fake capital isn’t capital at all. As I explained to listeners in the BBC programme last year
27:35 Suppose I go to my bank manager and I want to borrow some money to buy a house and I say, well, I want to borrow at 100% loan to value. I borrow what the house is currently worth. So I’ve got a London house worth £500,000. I say I have no equity, no money, can I just borrow £500,000… what is the bank manager going to say? He’s not going to look upon that too favourably, you would imagine. He wants a bit of equity, some loss absorption which consists of what I put up, my equity. But then I say, well look, house prices have always gone up – everyone knows they’ve always gone up – that house will have gone up 30%, 40% in 10 years’ time, when I repay the mortgage, than it is now, so can we regard that 100% loan to value mortgage as 60 to 70% LTV. You can imagine, you can imagine what the bank manager is going to say to that. But Matching Adjustment is exactly like that. If it’s absurd in the one case it’s got to be absurd in the second case, don’t you think?
The point of capital is that it has to be loss absorbing, in the bank manager sense. Capital that doesn’t exist is not loss absorbing, therefore capital that doesn’t exist is not capital.
If bank managers wouldn’t accept fake capital, why should pensioners?