Sir John Vickers kindly mentioned us last week (6 June 2019) in his keynote address to the 19th Annual International Conference on Policy Challenges for the Financial Sector in Washington. Most of the speech is about bank stress tests – he has long insisted that market-based measures should play a greater role in regulatory assessment than is current practice – but he mentions the insurance stress tests towards the end, citing the PRA’s current consultation on its stress test for insurers.
For the valuation of pension scheme liabilities, firms should assume that the discount rate would change by the level of any change in the risk-free rate plus 50% of the change in spread on AA rated corporate bonds. Under the proposed stress the risk-free rate decreases by 100bps and 50% of the spread on AA rated corporate bonds is an increase of 85bps. Therefore, both elements combined result in a 15bps fall at all tenors to the discount rate.1
As Vickers suggests (much more politely than us) this is a bit bonkers. For pension scheme with long dated liabilities, a 100bp drop in the discount rate would have a substantial effect on the liabilities, i.e. they would go up a lot. “But the assumed jump in credit spreads comes to the rescue, or 85% of it anyway, so undoes most of the liability stress. In economic terms this is very puzzling”.
Why have credit spreads jumped in the postulated scenario? Largely, if not wholly, because corporate credit risk has risen because of the stress. Indeed for non-sovereign bonds there is a 2 notch downgrade in option 1 of the stress test. If the insurer is holding corporate bonds, the risk that those assets will not pay out in full is therefore assumed to have increased. The value of those bonds in its asset portfolio will fall. But the obligation to pay pensions is not weakened by the rise in credit spreads, so it is hard to see why the insurer’s pension liabilities should be reduced – at all – on account of it. On the face of it, this curious regulatory accounting treatment of pensions also dilutes the stress test. It may also have implications for the measurement of insurers’ capital levels.
Right.
As it happens, Just Group’s recently published 2018 SFCR shows much the same effect, although on the balance sheet of the insurer rather than on its pension scheme. The infamous table S.22.01.22 (p.93) shows how removing matching adjustment ‘benefit’ impacts capital. Eligible own funds are £2,283m, but removing the MA takes away £2,722m, which on my arithmetic leaves them technically insolvent. But for 2017 the impact of removal was only about £2,000m. How so? Why the difference of £700m?
The answer seems to be on p.68, giving the value of the MA itself. They explain that the change was due to credit spreads and interest rates.
The Matching Adjustment for JRL and PLACL has increased (in bps) primarily due to increase in credit spreads on corporate bonds during the period (average bond spread increased by c.41bps in 2018). In addition, the spreads on the Lifetime Mortgage notes in JRL have also increased due to an increase in risk free rates in 2018, which reduces the Effective Value Test restriction on the economic value of LTMs on JRL balance sheet.
So MA to the rescue. Value of assets fall by £700m, value of liabilities falls by about the same amount, the world is safe again.
But as Vickers suggests, in economic terms this is very puzzling. Why did credit spreads rise 40bp in 2018? Largely, if not wholly, because corporate credit risk has gone up. Thus the risk that those assets will not pay out in full has increased. But (paraphrasing Vickers) the obligation to pay pensioners is not weakened by the rise in credit spreads, so it is hard to see why the insurer’s pension liabilities should be reduced – at all – on account of it.
- “Life Insurance Stress Test 2019“, PRA 2019