By DB and KD
Late in April, the PRA announced that it was planning a new life insurance stress test for 2019. At first sight, the test looks plausible: to stress AAA bond holdings by 150bp, going up to 400bp for unrated. However, the spread is divided into the Matching Adjustment, which is the spread deemed to represent illiquidity and which is therefore considered risk free, and the ‘fundamental spread’ (FS), which represents the true credit risk, as it were.
Only changes in the fundamental spread will impact the overall balance sheet and hence capital available. For example, if AAA portfolio spread goes up by 150bp, that will cause a large fall in the value of the portfolio. However, the FS only increases by 10bp, which is tiny. Therefore the MA will increase by 150-10 = 140bp (see the PRA’s Table 2), causing a large fall in the liabilities as well.
Almost all the impact of the stress on capital would be offset by the increase in fake capital produced by the Matching Adjustment cushion, so net capital would be barely affected. Consequently, the Matching Adjustment virtually guarantees that the life industry will pass the stress test without breaking into a sweat.
Such a ‘stress’ test is like an exam that is almost impossible to fail. What is the point?
If this is how the PRA are going to carry out the stress test, we would advise them not to bother, as the results would be a foregone conclusion and we wouldn’t want to see the PRA’s credibility undermined. The point of a stress test should be to determine if a firm can pass an adverse stress, not to virtually guarantee that a firm will pass regardless of what shape it is in. The problem is that the Matching Adjustment makes a mockery of the exercise.
If the PRA want a stress, they should strip out the Matching Adjustment and see how firms fare then. They might find one or two firms that would struggle.