Geeks only. The chart above shows why Equity Release Mortgages are both fascinating and toxic. We have discussed ERM valuation in a number of earlier posts, but ERMs are used as assets to back conventional annuity books under what’s called ‘matching adjustment’, but we need to consider both sides of the balance sheet, because the insurance industry has weird ways of valuing liabilities too.
The PRA (and IFRS) allows firms to discount insurance liabilities at the estimated rate of return on the assets, which is absolutely bonkers but the industry lobbied hard in 2012 for it and here we are. Thus there is a ‘matching adjustment benefit’ arising from the higher discount rate, which effectively creates equity on the books of an MA firm.
ERMs do not have a standard maturity profile, since they tend to pay off when the customer exits, either by death or into long term care, so they have a characteristic hump at about 20 years when borrowers are exiting at the highest rate, then decline gradually until there are just a few laggards who insist on occupying the property while 100 years old.
By contrast, annuities have to be paid for right away, so ERM firms solve this by matching the shorter term obligations with low investment grade corporate bonds, for which they also get MA ‘benefit’.
Unfortunately, while the regulatory Solvency and Financial Condition Report has copious detail on tiny stuff like deferred tax assets and liabilities, investment in joint ventures and so on, it has practically nothing on the stuff we would like to see, namely the maturity structure of the actual cashflows. All you get is ‘effect of setting MA to zero’ at the end of the report, which is about as helpful as a black highlighter pen unless you know the effect at each maturity point.
However, with a bit of sleuthing we were able to make a stab at it. We can estimate the profile of the obligations, assuming that they are annuities, and we already know the ERM profile because it depends on the exit rate we use to price the wretched thing. One minus the other gives us the required profile for the bonds required to fill the gap. Then we can work out (i) the value of discounting by the MA implied discount rate, which we can appropriately call a MAD, and (ii) the (present) value of discounting by risk free. Subtract the one from the other to get the MA benefit for each maturity, then calibrate the total effect for all cashflows to the ‘effect of setting MA to zero’ number in the SFCR.
What this shows is what you would expect. The bonds are needed only for maturities up to 20 years, thereafter the ERMs take over. Thus the MA benefit from the bonds is only about 25% of the total benefit. This also shows why ERMs are so toxic. At maturities under 20 years, ERMs are perfectly collateralised and therefore non-defaultable bonds, because the initial loan to value ratio is much lower than a conventional mortgage. These are combined with A and BBB corporate bonds under 20 years, which are not exactly non-defaultable, but are not junk either.
But after 20 years we get the ERM crossover effect. Unlike a conventional mortgage, the interest on an ERM compounds up, thus the loan value about doubles every 12 years, meaning it will reach the value of deferred possession of the property very quickly (there is a formula for the crossover maturity that depends on the deferment rate). So we leap from a fairly low risk portfolio with maturity less than 20 years, to equity for the remainder. Fascinating, although mostly toxic. It’s all good for 20 years, bonus earned, then the ghouls come out of the plague pit.
And because the ‘MA benefit’ (always put this in scare quotes) predominates at the longer maturities, the ERM providers are particularly vulnerable to regulators taking away what hath been given. Hence the squeakings of pain heard across actuary la-la land when the PRA published that paper, and which naturally led to this weird project sponsored by the Institute.