The Financial Policy Committee of the BoE has announced that they are getting rid of the affordability test for mortgage lending. To be precise, they are removing the requirement to stress test affordability at rates of up to 300 bps over the bank's current mortgage rate.
They started a consultation on this in February, and the industry was heavily in favour of getting rid of it. The consultation response is here. It doesn't include the text of the responses. It does include some very dodgy math about how the rule change is of no consequence because it only affects around 3% of renters (since many renters are no where near being able to get a mortgage, this actually seems like a lot).
I think that the affordability rule is good, and should have been kept, or modified, and the LTI rule, which was kept, is stupid, and completely inadequate, by itself, to do the job of making sure that lending is responsible.
Why the LTI test is stupid
Loan-to-income, as a ratio, is completely obsolete since sometime between the advent of married middle-class women working for a salary and the invention of calculators. Higher earners pay a higher rate of tax than lower earners. Conversely, they spend a lower proportion of their income on rent, bills, childcare. A two-income household and a single-income household earning exactly the same gross will take home vastly different amounts. Affordability will be very different for a couple without kids, with two kids in paid preschool, and with two kids in school.
The rule (limiting LTI to 4.5) only applies to 85% of the bank's mortgage book. Either these loans are fine, in which case why should they be restricted? Or the metric correctly indicates that borrowers are over extended, in which case why allow any of them?
But wait, this is a ‘macroprudential’ measure. This means it doesn't matter if those loans are dodgy: since there's only 15% of them, they aren't sufficient to make banks insolvent when they go bad, so they're ok. Obviously, this makes sense, since it's impossible for lots of heavily indebted homeowners defaulting on their over-priced homes to have a wider macro effect.
The basic mathematics of discrimination is well understood. If 14.9% of borrowers want loans in excess of 4.5 LTI then they can all have them with negligible effect on price. If 15.1% want them then the interest rates will decouple from the rest of the market. Is it a good idea for these borrowers to pay more, given the same overall credit worthiness, because of that calculation?
Further, since the denominator is the entire loan book, a bank which wants to make additional loans with high LTI can do so if they also make more with a low LTI. On the margin, yields will go down for borrowers with a higher income while they go up for borrowers with a lower income. Fine, whatever, but why is this a goal of regulation?
Banks are in favour of this rule, since the 15% part means they can do what they like. But intermediaries also love it (and hate the other rule). Brokers like it if it’s harder for some people to get loans, because those people are more likely to engage brokers. They like it even more if there are special rules and cases with some people eligible for some loans and not for others. Again, it tends to confuse or worry consumers, who then pay brokers for hand-holding. It also fragments the market and makes it less transparent. This favorizes brokers over price comparison sites.
Why the affordability test was (somewhat) good
It put together loan size, income, outgoings and interest rates in a single calculation, and thus got borrowers to think about what their financial situation will really look like. Households are not junk bond issuers, and neither interest coverage not leverage ratios matter as much as monthly cashflows, to the extent that they are predictable.
It linked the cost of a mortgage to the bank’s reversion rate, not just risk-free rates, and thus meant that the bank’s own rate affected borrowers’ understanding of what they would pay. Usually in the UK mortgages are marketed using the initial rate, and the initial term is quite short. Initial rates are set at ‘teaser’ levels, and offset by one-off fees. Rates past the initial period are usually higher, reflecting the borrower’s increased optionality, but also often a limited need to compete. Banks set them just low enough to discourage most borrowers from re-financing, higher if they don’t mind the book shrinking, higher still if many customers are trapped because of negative equity or changed personal circumstances. Many customers have assumed in the past that their post-initial period rate will be the initial rate, plus or minus changes in the BoE rate. In fact, while they were getting 2 year mortgages at mid-100s, some variable mortgage rates were already at 350 bps or higher, with usually no guarantee that they would evolve in lockstep with other rates. In fact, the FPC specifically pointed out, as a reason for getting rid of the test, that banks’ reversion rates haven’t gone down much as headline rates have cratered. In other words, because the rates that most borrowers will actually pay from 3 years into a 30-year mortgage are held, oligopolistically, at unfairly high levels, it would be inappropriate to use such rates in calculating whether or not people can afford to take out said mortgage.
The test is nice because one could change the exact parameters, while still preserving the properties above. It seems silly to have a test based on a band for future interest rates, keep it throughout a period of historically low rates, and then get rid of it when rates (start to) go up. Is the argument that now the risk-free rate is at 125, there’s no longer any room for it to conceivably go up by a further 300 bps? But if you really wanted to relax it for this or any other reason, it would be completely trivial to adapt the test. Just change the floor on the upwards shock in the stress test (100 bps), the additional 200 bps to be added on, the five-year period for the test, etc. I suspect the reason that the FPC didn’t do this is because there would have to be some quantitative evidence for such a change, whereas getting rid of the whole thing just requires a consultation, followed by the issuance of a piece of paper with ‘THIS IS GOOD FOR FIRST-TIME BUYERS’ written on it.
In conclusion, any time in the future when house prices drop, banks fail, and/or people can’t pay their mortgages, this piece predicted it and the prediction proves I am always right, invest in my hurriedly set up hedge fund the FPC and the PRA are depressingly rubbish and captured.