One of the jobs of the new Minister of Finance will be to decide whether or not to accept the Governor of the Reserve Bank’s request to add some sort of debt-to-income limit tool to the list of direct regulatory interventions that the government gives the Reserve Bank political cover to use. It does this under the Memorandum of Understanding on (so-called) macroprudential tools. I phrase things in that slightly awkward way because Parliament has delegated so much power to the Reserve Bank – probably without fully realising the import of several legislative changes over the years – that one unelected official, the Governor, does not actually need approval of the Minister of Finance, or of Parliament, to impose such intrusive direct controls.
To give some credit to the outgoing Minister of Finance, the Memorandum of Understanding framework, while legally non-binding, does more or less ensure that the current Governor would not use such a regulatory intervention without at least the political cover provided by allowing the inclusion of a debt-to-income limit on the list of approved tools. Longer-term, reform of the governance and regulatory powers of the Bank should include making decisions on the application of such controls formally a matter for the Minister of Finance, on the recommendation of the Reserve Bank.
The Reserve Bank has been at pains to claim that their successive waves of LVR controls have improved the resilience of the banking system. That claim is less well-founded than they would like people to believe. For example, shifting a large group of borrowers from say 81 per cent LVR mortgages to, say, 79 per cent LVR mortgages won’t make any material difference to the expected losses a bank might face in a severe downturn, but might actually modestly reduce the ability of a bank to withstand those losses (since loans with less than an 80 per cent LVR typically have lower risk weights). This risk is one my former colleague Ian Harrison has drawn attention to. In addition, the Bank has never presented any sort of analysis, not even impressionistic in nature, of what banks are doing instead of making high LVR housing mortgages. If their risk appetites haven’t changed, and the capital invested in the business hasn’t changed, the risks are likely to be developing somewhere else, perhaps somewhere rather less visible, on banks’ books. There are also ongoing questions about the evidence base behind the regulatory discrimination against those borrowing to buy a house for residental rental purposes. Before giving his imprimatur to the possibility of further Reserve Bank regulatory interventions, the new Minister of Finance might reasonably ask some harder questions about what has already been done. He might also ask some questions about when the drift towards ever more direct intervention – initially sold as quite temporary back in 2013 – might end.
Before approving the addition of any sort of debt to income limit tool to the approved list, it would also be worth the Minister insisting that the Bank’s background papers get public scrutiny. No doubt Treasury gets to see them and Treasury has had some serious questions in the past about proposed Bank interventions. But since the Governor says there is no urgency about using a debt to income tool, there can be no good grounds for not putting the background material out for wider scrutiny now, before the Minister makes his decision, not sometime – if ever – afterwards, when (with luck) the OIA finally gets the papers out of the Bank.
In particular, it would be good to see a careful assessment of the empirical evidence the Bank is using in support of its case for a DTI limit, on both soundness and efficiency dimensions (both important in the Reserve Bank Act). Along those lines, there was an interesting post out earlier this week on the blog of Richard Green a professor (in housing, real estater economics etc) at the University of Southern California.
In that post, he reports some interesting empirical work on a sample of 281000 fixed-rate mortgages purchased by Freddie Mac, one of the US quasi-government “agencies”, in 2004. He runs a regression model across two-thirds of these mortgages, using a range of variables to model the probability of subsequent default, including through the largest shakeout in the US housing market in many decades. His DTI term is not actually the ratio of debt to income, but the ratio of debt service to income, but clearly the two will be highly correlated, especially for these relatively high quality mortgages (ones that met US agency standards – “qualifying”), looking at all the mortgages across the same period of time.
The equation results are in Green’s post.
Note that while DTI is significant, it is not particularly important as a predictor of default. To place this in context, note that a cash-out refinance is 5.2 percentage points more likely to default than a purchase money loan, while a 10 percentage point change in DTI will produce a 1.3 percent increase the probability of default.
To be clear, increasing the total service burden from, say, 40 per cent of income to 50 per cent of income – a huge increase – produced a 1.3 per cent increase in the (always quite low) probability of default.
One reason he notes is measurement
First, while DTI is a predictor of mortgage default, it is a fairly weak predictor. The reason is that it tends to be measured badly, for a variety of reasons. For instance, suppose someone applying for a loan has salary income and non-salary income. If the salary income is sufficient to obtain a mortgage, both the borrower and the lender have incentives not to report the more difficult to document non-salary income. The borrower’s income will thus be understated, the DTI will be overstated, and the variable’s measurement contaminated.
More generally, and in the US context
The Consumer Financial Protection Board has deemed mortgages with DTIs above 43 percent to not be “qualified.” This means lenders making these loans do not have a safe-harbor for proving that the loans meet an ability to repay standard. Fannie and Freddie are for now exempt from this rule, but they have generally not been willing to originate loans with DTIs in excess of 45 percent. This basically means that no matter the loan-applicant’s score arising from a regression model predicting default, if her DTI is above 45 percent, she will not get a loan.
This is not only analytically incoherent, it means that high quality borrowers are failing to get loans, and that the mix of loans being originated is worse in quality than it otherwise would be. That’s because a well-specified regression will do a better job sorting borrowers more likely to default than a heuristic such as a DTI limit.
He tests this by applying his model to the one third of the sample of loans held back in the initial estimation.
To make the point, I run the following comparison using my holdout sample: the default rate observed if we use the DTI cut-off rule vs a rule that ranks borrowers based on default likelihood. If we used the DTI rule, we would have made loans to 91185 borrowers within the holdout sample, and observed a default rate of 14.0 percent. If we use the regression based rule…… we get an observed default rate of 10.0 percent. One could obviously loosen up on the regression rule, give more borrowers access to credit, and still have better loan performance.
And extending the point
Let’s do one more exercise, and impose the DTI rule on top of the regression rule I used above. The number of borrowers getting loans drops to 73133 (or about 20 percent), while the default rate drops by .7 percent relative to the model alone. That means an awful lot of borrowers are rejected in exchange for a modest improvement in default. If one used the model alone to reduce the number of approved loans by 20 percent, one would improve default performance by 1.4 percent relative to the 10 percent baseline. In short, whether the goal is access to credit, or loan performance (or, ideally, both), regression based underwriting just works far better than DTI overlays.
The current focus in the US isn’t on responding to a house price boom, but on access to finance (in a market still dominated by the government). But the sorts of questions posed by these sorts of results are just as relevant here as they might be in a US context. Perhaps here too, high debt to income borrowers might generally be better quality borrowers? How confident can the Reserve Bank be that an actual debt to income limit – as distinct from a pure hypothetical – will actually improve the resilience of banks – not just on the housing book, but overall? And even if there is some improvement in resilience, at what cost – recall the statutory efficiency mandate – in terms of access to credit would that gain come at?
Perhaps there are good answers to all these sorts of questions. Perhaps the Reserve Bank has access to other careful studies that produce different, and robust, results. But these are the sorts of questions the new Minister of Finance, and the public, should be asking in response to the Governor’s request for political imprimatur for adding another tool to his kit of potential interventions. And, more broadly, how confident can we be of any sustained gains from such interventions, as compared to the sure increases in resilience that would result from either higher risk weights on housing loans more generally, or higher overall capital requirements for banks (and non-banks regulated by the Reserve Bank)?