Following the recent coverage of our mortgage stress analysis (light it seems is now dawning on regulators, industry commentators and others that household debt is a real and growing issue), one question we get asked is – yes, but surely the banks have guidelines on affordability and serviceability, minimum assumed rate 2%+ above current rates, or 7%+?
So surely, households should have buffers as rates rise?
This is a great question, with a long history attached to it. A couple of years ago the regulators got a shock when them looked at bank practice, and started putting out more specific expectations on lending standards. Back in 2015, Wayne Byres made this point in a speech on lending standards. At its core was the observation that borrowers appeared to be able to get very different loan amounts from a selection of lenders, using the same base financial profile.
One significant factor behind differences in serviceability assessments, particularly for owner occupiers, was how ADIs measured the borrower’s living expenses (Chart 2a and 2b). As a regulator, it is hard to understand the rationale for large differences in what should be a relatively objective, and extremely critical, metric.4
Of major concern were a few ADIs who opted to make their credit assessment based on a lower level of living expenses than that declared by the borrower. That is obviously a practice that should not continue, and ADIs should be making reasonable inquiries about a borrower’s living expenses. In fact, best practice (and intuition) would be to apply minimum living expense assumptions that increase with borrower incomes; this was a practice adopted by only a minority of ADIs in our survey.
In 2014, the RBA made this statement in their Financial Stability Report.
Although aggregate bank lending to these higher-risk segments has not increased, it is noteworthy that a number of banks are currently expanding their new housing lending at a relatively fast pace in certain borrower, loan and geographic segments. There are also indications that some lenders are using less conservative serviceability assessments when determining the amount they will lend to selected borrowers. In addition to the general risks associated with rapid loan growth, banks should be mindful that faster-growing loan segments may pose higher risks than average, especially if they are increasing their lending to marginal borrowers or building up concentrated exposures to borrowers posing correlated risks. As noted above, the investor segment is one area where some banks are growing their lending at a relatively strong pace. Even though banks’ lending to investors has historically performed broadly in line with their lending to owner-occupiers, it cannot be assumed that this will always be the case. Furthermore, strong investor lending may contribute to a build-up in risk in banks’ mortgage portfolios by funding additional speculative demand that increases the chance of a sharp housing market downturn in the future.”
“A build-up in investor activity may also imply a changing risk profile in lenders’ mortgage exposures. Because the tax deductibility of interest expenses on investment property reduces an investor’s incentive to pay down loans more quickly than required, investor housing loans tend to amortise more slowly than owner-occupier loans. They are also more likely to be taken out on interest-only terms. While these factors increase the chance of investors experiencing negative equity, and thus generating loan losses for lenders if they default, the lower share of investors than owner-occupiers who have high initial loan-to-valuation ratios (LVRs; that is an LVR of 90 per cent or higher) potentially offsets this. Indeed, the performance of investor housing loans has historically been in line with that of owner-occupier housing loans.
The trouble is that the basis on which banks have been assessing available income has been too optimistic. This is because they are based their calculations on historic mortgage book performance, when incomes were rising strongly, and loan losses were very low.
But we are now in a new normal. We have flat incomes. We have rising costs. We have underemployment. We have low growth. But costs of living are rising (and higher for many than the ABS CPI figure would suggest). Affordability is not what it was. Lenders need to adjust, hard to do when home prices are so high.
Combined, many households are stretched, and the prospect of rising interest rates in these conditions are making things harder.
In addition, households have been willing to gear up with the prospect of future capital growth, so reach for the largest mortgage they can get. Perhaps sometimes they exaggerate their incomes, and understate their costs. This is true we think for households with larger incomes, and lifestyles. Some of these are now under pressure.
So, the root cause of the gap between theoretical affordability and real life is a serious one. Banks often set and forget loans, so do not revisit households finances unless there is a reset or a crisis. But for many, available incomes are falling (and bracket creep is not helping). Ongoing financial health-checks might be in order.
ASIC is rightly looking at this issue anew, but we fear the horse may have already bolted. APRA will tighten capital. Both will put upward pressure on mortgage rates, and test affordability further. This is a paradigm shift.
