“In past cycles, households have tended to default for three reasons: unemployment; family breakdown or health issues,” Mott points out. But this cycle is likely to look very different, with unemployment unlikely to rise from near historical lows.
Instead, what worries Mott is the speed of the rate rises, and how quickly housing debt has built up in recent years.
The inflation/rates story is well known; Tuesday’s predicted 0.5 of a percentage point rise would take rates to 1.85 per cent, from just 0.35 per cent in May, and ANZ forecasts rates hitting 3.35 per cent by the end of the year.
Warning from history
But Mott also provides some fascinating historical context on how debt has built up, by looking at the past 10 housing cycles to study the potential impact of rising rates on credit growth.
What’s particularly striking is the magnitude of growth in mortgage commitments in the two years preceding these 10 housing downturns; the 70.1 per cent increase in housing commitments in the two years before this current downturn is the second-biggest jump seen since lending data was first captured during the 1970s, and only beaten by the 131.5 per cent rise in the lead-up to the 1988 -89 housing downturn.
So, while the RBA has argued that mortgage borrowers look, on average, to be in a good position, with about 70 per cent ahead on their mortgage payments, Mott says this is meaningless.
“It is akin to the old saying that, ‘if you have your head in the oven and feet in the freezer, you feel OK, on average’. In banking, it is the tail that matters, the last 5 per cent to 10 per cent of borrowers.”
And this cohort, he says, have overextended themselves to get into the market since June 2020.
Borrowers at maximum
Mott uses Commonwealth Bank’s estimate that up to 10 per cent of borrowers have taken out their maximum possible mortgage over the past three years (that is, these borrowers can withstand 2.5 per cent of rate rises, but will have no excess cash), and estimates a similar or slightly larger proportion of borrowers will have gone very close to their maximum.
While he concedes the analysis is rough, he estimates that, in total, borrowers who have somewhere between $200 billion and $250 billion in mortgages will face severe stress if the cash rate hits 3 per cent later this year, as expected.
“If interest rates continue to rise sharply, and stay around these levels, there will be a ‘fat tail’ of borrowers who will simply not be able to afford to meet their repayments,” Mott says.
“For the first time in several decades, we are likely to see a wave of fully employed borrowers falling into delinquency as they simply can’t make ends meet.”
This period of mortgage stress would be compounded by the fact that $800 billion of fixed-rate mortgages taken out in the past two years at rock-bottom rates will start to expire over the next 18 months, with borrowers facing steep rises in borrowing costs; on a $1 million mortgage, annual interest payments of $19,000 may shoot above $50,000.
How does this all play into the major banks’ bad debt provisions?
Collective provisions stand at a historically low $17.4 billion, but if rates got to 3 per cent and stay there for a few years, Mott sees a scenario (although this is not a forecast) where that might need to rise by $16.4 billion by 2023- 24.
By way of example, he forecasts CBA’s bad debt charge rising from two basis points in the 2022 financial year to 30 basis points in 2024.
It should be noted there’s a lot of water to go under the bridge here. Perhaps most notably, the RBA would presumably react to housing market pain and the economic hit caused by mortgage delinquencies by cutting rates.
And to be clear, even if Mott’s scenario came to pass, there is no systemic risk to the banking system from these rising credit impairments; the “unquestionably strong” regime implemented over the past decade underpins the strength of the banks’ balance sheets.
But the pain would be felt in bank profits, where Mott sees other areas of concern.
His historical data suggests housing commitments generally fall from 20 per cent to 35 per cent during a housing downturn. But because of the rapid growth in the past two years, he estimates a 35 per cent fall is in the frame this time around. That could take mortgage credit growth from about 6 per cent in 2022 to 2 per cent in 2024 (within an overall range of between zero and 3 per cent growth).
Mott also remains concerned about costs. About 60 per cent of the banks’ costs come from wages, which are clearly rising, with IT and property costs creeping higher too. ANZ and NAB abandoned their formal cost-cutting targets earlier this year, and Mott has raised his cost estimates for the banks again as inflation bites.