The RBA Financial Stability Review (FSR) released in October 2018 presents the results of a study on the impact of APRA’s housing lending measures. The study articulates findings on the competitive impact that are intuitive, but which have never been so clearly acknowledged by a member of the Council of Financial Regulators.
The study clearly acknowledges the competitive impact of the investor growth measure, while also discerning some impact on (regional) house prices. Lending standards have tightened in many dimensions; for example, investor lending at LVRs above 90% has virtually dried up. In the meantime, owner-occupiers have increased their share of new lending. As regards payment stress, the study shows that the 90+ day arrears rate increases 60% for the group of borrowers who don’t manage to switch their interest only (IO) loan to principal and interest (P&I) before the expiration of the IO period. Perhaps unsurprisingly, the RBA does not highlight the anti-competitive effects of the measures as clearly as it could have, however, it is reassuring to know that the macro-prudential tools employed have had a measurable impact on housing lending.
The study lists 12 APRA and ASIC housing lending measures and announcements since November 2014 (Table 5.1) and focusses on the two measures that directly or indirectly affected investor lending:
- 10% investor growth benchmark (Dec 2014)
- 30% interest only limit (Mar 2017)
In our view, the most controversial measure was the investor lending benchmark because:
1) compliance was essentially impossible for ADIs with low mortgage origination;
2) anti-competitive effects were the strongest; and
3) it was presented as a benchmark despite effectively being a limit.
While the benchmark caused an immediate deceleration to investor lending, it did not have the same impact on overall mortgage lending, where owner-occupiers picked up the slack. The RBA doesn't comment on the possibility that the movements could be partly due to owner-occupiers taking out investor loans before the announcement, or investors applying for owner-occupier loans afterwards. Given the notable restatements of the share of investor loans in the back books of even the largest banks, the statistical accuracy of the numbers could have been scrutinized more.
The anti-competitive effects of the investor growth benchmark are spelled out in Graph 5.6 and the adjoining text, noting the following:
- Banks that had been gaining investor market share in 2014, i.e. the most constrained banks, suddenly had to stop competing for investor loans. Their aggregate investor market share is now 2 percentage points lower than it would otherwise have been.
- If APRA had structured the investor measure in a more competition-neutral way, these banks could have had a 30% larger market share in new investor loans. We would consider this a large competitive impact.
- The competitive impact was partially mitigated by the most constrained banks being able to increase their owner-occupier lending.
- Major banks benefited from the market share loss of the most constrained banks, which is natural given the largest mortgage lenders needed relatively less cushion to operate safely below the 10% growth benchmark.
- Non-ADI lenders have also seen higher growth over the same period, although this may mostly concern owner-occupiers not being able to obtain a large enough mortgage from an ADI.
Since 2014, interest only loans have become harder to obtain, which means that more borrowers are forced to convert to principal and interest at the end of the typical 5-year interest only period. The resulting stress is illustrated in Graph 5.7 which indicates:
- The group of borrowers who switch their IO loan to P&I before the expiration of the IO period are generally better able to service their loan. The past due rate of 0.2% for this group is 4 times lower than for P&I loans (0.8%) and does not increase in the year around the expiry of the IO period.
- The group who fails to switch their loan before the end of the IO period fares worse. The past due rate for this group starts to trend upwards three months before the IO period expires, suggesting that their arrears are not only due to the increase in repayment amount, but also to other factors. This group also experiences a 12-month increase in 90+ day arrears from around 0.8% to 1.3% (0.5 percentage points or approximately 60%).
Interestingly, the RBA researchers do not mention or seem to have adjusted for the counterintuitive effect that due to the increasing repayment amount, a switch to P&I will initially and instantaneously reduce arrears expressed in number of days of regular payments, the convention used in Australia. The conclusion for this might be that borrowers who switch to P&I after expiry of the IO period contain a proportionally much larger number of borrowers facing payment stress (for mortgages these are typically job loss, divorce and health issues), which have not yet been picked up in arrear statistics.
Regional price impact
The RBA has analysed the price development in regions with both a high and low share of investor loans and is able to quantify the impact of the lending measures on housing prices in Graph 5.9. They observe that since the introduction of the 30% IO limit, prices in low investor regions have increased 4%, while they have been flat in high investor regions (Graph 5.9).
The RBA attributes this difference predominantly to the differential pricing of investor loans, which has accelerated since the introduction of the IO limit (Graph 5.1). We consider the price impact welcome from a prudential perspective as this tempers systemic risks, although the RBA points out that the lending measures were targeting housing lending rather than housing prices.
We know that the RBA is close to these housing lending measures through its participation in the Council of Financial Regulators, so it is commendable that an organisation so close to the fire is able to investigate and acknowledge the controversial competitive impact. However, we note that the RBA seems to downplay the impact when it writes that the “measures had some impact on competition but it does not seem to have been permanent”. We would consider that a 25% lower investor loan market share for a group of ADIs is considerable, and so far there is no evidence that this group’s market share will revert to the levels it would have reached without the investor growth benchmark (RBA’s counterfactual scenario), nor do we expect this to happen anytime soon. Furthermore, the baseline assumption among economists is that temporary measures have temporary consequences. Calling out that the impact seems not to have been permanent without mentioning the temporary nature of the measure itself, might indicate some self-censorship.
Finally, since 2014 ADIs have effectively been subjected to an investor growth limit that was presented as a benchmark. We understand the RBA researchers are unable to draw attention to this seemingly Orwellian practice, but they are caught up in the euphemisms by describing the 30% interest only limit as a benchmark! Regardless of the need to keep these concepts separate, this should not diminish our appreciation for a study that sets the record straight on Australia’s induction into macro-prudential measures, and for APRA now taking steps to revoke this measure that has had such a marked impact on competition.
The bottom line is that the RBA rightly concludes that the macro-prudential housing policy measures have helped reduce financial stability risks, which ultimately benefits us all.
See RBA, Financial Stability Review, October 2018, pp. 75-88, https://www.rba.gov.au/publications/fsr/2018/ oct/pdf/05-effects-of-housing-lending-policy-measures.pdf.