I have posted a number of times on the extent to which the differential between mortgage risk weights applied to large and small banks is as big as is (repeatedly) asserted. APRA’s response to submissions on “Revisions to the capital framework for authorised deposit-taking institutions” (released 9 June 2019) has what I hope will be the definitive statement on the extent and justification for this difference.
I have copied the entire APRA comment on this differential below but the short version is that APRA does not see the difference being as large as it is claimed to be.
“When looked at holistically, the existing differential between the standardised and IRB approaches is small. While the precise calibration of the risk weights remains subject to further analysis, it is APRA’s intention that any differential in overall capital requirements, and hence any impact on pricing by standardised and IRB ADIs, will remain negligible.”
So far, it seems that the argument I have been making for some time on this question stands. Read on if you want more detail.
Risk weight differential in mortgage lending
For some time, there has been considerable interest in the impact, from a competition perspective, of the differential between standardised and IRB risk weights for mortgage lending.
Commentary on this issue has often focussed on the differential in average risk weights between ADIs using the two approaches. Superficially, this suggests a material differential exists. However, by only examining risk weights for on-balance sheet exposures, the impact of other important differences is ignored.
Beyond prescribed risk weights, differences in capital requirements for mortgage lending are driven by:
• differences in the credit quality of the underlying portfolio;
• differences in the ‘unquestionably strong’ capital benchmarks applied to standardised and IRB ADIs;
• differences in the treatment of credit conversion factors (CCFs) for standardised and IRB ADIs;
• the application of capital requirements for IRRBB to IRB, but not standardised, ADIs; and
• the requirement for an expected loss adjustment for IRB, but not standardised, ADIs.
Ignoring any differences in portfolio quality, each of the above factors serves to narrow any impact from standardised risk weights being higher than IRB risk weights. Under the current regulatory framework (i.e. before applying the proposals in this paper), APRA estimates that the impact of the overall difference in capital requirements on mortgage pricing is likely to be minimal – in the order of 5 basis points.
The analysis does not consider the operational costs arising from investing in developing and maintaining risk management systems to support IRB status, as well as data requirements.
Furthermore, the application of an additional capital buffer to those banks designated a domestically systemically important further narrows, if not completely eliminates, the overall difference for those banks.
APRA does not expect the changes proposed in this paper to materially change the above conclusions. When looked at holistically, the existing differential between the standardised and IRB approaches is small. While the precise calibration of the risk weights remains subject to further analysis, it is APRA’s intention that any differential in overall capital requirements, and hence any impact on pricing by standardised and IRB ADIs, will remain negligible.
From the outside 
2 thoughts on “Mortgage risk weight fact check – APRA’s perspective”