Via Banking Day:
ANZ in New Zealand has been rebuked by the Reserve Bank, with the RBNZ on Friday announcing that it “revoked ANZ Bank New Zealand Limited’s accreditation to model its own operational risk capital requirement due to a persistent failure in its controls and attestation process”.
ANZ NZ is now required to use the standardised approach for calculating appropriate operational risk capital.
From March 2019, this will increase its minimum capital held for operational risk by around 60 per cent to NZ$760 million, the RBNZ said.
ANZ NZ in fact alluded to this embarrassing turn of events two weeks ago in its disclosure statement for the six months to March 2019.
The bank wrote in that document that “in April 2019, the bank informed RBNZ that in the course of a self-review, the bank discovered that it had not been using an approved model for the calculation of the operational risk capital requirement since December 2014.”
The RBNZ made plain its annoyance on Friday.
“Accreditation is earned through maintaining high risk management standards, and comes with stringent responsibilities for the bank’s directors and management,” deputy governor Geoff Bascand said.
“The Reserve Bank’s role is to review and approve internal models. The onus is then on bank directors to ensure, and attest, that their bank is compliant with the Reserve Bank’s regulatory requirements.
“To do that, bank directors need to be satisfied that the internal assurance processes that sit behind the attestations are being adhered to.
“ANZ’s directors have attested to compliance despite the approved model not being used since 2014. The fact that this issue was not identified for so long highlights a persistent weakness with ANZ’s assurance process.”
So, while the RBNZ is busy radically lifting capital requirements for Australian banks, and censuring them for their sloppy and dodgy adherence to standards, the still split monetary and prudential authorities in Australia are being left in the dust.
As reported last week, there is not even any kind of consistency for Australian banks in the way they report their negative equity, nor therefore LVRs and capital:
ANZ’s negative equity data was the highest of the big four at 5 per cent, or 4.25 per cent when offset accounts are factored in. Macquarie said ANZ’s approach – which used combined customer debt secured by property divided by property value – was the most granular and the most “conservative”.
CBA uses a similar methodology for arriving at its figure, but unlike rival banks its published figure disclosed the number of loans rather than overall value. Benchmarking against the RBA data, Macquarie suggested negative equity could be as high as 4.2 per cent across the book.
NAB used a state-based pricing index “which in our view materially understates the extent of the issue”, Mr German said.
Macquarie found Westpac’s figures were also inconsistent, noting its negative equity exposure in Western Australia was just 4 per cent, compared to the RBA’s figure of 14 per cent.
This matters a lot because the valuation methodology tips into the loan-to-value (LVR) calculations that determine how much capital a bank must hold against each mortgage. As MB’s Deep T. has described for years, the bank’s use an extremely dodgy set of rules to reduce that capital and ramp up mortgage lending:
The definition of standard loan is as follows as referenced in APS112.
6. A standard eligible mortgage is defined as a residential mortgage where the ADI has:
(a) prior to loan approval and as part of the loan origination and approval process, documented, assessed and verified the ability of the borrowers to meet their repayment obligation;
(b) valued any residential property offered as security; and
(c) established that any property offered as security for the loan is readily marketable.
The ADI must also revalue any property offered as security for such loans when it becomes aware of a material change in the market value of property in an area or region.
Sounds reasonable enough, even the last sentence. But here lies the “arbitrage the system” opportunity that every banker dreams of. APRA in APS-112 have insisted that property is revalued by an ADI if “it becomes aware of a material change in the market value of property in an area or region”. I will express an opinion here that I’m sure APRA was trying to protect against properties falling in value but have left open the biggest game in town for all our banks.
Currently the major and regional banks use the services of the new technology of automated valuation models (AVMs) to revalue their mortgage books upwards to decrease their capital requirements. Has a light switched on yet as to why there has been such a dramatic rise in the use and media coverage of those property valuation spruikers. If anyone knows of one of those banks who do not use AVMs can you please let us all know.
So what is the effect of continually revaluing the properties upwards? Using our standard 95% LVR mortgage above, let’s assume a revaluation so that the LVR is now 85%. The capital required is now $100*50%*8%= $4 with a return on capital requirement of $4*20%= $0.80 pa. Quite a difference, and if we go one step further, then the result and achieve a 75% LVR then the capital is now $2.80 with a return requirement of $0.56.
So with roughly a revaluation of the property of 20% (ask any property spruiker, “That’s nothin’ mate!”) a bank can save itself $3.20 of capital per $100 of mortgage which can be recycled as capital to support another mortgage. Think about how that increase in both return on capital and funds allocated to another mortgagor slave is an absolute incentive for bankers to perpetuate the cycle up of house price valuations. Their reward? Huge bonuses based on what is in essence a positive reinforcement spiral where everyone pats each other on the back for what a great job they’re doing. Well at least, that is, until the money runs out.
What was that? Did you hear anything? Maybe the sound of a slow burning fuse? But I digress.
Under advanced Basel II methodology, banks are able to have their own approved internal models which replace the APS112 requirements. These models only make it much less transparent as to what the capital calculations are based on. Please see the CBA’s statement and make note of the amount of residential mortgages used to calculate capital adequacy on APS 330 Table 16a to 16d – Capital adequacy (risk weighted assets) ie $ 56,753M and the actual credit exposure on table APS 330 Table 17a – Total credit exposure excluding equities and securitization ie $326,384M to understand the extent of the difference between what is actually lent and the exposure used to calculate capital.
In this case for CBA it’s $269Bn. I cannot explain how the numbers have been calculated other than to point them out and that some internal model which updates valuations has been used. The benefits on the way up in house prices would seem to be huge. What would the effect be if house prices dropped 20%?, 30%?, 40%?
If you have followed all of this then you have my utmost compliments, but also, maybe your heart should also be jumping a beat or two. The emu in the kitchen is not what losses the banks will have to suffer if house prices reduce and borrowers default in large numbers, it is where do they find the capital to support all those loans now on increased LVRs that have not defaulted? As APRA have required, The ADI must also revalue any property offered as security for such loans when it becomes aware of a material change in the market value of property in an area or region. A positive reinforcement spiral on the way up and a negative one on the way down.
I think we can say with confidence that any reform for APRA and the RBA is now off the agenda. The most likely pressure to come to bear on APRA now is simply to loosen macroprudential policy and dodge any need for further banking reform by getting house prices rising again.