By Jenny Ruth
Oct. 1 (BusinessDesk) - The three independent experts' critiques of the Reserve Bank's bank capital proposals are largely supportive with the British expert, in particular, demolishing the commercial banks' complaints that the central bank wants them to hold too much capital.
David Miles of Imperial College, London, says criticisms that RBNZ failed to undertake a cost-benefit analysis "melt away in light of the procedure actually followed by the central bank."
RBNZ has said it will produce a formal cost-benefit analysis when it announces its final decisions, currently scheduled for early December, but governor Adrian Orr has said the bank has already effectively done one, even though the results haven't been so labelled.
While the RBNZ may have been overly conservative in some calculations, it was overly optimistic about, for example, the costs of a banking failure on the economy, Miles says in his paper.
"The analysis conducted by the RBNZ seems to have been done with care and in an open-minded way. The thoroughness of the analysis is impressive, given the relatively small team that worked on it," he says.
The Reserve Bank has proposed lifting the minimum amount of common equity tier 1 capital the four major banks have to hold from 8.5 percent of risk-weighted assets currently to 16 percent and to 15 percent for the smaller banks.
The big four banks account for about 88 percent of New Zealand's banking system.
The experts' views do contain some omissions. In particular, they fail to discuss the differing approach by the Australian Prudential Regulation Authority and the context that New Zealand's four major banks are owned by Australia's four major banks.
For example, while APRA is also requiring Australian banks to increase their capital, it is proposing to allow all the increase to come from hybrid or convertible securities, those which normally behave like debt but which can be converted to equity if a bank were to become stressed.
The RBNZ, by contrast, wants all the additional capital to be common equity.
The experts do consider the RBNZ's reasons for wanting to rely solely on common equity and generally approve of its arguments.
The American expert, Ross Levine from the University of California, Berkeley, says the central bank's arguments against allowing hybrids are "prudent and persuasive."
However, the Australian expert, Dr James Cummings from Macquarie University, suggests the RBNZ might look at whether the rules attached to hybrids "can be adapted to suit New Zealand's circumstances."
Miles queries estimates by some banks that the cost of tier 2 hybrids is 3-to-3.6 percent but that the cost of equity is about 14.5 percent.
"The reasons for the apparent enormous difference in cost are something of a mystery."
None of the experts discuss why the New Zealand subsidiaries have successfully passed a number of RBNZ stress tests.
Miles dismisses arguments that the RBNZ is requiring more bank capital than is the case in almost any other developed economy.
That argument that the proposals are "'out of line' is in itself not a powerful argument that the proposals go too far," he says, suggesting that other countries may not have "done enough of the sort of analysis" done by the RBNZ.
The RBNZ's own view is that although the new capital levels would be higher than in some other countries, they won't be out of kilter.
"From a total capital perspective, the RBNZ’s proposals are in line with other jurisdictions. However, we are proposing that New Zealand banks have a higher proportion of higher quality capital, which will make banks more resilient to losses," the central bank says.
Cummings says the proposals are "based on sensible analysis and advice in the New Zealand-specific context."
The analytical approaches RBNZ used "appropriately capture the relationship between bank capital and financial system soundness and efficiency."
Levine takes a more critical view but still says that "with few exceptions, the RBNZ conducted a sound analysis of bank regulatory capital requirements in New Zealand, employed appropriate data, methods and evidence".
"More generally, I found the RBNZ team working on the capital review first-rate analytically, dedicated to improving the banking system in New Zealand and ready to discuss and provide information on any consideration that I had."
All three experts support levelling the playing field for the smaller banks by reducing the capital advantage the four major Australian-owned banks have from being allowed to use their own internal models for assessing how much capital they need. All the other banks have to use standardised models.
In February, the RBNZ released data showing that ANZ Bank, New Zealand’s largest bank, needs to hold just below $3 of capital for every $100 of mortgages while the government-owned Kiwibank has to hold about $5.70.
It appears unlikely that ANZ's mortgage portfolio is so markedly safer than Kiwibank's.
Cummings approves of RBNZ's intention of making the four major banks report both under the internal models and using standardised models so that the differences will be readily apparent.
Such reporting will address a "gap in the literature" on whether internal models provide more accurate estimates of capital.
Miles says the RBNZ isn't alone in its concerns about "some surprisingly low assessments of risk-weighted assets for some banks", which internal models have been producing.
"Many central banks and other supervisors share the uneasiness about the IRB (internal models) approach," he says.
Levine says both the RBNZ's analysis and the international experience "indicate that banks not only have strong incentives to use their models to produce lower RWAs, they act on those incentives."
He also points out that RBNZ "does not have all of the data and programs to vet fully each bank's IRB model" and that the RBNZ has limited resources.
All the banks are supposed to get RBNZ approval for each of the internal models they use but it has come to light that both Westpac and ANZ have failed to get such approval over extended periods, in Westpac's case going back to 2008 when it first began using internal models.