The recent IMF financial stability assessment for Australia is confused on key points and poorly justified on some of its main recommendations.
There are twelve “high priority” recommendations. The bulk of the recommendations give the various regulators advice on how to do their jobs better – and it is unclear that the IMF is any better at regulation than our national regulators. One of these (#6) is purely bureaucratic, lobbying to help ASIC get more funding.
The recommendations which could have a significant impact are:
#8: Re-evaluate the merits of ex-ante funding for the financial claims scheme with a view to converting it to an ex-ante funded scheme.
#3 and #9: Introduce higher loss absorbency for systemic banks.
The financial claims scheme is designed to protect depositors in a crisis. The IMF would prefer that Australia impose a levy (tax) on deposits to build up a fund from which depositors could be paid if ever a bank was liquidated and there were insufficient resources to pay out its depositors. There are several issues. The government could impose a tax on bank deposits of the sort proposed but should recognise that part of the effect would be to make saving less attractive. The IMF suggests that the amount involved should “cover deposits of two to three mid-sized banks” which of course makes no sense in Australia, the amount would have to cover the potential failure of one of the major banks. And then (footnote 29) the report recognises that it is not clear what such a fund could invest in given the lack of government securities in Australia. It seems clear that the IMF team has not thought through the issues even though this is one of its high priority recommendations.
In recognition of the systemic importance of the major banks, the IMF praises the basic approach taken by APRA but says that, in addition, a higher loss absorbency provision should be considered. It suggests for example that to achieve a 99.95 percent probability of not defaulting on any payment would require the major banks to hold additional Tier 1 capital ranging from 1.4 to 5.2 percent of their risk weighted assets. This is both surprising and large. Most banks will already be using 99.95 percent as the standard for their economic capital risk models which makes it hard to believe that the could be so far away from holding an appropriate amount of capital.
When one goes to Box 3 (p23) to understand the calculation, all clarity disappears. It is based on a proprietary model (Moody’s KMV distance to default model) using parameters derived globally and applied to the Australian banks – which means that no one can say whether it makes sense in our context or not. Given the size of the US financial system it probably means we are using a model calibrated to the US for application to our economy and banking system (which are both fundamentally different). Since the relationship is non-linear, the IMF fitted a power curve to the relationship and drew inferences about how much more capital banks might need. Any result will be extremely sensitive to the particular non-linear function used. The IMF presents no sensitivity analysis and in the one chart presented, the curve does not match the data! This is very poor quality work on which to base two recommendations in the report.