ozrisk.net

As discussed below, quarterly disclosure of assets, credit quality and capital is supposed to be one of the “pillars” of the modern regulatory model governing Australia’s banking industry. Yet the mandated disclosure of this data is uneven, sometimes unavailable and, by and large, poorly read.

It was not meant to be this way. Theoretically, depositors and other investors in all banks were meant to take this information into account before making investment decisions. The fact that this is not happening means one of three things – it should be dropped as it is useless, it should be fixed or it just needs more time as people catch on.

The third possibility can be ignored, as the sheer difficulty of getting the data means that it is not probable that it will be done on a regular basis.

If we look at the second, then improvements clearly need to be made. Firstly, the target market for these disclosures needs to be determined – something that probably was not done before APS 330 was issued. In short – who is going to use this data? If it is the retail depositors, the mums & dads, then the data on the reports needs to be turned into something they can read and understand.

Some of the issuers of these reports are already doing this. Seventeen of the 121 issued reports feature explanatory information that may be accessible to the (relatively) unsophisticated user. Of the Big Four this includes CBA, the NAB and Westpac, with the standout being the Commonwealth. Their risk report seems to attempt to explain the whole 333 page Accord in each disclosure statement, the most recent of which was 87 pages long and includes a mass of detail.

Most of the disclosures, though, would have been meaningless to most mums & dads. To do as the Commonwealth does would also be a virtual impossibility for the small credit unions. Even to match, say, Westpac would be difficult.

Most of the disclosures simply take the bare compliance route – APS 330 mandates the tables to be included, so the data is in tabular format will either little or nothing by way of additional information. This is pretty difficult to read at the best of times.

Worse, and as is true of most disclosures, they really only are very useful when you can compare them to a competitor or to the institution itself, something that even the most sophisticated mum or dad would find difficult and time consuming.

If the target for the data is the wholesale market and credit risk analysts, the way that it is presented is also a problem. Given most institutions do not preserve a history on their websites, to get a meaningful data history means that an analyst would need to visit the websites of most of the 119 institutions that have one at least once every three months to get a full picture. For the four that don’t have a website an email or phone request would be needed.

The data would then need to be extracted from every report and then manually input into an analysis package.

This is clearly a lot of effort and, given it has not yet happened other than for this study, seemingly not a popular idea.

A possible technical solution to this would be for every ADI with a website to publish the data in a consistent way – possibly using XML or a similar method – to format the data for easy import into an analysis package.

Of course, this would mean than the 17 that are providing more and better information would have to be “dumbed down” to re-join the pack. Either that or they continue to produce the APS 330 formatted for general consumption as well as this other method.

A third possibility is that APRA itself publishes the report every quarter using data from its existing returns from each of the reporting institutions. This would have the benefit of a standard format and ease of finding. However, if APRA were to publish these reports then it would be difficult to avoid the impression that APRA has in some way endorsed the contents.

Perhaps they should be cancelled then – but this may be a breach of the Accord itself. There are two reasons why it may not be. The Accords were always meant to apply to “internationally active” banks, not to institutions with less than $1m in total capital. For the non-internationally active banks compliance with the Accord was a policy decision by APRA, not a requirement of the Accord.

Additionally, the Pillar 3 on which APS 330 is based speaks clearly about the concept of materiality – how it is the users that matter to this concept. The fact that there are so few, if any, users of these reports from the smaller financial institutions may well mean that the information is not actually material.

It’s clear that the APS 330 process, at least as it applies to the smaller institutions, needs to be re-thought. If we want to keep this form of market disclosure then it needs to be made useful to at least one group of investors or analysts. If the statements are not being used, then, perhaps, they should be dropped.

This is not coming out of retirement. I wrote this piece for Banking Day (an excellent way to keep abreast of banking in Australia) and I thought it may also be of more widespread interest. There will be one more piece tomorrow before I go quiet again.

Quarterly disclosure of assets, credit quality and capital is supposed to be one of the “pillars” of the modern regulatory model governing Australia’s banking industry. Yet the mandated disclosure of this data is uneven, sometimes unavailable and, by and large, poorly read.

Basel II Pillar 3 is the part of the international bank capital Accords that was intended to improve the discipline that the market imposes on banking institutions. It was meant to do this through increasing the amount and nature of disclosures that the banks (and other deposit takers) make on a routine basis, using the idea that more disclosure means there is more data and therefore better pricing.

There are, of course, two problems here. If the markets themselves are not pricing risk correctly (for whatever reason) then all the disclosures possible are not going to do the trick. The other problem is around whether anyone actually reads the disclosures.

The recent experience in Europe gives real pause on the first point. The Basel Accords themselves may be at least partially at fault here. The Accords effectively tell the banks that lending to their own government is so safe that they do not have to hold any capital at all against it. Most countries’ regulators also mandate that the bonds they receive for this lending are also always able to be sold instantly, and so count as being as liquid as cash.

Both of these now look, at best, overly optimistic.

What about the second, though? Does anyone actually read the disclosures and use them? If we turn to smaller Australian banks, building societies and credit unions, the answer seems to be a strong “No”.

Over the last few months these disclosures by all of the 123 banks (including the big four), building societies and credit unions operating in Australia have been collected as part of a research project. According to APS 330 all of them should have been reporting these every quarter since September 2008, making (so far) 12 reports in total.

There are also two differing types of disclosures – for listed banks or building societies extra disclosures must be made semi-annually and for unlisted institutions the extra disclosures are made annually.

There are still several faults in the process, with a couple not reporting at all and others not reporting fully or properly.

Both of the institutions not reporting were very small, but several of the other problems were more wide-spread. Of the 121 actually publishing, eight seemed to want to hide the results, making them virtually impossible to find on their websites, while six institutions had missed important data off the disclosures, including core numbers such as assets and impairments.

Three of the disclosures had clearly not been formatted for any sort of use.

Other problems were more widespread. For the June quarter end 14% (17 in total) of the disclosures were not out on time (40 business days after the relevant quarter end) – and this included three of the 12 banks.

The widespread issue, though, is that nearly 70% of the institutions are only making their most recent disclosure statement available – meaning that there is often no history or context, but, more importantly, for most of these the annual or semi-annual long disclosure is not on their website for at least half the year.

The other major problem is the sheer difficulty in harvesting the statements. The way that APS 330 mandates the disclosures is that they shall be put up on an institution’s website in a clear location. Formatting is meant to broadly follow the formatting in APS 330 – consisting of a variety of tables.

This was set up with the idea that an individual depositor could go to their institution’s website or offices and get a copy and then read it and understand their institution, and their risks, better. The problem here is that the basic format, and target audience, has not been thought through. There are not many depositors in the smaller institutions that would know the meaning of “Risk Weighted Assets”, “Tier One Capital” and the difference between 90 days past due or specifically impaired. Judging by many of the disclosures, not many of the preparers do either.

For your normal depositors, then, the disclosures are nearly useless.

For professionals, the situation is not much better.

Four ADIs do not have websites, making getting the data from them a process of calling or emailing them to ask for the data. For the others, it’s a matter of trying to find their website either by guesswork or using a search engine. If you are trying to get them all together and put them into a database it is a very time-consuming job.

The only people who may actually be able to use them, the wholesale providers of funds to the smaller institutions, would be able to get much better and timelier data direct from management.

There seems to be at least a few people reading, and using, the disclosure statements of the Big Four and Macquarie. This includes Banking Day readers, as the APS 330s for the majors are regularly reported there. However, it has not been noticed, least of all by APRA, that some institutions are not reporting at all, others are missing important data, several are late and most are not disclosing all of the data on a consistent basis.

It’s clear that the APS 330 process, at least as it applies to the smaller institutions, needs to be re-thought. If we want to keep this form of market disclosure, then it needs to be made useful to at least one group of investors – or analysts. If they are not being used, then perhaps they should be dropped.