The superannuation system is a black box to many.  To others, especially those with more exposure to its inner workings, it is hopelessly complicated, tangled and confusing.  Lifting the lid on such a system, to see how it works (and how it doesn’t), would be daunting from either perspective.  But that’s just what APRA’s recently released Reporting Standards attempt to do.  They are designed by APRA to enhance APRA’s supervisory capacity and to promote transparency and comparability within the superannuation industry.  And, if the Regulation Impact Statement (‘RIS’) filed by APRA is to be believed, those benefits outweigh the considerable costs that will be incurred by the industry in complying with the new requirements.

The problem is how to perform that calculation.  How can we know if all the time, expense and effort to collect, compile and publish the information will be worthwhile?  It should be a simple calculation: cost < benefit.  It turns out to be harder than it appears.

Start with the cost side of the inequality.  APRA’s RIS recognises that RSE licensees (aka fund trustees) will incur significant IT development costs to capture and report the information required by the new reporting standards.  The RIS also recognises that there will be indirect costs, such as process re-engineering and training, as a result of the new requirements, and that there will be flow-on implications (read, additional costs) for auditors, actuaries and outsourced fund managers.  There is also a cost the RIS doesn’t mention: the governance cost of distraction.  After all, the resources employed in the re-engineering process could have been deployed in other ways.  All up these costs run into the millions of dollars[i] for participants in the industry, and many tens of millions for the system as a whole.

That said, the RIS gives a misleading picture of the true costs of the Reporting Standards.  For a start, the RIS identifies that at least some of the costs arise because of the need to bring legacy systems up to scratch.  That’s not so much a cost of the regulation as simply the most recent reminder of both the failure of the government’s Product Rationalisation initiative and the chronic under-investment in capital expenditure by most participants in the superannuation industry.  It seems churlish to sheet home the costs of that expenditure to the Reporting Standards.  What’s more, APRA is characterising all the increases in cost as ‘regulatory’ when many of the changes to the Reporting Standards are driven not by changes in regulatory practice, but by legislative change.  Some, at least, of the changes are not within APRA’s discretion: failure to change would result in APRA failing to comply with its legislative duties directly.  The changes required to comply with the government’s broader Standard Business Reporting initiative are the prime example of this, but so too are the changes required to populate the fund dashboard.  That is another set of costs, then that should not properly be regarded as ‘regulatory’ except in the trivial sense that someone has to be on hand to enforce the legislative will of the parliament.

Some will no doubt ignore such quibbles as sophistry.  They will plough on with their simplistic cost-benefit analysis and remind us that the costs must be weighed against the value of the information to APRA.  With better quality information APRA expects that supervision of individual RSEs will be enhanced, especially those administering more complex structures that include many investment options and sub-funds.  It may also permit APRA to gauge the alignment between the RSE licensee’s business plan and the requirements of the members it serves.

The question is of course whether the information collected is capable of sustaining such expectations?  The data items are not so granular as to permit accurate portfolio evaluations.  No self-respecting investment bank, fund manager or even consultant would rely on summary data such as that required under the Reporting Standards. Indeed it is arguable that no self-respecting super fund board should rely on such data.  The stress testing they are now required by SPS 530 to perform would be trivial and misleading if it were to rely on such data.  So the data will not actually facilitate direct analysis by APRA.  It might prime APRA’s supervisory teams with questions to ask, but even that is doubtful.  The risks are after all, like the apocryphal devil, in the detail.  The Reporting Standards simply don’t require a level of detail that would permit the risks to be identified.

Nor is the data current enough to permit timely intervention in most cases.  The accumulation of CDO exposure by superannuation funds in the years leading up to the GFC might have been detectable (assuming it could have been detected inside the ‘fixed income’ classification), but risks such as those deriving from counterparty or currency exposures evolve continuously.  The inevitable delays in compiling, reporting and processing the data render naïve any expectation that APRA will be able to forestall those types of risks reaching critical levels either on a local or systemic level. 

Of course one thing that APRA will certainly do with at least some of the data is to package and republish it in various forms.  That presumably is what the reference to ‘transparency’ in the RIS[ii] is all about.  Transparency is on the whole a good thing.   It can have dysfunctional effects if it gives incentives to economic agents to ‘window dress’[iii] or behave in an otherwise unwarranted risk-averse manner,[iv] but on the whole it is usually regarded as accountability-facilitating and, as a result, legitimacy-promoting.  So APRA will provide some of the information to the Australian Bureau of Statistics for their publications. It will also use some of the information to populate the new fund ‘dashboards’ required for MySuper products and will no doubt adapt some of its own publications to report consolidated statistics based on this more granular data.  It may even permit carefully-managed access to some of the data to academics and other researchers to facilitate scrutiny of the system that is independent not just of the commercial participants but of the regulators enrolled in the system as well.

But how should the value of such transparency be measured?  There is no way to measure the benefit precisely.  Transparency may bring to light risks or misbehaviour that demand attention, but as Brandeis noted almost exactly a century ago (December 20 1913, to be precise), publicity and exposure change behaviour; they deter malfeasance.  Sunlight ‘disinfects.’  The quantum of the (counter-factual) risks and inefficiencies that would have occurred at a systemic level had the new measures not been introduced can only be guessed at.  Given the vast sums of money administered within the superannuation system, it seems reasonable to suggest that the benefit in terms of system integrity is greater than the costs incurred by super funds to comply with the Reporting Standards.   But by how much?  We will never know.