By Scott Donald, UNSW

SYDNEY - 16 May 2014: I’m all for recycling:  Paper, aluminium cans, PET bottles - all good.  But not everything deserves to be recycled.

Take the rediscovery by the Reserve Bank of Australia and the Grattan Institute of the OECD research into the costs of private pensions markets around the world.  There are several reasons why this research ought not to be recycled, at least from an Australian perspective.  Some of the reasons are well publicised; some less so.

First, and most obviously, any comparisons have to be between alternatives that are similar. This point is recognised in the RBA submission to the Financial System Inquiry.  Some describe this as trying to ensure you don't compare apples with oranges (or lemons?), but there is more to it than that.  Comparing the costs of systems composed of defined benefit schemes with a near-universal, choice-laden defined contribution private savings system (such as Australia’s) is like comparing the costs of travelling by public transport (everyone arrives at the same time and it is ‘cheap’, but you have to go where everyone else is going) with the costs of private conveyance (more flexibility in the hands of the driver but much more expensive).  The question of which is better is clearly very important for policy makers at a whole-of-system level.  However it does not tell you anything about the ‘efficiency’ of the system in practice because the level of individual utility (and also aggregate welfare) in the two alternatives is quite different.  Moreover, unlike the apples and oranges metaphor, there is an objective hierarchy in the comparison that is independent of personal preference.  Some people prefer apples to oranges and vice versa, but having flexibility and choice is unequivocally better that not having them.  They are worth paying for.  The key question is: how much are those attributes worth?  Different people will have different answers.

The second reason that the OECD’s analysis of costs should be interpreted carefully is that any attempt to measure costs in the superannuation system has to recognise that current regulatory settings make them almost impossible to measure accurately.  This is not simply the obvious point that organisations can hide disclosed costs using a variety of accounting practices and legal structures.  Obviously they can, recent initiatives to improve transparency in the system notwithstanding.  The point is that the would-be analyst has to recognise more subtle ‘costs’ such as the cost to the sponsor of a defined benefit fund of providing the funding guarantee or the differences in regulatory costs of differently designed systems (how do you compare the costs of regulating 300+ entities versus the implicit cost of the systemic risk from too-big-to-fail market concentrations).

But perhaps the most important problem with the OECD analysis is that the focus on costs itself is for most purposes misplaced.  Opportunities for efficiency gains (which is why the analysis is most often conducted), ought to focus on the presence of economic rent, not on costs.  Rent is the concept used by economists to describe the quantum of profit above that required to compensate for the risk undertaken.  Who then, if anyone, is extracting the rent (or in the more colourful argot of the tabloids, making the super-profit)?  This is the crux of the matter. 

It seems clear that there are no super-profits in the member benefit administration industry.  If anything the pursuit of ever-finer margins by funds at a local level has given rise to a system-wide risk in Australia that the provision of member administration services is not sustainable at current fee levels, especially given the appetite of many funds for the provision of choices to members.  Something has to give.  (Recall that low fees in the global custody market were recently found to have been supplemented through a range indirect ‘enhancements’ in the currency and cash management practices of a number of custodian banks.) 

It is probably also true in the asset consultant, audit or actuarial communities also, though in any case the ‘costs’ of those services are typically immaterial in the broader scheme of things.  It also seems unlikely that there are super-profits to be had in TPD insurance arrangements, given recent attempts at repricing by insurers.

What about asset management fees?  This is more complex.  Research in 2009 by Deloitte suggested that the fees paid by Australian funds for asset management are roughly commensurate with those paid in other countries, on a like-for-like basis.  However tax, liability and cultural factors cause Australian funds to be more attracted to equity investment than many of their international peers.  Alternative investment techniques, also, tend to be more expensive, as is active management.  So the risk-appetites of Australian funds tend to drive fees up.  Is this money well spent?  That is hard to gauge.  In order to answer that question you have to assess the success of funds’ investment strategies each against their unique objectives.  So a simple comparison of net returns will not of itself tell you whether the fees paid and costs incurred were worth it, you have to know what the objective was.  (By way of illustration, the performance of a number of large DB plans in the 2001-2006 period looked quite anaemic in absolute terms compared to other corporate fund peers, until you take into account the fact that they were ‘immunised’ using duration-matching bond portfolios and hence entailed almost no investment risk.)  Moreover, statistical performance measurement is a notoriously ‘noisy’ business.  You need a long time-series of observations (or very aberrant relative performance) to form any confident conclusions using statistical techniques when risk and estimation error are accounted for.  Then of course there is the issue of how to calculate accurately the costs of managing money internally, as many larger funds are increasingly doing.  What price management distraction, organisational risk and the like?  Beyond a certain scale, internal asset management seems cheap until something goes wrong.  And finally you have to recognise that a comparison of the fees paid to asset managers by Australian funds to international peers presupposes that fees are competitive in other markets.  If asset managers are able to extract economic rent in other markets, comparing ourselves to them should offer little solace.

So where to from here?  Is it all too hard? 

It is too important to put in the too-hard basket.  There are some things we can say. 

Any opportunities to reduce the costs of the superannuation industry must come from one or both of two sources.  They could come from re-configuring the architecture of the system to reduce the ability of the financial sector to over-specify the products it provides.  This one part of what the MySuper initiative is attempting to achieve.  The jury is still out on whether that initiative will be successful in achieving that objective, but a number of researchers are at least starting to address the question. 

Alternatively, cost reductions could result from price competition between service providers in some or all of the component services that are required to administer the system.  The de-entrenchment provisions in the Stronger Super reforms attempt to clear away some of the impediments to that.  This is reinforced by the express reference in the s52(6) investment covenant to costs as a consideration for trustees in setting investment strategies.  It is however not helped by an apparent lack of price sensitivity amongst fund members, which removes one source of potential pressure on trustees to push for lower costs.  Therefore whether the findings of Arnold and Liu in 2010 that entrenched arrangements appeared to be more expensive will be addressed by these initiatives likewise remains to be seen.  So to does the answer to the question of whether the trend towards internal asset management improves the costs incurred by funds and hence improves the overall efficiency of the system.

Finally, other obstacles to cost competition remain, most obviously the concentration of market power in Australian financial markets in the hands a few organisations (the banks, AMP and a small number of others).  These organisations dominate the advice, trustee, custody and asset management elements of the superannuation system.  Concentration undoubtedly provides some benefits to the system, for instance in terms of stability and regulatory focus, but it also has implications for competition, innovation and market sovereignty generally.  More work is needed in this area.  Urgently. 

Put simply, there may well be parts of the superannuation system that are too expensive but it is going to take much more careful analysis than we have so far seen to identify them accurately, and then further careful thought to decide what to do about it.