Speech ‘Supervisory Developments: The Reserve Bank's Perspective’
Talk to Banking '95 Conference
Sydney – –
I am pleased to be able to report that the Australian banking industry today is in pretty good shape. Recovery from its hangover from the 1980s is effectively completed – with capital ratios high, impaired loans much reduced and profitability healthy.
As they emerge from their recovery phase, banks have become more aggressively competitive. This has been particularly evident in new housing finance and some areas of corporate lending, and is another positive development – as long as the competitive juices are tempered by a good dose of realism and prudence – as, clearly, they were not in the 1980s. From time to time, we hear that banks are again compromising prudential standards in energetic pursuit of some types of business. This tendency will bear close watching, but we are yet to see evidence of it happening widely.
While prudential supervisors can rest reasonably comfortably these nights, supervision is certainly not standing still and the days remain busy. This is because the structure and risk profile of the banking system is changing rapidly and in significant ways.
Over the past 10 to 15 years banking (and the financial system more generally) has been dramatically transformed by deregulation, product innovation, the internationalisation of markets, and more volatile financial prices. Continuing change is assured by advances in communications, computing capacity and financial theory. Transaction and information costs will fall further, and the unbundling and repackaging of traditional services will increase.
Competition among various parts of the financial system has been intensified by a narrowing of distinctions among some products and easier access to overseas markets. In addition, securities markets and funds managers are tending to displace intermediation in certain market segments, with the prospect of greater reliance on wholesale funding by high quality borrowers and of better quality assets moving off banks' balance sheets through securitisation.
Such competition, and the lifting of regulatory restrictions on their activities, have raised both the willingness and capacity of banks to accept credit and other risks. In response to pressures and opportunities banks have also diversified into a wider and more complex range of products. An increasing proportion of their revenues comes from fee-based and other off-balance sheet activities, including advisory and funds management services, back-up facilities for capital markets participants, and proprietary trading in foreign currencies, securities and derivatives.
While the fundamental credit and market risks for banks in many of these activities are not new, they have become more difficult to measure and manage. Meanwhile some risks are newer: the marketing of more complex products, often combined with a shift in relationships from borrower/lender to adviser/agent, has altered legal responsibilities and increased ‘moral’ or ‘reputational’ risks. These have been highlighted in recent times by large compensation claims against banks in the United States and elsewhere.
The narrowing – but not elimination – of differences among certain products and the ambition of some institutions to offer one-stop financial shops have encouraged marketing alliances and financial conglomeration. (It is notable, however, that the early international euphoria for ‘bancassurance’ seems to have been tempered a good deal by hard experience.) Conglomerates pose a number of interesting challenges for supervisors, as well as for their own managers.
I would like to talk about two main topics – the Reserve Bank's general approach to bank supervision in this environment, and supervisory responses to the growth of financial conglomerates.
(a) The framework
I hope it is not too difficult to discern the general thinking behind our supervisory policies, but it is useful to spell this out from time to time.
At the outset, three basic points should be emphasised.
- Banking is too important not to be supervised. Partly, this is a matter of its size in the financial system: Australian banks control directly almost half of financial system assets, and indirectly another 12 per cent. Supervision is also important because of the ease with which problems in one bank can spread to otherwise sound institutions – innocent bystanders – either through its failure to meet obligations to them (including in the payments system) or through damaging depositor confidence.
- At the end of the day, no supervisory system can ensure that a bank won't get into serious trouble. Contrary to the ill-informed comment one sees from time to time, supervision in Australia does not and cannot purport to guarantee a bank against insolvency. Our slightly more modest aims are to reduce the likelihood of this happening, to protect the interests of depositors in the event of serious problems and, if a bank did fail, to seek to limit flow-on damage elsewhere in the financial system. It should go without saying that the main responsibility for the prudently profitable operation of a bank lies with its directors and the senior managers who oversee day-to-day operations. They do well to see the supervisor as an ally, but not an alibi for any lack of diligence on their part.
- Supervision can be costly. Within its limits, supervision undoubtedly promotes the soundness of banks and fosters public confidence in them. It can also, however, entail certain costs. These include the resources which banks devote to consultation with supervisors, to statistical reporting and extra audit fees, although I doubt that these are very significant in practice in Australia. (Most of the data required by a supervisor should be of intrinsic interest also to a bank's managers.) Larger costs would come from supervisory restrictions which curbed banks' flexibility to innovate and adapt, to take or create new business opportunities. Such consequences might be justifiable in the interests of stability, but the potential damage needs to be recognised and avoided wherever practicable. In other words, supervision should interfere with the commercial ambitions of banks only when there is clear prudential justification.
Within this broad framework, there are three key elements in our supervision – one which has been a cornerstone for some time and two which are becoming more important as the complexity of banking increases.
The first is capital adequacy. Banks need a level of capital commensurate with the risks of loss in their business. A good level of capital cannot prevent a bank being weakened by poor management or bad luck, but it is a source of market confidence and will give a breathing space for survival which is not available to less well capitalised banks. Since the mid 1980s Australian banks have had to hold a minimum amount of capital in proportion to the credit risks they carry. Soon there will also be capital requirements for market risks – that is, risks from volatility in interest rates, exchange rates and equity prices.
The second element is good information. The experience of the 1980s drove home to many banks – and their supervisors – the pressing need for more accurate, relevant and timely information on asset quality. Again, better information will not necessarily stop a bank getting into difficulty, but it could help to prevent a blind accumulation of problems and should assist in identifying weaknesses before they become full-blown. The earlier a weakness is identified the easier it is to organise a remedy – which might mean closer nursing of impaired loans, or topping up capital. One important example of our work in this area was the design, with banks and their external auditors, of more objective and consistent reporting guidelines for impaired assets.
Information about banks is important in another way – helping the market to differentiate between the well-run and the poorly managed so it can apply appropriate disciplines to the latter. While we do not believe supervision can rely as heavily on public disclosure as some have advocated – because that entails unrealistic expectations of depositors – we do support high standards of disclosure about banks' activities. Our guidelines for reporting impaired assets are being carried through to their public accounts. We also support improvements in banks' public reporting of derivatives operations.
Information plays a key role in one other respect. This is in our requirements that banks disclose to depositors/investors how far they can and – the key point – cannot rely on the support of a bank itself when they deal with an associate of that bank. Such requirements have become essential for letting banks diversify into such ‘non-banking’ activities as funds management.
The final element of growing importance in bank supervision is a focus on internal risk management systems. Management systems and information are clearly two sides of the one coin – accurate, timely information is both a key input to and an output of a useful management system.
We have long recognised the need for robust systems to manage risks but our interest has been sharpened for a couple of reasons. First, the biggest bank losses in the early 1990s were closely correlated with internal systems which were either poorly designed or had broken down in the late 1980s. As a result, in the past couple of years we have spent a lot of time looking at credit management systems and developing benchmarks for ‘good practice’. One outcome is that we now expect banks generally to use a grading system to give a snapshot of the overall quality of their loan portfolios and to track changes in that.
The other reason for focusing on management systems is that the complexity of risk assessment in the more sophisticated banks is making simple rules of thumb less appropriate in determining what, for instance, are adequate amounts of capital or liquid assets. On the other hand, relatively simple approaches can remain suitable for banks with a narrower range of activities.
One way for us to get better acquainted with banks' risk management systems has been the program of specialised visits which commenced in 1992. These visits help us to assess whether systems for monitoring and controlling risk are prudent, operating effectively and generally consistent with good industry practice; they also strengthen our ability to direct in-depth investigations into the affairs of a bank should serious doubts arise as to its soundness. Visits conclude in a session with bank management where we seek to clarify issues which have arisen and discuss our observations, including the need for corrective action if deficiencies seem present.
Until recently our on-site work focussed on systems for managing credit risk, but at the end of 1994 we began a parallel program related to trading and treasury activities. By the end of this year, we expect to have visited all banks at least once and to be into a second round on asset quality; we also expect to have completed market risk visits to banks with substantial treasury operations.
The importance of strong control systems has been emphasised in supervisory consideration of derivatives activities, which are both complex and involve the potential of large losses accumulating rapidly. The recent Barings collapse appears to have been caused more by systems failure than by any newly discovered dangers in the relatively simple derivatives on which the losses were made. There are now international benchmarks of best practice for management of derivatives. Our 1994 survey indicated that Australian standards measure up quite well, but there is always room for improvement.
(b) Two cases
Let me illustrate how these general features of our supervisory approach apply to a couple of current policy issues.
We have been reviewing guidelines on banks' involvement in securitisation. Our 1992 guidelines restricted significantly the extent to which a bank could provide credit enhancement and liquidity support to securitisation schemes for assets originating on that bank's balance sheet. Our concern was with obligations – both legal and commercial – which might fall back on a bank if investors were disappointed in the performance of a securitisation vehicle with which a bank had close association. Banks argued that these guidelines effectively denied them the benefits of securitising assets.
The outcome of our review, almost finalised, will be guidelines which:
- lessen the previous constraints on securitisation, while
- more clearly specifying the capital requirements against risks to which a bank is exposed by the legal structure of a securitisation vehicle, and
- tightening disclosure requirements to ensure that investors understand they have no claim, beyond what is specified in that legal structure, against the bank itself if investments perform poorly.
Another topical policy area, both in Australia and internationally, is market risk. As noted earlier, it is intended that banks be obliged to hold capital against their market risks, but the measurement of those risks and capital requirements is complicated, more than for credit risk. Originally, the Basle Committee on Banking Supervision proposed that a standard model be imposed on all banks for these purposes. Banks and many supervisors (including ourselves) argued against this on the grounds that some banks already used very sophisticated models to manage market risk. Imposing on these banks the cost of parallel calculations with an inferior standard model for regulatory purposes made little sense.
When the Basle Committee issues revised guidelines later this year they will almost certainly include room for regulators to recognise banks' own systems, as long as they come up to certain minimum standards, in determining how much capital has to be held. This approach would require us to be knowledgeable about the relevant systems and the visits described above will be helpful for this purpose. We would not see this as getting us too close to banks' day-to-day operations, with all the attendant risks. Rather, it would be consistent with reassuring ourselves that the management and Boards of banks are taking their responsibilities seriously.
In supervision, the Reserve Bank's direct interests are very much confined to banks and their depositors. While banks have been allowed to expand into business such as life insurance and funds management through subsidiaries, this is only on the basis that:
- the bank's direct exposures through credit or liquidity support are strictly limited and subject to capital requirements, and
- the bank and its subsidiaries (or other associates) must be clearly differentiated for customers.
These are the main elements of our so-called ‘separation policy’ which has firm basis in the provisions of the Banking Act.
It would, however, be totally unrealistic for us to take no further interest in the subsidiaries of a bank since it is inconceivable that a bank would not be tainted, at least, by a serious problem in a close associate. More broadly, we are keenly concerned with financial conglomerates because we have an interest in the health and efficiency of the financial system as a whole.
Financial conglomerates have always been with us, but they are becoming more common and are tending to bring together a broader range of activities than in the past. Of the thirty largest financial entities in Australia, accounting for three quarters of financial system assets, virtually all are conglomerates.
For the Bank and other prudential supervisors, the main concerns about conglomerates are:
- they might be more difficult to manage than institutions with a more specialised focus;
- they can be less transparent than simpler institutions – for instance, assessment of financial strength can be complicated by intra-group shuffling;
- problems in one part of a group can be communicated directly or indirectly to otherwise-healthy parts – contagion risk or ‘guilt by association’;
- problems of transparency and contagion can be a particular worry when significant parts of a group are not overseen by any prudential supervisor or regulator; and
- the priorities of supervisors with responsibilities for different parts of a group might not coincide.
We are fortunate that Australian conglomerates have some features which lessen these concerns:
- most are engaged only in financial activities;
- these activities are mostly supervised, and only a small number of official agencies is involved; and
- most of our conglomerates are dominated by one member, usually a bank or an insurance company. For instance, of the 16 largest conglomerates which include a bank, the bank itself accounts for more than 75 per cent of total Australian assets in 11 cases. Similarly, in the three largest conglomerates not including a bank, a life insurance company has more than 75 per cent of total assets.
There are no international standards for supervising financial conglomerates but some commonsense principles are widely accepted:
- the need for transparency of structures, so that supervisors can track intra-group transactions and understand lines of management accountability;
- ‘fitness’ tests for dominant owners and senior management;
- that all significant parts of a conglomerate should be supervised by an agency with adequate authority; and
- the critical importance of co-operation among the regulators involved with different parts of a group, including willingness to exchange information.
Australia rates fairly well on these criteria. As I noted earlier, conglomerate structures are, in the main, relatively simple and well-covered by supervision while ownership tests for major institutional groups are quite stringent.
As to co-operation among regulators, we have the Council of Financial Supervisors, established in 1992. Its members are the Reserve Bank, the Insurance and Superannuation Commission, the Australian Financial Institutions Commission and the Australian Securities Commission. Together, these agencies have authority over institutions managing about 95 per cent of financial system assets.
Council to date has focussed on building effective liaison among its members to supplement the bilateral contacts which already existed. It has become a useful forum for discussion and co-ordination of supervision policy. It has also agreed on broad principles for supervising conglomerates. These principles, inter alia, commit Council members to liaising with each other when a problem affecting any entity in a group is judged likely to impact on other parts. The supervisor of the parent (or largest) entity in the conglomerate would co-ordinate – that is, take on something akin to a ‘lead regulator’ role.
Consistent with the principles, Council has settled guidelines for information-sharing among its members and is pursuing legislative changes to make this easier to do. Alongside these multilateral initiatives, pairs of Council members have reached understandings on information exchange and on supervision of particular institutions.
As I indicated earlier, most Australian conglomerates are presently headed by a substantial supervised financial institution. An important item on the Council's work program for this year is to review the pros and cons of conglomerates headed by special purpose holding companies. Council wants to assess, in particular, whether that structure makes it easier or more difficult to limit contagion problems and enhance transparency of intra-group dealings. A key question obviously is what level of supervision should apply to holding companies.
An issue for the Council's future attention is how best to go about assessing the overall health of a financial conglomerate. In this vein, an international group of senior regulators has been looking into the technical difficulties of assessing the capital adequacy of ‘bancassurance’ conglomerates. While it is true that some banking and insurance products are becoming more alike, the ‘blurring of differences’ between the two sectors is often overstated. The characteristic features of a bank balance sheet and an insurance company balance sheet remain very different, and the relevant supervisory techniques are correspondingly dissimilar.
In passing I should note that the Council's interest has not been confined to conglomerates. It is, for instance, overseeing a review of how far disclosure standards for similar products from different sorts of institutions might be brought into line. It is also looking to ensure that the various studies relating to derivatives are co-ordinated.
From time to time, interest is expressed in setting up a ‘mega-regulator’ to deal with the challenges posed by financial conglomerates. It is true that the current supervisory arrangements can look somewhat untidy. But that doesn't mean they are ineffective (and there is not much justification to pursue tidiness for its own sake). While historical experience is hardly conclusive, it is worth noting that none of the financial ‘basket cases’ in Australia over the past decade can be blamed in any degree on a lack of co-ordination among the supervisors of a financial conglomerate.
There is no strong support internationally for having one big regulator. And the handful of such agencies that do exist (in, for instance, Scandinavia and Canada) are organised into a banking unit, an insurance unit and so on, so that issues of co-ordination and information flows have still to be addressed. One large bureaucracy will not necessarily handle these any better than two or three smaller ones.
My impression is that the Council of Financial Supervisors is working well and dealing effectively with the current challenges. It is a clear advance on co-ordination arrangements in most other countries. Meanwhile, its flexibility and relative informality seem well-suited to this period of rapid change in our financial system. It follows, of course, that the form and role of the Council must be able to evolve, as the shape of Australia's financial system is remoulded further by the forces I described at the outset.