Macroprudential Analysis and Policy in the Australian
Financial Stability Framework – September 2012
4. System-wide View

The Australian agencies see a system-wide view as an essential part of effective prudential supervision, inextricable from the supervision of individual institutions.

APRA is the supervisor of financial institutions in Australia. In carrying out its duties, APRA takes an industry-wide, or systemic, perspective, as is consistent with its mandate. APRA makes its systemic mandate operational through a number of elements of its supervisory practice. APRA's risk-based approach subjects institutions that pose greater systemic risks to more intensive supervision, and potentially higher capital or other prudential requirements. APRA uses tools such as industry-wide stress tests, horizontal reviews and thematic analysis of emerging risks to inform its supervisory focus and actions. Assessments of systemic risks can motivate sector-wide prudential action, either of a supervisory or policy nature. APRA supervisors identify, monitor and analyse a wide range of market and financial developments and other environmental factors that may impact financial markets and make use of such information in the supervision of entities. Information and insights gained from a wide variety of sources are used.

Such a system-wide view involves looking both at the whole system as a single unit and at the way interactions of different parts of the system might feed back onto others. For example, decisions by lenders will affect their counterparties, and through the effect of credit supply on asset prices, will also affect others who are not their customers but own those assets. One of the ways these interactions can be traced is through system-wide or industry-wide stress testing (see Box A). Taking a system-wide view also involves a recognition that financial instability will occur long before the median or average member of a particular sector becomes distressed: risks manifest in the most vulnerable segments. Therefore it is important to analyse both aggregated and disaggregated data.

In addition it is important to draw out the underlying behaviours that are generating those data. These behaviours may point to potential vulnerabilities or future risks, for example those arising from over-exuberance, excessive risk-taking and rent-seeking behaviour. The observations of prudential supervisors at APRA about the attitudes and behaviours of executives and boards of supervised institutions can be a useful input to this analysis, complementary to the quantitative analysis done at both agencies.

The financial system involves a number of complexities and uncertainties, including: the diversity of entities in the system; the importance of feedbacks and interactions between those entities; and the role of unobservable characteristics and decisions, such as risk appetite and expectations. For this reason, the Australian authorities do not believe it is feasible or desirable to confine their analysis of the system to a single style of analysis or statistical model. Academic contributions to the understanding of financial systems continue to evolve; it makes sense to incorporate new developments and thinking over time, rather than restrict analysis to ideas that can be absorbed into a single framework. An implication of this view is that the agencies place little emphasis on detecting asset price misalignments based on the output of particular valuation models in isolation.

Box A: Stress testing

Stress testing is used by APRA as part of its regular supervisory activities. APRA's use of stress testing extends beyond the prudential requirements for regulated institutions to undertake their own stress testing. APRA conducts its own stress testing to better understand the vulnerabilities of regulated entities, industries and the financial system. These stress tests can involve cross-industry participation, so that the implications from a common, severe but plausible scenario can be explored. The information obtained from stress testing is used by APRA to consider actions that may be necessary to address any concerns it has.

APRA works with the RBA and the Reserve Bank of New Zealand (RBNZ) in the development of stress scenarios to be used in system-wide stress tests of banks. The involvement of the RBNZ reflects the dominant market presence of Australian-owned bank subsidiaries in New Zealand. Where relevant, the RBNZ also works with APRA on the analysis of regulated entity stress test submissions.

In recent times, APRA's approach to authorised deposit-taking institutions (ADI) stress testing has involved a two-phased process. In Phase 1, participating ADIs are asked to consider the financial and regulatory implications of a macroeconomic scenario by processing the stress through their own models and application of their judgement. In Phase 1 the participating ADIs are asked to consider the stress without application of any mitigating actions – in essence the ADI is on autopilot throughout the event and does not take action to reduce or avoid losses.

Detailed, standardised template information is sought from the participating ADIs to garner their perspective on the impact of the scenario on a range of different portfolio areas. In particular, granular information is gathered with regard to the behaviour (credit migration, probabilities of default and loss given default) of key credit portfolio segments. Information is also collected on trading and fair value outcomes, liquidity consequences and broad balance sheet, profit and loss and overall regulatory impacts.

It can be expected that portfolio differences result in varied outcomes in Phase 1 of the stress test. Differences in judgements and methodologies also contribute significantly to variations. To control for the variability of judgements and methodologies applied in Phase 1, APRA provides ADIs granular risk estimates (credit migrations, default rates, loss given default estimates etc) to be applied under a defined methodology in Phase 2. By controlling the risk inputs and the methodologies used APRA can better understand the consequences stemming from portfolio differences.

Phase 2 of the stress test essentially has two components. The first component of Phase 2 seeks from participating ADIs the regulatory and financial outcomes that result from application of the APRA stress methodology and risk estimates but before mitigating actions. The second component involves obtaining detailed information on the mitigating actions that ADIs would undertake to reduce the financial and regulatory consequences of the scenario.

In applying a two-phased process APRA believes value is obtained through:

  • observing the variability in regulated entity stress testing practices (and necessarily linked capabilities such as IT flexibility) with an aim of driving longer-term improvement;
  • gaining perspectives on the stress scenario judgements of regulated entities and, by implication, the maturity of their risk culture;
  • considering the preparedness of regulated entities to deal with adverse circumstances;
  • understanding the potential for regulated entities to apply mitigating actions that appear rational at an individual entity level but can, on a system-wide basis, result in deepening the stress (e.g. cutting credit growth); and
  • identifying entity and system vulnerabilities and initiating appropriate action.

As is the case with overseas regulators, APRA has stepped up its stress-testing activities since the global financial crisis. APRA's stress testing work, like most prudential supervisory activity, is typically undertaken away from the public gaze. Such confidentiality allows APRA to explore the boundaries of viability of regulated entities without creating unnecessary public concern.