The architecture of group supervision

John Trowbridge, Executive Member - Extract from Global Perspectives on Insurance Today

The effective supervision of financial services groups has become one of the more important regulatory challenges emerging from the global financial crisis (GFC) during the period 2007 to 2009. Such supervision was not well advanced beforehand and the GFC has emphasised the gap that it represents in the armoury of many prudential supervisors around the world.

Although most of the causes and effects of the GFC were within the banking industry, the insurance industry did gain considerable prominence due to the travails of AIG, the US-based insurer that owned a "rogue" unregulated subsidiary, AIG Financial Products, underwriter of credit default swaps. This subsidiary demonstrated the interconnectedness of the insurance industry with the remainder of the financial sector and would have brought down the whole AIG group had the US government not intervened.

This chapter explores the need for effective group supervision, the associated design questions and some of the issues that arise from efforts to implement group supervision. While it concentrates mainly on insurance, it also refers to the supervision of banking groups and of conglomerate financial groups.

Terminology

• "supervisor" and "regulator" can often be used interchangeably,

but

• regulation and supervision are distinctly different functions:
  • regulation represents the rules imposed on companies by the regulator, whereas
  • supervision relates to the activities designed to enforce the regulations.

For the purposes of this exposition –

• "scrutiny" is used whenever the need arises to refer to the supervisory function separately from the regulatory function.

Because the author is an Australian prudential supervisor, many of the experiences and examples in this chapter relate to Australia, where HIH Insurance, a major supplier of liability insurance, collapsed in 2001, causing considerable financial, economic and political disruption at the time. This event had major ramifications for prudential supervision in Australia as the 2003 HIH Royal Commission report[2] made far-reaching recommendations aimed at strengthening both regulation and supervision of Australian insurance. The recommendations included in respect of groups that –

"as a matter of high priority, the Australian Prudential Regulation Authority develop and promulgate a standard for the effective regulation of authorised insurers that operate as part of a corporate group" and that it "monitor the financial condition of corporate groups, including those with foreign operations." (HIH Royal Commission, Recommendations 38 and 39).

As a result, the Government legislated to give APRA the power to authorise non-operating holding companies (NOHCs) for general insurers in 2002[3] and life insurers in 2009[4]. APRA has subsequently responded by introducing a comprehensive approach to the supervision of general insurance groups and, at the time of writing, is in the process of extending the same approach to life insurance groups and cross sectoral groups.

Background - Insurance industry evolution 1970 to 2010

In the 1970s, both the life and general insurance industries were fragmented in many countries, often characterised by modest market shares for even the larger companies. For example Australia had more than 50 active general insurance companies and the top five companies had no more than 30% aggregate market share whereas by 2010 the top five companies wrote more than 90% of most market segments. The life insurance industry has consolidated in similar fashion.

Other features of the industry were -

• many companies had a home base in one country and numerous foreign subsidiaries or branches. This applied to life insurance, general insurance for direct business and reinsurance
• there were many composites (life and general)
• there were few countries where banking could exist within a life insurance, general insurance or composite insurance group
• most companies throughout the world operated through intermediaries rather than selling direct to the customer.

As well as being an industry that was fragmented locally and globally -

• most organisations had simple structures, even if they operated internationally
• the scope of regulation was limited and supervisory activities were at a low level
• the range of corporate activities was normally concentrated on underwriting insurance and doing nothing else other than managing an investment portfolio.

The life industry was fairly well regulated, based on the previous century's development in conjunction with the actuarial profession. As a result life companies were generally reasonably well managed financially, although in many markets there were mutuals which had built up big surpluses in earlier decades and which were therefore able to use this surplus to support the cost of new business and still maintain their financial soundness. Many companies wrote participating business in which all the risks were pooled and shared amongst the policyholders rather than exposing the shareholders.

It was rare to see life insurance companies over-reaching themselves, although there were two examples in Australia in the 1980s when aggressive growth and high expenses led to their demise. On the other hand general insurance companies were often plagued by the difficulties of managing volatile business and indeed many of the failures that occurred in the second half of the 20th century were the result of mismanagement of long tail business. This problem often caused governments to make many different arrangements, including movement of the statutory or compulsory classes (usually workers compensation or motor liability) between the public sector and private sector, sometimes more than once.

There tended to be excessive capital in many companies. Today that would be called lazy capital and, by comparison with today, one could probably say that their shareholders were undemanding. However, one mustn't ignore the prominence of mutuals, especially in the life insurance industry, which tended to have high capital ratios because of their lack of access to new capital. Note that earnings on the excess capital were often used to subsidise to some extent the remainder of the business, putting considerable competitive or price pressure on shareholder companies.

During the last 40 years there have been a number of failures, much consolidation within the industry by acquisition and merger, many demutualisations, and increasing emphasis on shareholder returns and the effective use of capital. Further, there has been an increasing variety of corporate structures and business models, together with some expansion in the range of corporate activities undertaken by insurance companies or insurance groups. In some cases these activities have been outside the insurance industry, usually but not always in the financial sector, in various other forms of business activity.

By 2010 we have a relatively small number of insurance institutions that are rather larger than the 1970s variety. They are geographically diverse but their size also represents a concentration of capital and in effect a concentration of risk. Offsetting this concentration to some extent is the diversification that occurs when portfolios are more widely spread. The industry is now less fragmented, more concentrated and -

• structures are more complex, financially and legally
• some international companies have large market shares in several jurisdictions and have branches and solo entities in many countries
• some companies have a diversity of activities, not only in direct insurance and reinsurance but in distribution ownership and, in some cases, involvement in other financial sector businesses
• solo regulation has developed strongly
• supervision is generally more intrusive and more effective.

Notwithstanding all these developments, one simplification has occurred: there are fewer insurance organisations writing both life insurance and general insurance. In 1970, many industry participants operated in both industries. While that continues to be the case in reinsurance, in direct insurance there have been many organisations that have withdrawn from either life insurance or general insurance in order to concentrate on just one arm of the industry.

In the same time period, the financial services industry generally has been liberalised. In 1970 there were few countries where insurers could engage in banking or banks could engage in insurance. The funds management industry did not really exist as a separate industry but proliferation and development of unit linked investment products in the last 30 years or so has spawned a large and relatively new funds management industry. As a result, the life insurance industry in 2010 is only part of the industry that it not just dominated but monopolised in 1970.

Regulation and supervision have advanced considerably in the same period. In the 1970s there were simple capital rules, no thought of group supervision, and limited attention was paid to risk management and governance.

The fundamental questions are the same today as they were in 1970, being about capital, liability measurement, how to underwrite effectively and how to manage underwriting risk, investments, pricing and profitability.

Nevertheless there were parallel advances for underwriting companies, investors and supervisors. Generally speaking -

• underwriters now
  • do better and more technical pricing
  • are more technical and more professional in reserving for liabilities
  • make widespread use of actuaries to support pricing, financial management and risk management
  • pay greater attention and give better responses to the interests and needs of both investors and regulators
• investors have become more discerning -
  • they want capital efficiency
  • they want an emphasis on underwriting, not on gaining extra profits from the investment of policyholder and shareholder funds
  • they want to see some stability of earnings and regular growth in earnings.
• supervisors have overseen a maturing of solo supervision and have also -
  • developed capital requirements that are more risk sensitive
  • introduced substantial risk management and governance requirements
  • upgraded the requirements for, and scrutiny of, both reinsurance and investments.

In the field of regulation and supervision, life insurance has evolved and strengthened. General insurance has evolved from a relatively low level of regulation compared to life insurance and banking, and in some cases from no prudential regulation at all. For example, in Australia there was no general insurance regulation until 1974[5] and in many European countries, reinsurance was still unregulated at the end of the 20th century[6]. Nevertheless in the last 40 years, general insurance has tended to follow the life insurance precedent of introducing actuarial control, particularly over claims liabilities and capital management, and also increasingly effective capital standards. It can also be said that solo supervision has matured.

Incidentally, one area that has not evolved very much in this period is accounting standards. There are still major differences in balance sheet structure and profit measurement in different jurisdictions.

The need for effective group supervision

During the evolution of the insurance industry as explained above, there has been a proliferation of corporate structures, business models and market developments. There are many kinds of group structures and more financial sector businesses operating within group structures.

Contagion risk exists within these corporate structures and experience is demonstrating that generally these risks are not as well understood as they might be, either by company boards and management or by supervisors.

There have been some well-publicised examples of life-threatening contagion within financial groups in the last decade and many less publicised examples as well. Three cases of interest that illustrate the consequences of such contagion risk are -

• HIH, an Australian general insurer that grew very quickly in the 1990s by aggressive growth of its portfolios and an ambitious local and international acquisition strategy, and then collapsed in 2001 as the country's largest ever corporate failure. Many policyholders lost part or all of their premiums and many claimants were not paid when the company failed.
• AMP, Australia's largest life insurer for more than a century, made a commercially successful acquisition of Pearl Assurance in the UK in the early 1990s. It subsequently demutualised and then made a leveraged acquisition of NPI in the UK, secured in part by the Pearl. The group saw a significant weakening of its capital position as a result of the UK stock market downturn of 2002, which saw NPI and Pearl both trying to call on the same capital. The group maintained the integrity of its Australian business only by divesting its UK businesses.
• AIG, the US insurer with more than 150 insurance subsidiaries, life and general, and a collection of unregulated subsidiaries including AIG Financial Products. Its losses during the GFC were very great and exceeded the parent's ability to finance it, partly because the assets of the insurance subsidiaries were quarantined by insurance supervisors around the world in order to protect policyholders. The US government lent AIG very large amounts and also injected equity to ensure AIG's survival pending divestments which, at the time of writing, are in the course of execution.

These cases illustrate some of the contagion problems which can occur in financial groups and which, in each case, were either not foreseen or not well managed at the time. For example -

HIH: the company, a liability specialist, had consistently engaged in under-reserving of its claims in the 1990s and as a result also underpriced its products, thereby generating illusory profits for several years. These profits, the apparently sound balance sheet and the company's low prices facilitated rapid growth of existing portfolios and expansion into new markets. When the inevitable crash came, it was bigger than anyone had foreseen because a myriad of intra-group transactions in the 1990s, followed by a range of questionable reinsurance and other transactions in the "death phase" as senior executives "gambled for resurrection", severely compounded the earlier weaknesses.

Poor financial management at all levels combined with unwise or unsound intra-group transactions were fatal.

• AMP: the Australian parent owned several life companies in the UK but the fateful transaction was the purchase of NPI in 1999. It was financed by some borrowings and supported by guarantees from Pearl Assurance, already owned by AMP and itself financially sound. However the leverage on capital inherent in both the NPI financing and the high allocation of shareholders funds to stock market investments caused near fatal capital losses when the UK stock market fell 43% in 2001 and 2002.
  Adverse market conditions damaged two subsidiaries within the group during an ambitious growth phase and, alongside some gearing and some intra-group guarantees, weakened the capital position of the group, forcing divestitures.
• AIG: the credit default swaps business, AIG Financial Products, was an unregulated business which had grown very quickly and which, when viewed with hindsight following the extraordinary conditions of the GFC, was badly under-capitalised and taking excessive risks. These matters were not visible to insurance supervisors at the time.
  Adverse market conditions caused losses in an under-capitalised and unregulated subsidiary which were beyond the capacity of the parent to support. The group’s collapse was prevented only by emergency government funding pending forced divestitures.

The general conclusions from these cases that are relevant to the management and the prudential supervision of corporate groups in the financial sector are –

1. contagion risk of various kinds can exist within corporate groups,
2. complex corporate structures and financial arrangements can mask some contagion risks,
3. boards and management of groups need to understand these contagion risks well and manage them effectively, and
4. prudential supervisors need to work out how to maximise the protection of policyholders and depositors against such risks.

The AIG case has particular interest and has caught the attention of insurance regulators around the world because it has demonstrated two critical features of insurance regulation as it applies in 2010 in most jurisdictions around the world –

• The first feature is the extraordinary strength of solo supervision: for more than 150 AIG subsidiaries in many different countries, not one policyholder was left stranded, notwithstanding the difficulties of the parent company.
  The explanation is that solo supervision of insurance companies is functioning well and the evidence is that the assets of the AIG insurance subsidiaries were successfully quarantined by regulators within those subsidiaries, out of reach of the parent. This is to the credit of insurance supervisors in the US and in many other countries.
• The second feature is the absence of effective group supervision (because one subsidiary, an unregulated one, was able to damage the group severely). That is the subject of this paper.

Hence solo supervision, i.e. supervision of insurers at a solo entity level, does work.

The challenge is to maintain the integrity of the solo supervision of insurers while also carrying out effective group supervision that protects the integrity of the group including all regulated and unregulated subsidiaries.

The AIG case demonstrated this need clearly but at the same time it and the other cases already cited demonstrate that the risk attributes or exposure to risk of groups as a whole cannot be ignored. The lessons from AIG should now lead to a range of regulatory initiatives that will include, in time –

• either more effective supervision of specialist insurance groups and cross-sectoral financial groups or restrictions on the scope of activities of specialist insurance groups, and
• for financial sector groups, greater emphasis on and scrutiny of intra-group transactions, capital management, governance and risk management practices.

Designing group supervision

The prudential goal of group supervision is to ensure that the group is financially sound and that group activities and inter-relationships do not adversely affect the financial soundness of the regulated entities within the group.

Expressed another way, the primary prudential aim of group supervision is to minimise contagion risk to regulated entities within a group from other group operations and thereby reduce the likelihood of failure of the group's regulated entities.

The supervisor‟s position is based around protection for policyholders from contagion risks within the group, and the three main such risks are:

• Inappropriate capital or ownership arrangements at group level (e.g. debt financing or, as is often the case with venture capital, short term ownership)
• Unregulated entities in the group that can compromise the solvency of the group.
• Intra-group transactions such as loans, guarantees, double leveraging, etc.

It is also relevant to note that the introduction of group supervision can be seen by financial groups as just another regulatory burden. In addition to improved risk management, however, there are three potential benefits that insurers can gain from group supervision -

• mobility of capital;
• recognition of diversification benefits; and
• reduced or streamlined regulatory requirements

It is useful to describe the current state of play in order to create a reference point for understanding design issues associated with group supervision.

What do we have today?

Supervisors -

• a comprehensive understanding of what needs to be done to protect policyholders in solo entities
• good regulation in most countries for the solo supervision of insurance companies
• regulations that are increasingly principles-based rather than prescriptive
• an increasing number of staff of regulatory agencies who are competent supervisors in dealing with individual companies.

Industry -

• the capability generally to meet regulatory requirements at solo level
• entrepreneurial instincts which often push the boundaries of existing business models and scope of business activities
• fewer mutuals, because many have demutualised
• insurance groups undertaking an increasingly wider range of non-insurance activities
• the existence of some financial conglomerates.

What do we not have today?

Supervisors -

• there is no comparability of balance sheet and profitability measurement and no uniformity of capital requirements for insurers across different jurisdictions group supervision is under-developed, and in many jurisdictions there is limited experience and understanding of the full ramifications of group supervision, and
• in many jurisdictions, there are insufficient legislative powers to undertake effective group supervision.

Industry -

• without prudential regulations in place, financial groups generally lack the discipline and incentive, and in some cases the skills and experience, to introduce high standards of governance, capital management and risk management at group level.

The "prudential hierarchy"

Parliament or government promulgates legislation whose primary role is usually to give supervisors the powers they need to make regulations and to enforce them. Legislation may also prescribe some aspects of the regulations.

Regulations are often designed or issued by the supervisor rather than the legislature. They generally have the full force of law but also contain many details that are usually more readily modified than legislation.

Supervision is the activity of trying to enforce regulations. Prudential supervision is preventative (before failure) rather than punitive (after failure), the latter approach being more common outside the prudentially regulated sector.

The primary conditions for successful prudential supervision, of both solo entities and groups, are –

• appropriate legislative powers
• suitable regulations, and
• effective supervision.

So how is group regulation done, what are the legislative pre-conditions, what is the nature of the regulations needed, how will the groups themselves and their subsidiaries be affected, and what can be achieved by supervisors?

Legislative powers

Adequate legislative powers are the starting point for supervisors.

Groups that are predominantly insurance groups or which are conglomerate groups that include one or more insurers are not, in most jurisdictions, being successfully supervised as groups. There are several reasons for this situation but it is a common feature of all too many jurisdictions that the supervisors do not have the legislative powers they need to undertake effective group supervision and therefore also lack the necessary supervisory tools. And these problems exist at two levels: in individual jurisdictions and across borders.

As is the case for solo entities, the supervisor needs to have the authority or the power to set standards for all aspects of capital, governance and risk management and also to enforce adherence to the standards.

Incidentally "all aspects" needs to include unfettered access to information and, in relation to capital, the construction of the company's balance sheet for prudential purposes. The supervisor needs to be able to allow or disallow particular assets or liabilities and to determine their value, sometimes referred to as applying "prudential filters" to the balance sheet. In other words, the supervisor needs to be able to modify the balance sheet in the published accounts, which meet accounting standards, so that each item in the balance sheet is measured appropriately for prudential purposes. This may mean, for example, writing down some or all forms of intangible assets and increasing the value of policyholder and claims liabilities.

If our reference point is a supervisor that has all of the legislative powers needed for the successful supervision of solo insurers (and banks), we can also assume that the supervisor has adequate powers to supervise those groups which have as their parent company a licensed operating insurance company or bank. We make this assumption because, by having the power to set standards for capital, governance and risk management of the parent, including the power to assess or reassess the values of any assets and liabilities of the parent (which would include in such a case all the subsidiaries, regulated and unregulated, local and foreign), the supervisor can be expected to have at his disposal all the requisite prudential tools for group supervision.

Hence the real test for group supervision arises where the group has as its parent a NOHC (non-operating holding company), or alternatively a company that is an operating company but operating in an industry or in a manner that falls outside the power or scope of the prudential supervisor.

Where there is a NOHC, the preferred position is that the supervisor has the power to apply to the NOHC the full suite of prudential standards and tools available for solo supervision. This would include the power to obtain through the NOHC any information that the supervisor might wish to have about any subsidiaries in the group. This information access is essential, for there are two kinds of relevant subsidiaries that are unregulated, namely companies whose business is not the underwriting of insurance policies or the acceptance of bank deposits, and insurers or banks which are regulated in a foreign jurisdiction and which, ipso facto, are not regulated by the group‟s home jurisdiction.

The actual position for NOHCs is less than ideal in most jurisdictions. For example –

• in Europe, NOHCs are not regulated but the regulators of solo entities have full access to information from upstream companies including a parent NOHC;
• in Switzerland, NOHCs themselves are not regulated but the supervisor has full information access and also the power to apply capital requirements, risk management and governance standards to the group as a whole and, as a consequence, is able to enforce group prudential requirements through the individual regulated entities.

In many other jurisdictions, the influence of supervisors over NOHCs, directly or indirectly is rather less than in the European Union and in Switzerland. By contrast, in Australia the supervisor has all of the requisite powers, thereby enabling the supervisor to introduce comprehensive group supervision.

Where the parent is a company that lies outside the reach of the prudential supervisor, there are two possibilities to be considered -

• One is the case of a local parent company which is an operating company and is unregulated (for example a funds manager, retailer, building company, etc.).
• The other is the case of a foreign parent company, usually an insurer (or bank), whether it be an operating or a non-operating company (note that a foreign parent licensed in its home jurisdiction is still an unregulated parent in the local jurisdiction).

In both cases a good solution is to require the financial group to insert a NOHC between the group and its parent, which the legislation also needs to permit. This is far more than for the sake of good order, because it allows the financial group to be identified and supervised as an integrated whole that is financially and legally separated from the unregulated parent.

Group regulation – insurance groups

If there is legislative power to apply the full suite of prudential standards to NOHCs, regulations relating to groups can be developed which enable supervisors to deal with the potential contagion risks within groups.

Supervising insurance groups on a consolidated basis is a viable approach for understanding and minimising contagion risk within such groups. The capital base of a specialist insurance group (or specialist banking group) can be assessed on the basis that the group is treated as if it were a single consolidated entity. This works because in a single industry the capital requirements of each insurer (or bank) in the group can be added together to determine the group capital requirement, to be compared with the group capital base derived from the consolidated accounts.

In some insurance jurisdictions, a single industry group means general insurance only or life insurance only, so that a group containing both would be classed as a conglomerate group (see next section below). Consolidation will not always work for conglomerate groups and also will not work if the supervisor does not have full regulatory power over a parent NOHC.

As with all prudential regulation, there are many aspects to be catered for in standards. Instead of explaining the possibilities abstractly, it is expedient to explain some of the more significant aspects by reference to the group-specific requirements that APRA has introduced for general insurers licensed in Australia. They are –

• Definition of a group: an insurance group is defined so as to include all entities in the group conducting insurance and related business, both locally and internationally. The group contains all insurance subsidiaries (both domestic and international) and any other controlled entities integral to its insurance business including related service entities.
• Non-consolidated subsidiaries: Individual subsidiaries that are not associated with the insurance business can be deconsolidated for group supervision purposes. In such cases, the value of their equity exposure is deducted from the group's capital base.
• Capital upgrades: when determining the capital base of the group, individual components of capital measured in a subsidiary, regulated or not, must not be upgraded to a higher category of capital when included in the measurement of the group's capital base (for example, debt in the parent cannot be treated as equity in a subsidiary).
• Intra-group capital transactions: any component of capital arising from intra-group transactions may be excluded from an insurance group‟s capital base if that component is assessed as not contributing to the financial strength of the group.
• Unregulated subsidiaries that are under-capitalised: where a subsidiary within the group is under-capitalised, the parent may be required to deduct from its capital base an amount to cover the deficiency.

Effective treatment of unregulated subsidiaries is important, as the AIG case has demonstrated. APRA deals with unregulated subsidiaries in two steps:

- information on the subsidiary has to be made available to the supervisor on request; and

- if APRA believes it is under-capitalised or otherwise a threat to the financial strength of the group, the parent is obliged to hold enough extra capital to offset the under-capitalisation.

Group regulation – conglomerate groups

Many aspects of the regulation of specialist insurance (or banking) groups and conglomerate groups are the same. Those features which are different, however, are so different that full conglomerate supervision is yet to be introduced in any jurisdiction. The jurisdiction that is closest to its introduction is probably Australia where, at the time of writing, APRA has released a discussion paper with comprehensive proposals[7]. Subject to consultation, full conglomerate supervision is expected to be introduced in 2012.

The complications of conglomerate groups relative to specialist insurance groups or banking groups revolve around matters of capital such as –

• inability to consolidate, in relation to both the group‟s actual capital and its required capital, consequent upon the different nature of the capital requirements in different industries;
• the need to identify and deal with unregulated subsidiaries that are material such that it is not appropriate simply to deconsolidate them or treat them as excised from the group -
  • it is necessary to ensure that no such subsidiary is under-capitalised, now or in the future and
  • it may also be appropriate to recognise surplus capital if any exists in a material unregulated subsidiary.
• the complications around quality of capital (Tier 2 capital and hybrid Tier 1 capital) when two different industries are involved; and
• capital management which, as a process and as a board responsibility, is more demanding and more complex than for solo entities or specialist groups.

Each of these problems is soluble, however, and the Australian approach is illustrative. APRA has proposed –

• in lieu of consolidation, a "building block" or "legal entity aggregation" approach for required capital and either an aggregation or consolidation approach for actual capital;
• for material unregulated subsidiaries, that groups undertake their own assessment of capital adequacy across the group and then subject that assessment to scrutiny by APRA;
• to deal with quality of capital, that only equity-equivalent (Tier 1) capital be recognised for group capital purposes, to be compared to a tier 1 equivalent required capital, and
• on capital management, that the board make its own internal assessment of capital adequacy, including target levels of surplus or capital for the group, and also undertake a "transferability assessment" regarding the disposition of surplus capital within the group.

Limited legislative and regulatory powers

In jurisdictions where the ideal set of legislative powers is not granted to the supervisor, or where regulatory requirements are limited by legislative or other factors outside the control of the supervisor, some trade-offs and compromises will be needed pending improved powers.

Group supervision entails more risk for regulators, for not doing it effectively or getting it wrong may be worse than not doing it at all. So regulators and legislators need to work out how far they are prepared to go -

• do they restrict the activities of the group (for example, to insurance businesses only) in order to facilitate and ensure effective group supervision, or
• do they design group regulations to deal with companies with multiple subsidiaries in one or more industries, i.e. avoid restricting the business activities of companies that engage in insurance while simultaneously protecting the interests of policyholders?

This latter alternative sounds simple and of course it supports entrepreneurial and other activities that work in the direction of industry development, whereas the former would constrain the business activities of insurance companies and limit the evolution of insurance groups and other cross-sectoral financial groups.

The evolution of financial services, the expectations of most communities and the ambitions of most governments will lead to the latter direction, i.e. regulation of groups with multiple subsidiaries including non-insurance activities that operate alongside insurance activities. It is therefore imperative that regulators find a way of going down this path. The challenge for regulators is to continue to protect the interests of policyholders in regulated insurers without contamination or contagion from other parts of the group for as long as it takes to introduce an effective group supervision regime.

Note that the matter is complicated by the existence of both branches and subsidiaries within many group structures. Effective regulation of branches that are in a different country from the company's head office is not a straightforward matter.

Supervision or scrutiny of groups

One may ask: why is group supervision so difficult and how different is it from solo entity supervision? The general answer in 2010 is that, in most jurisdictions, there is a lack of experience and a lack of precedent for doing it. So before looking at some of the details, it is worth revisiting some of the principles for effective prudential supervision or scrutiny. These principles apply equally to solo entities and groups.

Once the regulations are in place, scrutiny of a group is not very different from scrutiny of a solo entity, just more demanding. Clearly also it is easier to achieve in cases where the group is prudentially supervised by only one agency. In many jurisdictions there is only one agency and one set of regulations applicable but in some there are several. Fragmentation of regulation and supervision is problematical, as demonstrated by the AIG case, where the group supervisor, OTS, was separate from the New York Insurance Commissioner and other state commissioners responsible for US insurance subsidiaries of AIG.

Furthermore, where groups have operations in more than one jurisdiction, effective supervision necessarily involves some level of recognition and/or interaction of supervisors across borders.

Because of the importance of effective supervision or scrutiny, and the fact that the GFC brought to the surface supervisory limitations in some jurisdictions, it is worth describing here some of the characteristics of effective supervision. One very good source is the HIH Royal Commission, which in 2003 included the following recommendations for APRA –

• "develop a more sceptical, questioning and, where necessary, aggressive approach to its prudential supervision of general insurers. Consultation, inquiry and constructive dialogue should be balanced by firmness in its requirements and a preparedness to enforce compliance with applicable standards.
• "develop and review processes, guidelines and training to assist its staff in considering the appropriate approach to take towards supervised entities in different situations.
• "develop systems to encourage its staff and management continually to question their assumptions, views and conclusions about the financial viability of supervised entities, particularly on the receipt of new information about an entity.
• "develop an internal system for tracking all relevant information concerning regulated entities.
• "develop mechanisms for investigating the reinsurance arrangements of authorised general insurers on a random but frequent basis."

(HIH Royal Commission, Recommendations 26 to 30).

These recommendations give depth to a more abbreviated but consistent principle enunciated by the Financial Stability Board in its 2009 "Principles for Sound Compensation Practices" regarding effective supervisory oversight –

"Supervisory review … must be rigorous and sustained, and deficiencies must be addressed promptly with supervisory action."

Note also the importance of the primary governance principle for effective supervision, namely that the supervisor be able to hold the board of a regulated entity (solo or group parent) directly accountable for meeting the supervisor's standards and requirements.

Living up to these standards may not be easy, but failing to live up to them almost certainly increases both institutional and systemic risk, independently of the quality and scope of regulations.

The accountability generated by high standards of scrutiny will always be worth more prudentially than the introduction of, for example, higher capital requirements. That is emerging as one of the more important prudential lessons of the GFC.

Some lessons learned about group supervision

Australia has had group regulation of banks since 1999. It has been a relatively simple regime to supervise, however, because until recently all banking groups have had their major operating company as the parent company within the group. As a result, most of the business of the group was done in the parent company and all the other businesses, whether banking or not, were subsidiaries of the operating company.

Nevertheless there is one overriding lesson from the Australian experience with larger banking groups. It is that the measurement of actual capital is a very difficult exercise and nothing can be taken for granted.

The Australian experience in introducing group supervision for general insurers in 2009 is instructive. The market is not large but it has about 130 authorised solo insurers and some 30 groups including several that are "one on one", i.e. the group comprises one operating insurer owned by a non-operating holding company. APRA decided that every solo insurer not owned directly by another insurer would need to be subject to group supervision if there is no upstream entity, i.e. in all cases except where the insurer is owned by a listed company (which by definition has multiple shareholders and dispersed ownership) or is a mutual (i.e. owned directly by its policyholders) or is owned directly by a foreign parent that is an established insurer. If the insurer is privately owned by a small number of owners, APRA may require the establishment of a NOHC in order to have a clear view of the capital structure.

The reasons for this "one on one" approach are best explained by some examples -

• APRA has discovered that in some groups the solo entity was equity funded by a parent that was itself debt funded. This is an unsatisfactory position which needs to be rectified by the parent restructuring its capital. In other words APRA has rejected the "ring fencing" approach which, not surprisingly, is widely advocated by the industry but which in principle is unacceptable prudentially because it usually means forgoing group supervision in favour of solo supervision only.
• Some groups have complicated legal structures with various corporate goals such as debt funding, moving profits or costs to other entities in the group for tax or other purposes, etc. In some such cases the best solution for group supervision is to restructure the group. It is highly desirable that in this situation the regulator has the power to oblige the group to restructure. One of the principles here is that capital cannot be upgraded from debt to equity or, expressed another way, the capital in the insurance business needs to be totally unencumbered, so that it is available to absorb losses, with no recourse of any kind being available to the owners.
• Some groups sit inside a larger group that undertakes other activities that are linked, perhaps only loosely, to the insurer's activities. The prudential supervisor is interested in the integrity of the capital position of the insurer and therefore in the corporate and financial integrity of the group as a whole. Some insurers have a parent with unregulated subsidiaries that are reliant on the parent for their solvency and which are also suppliers of services to the insurer. In these situations there may be intra-group debts, guarantees or other transactions which need to be understood.

Insurers typically respond in cases like these by proposing a stronger "ring fence" for the insurer. APRA can be expected to reject this argument on the grounds that it wishes to be privy at all times to all material intra-group transactions, present and future, because potential contagion risk within groups needs to be subject to scrutiny.

The overall effect of APRA's introduction of group supervision for general insurance has been a "clean-up" of the corporate structures and capital structures across the industry, as well as reduced risk across the industry.

To summarise the situation, there have been several important learnings from APRA‟s experience in introducing full group supervision for general insurance, which commenced only in 2009, a year of implementation for both the industry and the supervisor.

A surprisingly large number of unforeseen implementation issues arose, with many different possible legal structures exhibiting in some cases complexity around capital, difficulties around governance and increased risk across the group.

Noting that all general insurance groups in Australia are now headed by authorised NOHCs, the main observations following the implementation are -

1. The pre-existing group structures, both legal and financial, which had not previously been visible to the regulator, showed an astonishing variety.
2. The legal structures of some groups were unsuited to group supervision and needed to be modified; some groups needed a degree of formal restructuring and others needed to do some serious thinking about governance and capital management.
3. Restructuring of the group can be beneficial for the group as well as for the supervisor -

• restructuring is needed only if it is complexity in the group that creates difficulty or lack of clarity around one or more of governance, capital structures, financial management and risk management
• experience shows that simplifying the legal structure to overcome these difficulties and to bring clarity and transparency is often welcomed by the board, for it improves the board‟s ability to do its job effectively just as it benefits the supervisor, i.e. both the board and the supervisor find that greater transparency and greater simplicity of the legal structure and financial structure reduce risk and increase the ability of the board and the management team to understand and manage the business as a whole.

1. Because of (2) and (3) above, the very introduction of group supervision sometimes has an immediate risk reduction consequence for the group.

If there is one overarching observation that can be made from the APRA experience, recognising that APRA has the benefit of adequate legislative powers to introduce comprehensive group supervision, it is that -

Applying carefully a set of principles-based group regulations aimed at protecting policyholders against contagion risk within the group, supported by effective scrutiny by the supervisor, can achieve the goal of applying successful supervision to a wide variety of business models, legal structures and financial arrangements within groups.

Footnotes

1. This is the author's original manuscript before being edited to appear in Global Perspectives on Insurance Today (editors Cecelia Kempler, Michel Flamée, Charles Yang, Paul Windels) to be published in 2010 by Palgrave Macmillan.
2. See the testimony of Baxter (2010) to the "Committee on Government Oversight and Reform; U.S. House of Representatives", for a summary of the near collapse of AIG and the Federal Reserve's intervention to prevent the group's collapse.
3. The HIH Royal Commission was established to investigate the 2001 collapse of HIH, which represented the largest corporate failure in Australian history. The HIH Royal Commission delivered a three-volume report to the Australian Parliament on 16 April 2003 detailing extensive recommendations for improving regulation and supervision of Australian insurance companies.
4. General Insurance Reform Act 2001 (Cth) amended the Insurance Act 1973 (Cth) and came into effect from 30 June 2002.
5. Financial Sector Legislation Amendment (Enhancing Supervision and Enforcement) Act 2009 (Cth).
6. Regulatory power was given under the Insurance Act 1973 (Cth) to the Insurance Commissioner, with regulatory power ultimately transferred to APRA in 1998.
7. Policy developments have been made by European Commission (EU) in the form of Directive 2005/68/EC (16 November 2005).
8. APRA (2010).