Address by Deputy Governor Matthew Elderfield to the European Insurance Forum 2013

09 May 2013 Speech

It is a great pleasure to be invited here to speak at the European Insurance Forum for another year. My topic, unsurprisingly, will be the regulatory agenda facing the insurance industry. I approach this, however, with an appreciation that for many in the industry this regulatory agenda – at a European level at least – has a high degree of uncertainty, probably more so than when I last spoke here. This is, of course, due to the difficulties facing final agreement on the Solvency II package of directives and regulations, so let me in fact start there, with a discussion of the prospects of Solvency II. And let me introduce a theme for these remarks: that while the exact implementation timeline of the directive is uncertain, frustratingly so for both authorities and industry, European regulators are nevertheless strongly committed to using this intervening period to prepare for the directive and also to make sure that the debate on the exact shape of future European standards does not distract from some pressing supervisory issues.

The drawn out process of Solvency II negotiation and implementation has perhaps naturally resulted in a combination of fatigue and exasperation, with the high costs of preparation compounding concerns, and indeed with those costs being multiplied by the lack of a clear timeline and all the challenges that poses for cost effective project management; more of that in a moment. However, it is important to remind ourselves of the compelling reasons that led to the Solvency II project and the need for reform of the existing regulatory framework for insurance. To be blunt: it is unacceptable that the common regulatory framework for insurance in Europe in the 21st-century is not risk-based and only takes account, very crudely, of one side of the balance sheet. The European Union urgently needs a new regulatory standard which differentiates solvency charges based on the inherent risk of different lines of business and which provides incentives for enhanced risk management. It urgently needs a framework that takes account of asset risks in an insurance company. It urgently needs a framework that encourages better governance and management of risk. And it urgently needs a framework that provides better disclosure to market participants – on a consistent basis – of the health of insurance companies.

Solvency II, imperfect though it is, is designed to do all that. And in the absence of Solvency II, we not only have the problem of a protracted period of living with the greater imperfections of the current regulatory framework, but we also have the serious risk of fragmentation and extra cost to industry and consumers if each regulator in each jurisdiction goes off and does its own thing. Solvency II does indeed have its imperfections: I have had long-standing concern regarding the potential for excessive reduction of solvency buffers due to over optimistic calibration of internal models, for example, and I would like share a few thoughts on this in a moment. Also, while a more complex framework is almost certainly inevitable as a result of the need to develop a risk-based system, this complexity has clearly gone too far in some areas - and makes implementation very difficult for smaller companies. All that said, in my view, the benefits of moving ahead with Solvency II outweigh the costs and risks of retaining the existing EU framework and the development of a myriad of competing national regimes. That way means fragmentation of the single market and, in the long run, increased costs for international insurance companies and their customers.

As you know, there is a protracted delay in concluding negotiations on the Omnibus II directive, which is the last piece of the legislative jigsaw before Solvency II itself comes into force and with it the necessary pieces of level II technical standards and guidance. This protracted delay is a source of considerable concern to regulators and industry alike. It is disappointing that the political process has in effect stalled and that the entire framework is essentially stuck on one issue, albeit a very important one. The open issue, as you know, is the treatment of long-term guarantees. In this area, the political institutions of the European process, namely the Council, the European Parliament and the Commission have reached an impasse and have commissioned a quantitative impact study on the issue. I'm grateful for those Irish companies that have taken part in the exercise. But frankly, at the heart of the issue are strongly held views which will require a political compromise. The results of the quantitative impact study will not automatically provide an answer to this.

The heart of the issue is whether long-term guarantee products should be subject to the same market consistent evaluation framework as other insurance liabilities under the directive. After considerable debate there is acceptance, but by no means universal acceptance, of a compromise whereby certain annuity products that cannot be exited early and which have clearly defined and matched assets can effectively be carved out from the full rigour of market consistent valuation, by allowing a matched premium adjustment. However, more contentious is the question of whether this more liberal treatment should be extended to a wider range of long-term guarantee products where asset matching is less clear and, crucially, the ability to exit a policy is permitted. The advocates of a wider application of the matched principal framework argue that the contractual ability to surrender a policy should not be the only concern and that companies can model the likely behaviour of policyholders. Detractors worry that early termination of policies can crystallise impaired market values in matching assets. Underlying this debate is the suspicion that if some insurers apply a market consistent valuation approach to the back book of existing guarantees in current market conditions then this will lead to severe strains on solvency levels.

It is important that a sensible solution be found to this issue if we are to get Solvency II concluded. The treatment of long-term guarantees is however an issue of considerable prudential concern, as will be evident when I shortly explain the work of EIOPA on long-term low interest rates. Nevertheless, my personal view is that prudent insurance regulators could live with a more flexible approach to allow matched premium adjustments for a wider range of long-term guarantee products on three conditions. First, that the regime would only apply to back books and would therefore be transitional in nature. Second, that insurance companies be required to make a pillar III disclosure setting out their solvency position both with and without the matched principal adjustment. And third, that national supervisory authorities have flexibility to impose supplementary pillar II charges to take account of inadequate solvency buffers on a case-by-case basis. This framework would, I believe, allow a transition period for back book guarantee products while also providing a clear line of sight as to the solvency position of individual firms alongside the necessary supervisory tools to deal with particular problem cases.

One way or another it is important that a balanced compromise is found and the directive package can be brought to a swift and successful conclusion. With European elections looming next year, it is important that this process concludes in the autumn at the latest. Like those in the industry, the regulators around the EIOPA table find the delay in concluding the political negotiations frustrating. It is important that we have a credible new timetable for implementation. While a significant portion of the costs of Solvency II implementation involves a necessary and welcome upgrading of risk management systems and techniques, it is equally clear that the level of cost has been exacerbated by the uncertainties in the implementation timeline and the delays in the political process. These unnecessary cost overruns are insupportable. This autumn, the political process has one further chance to put this right.

When it became clear that the Solvency II process had become temporarily stuck towards the end of last year, the Central Bank engaged in a process of consultation with the principal insurance trade bodies in Ireland, including DIMA, the Irish Insurance Federation and Financial Services Ireland. We had a very frank and open discussion around the practical challenges caused by the delay in the directive and sought input from industry as to how to navigate through these issues. I was very pleased by the constructive engagement we had from industry at the time. There was quickly a meeting of minds on a central proposition: That it did not make sense for the Central Bank to move ahead in isolation on aspects of Solvency II implementation but rather that we should support the European-wide effort to continue to progress Solvency II. That has indeed been our approach and in fact is very similar to the approach taken by other regulators around the EIOPA table. There is a common and welcome view that it would be counter-productive for each authority to head off in its own direction to fill in the gap while Solvency II is concluded. Instead, there was a strong consensus in favour of actions at a European level to continue the preparations for the package of directives. This has, as you will know, manifested itself in EIOPA's initiative to publish interim guidelines to help prepare for the directive, so let me take a little time to describe these and the Central Bank’s approach to their implementation.

EIOPA has on 27 March published a consultation paper setting out its proposals for interim guidelines covering aspects of the Solvency II package relating to system of governance, forward looking assessment of the undertaking’s own risks (based on ORSA principles), pre-applications for internal models and submission of information to national competent authorities. This package is an important initiative in a few respects. It provides a foundation for the ultimate implementation of the Solvency II framework and ensures effective preparation of industry and regulator alike. Much of the Solvency II framework is already clear, so it is possible and indeed prudent to begin this preparation already, even if there are important loose ends to be resolved in the political process. In taking this action, EIOPA is also crucially maintaining momentum in the overall Solvency II implementation process and is helping "fill the gap" created by the hiatus in the political negotiation - an otherwise uncertain period where national authorities could have headed off in different directions. In this respect, I think the EIOPA initiative should be strongly welcomed by regulators and industry alike and I would like to acknowledge the leadership that Gabriel Bernardino, EIOPA's chairman, Carlos Montalvo, EIOPA's Executive Director, and their team have shown. It would be easy to sit back and wait for the political process to restart, but this poses significant risk of drift and divergence in the single market.

I would hope that the implementation of the system of governance provisions and own risk assessment of the interim guidelines will be manageable and useful transitional steps towards full Solvency II adoption.

There are also proposed interim guidelines concerning Pillar 3 reporting by insurance companies. The guidelines suggest a one-year period of parallel reporting, to help prepare for Solvency II and effectively test the quality of reporting arrangements in firms. Certain quarterly reporting obligations would also be phased in during this period. Helpfully, the guidelines provide various exemptions from reporting on proportionality grounds for smaller insurers. The one-year period of parallel reporting seeks to strike a balance between the goal of adequate preparation (where regulators typically see early problems with reporting for new regimes and therefore the need for a period of debugging) as opposed to the cost of dual requirements. The Central Bank is sensitive to this issue and the need to strike a careful balance. We will expect best efforts, especially from the High Impact insurers, but understand that there may be very good reasons for firms not being fully compliant during this period as insurers prepare for all the aspects of Solvency II. The use of statutory powers including sanctions will only be considered in extreme cases and as a last resort.

Indeed, this will generally be our approach to the implementation of the interim guidelines as a whole. We will certainly take the approach in this interim period of working collaboratively with industry on implementation. We will adopt a proportionate approach and our thinking will be informed by our existing risk-based supervisory framework. We expect firms to make their best efforts and show a willingness to respond to supervisory feedback on their implementation plans and content, and be transparent and up front with us in terms of the challenges they face with full implementation. As in reporting, the use of statutory powers including sanctions will only be considered in extreme cases and as a last resort. We will elaborate further on our approach to the implementation of the EIOPA guidelines at an industry briefing on May 24th, which no doubt many of you here plan to attend.

The interim guidelines also speak to the arrangements for continuing internal model assessment and preparation. It is not practical to try to bring the model approval process to an early conclusion with the final Pillar 1 provisions still uncertain and the full legal framework incomplete. Nor is it sensible to do a hard stop on model approval work, to be recommenced at a later date, for fear of losing momentum and later going over old ground. Instead, our approach is to re-programme the level and intensity of engagement on internal models with firms in order to seek to maintain progress. This is not ideal, in terms of protracting the overall process but should allow better development of models against regulatory standards in the interim period and also a less intensive level of engagement at the end of the process.

Stepping back from the logistics of internal model approval programme management, there is still an important broader policy question in this area. As I have discussed at previous forums, the Central Bank (in common with a number of other supervisors) is concerned to ensure that the design of models enables a prudent recognition of diversification effects within insurance company balance sheets, but does not go so far as to erode solvency buffers through over-optimistic correlation assumptions that do not stand up in times of stress. This is partly a design weakness in the text of Solvency II, as the directive does not impose specific constraints on correlation and diversification recognition in internal models per se, but requires this to be thrashed out between firm and supervisor in the approval process. (In contrast, the banking regulatory framework provides such hard, binding constraints on diversification in internal models.) As I have explained previously, I'm concerned that this case-by-case approach is vulnerable in a number of ways. Supervisory quants are at risk of being outgunned by industry quants in the minutiae of a correlation debate. Supervisors are exposed to the weakest link of an overly generous approval in some perhaps under-resourced jurisdiction which sets the bar too low and creates competitive equality pressures.

Fundamentally, what is needed is a broad agreement on the outer bounds of acceptable levels of diversification in the model approval process. One approach being explored actively in another jurisdiction is that of setting a floor at some proportion of the MCR derived under Solvency II, when agreeing the SCR output from an internal model. Pending active development at a European level of an agreed approach to ensure a level playing field, the Central Bank’s message to Irish firms is to take a conservative approach to the recognition of diversification and to think very hard about how well this will hold up in extreme tail event stress scenarios.

While the interim guidelines covering models, system of governance and reporting are at the top of EIOPA’s agenda in terms of standard-setting, the key supervisory issue at the top of the EIOPA list of concerns is the impact of long-term low interest rates on the insurance sector. One of the principal instruments deployed by monetary policy authorities in combating the current financial crisis has, of course, been the policy of maintaining low interest rates for a sustained period of time in light of price stability and in order to address economic weakness. However, this has the potential to have adverse side-effects for insurers. Of greatest concern to EIOPA, and indeed the Central Bank, is the potential adverse impact on life insurers with long-term guarantees to policyholders which exceed the available yield on assets in current market conditions. We need only look back to the Japanese experience of the 1990s to see the severe damage that this can do. However, general insurers are not immune from the side-effects: low yielding investment portfolios expose uncompetitive combined ratios, and therefore put extreme pressure on underwriting performance at the time of continuing soft markets.

These considerations have led EIOPA to recently publish an opinion setting out its concerns on the prudential implications of a sustained period of low interest rates. This encourages supervisors to take a range of measures to monitor and tackle the potential risks that are building up. This issue is relevant to the Irish market, especially because of the concentration of variable annuity writers that are based in Ireland and operate internationally.

To my mind, the issues for this particular business can be conceptually divided between a front book and a back book problem. For newly written business, it is important that firms are adjusting their product design and improving their risk management capabilities. We are largely seeing this happening already for the Irish based VA writers. And our regulatory actions in the past few years have had this in mind. For example, the Central Bank tightened solvency requirements on newly written variable annuity business at the start of 2011 and has also developed a more intense supervisory program of reviewing hedging efficacy at VA writers. We now have a good understanding of dynamic hedging operations and conduct quarterly reviews of hedging efficiency with firms. There has been good engagement with firms as part of this process, for which we are grateful. While the largest international VA writers are taking measures to adapt product design and pricing in order to de-risk the front book, it is however not yet clear whether this is a widespread enough development across European markets for guarantee products more generally.

While these measures help deal with front book problems (and would therefore accommodate a more conservative and market consistent valuation approach under Solvency II), it is clear that back books of VA and guaranteed business are likely exposed to a long period of low interest rates. These back book risks could also be amplified by adverse trends in terms of surrenders and longevity, causing a build-up of solvency problems. The Central Bank believes that the most effective way to assess these back book risks is through robust stress testing and we plan to roll out such stress tests with those firms that have potentially vulnerable back books. We would expect management and boards to already be thinking hard about these risks. They should be looking beyond their current non-stressed solvency buffers and contemplate what management actions may be necessary as a prudent measure for such stress scenarios. This will clearly be a point of discussion following our own supervisory exercises in this area.

This Central Bank initiative on stress testing fits clearly within the EIOPA opinion on long-term low interest rates, which calls for such action by supervisors. On the purely domestic front, there is one further development I would like to share with you and then I will say a few final words about our early experience of implementing PRISM, our new risk-based supervisory framework at the Central Bank.

The most disruptive development in the domestic Irish insurance market in the past 3 years has, of course, been the failure of Quinn Insurance Limited following serious persistent breaches of solvency requirements. There has been much said and written about Quinn Insurance already and I do not propose to look backwards and review the specifics of the case. However, one clear supervisory concern relates to the emergence of the significant deficit in the company’s reserves despite unqualified audit and actuarial reports, which in turn has required a significant call on the Insurance Compensation Fund. While we accept that Quinn Insurance is not a proxy for the insurance industry as a whole, lessons must be learned by the Central Bank and the industry in relation to the assessment of the adequacy of both reserving and pricing. Both areas have come under additional supervisory scrutiny from the Central Bank during 2012 and 2013 and firms can expect this to continue further in the future with the introduction of specific governance and oversight requirements. This new approach to assessing reserving adequacy will be closely integrated with our risk-based supervisory framework, which we call PRISM. This has now been in place for insurers for more than a year, so let me take an opportunity for a few reflections.

Our risk-based supervisory framework starts with an allocation of resources based on impact: that is, for insurers, the impact of failure on policy holders. This is used to calibrate the level of engagement with firms and the frequency and intensity of risk assessment. It prompts front line supervisors to conduct a systematic assessment of a range of risk categories for higher impact firms, leading to a formal scoring which is communicated to the firm in question, along with a risk mitigation plan setting out a "to do" list of actions to address areas of concern. Not all issues identified in an assessment merit further action - this depends on their magnitude and their probability of crystallising - and firms have an opportunity to both correct factual errors in a risk assessment and suggest alternative mitigation actions.

Looking back over more than a year's operation of PRISM for insurance companies, it's clear that we are still some way from fully embedding our new approach and further refinements and fine tuning should take place. But, I'm pleased that the framework is delivering in three critical respects. First, that it is indeed delivering its core objective of requiring front line supervisors to think critically about the risks present in insurance firms and to do so in a structured and consistent manner. This is therefore flushing out issues and concerns, improving standards of risk management and compliance.

Second, the process is resulting in mitigation plans that seek to tackle the risks identified in a time-bound and decisive way. One of the criticisms of supervisors in the pre-crisis period relates to a tendency to keep studying a risk, kicking the can down the road, rather than being more decisive in terms of actions. By and large, the mitigation plans are designed to do this, although clearly we will need to go through a cycle of assessments to see how well they have achieved their desired outcome. I would hope these plans provide clarity to a board or senior management in terms of where the Central Bank is coming from and can be part of a "no surprises" approach.

Thirdly, the PRISM framework includes an important role for risk governance panels. These are challenge and quality assurance forums, chaired by more senior supervisors and involving experienced risk advisers from industry or regulatory backgrounds. I'm pleased by how these are working, in terms of providing challenge to supervisory teams, both to be more robust in some cases, but also more proportionate and focused in others.

It is, however, important that the Central Bank is prepared to fine tune and continuously improve our supervisory framework. Part of this will mean a continuing investment in staff training and development: our staffing numbers have levelled off (and indeed will be somewhat lower than we projected necessary a few years ago), but we know we need to work harder to improve our understanding of business models, market developments and financial analysis, to balance out more established audit and compliance skill sets. We also plan to take stock of PRISM implementation before too long.

Good supervision involves a commitment to continuous improvement, keeping up with best practice supervisory techniques and also improving skills and industry knowledge in front line supervisors. The last few years has seen, I believe, a significant upgrading in supervisory approach and capability in Ireland. As these are probably my final public remarks on the subject while at the Central Bank, I would like to comment that while the tide of regulation - that is rule-making - may naturally ebb and flow over the years, it is important to maintain a constant commitment to strong supervision.

Before I develop that thought, I think it is only appropriate that I recognise the big effort that most management teams and Boards of Directors have taken to step up and adapt to the significant changes that are happening internationally and here domestically. That’s not only the insurance industry – which has put in a lot of effort to prepare for Solvency II – but other firms in other sectors too who have, for example, implemented important strengthening of rules on corporate governance and fitness and probity, and who are now subject to more rigorous supervisory assessment. I recognise that this has involved a lot of work on the part of industry and for the most part we have been pleased by the engagement we see and by the improved standards of compliance that are resulting; most firms’ management and Boards “get it” and realise that, post crisis, higher standards are here to stay and that Ireland’s reputation as a place to do business depends on good regulation and strong supervision.

So, let me finish by setting out my personal thoughts on the four key elements that need to be preserved to ensure strong supervision continues:

  • Element one is an adequately resourced and high quality supervisory staff. In this context, concerns over costs (in the aggregate to industry or at the level of pay that will be required to retain staff when austerity eases and markets return) need to be tempered by recollection of the terrible costs to Ireland's taxpayers and society of financial failure. The €1.65 billion projected insurance guarantee scheme support for Quinn Insurance is in the order of 80 times the current year costs of supervision of the insurance industry. And the €64 billion capital investment from the taxpayer into the Irish banking system represents a staggering 1,409 times the current year costs of banking supervision;
  • A second element is a strong set of powers, such as is currently being adopted in the Dail. These are important to allow effective legal powers for supervisory intervention and require an ongoing commitment to periodically update them in light of emerging supervisory best practice and industry developments;
  • A third element, which I've already spoken about, is a supervisory philosophy that encourages challenge of firms and that problems are tackled decisively and definitively, rather than being allowed to fester, even if that is uncomfortable or inconvenient; and
  • the final element is a supervisory institution that is independent. This does not mean immunity from feedback or accountability. But it does involve being truly independent in its supervisory decision making from industry or political intervention.

It was not so long ago that a now notorious banker complained on the radio about regulation saying, “Its time to shout stop: The tide of regulation has gone far enough.” At some point the debate about over-regulation and supervision will return, if indeed it has ever gone away. That is healthy and reasonable. But I would hope that debates on regulation and supervision are done transparently and that they are given short shrift if they take place as a vaguely articulated concern about burden and competitiveness without being grounded in specifics to ensure an informed debate on policy. And when this debate does resurface, it would be sensible to cast a watchful eye on whether adequate resources, best practice powers, a challenging approach with firms and genuine independence are being maintained, so that strong supervision survives this ebb and flow intact, through bad times and, in the future, good ones too.

Thank you.