Insurance in a wider financial context
Speech by Callum McCarthy, Chairman, FSA
at the Insurance Sector Conference
8 April 2008
You have spent the day discussing the issues affecting insurance firms and their customers, and have covered a wide range of issues – solvency standards, climate change, treating customers fairly. This range of issues matches the breadth of business and experience represented by those attending this conference – brokers, aggregators, risk-carriers, in both retail and wholesale insurance. In talking to you, I am de facto addressing "Insurance in a wider financial context", the title to which I have been asked to speak.
What I would like to do is to consider – you will be relieved to know, at the end of a long day, briefly – two aspects of the wider financial context, one immediate, the other long term. The immediate issues are those which arise from the present turbulence in financial markets, particularly the debt and funding markets. How have these affected the insurance sector? How has the insurance sector contributed to the present problems, either positively or negatively? What else can we expect? The longer‑term issues that I want to consider are those posed by increasing life expectancy, a reminder, whatever immediate capital market issues prevail at any time, that fundamental assumptions about morbidity and mortality remain at the heart of at least one part of the insurance industry.
Present market conditions
Let me start with present conditions. The fragility of financial markets is a very real and present condition. Since August 2007 we have seen a marked and sustained reduction in a number of funding markets and instruments, both in the UK and globally. Retail MBS issuance in the UK – which amounted to about 15 per cent of global Retail MBS issuance in 2007 – has been almost totally absent in the UK this year; the covered bond market is similarly dormant. Global markets show comparable inactivity: globally, Retail MBS issuance in 2008 is a tenth of that in 2007, and LBO issuance in 2008 is a quarter of 2007. The money markets show continuing pressures, with three‑month £ LIBOR:OIS spreads, which normally are below 20 bps, now approaching a full percentage point. This, I would point out, is largely a financial, not an economic, set of problems, as these difficulties have originated in the financial sector, not the real economy which for the most part remains strong. World economic growth was high in 2007; bad debts and provisions were low; and yet a problem which had its origins in the US mortgage market has spread to cause widespread anxiety and lack of confidence in many parts of financial services, and may yet have wider effects on the real economy.
In August 2007 when events came to a head in the UK, the immediate focus of the authorities, and that of the media and politicians, was very much on the banking sector. But as we have seen in more recent months, the insurance sector is not immune from the market turmoil. Indeed, firms that provide credit insurance to lenders or bondholders – the financial guarantee insurers – or 'monolines' as they are in some ways deceptively referred to – have emerged as a source of considerable contagion between the banking and insurance sectors. Very briefly the deterioration in the structured credit markets increased the potential for higher than expected claims to emerge. This in turn put a question mark against monolines' level of capital, in some cases to the extent that rating agencies have downgraded their credit ratings. This affects not only monolines' ability to continue to write business, but also wider market confidence, as the ratings of the securities insured by monolines are also at risk of downgrade.
Let me highlight what I think are two important aspects to the present position which have wider application and significance.
The first of these aspects is the role of the credit rating agencies. There are clearly issues of potential conflicts of interest in those giving ratings being rewarded by the institution receiving them, as well as other specific issues. From a regulatory perspective the issues are being examined in Europe within CESR, in the US by the SEC and globally within IOSCO.
My own concerns, apart from the important issue of managing effectively conflicts of interest which cannot be eliminated (not a problem specific to CRAs, but rather one endemic in financial services), are that we deal with some of the major things that went wrong in those responsible for carrying out the ratings and in those using ratings: the overtrading by CRAs which inevitably led to a lowering in the standards of the rating process; the failure by CRAs to recognise the degradation of data as mortgage origination standards in the US declined; and – probably the most dangerous failure of all – the over reliance by investors on ratings of products whose risk characteristics they did not properly understand.
The second aspect is international cooperation and action. The situation with the monolines is one example of where we have relied on informal links with other regulators – in this case the US – to deal with problems. All the monoline insurers authorised to do business in the UK are part of the US monoline groups. This means that their fortunes are inextricably linked to those of their parents, and are therefore influenced by the actions of the US supervisors. So we are collaborating as closely as we can with the US, dealing with the issues that come up on a firm-by-firm basis. We also have good relationships with the US parents of UK monoline subsidiaries, and their willingness to share information means we are able to closely monitor the situation and the likely implications of US developments on UK monolines and the wider UK market. But I think it is extremely important that we build on these informal arrangements and strengthen them, so that we can react quickly, and that regulatory cooperation is part of an established framework, rather than reliant on good relationships between individuals.
In this, the experience of international cooperation between national supervisors in respect of the monolines is a microcosm of a wider issue, namely how in a world in which large complex financial institutions, operating globally, are of increasing importance do we establish effective, efficient and consistent regulatory policies and practices. The present market problems have given a welcome impetus to devising answers to the set of questions this gives rise to.
Market turbulence – traditional insurers
Let me turn to the impact of recent market turbulence on more traditional insurers. Our investigations to date suggest that your overall direct exposures to sub-prime are somewhere between low and insignificant. In many firms exposures to asset backed securities have grown in recent years, but remain a relatively small component of total assets. And exposure to sub-prime seems to be, in most cases, a relatively small proportion of asset-backed securities. There is also no indication, to date, that any firms have significant exposures to structured credit products – for example CDS-type obligations – through their underwriting or investment activities. Now, the alert among you will have noticed what I just said contained phrases like 'seems to be' and 'to date'. This is for two reasons; one, because I think that in the current market conditions it would be very dangerous to be complacent and there is a need to constantly review and revisit our assumptions as market developments unravel; and two, because firms have not always supplied the information we have requested on a consistent basis – and indeed some firms have been slow to provide it - a failure which we are working to rectify. Any firm which has been backward in providing the information the FSA needs to assess risks should expect severe and constant pressure to make good its position.
Irrespective of holdings of structured products, it is clear that some of you will have seen falls in the value of your equity, bond and property portfolios, rating downgrades on some investments and lower (total) returns on investment – all of which will affect profits in respect of 2007/08. I think we can expect uncertain and turbulent market conditions to continue for the foreseeable future. In this climate, it is extremely important to identify severe but plausible stress and scenario tests. It is my impression from talking to colleagues that in recent months the ICAS framework has proved itself a very useful tool in understanding key risks and in identifying vulnerable or outlier firms. We shall be encouraging firms to further develop their stress and scenario testing. First, we expect firms to be continuously reviewing their stress test assumptions, both for their ICAS calculations and, just as importantly, for their own internal risk appetite calculations. Second, we look to firms to carry out a more extensive range of scenario testing, thinking carefully about the 'ripple effect' that in stressed conditions one event could have on a whole range of future events. Firms should be thinking through for themselves the scenarios that they feel they would need to survive. Firms should also think the impossible: that is, to identify those scenarios that they would not survive.
A third issue is that of liquidity. Events of recent months have had a significant impact on liquidity in the banking sector but what of the insurance sector? In a recent survey of UK life insurers you told us that, even in the exceptional circumstances in the last quarter of 2007, you did not experience liquidity pressures, with the average exposure to illiquid assets of below 5% of total admissible assets. While we take some confidence from this, liquidity is an issue we will want to keep under review, especially as assets that are liquid in normal markets might become illiquid in some stress scenarios, and you can expect it to be a more prominent feature of ICA reviews in the future.
The majority of these illiquid assets held are in property, and while many property funds continue to operate as usual, some funds are experiencing liquidity difficulties. Three main factors have combined to create this situation: first, the significant fall in commercial property values in recent months; second, the illiquid nature of some of the underlying investments, particularly relevant for those funds investing directly in bricks and mortar; and third, the significant growth in redemptions in these funds. This combination of factors has led to some funds having to defer redemptions, affecting both institutional and retail investors. Although this scenario is recognised in our rules, it is nevertheless an uncomfortable position to be in. Our priority is to ensure that consumers are treated fairly: that is, that the interests of redeeming investors are balanced against those of investors remaining in the fund; and that in your communications with investors and intermediaries, and in promotional material, you provide a true reflection of the risks inherent within these products. And of course we would expect you to communicate important issues, such as the suspension of redemptions, to investors in a timely and balanced manner.
The final issue I want to touch on is valuation. Valuation is an important aspect of firms' monitoring of capital adequacy and we have tightened our focus on how firms are valuing illiquid instruments. Our work will cover several aspects of valuation, including: the results of recent independent price verification processes; valuation methods for loans and securities that have recently proved problematic to value; questions of model calibration for mark-to-model products; recent changes in valuation adjustment methods; and interactions with auditors with respect to recent valuation challenges.
Longer-term risks – access to capital
I have set out ways in which the insurance sector may be affected by the present market difficulties. I should also make clear that the insurance sector has also made some contribution to the spreading of risk which is a benefit of the originate to distribute model of banking – though, as I said earlier, the scale may be limited. The deficiencies in that model have been much discussed, and clearly need correcting, which will be a cause of much work over the coming months and years. But the model has also helpfully meant that risks and losses which, had they lodged exclusively within the banking sector, would have proved even more damaging than has occurred have in fact been spread more widely. You may not appreciate it, but the insurance sector contribution to this has been useful.
Long‑term issues
Let me turn – with some relief I should admit – from the immediate issues of market turbulence which have taken up so much of our recent efforts, to longer‑term issues which will continue to shape the industry long after the financial problems of 2007 and 2008 have been relegated to the PhD theses of future economic scholars. One of those issues, the impact of climate warming, has already been discussed at this conference, in the session addressed by Dr Evan Mills, so I will be very brief in my remarks on this enormous subject. The issues raised for the industry and its customers are various and often of great importance: for example, weather‑related losses – still, as throughout human history, one of the greatest sources of damage – may change both in terms of the scale of expected losses, and in the variation of potential losses – a combination which has obvious and difficult implications for those in the UK living on flood plains (or, indeed, for conference halls, or centres of government, built upon flood plains). Understanding the changes in risk associated with global warming, and pricing those changes efficiently, will be a major and difficult task for providers of cover, and for those who have to assess the required capital for those providers.
The second long‑term issue which will determine the shape and success of the life industry is of course longevity risk.
Longer term risks – longevity risk
In the life insurance sector, ageing population and improving life expectancy give rise to longevity risk, and consumers are living longer (on average) than they themselves or life insurers had initially expected.
Firms need to make longevity projections to correctly price the risks to their business and assess how this affects policyholders both today and in the future. However, there is considerable uncertainty in assessing future mortality trends as the available information is limited and trends are slow to become apparent. Past, and as I think the actuarial profession would acknowledge, current projections underestimate future improvements in mortality and life expectancy. And while there have been proposals to address the risks arising from uncertainty in the projections, as yet there is no consensus as to what extent the mortality projections should be adjusted.
While there is general uncertainty over mortality improvements in the general population, developments in medical science and technological advances mean that mortality for older age groups (for example aged 75 and over) could also start to improve more rapidly in the near future. Work we recently carried out found evidence that past trends in mortality are not stable over time. In the early years of our sample, mortality improved at a faster pace for younger ages than for older ages. However, as time passed, improvements for younger ages subsided while improvements for older ages accelerated. This illustrates how adapting a model to factor in future improvements for older ages can yield considerably higher estimates for life expectancy. We hope that our study will contribute to the debate in this area and I would encourage you to contribute to discussions over how the methodology for capturing this aspect of longevity risk can be improved.
Another area of uncertainty is around how the social and economic background of policyholders affects future mortality improvements. Mortality rates for the higher socio-economic groups have improved more rapidly in the past. It could be argued that higher socio-economic groups have already benefited from improvements due to lifestyle changes and that the lower socio-economic groups are more likely to experience higher improvements in the future. If firms were to presume – falsely – that current trends continue and higher socio-economic groups see faster improvements in mortality when pricing their annuity products, firms could make a loss on annuities sold to lower socio-economic groups.
Annuity providers in particular are exposed to longevity risk, and this risk is magnified in the current economic climate of relatively low interest rates. Pricing errors could give rise to substantial losses, potentially losses that are large enough to eliminate any contingency margins firms have allowed for. So alongside making improvements to models to capture longevity risks, firms need to ensure that capital reserving reflects the true extent of the uncertainty behind future longevity estimates.
Conclusion
I am not sure if I have delivered on my aspiration to be brief. I am conscious that I have done little more than sketch out some of the issues which will determine the future shape and direction of the insurance industry. The issues are both the immediate issues we face in dealing with a set of financial problems which make today's capital markets fragile and the longer‑term issues of global warming and life expectancy which will determine the decisions which future generations of both general and life insurers will be called upon to take. Yours is indeed an industry with a wide context – financial and societal – in which to operate. I hope today's discussions will have helped you, and the FSA, to go about your and our work more effectively.