Actuaries and the Financial Management of General Insurers (1981-1994)
In the 1980s and 1990s, technical actuarial research in general insurance broadened beyond the assessment of claim reserves and solvency margins to consider the wider horizons of the financial management of insurance firms. Sidney Benjamin was ahead of his time in this field as he was in maturity guarantee reserving, and his 1976 paper ‘Profit and Other Financial Concepts in Insurance’ was an early forerunner to this strand of thinking. But the 1981 Journal paper by Abbott, Clarke and Treen, ‘Some Financial Aspects of a General Insurance Company’, with its focus on the measurement and management of shareholder returns, arguably marked the real departure point from traditional actuarial ground.
Like Benjamin in 1976, Abbott et al. advocated the use of discounted reserves, providing inflation was consistently applied to the claims projections. The paper particularly focused on how undiscounted reserves resulted in a misleading profit emergence pattern over the life of the business. It highlighted how material this effect could be in the high inflation environment that prevailed at the time the paper was written and how changes in inflation and interest rates over the run-off of the business could further distort the emergence of profits when using the undiscounted reserving approach. The paper also discussed how to assess the return on shareholder capital that had been earned by a general insurance business. It did not, however, venture to answer what return ought to be required by general insurance shareholders other than to suggest that an arbitrary amount in excess of the real risk-free interest rate should be a form of profit objective.
A short paper by G.C. Taylor published in the Journal in 1984 considered the capital that should be required to write new business and from where this capital should come. In particular, in general insurance in the 1970s, the notion had developed that insurance business should be self-financing: that is, solvency margin requirements should be funded entirely from premium loadings. Taylor argued that solvency margin requirements should be considered as part of the working capital needs of a general insurance business and should therefore be funded by the shareholder rather than the policyholder. He further argued that whilst these funds could be invested and would generate investment return for shareholders, this return would be inadequate compensation for shareholders due to effects such as double-taxation. Hence premiums should be loaded to generate an acceptable expected return on that capital (which would be less than the loading for full self-financing). He then developed some algebra to show how to calculate this premium loading, though the required return on shareholder capital was taken as an arbitrary parameter that was ‘dictated by the equity market’. This work can be seen as a step towards a more economically coherent perspective on insurance premium setting. The concept of loading premiums to include a cost of capital has endured.
A 1990 paper by Daykin and Hey, two leading general insurance actuaries that we already met earlier in our general insurance discussion, discussed how the stochastic asset-liability cashflow modelling that had been developed by the Solvency Working Group in 1987 for the purposes of solvency assessment could be extended for use in the wider financial management of a general insurance firm. The 1987 approach to solvency assessment had explicitly assumed that the insurer ceased to write new business. The 1990 paper considered how the simulation model could be extended to include the long-term projection of new business and its financial impacts. This naturally required a modelling framework to describe the behaviour of new business in terms of business volumes, profitability, competitiveness of pricing relative to the wider market, and how these variables behave jointly with each other and with the other stochastic variables and outputs in the model. For example, the business may price more aggressively when its balance sheet is strong relative to the required solvency margin. Daykin and Hey produced results for the ten-year stochastic projection of asset and liability cashflows, profits and balance sheets under a variety of different assumptions for new business writing and pricing strategies. Their objective was to show that technical actuarial skills could be employed as an intrinsic part of business strategy development and planning in a general insurance firm.
Daykin and Hey also considered the potential use of financial economics in general insurance business management. Once again, this topic of general insurance research had first been explored overseas—in this case in the USA. The particular stimulus for the use of financial economics in US general insurance had been in determining the return required by investors for funding general insurance business. This had relevance beyond the commercial management considerations of a given company—in the USA, premium rates were controlled by government regulation, and this required an objective and rigorous framework for establishing a ‘fair’ premium rate. This, in turn demanded an assumption about the level of return reasonably required by the providers of insurance capital.
The application of the Capital Asset Pricing Model to determine required returns on general insurance equity capital was initially explored in the USA in the late 1970s and early 1980s. This work highlighted that where general insurance claims are uncorrelated with market returns, no risk premium should be required by shareholders for bearing this (diversifiable) risk. Daykin and Hey expressed discomfort with this implication for general insurance required returns, arguing that it ‘suggests that the CAPM is missing some important aspects’. A natural candidate for these missing aspects was the frictional cost of capital argument that had been presented by Taylor in 1984, but this line of thought was not pursued by Daykin and Hey.
Daykin and Hey also briefly considered the potential use of Merton-style option modelling of an insurer’s capital structure, where shareholder equity is modelled as a call option on the net assets of the general insurance company.
This had also recently been explored in the USA. By explicitly capturing the leverage implied by insurance business (where policyholders are essentially debtholders), higher required shareholder returns (and hence premium rates) could be implied, which Daykin and Hey welcomed. They suggested that their simulation modelling approach could be used to analyse more realistic forms of optionality in shareholder returns than the simplistic analytical models presented in the literature.
Daykin and Hey widened the actuarial perspective on general insurance from solvency to include profitability, return of capital and new business pricing strategy. Later in 1990, a paper by Ryan and Larner was published in the Journal that widened the actuarial horizons of general insurance further to include company valuation. This paper discussed the application to general insurance of the appraisal value method that actuaries had recently started to apply in merger and acquisition work in both life and general insurance business. The essence of the appraisal value was that it involved an explicit projection of the cashflow earnings of the business, using assumptions set following actuarial investigation. These cashflows would then be discounted at appropriate risk discount rates to obtain the company’s appraisal value.
Ryan and Larner considered the appraisal value in three components: the adjusted net asset value (the current balance sheet net asset value, adjusted for the cost of those assets being ‘locked-in’ to the insurance balance sheet); other value arising from past written business (expected release of surplus from insurance reserves); and the value arising from future written business. The adjusted net asset value would be based on the market value of assets, and the deduction from the net asset value for capital ‘lock-in’ would be based on the shareholders’ cost of capital. They explained the rationale for the cost of capital as follows:
They [shareholders] will require a larger return on their funds [if invested in an insurance operation] than if they invested them separately. This arises, partly because the capital is being exposed to the risk of loss in the insurance business and partly because it could be used elsewhere____We use the term “cost of capital” to mean the value of the shortfall in net earnings between the risk return required by shareholders and the actual [expected] investment return.
Their cost of capital represented the incremental additional return that arose from the consequences of investing via an insurance entity rather than directly in financial markets. Taylor had taken a similar position in his 1984 paper on loading premiums for the cost of capital, though he had argued that this additional cost arose from ‘frictional’ sources such as double-taxation rather than because of exposure to insurance risk.
The risk discount rate would reflect the riskiness of the shareholder cashflow streams. The authors suggested different risk discount rates should be applied to different elements of the projected earnings stream to reflect their risk characteristics (for example, a higher discount rate would be applied to earnings from future business than that used for future profits emerging from existing business). Whilst the authors were aware of financial economists’ approaches to valuation, some of their suggestions were not necessarily consistent with those ideas. In particular, their suggestion that the risk discount rate should be a function of the extent to which the business was exposed to diversifiable insurance risk ran contrary to a fundamental principle of financial economics. This idea ran throughout the paper—for example, later the paper suggested that a well-diversified insurance business could command a lower risk discount rate and, hence, higher appraisal value than a less-diversified business.
A 1994 Journal paper by Bride and Lomax provided a perspective on the financial management of general insurance business that was more closely aligned to financial economics. Their starting point was that the appraisal value implementation that had been developed by actuaries to support shareholder valuations in merger and acquisition activity ‘fails to capture the operational dynamics of non-life insurance business and obscures the issues surrounding the nature of shareholders’, policyholders’ and other creditors’ claims on the assets of the firm’. However, they did not propose to reject the entire framework of discounted cashflow projections of the appraisal value— rather, they attempted to show how the risk-adjusted discount rate and cost of capital adjustments could be rigorously set.
They emphasised that the idea that shareholders did not require a reward for bearing diversifiable risk was the most fundamental and least controversial result produced by financial economics. This had implications both for the setting of the risk discount rate and the cost of capital adjustment—Ryan and Larner had argued that one of the reasons for the cost of capital adjustment was that shareholders required compensation for their exposure to the (diversifiable) risk of future insurance losses. This was categorically rejected by Bride and Lomax.
They also highlighted what they viewed as other actuarial misconceptions in valuation. In particular, they argued that the choice of asset strategy of a general insurance company should have no impact on the value of the firm: ‘an investment in gilts and an investment in equities have equal risk-adjusted total yields from the perspective of a shareholder with a diversified portfolio of assets’.
In short, Bride and Lomax’s 1994 paper was the general insurance equivalent of Exley, Mehta and Smith’s 1997 defined benefit pension paper. It attempted to bring actuarial practice on the measurement and management of shareholder value in general insurance fully in line with core financial economics concepts. This was arguably less of a challenge in general insurance than in defined benefit pensions, with its legacy of 100 years of satisfied actuarial service. But the authors still wryly noted in their response to the Staple Inn discussion that ‘the subject of financial economics continues to raise temperatures in the profession’.
Actuaries in general insurance as well as pensions found it difficult to accept many of the insights and implications of financial economics. A reflex to reject them as academic theories based on unrealistic assumptions remained prevalent. In opening the discussion, Duffy commented:
It is questionable whether the shareholders, let alone the management, of the
major UK composites and other non-life firms follow the precepts of financial
The influential general insurance actuary J.P. Ryan expressed his opinion at Staple Inn in unambiguous terms: ‘The authors use a number of oversimplified economic assumptions when applying their model. It is largely these oversimplified assumptions that give rise to some of the odd results in the paper.’
Incorporating financial economics into the practices of any British actuarial field was no straightforward task in the 1990s. But in general insurance, as in life assurance and defined benefit pensions, the genie was out of the bottle. The process of embracing and incorporating the financial economic insights of the previous half-century into core actuarial thinking and practices was irrevocably underway.
-  Abbott et al. (1981).
-  Taylor (1984).
-  Taylor (1984), p. 178.
-  Daykin and Hey (1990).
-  See, for example, Biger and Kahane (1978); Hill (1979).
-  Daykin and Hey (1990), p. 197.
-  Doherty and Garven (1986).
-  Ryan and Larner (1990).
-  Ryan and Larner (1990), pp. 603-4.
-  Bride and Lomax (1994).
-  Bride and Lomax (1994), p. 363.
-  Bride and Lomax (1994), p. 388.
-  Bride and Lomax (1994), p. 438.
-  Duffy, in Discussion, Bride and Lomax (1994), p. 421.
-  Ryan, in Discussion, Bride and Lomax (1994), p. 429.