The Equitable Distribution of Surplus (1800-1944)

Above, we identified that, historically, a central purpose of the actuarial valuations of assets and liabilities has been to determine the amount of surplus that emerged between two valuation dates, and hence what was available to be distributed to with-profit policyholders as bonus. The measure of the emergence of surplus was not necessarily intended to most accurately capture the economic profit that had emerged over the period. There was a (sometimes implicit) objective to distribute more or less than the amount of profit that had emerged—to attempt to ‘smooth’ volatile returns over time, or simply because there was a view that some surplus should be retained as a contribution to the estate. At the Equitable, William Morgan had established in 1800 that a maximum limit of two thirds of the surplus that had been assessed to have emerged between valuations could be distributed to policyholders.

But this was never a universally accepted standard and a number of actuaries felt that the Equitable had inequitably under-distributed bonus. Charles Jellicoe, the first editor of the Journal of the Institute of Actuaries, argued in 1851 that the whole of the earned surplus should be distributed to the policyholders, reasoning ‘The practice of reserving a part of it seems unnecessary, for there is not only the excess included in the future premiums, but very generally a subscribed capital to fall back upon, beside the fund reserved for the liability.’[1]

How much of the surplus should be distributed to the current generation of policyholders, and how it should be distributed amongst those policyholders, were decisions primarily driven by a consideration of what was equitable. The concept of policyholder equity has long been amorphous and ambiguous, especially for British life offices. What exactly was meant by equity in the context of a pooled system of investment such as with-profits? When unexpected events happened, how much cross-subsidy between the lucky and unlucky policyholders was ‘equitable’? As R.H. Storr-Best put it in 1962: ‘The problems [with equitable surplus distribution] that have arisen result from the conflict of these two ideas—on the one hand, the averaging of experience and on the other the return to the policyholder of what is his due. The interpretation of equity depends on the emphasis placed on each idea’.[2] Thus policyholder equity was a matter of interpretation, precedent and, ultimately, actuarial judgement.

As well as considering how much bonus should be distributed to each policyholder, the form the bonus took was another important element in surplus distribution. Chapter 2 discussed how William Morgan, in accordance with Richard Price’s advice, pioneered the distribution of surplus through an irreversible addition to the sum assured (instead of by cash dividend). This practice was adopted by the significant majority of British with-profit offices during the nineteenth century, but not universally. Of 59 British life offices writing with-profit business in 1870, 47 were found to be distributing surplus by reversionary bonus, with the remaining twelve doing so by cash dividend or by abatement of future premiums. Around half of the offices distributing by reversionary bonus did so by a uniform rate applied to all with-profit policyholders’ sums assured. This was clearly the simplest mode of bonus, but it constrained the amount of variation in the relative size of bonus that could be distributed to different policyholders. The remaining half varied the reversionary bonus according to age and/or the duration of the contract.[3]

By the third quarter of the nineteenth century, the equitable distribution of surplus had become one of the most active areas of actuarial research. It featured heavily in the early volumes of the Journal of the Institute of Actuaries, and was a favoured topic of Charles Jellicoe, its first editor (Jellicoe went on to become President of the Institute between 1860 and 1867). Jellicoe held the view that bonuses should be distributed in close accordance with the profit that had been contributed by the specific policy: ‘The surplus then, being satisfactorily ascertained, it appears only consistent with justice and proprietary that it or its value, as exactly as possible, should be returned to the contributors.’[1]

This view was concurrently becoming an established principle in the USA. The US with-profits system was still nascent at this time. Considerable freedom therefore existed to devise a surplus distribution system without the encumbrance of precedence or pre-existing policyholder expectations. Sheppard Homans, the chief actuary of Mutual Life Insurance Company of New York, one of the largest US life offices, played a leading role in the development of the theory and practice of equitable surplus distribution in the USA. Homans would go on to have a distinguished actuarial career, becoming the first President of the Actuarial Society of America in 1889, and, as we saw earlier in this chapter, he published the first mortality table based on US life office experience in 1868. He was invited to present a paper to the Institute of Actuaries on his thinking on surplus distribution. His paper, ‘On the Equitable Distribution of Surplus’, was published in the Journal in 1863.

Homans’s paper starts by noting that the experience of the previous five years at the Mutual Life of New York had been very profitable—mortality experience had been around 70 % of the Carlisle table (and the table approximated to the Mutual Life’s premium mortality basis), whilst investment returns had been around 6.5 % compared to a premium basis of 4 %. He noted that this size of surplus was unexpected and could not be expected to regularly recur. Homans’s approach to determining how surplus should be allocated amongst policyholders started with the principle ‘that each participant should be benefited in proportion to the excess of his payments over and above the actual cost of insurance’. Homans did not dwell on this objective, writing that it was ‘seemingly recognised by all actuaries’.[5] Whilst that was something of a simplification, his principle was consistent with the views published by Charles Jellicoe over the preceding decade.

Homans’s principle still left plenty of latitude. Any insurance relies on pooling and cross-subsidy, so the calculation of an ‘actual cost of insurance’ experienced by an individual depends on a definition of what risks are pooled by whom over what period. The bonus allocation system advocated by Homans involved splitting the life fund into separate ‘sub-funds’ with each sub-fund including a cohort of policies of lives of the same age of policyholder and same policy duration. The cost of mortality insurance was assessed within each of these sub-funds based on the mortality experience of the cohort. A surplus calculation was then made for each based on this assessed cost of insurance, together with the fund’s earned rate of interest and the change in the cohort’s liability reserve over the assessment period. He did not suggest any smoothing of the mortality insurance cost over time or any cross-subsidy between cohorts: if the experienced mortality rate in a given year of the cohort of 50 year-old policyholders with a policy duration of two years was half of that anticipated in the premium basis, and the experienced mortality rate of 61 year-old mortality rate with policy duration of one year was the same as that anticipated in the premium basis, the first cohort would receive a contribution to dividend from this particular source whereas the second cohort would not.

Homans’s contribution method was first implemented by himself at Mutual Life in the 1860s, and was then widely adopted by the US life offices. It did not have a direct impact on bonus practices at the British offices at the time. The reversionary bonus system had never aimed to deliver the granularity of cohort-specific surplus distribution that Homans’s system achieved. Over the subsequent decades, the British method of bonus distribution moved further towards the uniform reversionary bonus system, and hence further from a method that could deliver the level of granularity in surplus distribution that was implied by the contribution method (and that was present in the USA). The British actuarial profession, however, continued to wrestle with its own interpretation of equitable distribution of surplus over the following decades. The contribution concept constantly resonated in the profession’s collective conscience as an important reference point during these deliberations.

Surplus distribution was an ongoing topic of actuarial thought through the final quarter of the nineteenth century. H.W. Andras, the actuary of the University Life Assurance Society, presented a notable paper to the Institute in 1896 that shone further light on the performance of British bonus methods in the context of the contribution concept.[6] He analysed the uniform reversionary bonus system and its practice in Britain at the time, considering how well it could achieve equitable outcomes for policyholders. Andras concluded that:

This system of bonus distribution ... although having the advantage of simplicity, is of too rigid a nature to be suitable to the varying circumstances of life offices, if we regard as important the equitable distribution of profit amongst those who have contributed to it [emphasis added].[7]

Andras’s conclusion was partly based on the difficulty of using a uniform bonus rate to distribute profit equitably when the in-force policies had been written at different premium rates. This was one of the reasons why British with-profit rates were subsequently so ‘sticky’ through the first half of the twentieth century, despite the substantial changes in interest rate environment over that period. However, Andras balanced his conclusion by noting that ‘in offices experiencing average results ... it [the uniform compound bonus rate] would give bonuses approximately the same as the contribution method, so that, in such cases, the merit of simplicity may be combined with the principles of equity’.[7]

As ever, the record of the Staple Inn discussion of the paper provides an insight into the state of thinking amongst the leaders of the profession. G.F. Hardy argued:

[The American contribution method] was a highly scientific plan, but on the whole was not practicably workable — at all events in this country — the main objection being that the whole of the mortality fluctuations at individual ages reappeared in the shape of fluctuations in the bonus . In the present state of our knowledge it was not known what the true rates of mortality were, and that was a great argument in favour of adopting a simple method like the compound reversionary bonus method, instead of the much more complex contribution plan.[9]

The above comments highlighted the fundamental tensions: the large amount of calculation and complexity the administration of the contribution method entailed was unappealing to British actuaries, but they were naturally anxious that their uniform bonus rate approach would not be so blunt a tool as to be self-evidently inadequate for delivering equitable treatment amongst policyholders. Furthermore, the British system embraced crosssubsidy between current policyholders and across policyholder generations in a way that US with-profits had never done—stability in bonus rates was a fundamental feature of the British with-profit system.

The challenges of ensuring an equitable distribution of surplus further increased for the British system of with-profits in the twentieth century, as with-profits business was increasingly run on steroids. By the second quarter of the twentieth century, ‘bonus loadings’ in British with-profit premium rates tended to be significantly larger than elsewhere in the world. The bonus loadings implied by the difference between with-profit and without-profit premium rates tripled between 1870 and 1940.[10] And as was discussed in this chapter, with-profit funds’ allocations to risky assets such as equities and real estate increased significantly over the second quarter of the twentieth century. Obtaining equity amongst policyholders was further complicated by changing product mixes within with-profit funds (with endowment assurance becoming increasingly popular relative to whole-of-life assurance). These features ensured that the equitable distribution of surplus remained one of the major topics of British actuarial thought during the first half of the twentieth century.

Alongside these changes in with-profit funds’ bonus loadings, investment strategy and product mix, with-profit business was also exposed to exceptional economic variability in the first half of the twentieth century, particularly in the aftermath of the First World War. With-profit funds tried to maintain very stable with-profit premium rates through this period, even in the presence of unprecedented gyrations in long-term interest rates. This meant that the ‘bonus-earning’ ability of policies written at different times could be substantially different. The historical priority given to stable bonus rates over contribution equity became an increasing concern, as was tactfully noted in Lochhead’s actuarial student textbook:

The question has been raised whether the subordination of other considerations to the stabilizing of bonuses is always perfectly fair as between policyholders of different generations.[11]

In his presidential address to the Institute of Actuaries in October 1943, H.E. Melville questioned whether current methods were generating equitable distributions of surplus. This prompted a number of papers on the topic to be presented to the Institute and Faculty in the subsequent few years, the most important of which was T.R. Suttie’s ‘The Treatment of Appreciation or Depreciation in The Assets of a Life Assurance Fund’.[12]

Suttie’s objective was ‘to discuss how appreciation or depreciation [in fixed interest securities] should be treated in order to obtain the maximum practical degree of equity between the different generations of policyholders in an office which distributes by means of a uniform reversionary bonus’.[13] His analysis included a historical review of the twentieth-century performance of the ‘always write down, never write up’ asset valuation practice from the perspective of policyholder equity. He focused on two periods of volatility. First, in the inflationary aftermath of the First World War:

During the war of 1914—18 and the years immediately following, there was a very substantial rise in the level of interest rates and consequent depreciation of investments. It would appear that in many cases life offices wrote down their assets to their new market values and continued to make their valuations on the same net premium basis as had been used in the past ... this procedure benefited the more recent policies at the expense of the older. Further, if the level of premiums charged for new policies remained unaltered, it benefited policies entering after the depreciation occurred, while if the premiums were reduced on account of the increase in the rates of interest obtainable, new policies were even more favourably treated.[14]

And second, the impact of the Great Depression:

In 1932 ... there was a reduction in the level of interest rates and consequent appreciation in investments, but it would appear the majority of offices did not write up the assets to their new market values so that the difference between book values and the market values constituted a hidden reserve which was not made available for distribution in bonuses. This again benefited the more recent policies at the expense of the older policies, and greatly benefited the policies entering after the appreciation unless the premiums for new policies were substantially increased.[14]

Suttie particularly focused on the high interest rate scenario. He considered how current bonuses and assumed future bonus rates (in the context of a bonus reserve valuation) should be impacted by a substantial rise in interest rates, given the objective of ensuring equity between different generations of policyholders. He used a model office to examine how different allocations to current and assumed future bonus rates impacted on the different policyholders in his office. He concluded:

The most equitable method of dealing with [bond asset] depreciation under a uniform reversionary bonus system is to write down the assets to their new market values and to increase the valuation rate of interest to that obtainable under the new conditions, valuing as a liability a future rate of bonus at the rate which the existing premium scale will support under the new conditions ... If these arguments are accepted, depreciation will involve a decrease in the rate of bonus declared in respect of the valuation period during which the change in value occurs . while, if conditions again become stable, future bonuses will be at a higher rate than in the past.[16]

But he went on to note that ‘Such abrupt changes in the level of bonus rates are contrary to the usual practice of offices and it would undoubtedly be difficult for one office to follow such a policy.’[16]

Suttie’s analysis was essentially showing how a contribution approach to bonus distribution would behave through time in an interest rate stress scenario. Perhaps predictably, the discussion following his presentation of the paper cleaved to the more conventional route of setting the current and future bonus rates at equal levels that would result in a reduction in the bonus reserve valuation that absorbed the asset depreciation. In essence, his peers advocated spreading the recognition of the market value impact of the interest rate increase over the full run-off of the existing business. Concerns were voiced over the variability of bonus that would be implied by Suttie’s proposed approach:

He did not know what would be said by the holder of a policy nearing maturity if, just before maturity, there were a fall in prices and, as a consequence, a reduction of the sum payable at maturity. He felt that such a type of distribution would be bound to cause dissatisfaction.[18]

Stability in bonus rates still trumped all else in British actuarial thinking on surplus distribution. Unlike in the United States, cross-generational smoothing had become a fundamental feature of with-profits. This cross-subsidy could apply not only between different generations of existing policyholders, but also between policyholders and the estate, as was expressed at the Staple Inn meeting:

With most offices, part of the profits of the past had been used to create contingency reserves which should be drawn upon to deal with the depreciation of the assets ... the aim should be to smooth out the effect of the depreciation by declaring a current bonus at a reasonable level and by rebuilding the contingency reserves gradually in the future [emphasis added].164

The British actuarial profession’s interpretation of equitable bonus policy remained much broader, and its implementation more heuristic, than the contribution method that dominated bonus methods in the USA. This was partly a function of legacy: the British life offices, arguably and perhaps ironically due to a lack of policyholder equity in the past, had accumulated significant free reserves in their funds. These estates permitted the actuary a considerable latitude in bonus policy. The inevitable arbitrariness entailed in deciding what to do with the estate provided another reason not to be overly meticulous about how to distribute the surplus generated by existing policyholders. The British approach relied heavily on the judgement and expertise of the actuarial profession, and the public’s trust in it. In that sense it was perhaps an approach that reflected the time and place of Britain in the first half of the twentieth century—a class-based society that placed unquestioning trust and deference in its institutions and professions. Times would change.

  • [1] Jellicoe (1851), p. 26.
  • [2] R.H. Storr-Best in Discussion, Cox and Storr-Best (1962a), p. 40.
  • [3] Cox and Storr-Best (1962a), p. 47.
  • [4] Jellicoe (1851), p. 26.
  • [5] Homans (1863), p. 122.
  • [6] Andras (1896).
  • [7] Andras (1896), p. 359.
  • [8] Andras (1896), p. 359.
  • [9] In Discussion, Andras (1896), p. 368.
  • [10] Cox and Storr-Best (1962a), p. 66.
  • [11] Lochhead (1932), p. 75.
  • [12] Suttie (1944). See also Anderson (1944).
  • [13] Suttie (1944), p. 203.
  • [14] Suttie (1944), p. 214.
  • [15] Suttie (1944), p. 214.
  • [16] Suttie (1944), p. 213.
  • [17] Suttie (1944), p. 213.
  • [18] M.E. Ogborn, in Discussion, Suttie (1944), p. 220.
< Prev   CONTENTS   Source   Next >