Proposals for the Statutory Basis of Valuation of the Liabilities of Long-Term Insurance Business

OF THE DISCUSSION Mr P. A. C. Seymour, introducing the paper, said that the Working Party had been asked to consider possible modifications of the valuation method embraced in the six principles which, when combined with assets taken at market values, would ensure a reasonable standard of adequacy in times of rapidly changing interest rates, and which could be expressed in statutory rules. Mr Seymour said he would mention three points which arose in the Faculty discussion the previous week. In Table (a) of Appendix 2 asset values were tabulated based on discounting the net interest receipts and the gross redemption value at the net valuation rate of interest. It had been suggested the previous week that in practice the market value of the stock would be lower, being determined by discounting the gross interest and redemption proceeds at the gross rate of interest. While that might be true, it was not really relevant. As stated in 5.4.8 the valuation interest rate should be based upon the net redemption yield, not upon the gross redemption yield net of tax;. As could be expected, there had been some speakers who favoured the gross premium method. As had been demonstrated in Appendix 5, a gross premium method was extremely sensitive to a difference in the interest rate between the premium basis and the valuation basis. Furthermore, there was great difficulty in deciding what should be laid down in the regulations as a reasonable allowance for future bonuses and expenses. One suggestion to deal with the negative value problem, which had been heard from several gross premium supporters, was that a lapse assumption should be built into the valuation. It had been claimed that such an assumption had the effect of reducing the mean term of the liabilities, making it more feasible to match and reducing the negative values arising. Apart from the difficulty of deciding on what lapse assumption to make, there were other dangers which might be illustrated with a topical example, the income bond. Mr Seymour felt that the correct matching procedure was to invest in negotiable securities yielding the required fixed income and redeemable on the maturity date for the required amount. The surrender values should then be based on the value of such assets at the time of surrender. It seemed quite wrong to guarantee surrender values in sterling, and then invest some of the assets shorter to cover an estimated lapse rate. If more lapses occurred than expected there was mismatching to the surrender values, and if less lapses occurred there was mismatching to the maturity values. Furthermore, selection against the office would tend to aggravate losses. The point was that it was impossible to match surrender and maturity values at the same time; lapses should therefore be ignored in the matching policy. At the Faculty meeting it had been stated that practicalities should not be forgotten. It had, for example, been suggested that condensing a valuation into a single present value was not very informative. While it would be preferable to conduct a valuation on an emerging cost basis looking at the projected revenue account year by year, it was very difficult to imagine a set of regulations based on such a method. Furthermore the computer systems required would be complex and costly. Mr Seymour apologised for the fact that the working party was not able to cover any special classes of policy in detail. They had felt that it was better to centre on the main problem, namely, whether the net premium method was suitable for the most common classes rather than divert energies in other directions. Since the original Bill almost two years previously, a number of companies had got into difficulties, and there might be great pressure to finalize the necessary regulations. Mr B. J. Brindley, in opening the discussion, said that the paper provided a useful summary of the background and circumstances of solvency valuation. It also gave some interesting ideas for possible adjustments of the net premium method of valuation. He endorsed what Mr Seymour said about the terms of reference. They were very specific on both sides of the valuation balance- sheet. On the assets side they were restricted to market values. On the liabilities side they were 90 Proposals for the Statutory Basis of Valuation asked to consider the desirability and possibility of a modification of the method of valuation embraced in the six principles. Mr Brindley wished to begin by considering the liabilities side of the valuation balance sheet, and take in turn each of the six principles to see what the paper had to say about them. To deal with the less difficult first, in 5.7.1 the paper pointed out that the second principle, a maximum 3% zillmer adjustment, ‘seems quite adequate’. That seemed a very reasonable figure. The third principle stated that the net premium should be tested against the office premium, and that the difference should be sufficient to meet renewal expenses, including an allowance for inflation. In 5.8.1 and 6.4 the authors commented on the difficult position of an actuary when considering the possibility of runaway inflation. Such a situation would, however, have such widespread implications for the whole financial sector of the economy that other, much more serious effects, would occur first, such as wholesale surrender. Perhaps the Department could give some guidance as to a ‘reasonable’ inflation rate. The fourth principle was clearly acceptable, namely, that recognized tables of mortality and morbidity should be employed. The importance of the sixth principle, namely, that surrender values which were guaranteed should be covered by market values, had been very firmly demonstrated by the industry’s recent unhappy experiences. There were very real problems created when market values fluctuated, from which the subject of matching followed. The bulk of the paper was directed towards the first and fifth principles, namely, that the basic valuation method should be the net premium one, and the rate of interest should be that earned by the fund on market values. In his original paper (J.I.A. 92, 75) which set out those principles on solvency valuation, Mr Skerman was very specific concerning one reason for choosing a net premium valuation. That was to try to ensure that the amounts included in the future premiums for future bonuses should be available for future profits and not capitalized. That idea contrasted sharply with a minimum solvency criterion for a typical with-profits office where ‘minimum’ meant the ability of such an office to meet its strictly contractual liabilities only. Mr Brindley said he had tried to examine the first and fifth principles to see whether they met that requirement (see tables below). Companies A and B had identical liabilities, and were well-established mutual companies with a 41⁄2% compound reversionary bonus. Those liabilities would be typical of a high proportion of U.K. companies. Company A pursued an investment policy of complete immunization and Company B immunized the contractual benefits, both by investing in cash and irredeemables. The figures in the 6% columns showed that the proposed rules did not satisfy the criterion of protecting policyholders’ future bonus expectations. That was on the basis of V1. Since V2 and V3, which were given in 5.2.3 and 5.2.4 gave weaker reserves, they would be even more inappropriate.