Equities and With-Profits: Smoothing and Guarantees (1952-1976)
Chapter 3 noted how actuarial thinking on life office investment strategy began a journey in the 1920s, away from the highly conservative nineteenth- century tenets of Bailey. By the late 1940s, the Dalton ‘cheap money’ era had further prompted British life offices to move materially into equity and property investments in an effort to maintain their investment yields in excess of those that had been guaranteed in with-profit premium rates. This asset allocation trend continued unabated into the 1950s and 1960s. By 1951, 20-25 % of life office assets were invested in equities and property;1 by 1961 this allocation had grown to over 30 %.2 Life offices had a mixture of both with-profit and without-profit liabilities. The without-profit business was effectively backed by bonds, and so the with-profit business’s equity/property exposures were significantly higher than the above figures. By 1981, Frank
Dodds (1979), p. 176. Dodds (1979), p. 50.
Redington noted that for some British with-profit funds the equity/property allocation was 100 %. In the decades following the Second World War, British with-profits business had transformed into something different, and the actuarial thought of British life actuaries had to catch up.
There were a couple of key (and related) questions that required actuarial answers. How did substantial equity/property investment impact on life office solvency assessment? And how should with-profit bonuses be distributed in the presence of significant equity/property investment? Most actuarial energy was initially invested in the second of these questions. In the 20 years following the end of the Second World War, equity markets performed remarkably well (especially in nominal terms). This was driven by both a high rate of dividend growth and a fall in equity market dividend yields. Between 1950 and 1960, dividend pay-outs more than doubled and market values increased by a greater amount. Long-term interest rates steadily increased during this era (from a low of 2.6 % in 1946 to 6.8 % by 1966). Life office solvency was therefore generally not felt to be a major concern. Conversely, there was a pressing need to deal with the embarrassment of riches that with-profit funds had found themselves holding. This was made particularly pertinent by the increasing competition from other forms of savings vehicle such as unit trusts.
The crux of the bonus distribution challenge was that the reversionary bonus system did not recognise any equity/property capital gains until they were crystallised in asset book values, and the writing-up of book values prior to the sale of the asset was generally viewed as actuarially distasteful. In the competitive environment, this was a fundamental issue. The only way equity returns naturally fed into the assessment of distributable surplus was through the dividends received. By 1960, the ‘reverse yield gap’ had emerged: equities generated a lower (initial) income than risk-free bonds. Distributable surplus would therefore be reduced by increasing equity allocations, even following a period when the equity holdings had materially outperformed bonds in terms of market value growth. Moreover, during this time, market practice was such that actuaries were under significant pressure not to reduce bonus rates other than in the most exceptional circumstances. This further incentiv- ised the actuary against releasing unsustainable rates of equity market value increase into distributable surplus. As a result, with-profit pay-outs started to significantly lag comparable unit trust pay-outs, and life office estates increasingly bulged with undistributed and unrecognised equity market value gains.
This situation was unsustainable—the with-profit bonus system was not fit for the purpose of distributing surplus from equity-dominated asset allocations. By the mid to late 1950s, British life actuaries started to recognise that a more flexible bonus system would be necessary in order to distribute equity capital gains equitably amongst the policyholders whose premiums had generated them. It was crucial that this was done in a way that did not create the expectation that such distributions could be predictably delivered every year. This required two new devices: a special or terminal bonus that was distinct from the reversionary bonus and that did not have the same policyholder expectations for stability attached to it; and an acceptable method for writing up book values of equity/property assets to market values in order to transfer some gains into distributable surplus. It is difficult to overstate how counter to traditional British actuarial principles these developments were. But several papers were published between 1959 and 1976 that progressively established the new orthodoxy.
An Institute paper presented to Staple Inn in 1959 by the Equity and Law actuary Norman Benz was an important early landmark in establishing these radical ideas in mainstream actuarial thought. Those ideas were given further support in another Institute paper written by PE. Moody in 1964. These papers reflected the growing recognition that intergenerational policyholder equity and competition in the long-term savings market both demanded that policyholders receive the capital gains generated by their investments. This naturally led to an increasing embrace of the contribution principle of bonus distribution that had been well-established in the USA for so many decades. For example, in an influential 1968 paper, Skerman argued that ‘the distributed profits should be allocated between policyholders in relation to the contribution which their policies have made to them’.
However, in the 1950s and 1960s (and beyond), there was still a prevalent view amongst British actuaries that with-profit funds should deliver something more stable than what was directly generated by the market returns of their invested premiums. Both Moody and Benz advocated a smoothed distribution of equity gains to policyholders. The concept of using the with-profit fund’s estate to deliver a smoothed equity return to policyholders would remain a permanent central tenet of British actuarial thought. The 1976 report of the Institute of Actuaries’ Working Party on Bonus Distribution with High Equity Backing stated that ‘the general view of the working party was that conventional with-profits business should smooth out the fluctuations in investment return’ and that ‘the estate should be used to ... level out fluctuations in experience .
Benz and Moody explored how to smooth the distribution of equity capital gains to with-profit policyholders. In both papers there was a particular wariness about distributing equity capital gains that arose through falls in yields rather than through growth in dividend pay-out levels. Both Benz and Moody advocated the idea of using an actuarial estimate of the future dividend growth rate to determine a steady writing up of asset book values and hence release to distributable surplus, irrespective of how the market yield was behaving. In his paper, Skerman agreed that market value changes arising due to changes in equity dividend yields should not enter into distributable surplus. In the Staple Inn discussion of Skerman’s paper, A.C. Stalker advocated the use of the historical average dividend yield to determine equity book value write-ups. Whilst such approaches could be substantially dislocated from market value movements, the implications of the Fama and French empirical equity studies that would emerge some 25 years later arguably offer some degree of intellectual support for these ideas. But such an approach was not without practical difficulty: did with-profit funds have the capital to underwrite larger distributions than those implied by market value gains when equity market yields were unusually high? Did policyholders understand that they may not get a material portion of ‘their’ equity capital gain when markets performed very well?
Whilst Benz and Moody advocated a quantitative method for developing a smoothed distribution of the equity capital gains, Skerman did not welcome such a mechanical approach to valuing assets. Instead, he argued that the release of equity gains into distributable surplus should simply be a matter of actuarial judgement as part of the actuary’s role in the ‘proper steering’ of the with-profit fund. Such an approach placed a huge amount of discretion in the hands of the actuary, even by British actuarial standards. In such a system, an actuarial discipline that balanced policyholder security against intergenerational equity and prudence against competitive marketing pressures would be critical to the long-term success of with-profits as a popular and sustainable long-term savings vehicle.
Life office practices evolved concurrently. Leading British with-profit funds started to use terminal bonus as an additional degree of freedom to deliver more equitable pay-outs to with-profit policyholders in the late 1950s. The Prudential, with Frank Redington as its chief actuary, was a leader in this innovation, introducing terminal bonuses in 1956. The terminal bonus was an almost universally adopted mechanism amongst British with-profit offices by the end of the 1960s. However, such was the success of equity markets during this era that a real risk emerged of the terminal bonus becoming established in the minds of policyholders and salesmen as just as stable and predictable as the reversionary bonus. The ‘reasonable expectations of policyholders’ were enshrined in UK legislation by the Insurance Companies Act of 1974, and this further concentrated actuarial minds on bonus policy. But the extreme equity market experience of 1973 and 1974 prompted several offices to reduce terminal bonuses and one or two offices even reduced them to zero. This suggests the actuary was able to use the terminal bonus mechanism broadly as intended.
What of actuarial thought on the appropriate level of equity allocation in with-profit funds and its implications for solvency? When equity investment was first introduced to UK with-profit funds in the 1920s and 1930s, this topic received little consideration. Equity allocations were small and solvency buffers were usually large, and there was little discussion of what upper limit should be placed on the equity allocation. Raynes’s influential papers of 1928 and 1937 said virtually nothing on this question. In the closing remarks in the discussion of Raynes’s 1937 paper, something is finally mentioned in relation to liabilities when George Recknell, the actuary at Keynes’s National Mutual, suggested:
Concerning the percentage of ordinary shares which should be purchased there was something to be said, as a very rough rule, for investing on more or less conventional lines to the extent of the funds needed to support the contractual liabilities, while leaving the surplus free to invest in ordinary shares and real estate.
Haynes and Kirton’s 1952 paper had similarly advocated Recknell’s form of asset-liability matching: fixed liabilities (sums assured, including accrued reversionary bonuses) should be matched by bonds, and the excess funds could be invested in equities and property. A more ambitious discussion of the equity asset allocation in a with-profit fund was developed in 1957 in an influential Institute paper written by J.L. Anderson and J.D. Binns, who were, respectively, the actuary and investment secretary of Scottish Widows at the time. Anderson and Binns suggested that with-profit funds should invest more in equities than the maximum level implied by the above principle that guaranteed benefits were matched by bonds:
Suppose that k is the maximum depreciation on present market values which is envisaged. Estimate the value of a function we shall call the “remainder” R, i.e. the excess of the total assets at market value over the liabilities on a gross premium basis ... It can then be argued that R/k can safely be invested in equities provided the balance of the fund is reasonably matched.
The full-matching approach advocated by Recknell, Haynes and Kirton was equivalent to setting k = 1, i.e. assuming equities could fall to zero. By setting k to less than 1, the with-profit fund opened up the possibility that poor equity market performance would result in the office failing to meet its guaranteed liabilities: of course, the probability of this possibility depended on the assumed size for their k parameter (and on how quickly the fund could shift out of equities if it wished to as market values fell). Their logic could be viewed as a precursor to a modern ‘portfolio insurance’ strategy, which is typically couched in the logic of Black-Scholes’s dynamic hedging (though Anderson and Binns gave no indication that it was their intention for with-profit funds to reduce their probability of insolvency through very frequent rebalancing of the equity allocation to the new level implied by market value changes).
Anderson and Binns noted that the Dow Jones Industrial Index had fallen by over 80 % during 1929-1932, but dismissed this period as exceptional, and suggested using a k parameter value of 0.6. In the discussion of the paper that followed at the Institute meeting, their proposal excited little alarm, except for one lone voice, S.H. Cooper, who noted:
If he had correctly understood the ... remainder, it followed that any depreciation in excess of the limit envisaged would absorb the whole of the free reserves of the fund ... and would encroach upon the cover required for the basic liabilities of the fund. In that context it seemed to him that they should be concerned with possibilities rather than probabilities, and he had been rather surprised to see the authors dismiss the American experience of 1929-32 as exceptional.
Despite such occasional protestations, life office equity/property allocations continued their inexorable rise in the following years. By 1976, allocations to equity and property were around 50 % of life office assets. This excited little actuarial concern. Conventional actuarial wisdom embraced equities as an inflation-proof long-term asset which life offices were uniquely equipped to hold on behalf of policyholders due to their large estates and the ability it gave them to protect policyholders from equity markets’ irrational short-term market movements. In concluding his history of the first 200 years of Equitable Life, Ogborn in 1962 mused:
What seems important at the time of writing is the greater freedom of investment which the life offices have felt both necessary and desirable, in the current economic climate, for the purpose of giving their policyholders some share of profits from the expansion of industry and some protection against the possible effects of further inflation, should it recur.
By the 1960s, British life actuaries had learned to feel quite sanguine about the financial market risk exposures that their businesses were underwriting in ever-increasing scale. This contrasted with the technical sophistication that was developing in the pricing and management of financial market risk outside the actuarial world. An interesting next 40 years would ensue for the profession.
-  Redington (1981), p. 378.
-  Benz (1960).
-  Moody (1964).
-  Skelman (1968).
-  Kennedy et al. (1976), p. 12.
-  Skelman (1968), p. 72.
-  Kennedy et al. (1976).
-  G.H. Recknell, in Discussion, Raynes (1937), p. 505.
-  Anderson and Binns (1957).
-  Anderson and Binns (1957), p. 125.
-  S.H. Cooper in Discussion, Anderson and Binns (1957), p. 143.
-  Dodds (1979), p. 50.
-  Ogborn (1962), p. 256.