Actuarial Thought on Investment Strategy, 1858-1952

British actuarial thought on investment strategy for life offices began to emerge, somewhat tentatively, in the third quarter of the nineteenth century. Barely a handful of investment papers were published in the Journal of the Institute of Actuaries and Transactions of the Faculty of Actuaries in the second half of the nineteenth century. These papers, however, tended to be written by senior and influential actuaries. They were leading the profession into a new field that would become ever more significant in twentieth century actuarial thought and practice.

There were several drivers for the increasing interest in this topic in the mid-nineteenth century. The experience of the previous decades had imbued actuaries with a growing sense of confidence in their modelling of mortality rates. A long track record of success in predicting mortality rate behaviour and profiting from it had been established. The plagues and epidemics of the seventeenth and eighteenth centuries were, by then, largely distant memories. Mortality behaviour appeared to be governed by laws that were well understood by the actuarial profession. With this mastery came the realisation that the increasingly competitive life assurance landscape could not yield the level of mortality profits that had been earned on assurance business over the previous 100 years. Mortality protection was becoming commoditised.

This was accompanied by an increasing sense that canny investment decision-making could be an important source of profit and differentiation for life assurers. In 1858, actuaries would read in the Journal:

The two principal sources of profit to an Assurance company are the selection of lives and the accumulation of the excess of the premiums at a higher rate of interest than is assumed as the basis of the Company’s operations ... With regard to the [selection of lives], it is impossible not to feel that the margin of profit has very much diminished of late years . This makes the question of the rate of interest obtained from investments so much the more important: it is, in fact, the principal source of profit.[1]

And in 1862:

Of the two main elements on which all life assurance transactions depend — the rate of mortality and the rate of interest — the latter, I think, affords more scope for the exercise of judgement and skill than the former ... there can be no doubt that the amount of interest realised on the assets can be materially influenced by

the degree of judgement and knowledge brought to bear upon the subject.[2]

These statements could be viewed as the very first steps on a long slippery slope that led to twentieth-century British actuaries taking actions which suggested that they regarded geared exposure to financial market risks as an equivalent and equally sustainable form of risk taking to the underwriting of largely diversifiable mortality risk. That was not, however, the vision of the nineteenth-century British actuarial profession. Rather, these early considerations of the asset side of the balance sheet can be viewed as a natural broadening of the actuarial horizon as the skills, experience, confidence and influence of the profession grew. It was also perhaps inevitable in the context of the long-term trend in British life assurance whereby the products’ investment element became ever more significant relative to the protection element.

In the early nineteenth century, British life offices invested in two main asset classes: long-term government bonds (usually consols, which were perpetuities) and mortgages on freehold property (including agricultural lands). Long-term government bonds served two purposes for life offices. First, whilst the emergence of concepts such as duration and interest rate immunisation were many years away, there was an intuitive understanding that the longterm nature of the liabilities demanded that some funds should be invested in assets that paid similarly long-term cashflows. Second, in this era, the possibility of an epidemic or plague, whilst viewed as increasingly unlikely, was still a credible contingency. Funds needed to be readily convertible to cash in such circumstances, and so the liquidity offered by government bonds was valued.[3] The main attraction of mortgages was that they offered higher yields than government bonds—high-quality mortgages typically yielded between 50 and 100 basis points more than long-term government bonds.[4] The mortgages, however, offered neither liquidity nor long-term fixed cashflows, and these characteristics constrained the amounts that were invested in the asset class.

By the middle of the nineteenth century, the concerns about the possibility of epidemics had receded and with it so did the requirement for asset liquidity. The amounts invested in mortgages correspondingly rose. This investment strategy stands up well to modern scrutiny: it was a strategy that paid respect to the liabilities’ interest rate duration and liquidity characteristics and which aimed to obtain additional yield through reducing excess asset liquidity rather than by materially increasing market or credit risk.

The first actuarial paper on life assurance investments was published by Samuel Brown in 1858.[5] Brown was a senior figure in the actuarial profession of the time. He was actuary of the Guardian Assurance Company, a well-established British life office, and he would go on to be President of the Institute of Actuaries between 1867 and 1870. Brown did not propose a doctrine for investment strategy. He restricted himself to providing a commentary on the main asset classes that life assurance funds invested in, and undertook a small survey of the asset allocations of four major British life assurance offices of the time (Eagle Life, London Life, Rock Life and Metropolitan Life). The survey showed that, across the four offices, the amounts invested in mortgages varied between 46 % and 79 % of their total assets.

By the 1850s, low interest rates had rendered long-term government bonds increasingly unattractive—the consol yield had fallen from almost 5 % in 1810 to barely above 3 % by 18 5 5.[6] The 3 % yield level was important to actuaries because virtually all with-profit business had been written using a 3 % interest rate in the premium basis. If funds could not generate a 3 % return, the life offices’ profitability, and ultimately their solvency, was threatened. This fall in yield triggered significant reductions in allocations to consols—by 1858 the allocations had been reduced from perhaps as much as 50 % earlier in the century to between 5 % and 15 % across the four offices. The funds were mainly switched into high-quality bonds (such as colonial sovereign bonds) and debentures (mainly on railways and docks). These asset classes had increased in size and liquidity over the first half of the nineteenth century with the continuing progress of the industrial revolution and the British Empire. Debentures typically offered between 50 and 150 basis points more than consols.[7] They also provided more liquidity and more duration than mortgages, but were accompanied by a more material default risk. This is arguably the first example of a phenomenon that reoccurs in actuarial history: long periods of falling long-term interest rates (for example, in Britain in 1930—1945) tended to compel institutions to bet their way out of their potential unprofitability by increasing investment risk rather than locking in a loss.

This brief sketch of the nineteenth-century asset allocation practices of British life offices highlights a number of asset-liability management considerations that were the natural domain of the actuary: how much asset liquidity was required by life assurance liabilities? How much investment and credit risk should be sought by the asset strategy and how did this choice interact with pricing, reserving and bonus policy? Did the very long-term nature of the liabilities create particular forms of risk that would help determine a desirable asset profile? The history of actuarial consideration of questions such as these is explored below.

  • [1] Brown (1858), pp. 241-42.
  • [2] Bailey (1862), p. 143.
  • [3] Deuchar (1890), p. 451.
  • [4] Brown (1858), pp. 245, 253-54.
  • [5] Brown (1858).
  • [6] Homer and Sylla (1996), p. 182.
  • [7] Brown (1858), pp. 253-54.
< Prev   CONTENTS   Source   Next >