1. After the discussion on Sir George (then Mr) Epps's paper on ‘Superannuation Funds’, the President in proposing a vote of thanks said: ‘It is of the highest importance that we should exchange views with one another and try to come to a common opinion on matters of practice so that the world at large will bring against us no charge of inconsistency in the advice which we respectively give to those who consult us.’ In 1921, when Epps's paper was presented, actuaries were faced with the position that current yields on new investments were very much higher than had been anticipated in earlier valuations, while most of the older funds had suffered very heavy depreciation in values of existing investments. Today the circumstances are reversed. The rate of interest which can be earned on new investments is very low, but there has been substantial appreciation in values of existing investments. It is in the hope of provoking an exchange of views on the best method of dealing with this present situation that this paper has been written.

(1) The proper method of taxation of annuities has been a subject of contention for many years, and it seems a suitable time to discuss what are the correct principles to be applied.

(2) It seems illogical that life annuities should be taxed on a different principle from annuities payable for a fixed period not dependent on life. It also seems wrong that, whereas a person who has saved capital on which to live in old age may draw on the capital from time to time without being taxed on it, the same person if he buys an annuity so as to equalize the payments throughout the remainder of life has to pay tax on the capital content of the annuity.

(3) Various methods(1) have been devised in other countries as practical measures giving some allowance for the capital content of annuity payments, but those methods that I have seen suffer to my mind from being founded on arbitrary rules rather than on principle. It should be noted that the practice of each country is bound up with the definition of taxable income and the treatment of wasting assets in that country. In Great Britain it is possible so to word an annuity contract that part of the annuity is not taxable, but personally I think the question of principle should be faced, and not avoided by rewriting the contracts in some other form.

Mortality statistics have now been available for the population of Great Britain for just over a hundred years and during this period there has been a steady and sustained decrease in mortality at all ages and in both sexes, though the extent of the decrease has been progressively less with increasing age.

The main part of the improvement in mortality during the latter half of the nineteenth century was due to improvements in sanitation and hygiene and above all the provision, at any rate in cities, of a reasonably safe water supply. Cholera which occurred in severe epidemic form in London between 1832 and 1866 had disappeared completely well before the beginning of the twentieth century. The widespread epidemics of typhoid fever which occurred as a result of water supplies becoming contaminated with sewage ceased when adequate precautions were taken to safeguard the sources and transmission of water in cities. It is true that small and localized epidemics of typhoid still occur, but these are all found to be due to contamination of food or water by ‘typhoid carriers’. A recent epidemic in Wales, in which some 150 cases of typhoid occurred, was shown to have been caused by the eating of ice-cream which was handled and sold by a man in perfect health but who, following an attack of typhoid many years previously, had continued to pass virulent typhoid organisms in the stools. Fortunately the notification of cases of the disease enables the public health authorities rapidly to track down carriers in this type of epidemic.

In this paper, everyday actuarial data have been examined by means of standard statistical techniques. The data in question have been drawn from the publications of the Continuous Mortality Investigation Committee, and the methods of statistical analysis have been applied to investigate the nature and extent of the heterogeneity which may exist in those data.