« AnteriorContinuar »
The English Reg.-Gen. said in his 5th Rep. (1843):
The census has been taken with great regularity in the U.S.-it was part of the constitution of the States that it should be so taken for political purposes-and the ages are properly distinguished; but abstracts of the registers of deaths have only been pub. in Philadelphia, New York, Boston, and haps a few other of the older cities. No National Life Table can therefore be formed until, in addition to the census, a careful system of registration of deaths is introduced.
ELLIOT'S TABLE (1857).—About 1857 a table of mort. was constructed by Mr. E. B. Elliot, based upon the census returns of 1850 for the State of Massachusetts. This table was pub. in the Proc. of the Amer. Asso. for the Adv. of Science, held at Montreal, in 1857. It was understood to form part of an orig. series which had been prepared for the New England Mut. Ins. Co. of Boston, U.S., from extensive and trustworthy European and American data. The following abstract of this table, comparing its results with those of the Carlisle Table, will be sufficient for our present purposes:
It appears that from about age 5 to age 15, lower rates of mort. prevailed in Massachusetts than is generally the case in European communities; that from age 15 to various ages between 35 and 50, the Massachusetts rates are much higher than, after which they again fall somewhat below, the European rates. Under the age of 5 years, the mort. in Massachusetts seems more intense than in Europe generally; from 3 to 15 it approaches closely to that of Sweden; from 17 to 45 nearer to Belgium, though higher; and after 45, nearer to, but lower than, the average English rates. As a whole the mort. of Massachusetts is better represented by that of England than of any other European state.
MUTUAL BENEFIT TABLE (1858).—In 1858 the directors of the Mut. Benefit Life Ins. Co. of New Jersey, pub. the mort. experience of the office for the 12 years it had then been in existence. Its members had averaged 4,000, until the later years, when they had been about 5,000. In the observations, lives were carefully distinguished from policies; and pains were also taken to ascertain with accuracy-(1) the number of persons who, at each age of life, took or renewed a pol. by payment of prem.; (2) how many died within the year the policy had to run. Each age therefore included new or recent selections, as well as those of earlier date, added together in each successive year of the Co's. existence. The experience had been very favourable, much of which was due to the effect of recent selection. The following is an abstract of a mort. table deduced from the experience named, prepared, we believe, by Mr. Sheppard Homans. The number of actual deaths are contrasted with the number of "probable deaths" as shown by the Carlisle and the Experience Tables.
HOMANS' TABLE (No. 1) -In 1859 there was pub. by the Mutual Ins. Co. of N. Y. an elaborate report on the affairs of that Co., which had been prepared by Mr. Sheppard Homans, its actuary. The general features of that report it is not our present purpose here to deal with. They will be discussed, with the details of some previous reports on the affairs of the same Co., by its previous actuary, Professor Gill, under UNITED STATES; where we propose to give a pretty complete outline, not only of the growth and development of ins. in all its branches, but also of MORT. OBS. made in that great country, and of many incidental matters. Forming part of the report under notice was a TABLE of MORT. constructed by Mr. Homans, upon the mort. experience of the Mutual. It is of this table and the facts immediately connected with it that we have now to speak. The table is based upon the experience of the office during a period of 15 years ending 1858, that is, through the entire duration of the office. The number of lives existing at the date of observation was 10,387, of which 9,485 were "whole life," 572 "seven years" pols.; 174 for other short terms, and 156 were "endow." pols. It will not fail to be noted that the number of lives here under observation was greater than the entire pop. of either Northampton or Carlisle, at the respective dates of the observations which have been named after them. The following explanation given by Mr. Homans must not be overlooked: With the exception of this single assumption, that the office age is the real age, the results which are now presented are deduced from the most rigorous calculation. The number of lives exposed to mort. are carefully separated from the number of policies, and no care or labour has been spared in making the various observations and deductions as complete and accurate as possible.
The amount insured under the above policies was (taking 5 dols. to the £) £6,094,739, while the bonus add. thereon amounted to £656,446, making a total of £6,751,185. We now reach the table, which Mr. Homans calls "an adjusted rate of mort. according to the general experience of the Mut. Life Ins. Co. of N.Y. for the 15 years ending Feb. 1, 1858." We propose here only to give an abstract of the table, because we shall have to give it in detail hereafter, when it shall have passed through all the "adjustments," graduation," and other crucial tests, which it is the fate of such tables to encounter. The orig. is placed for purposes of comparison alongside the table on which the operations of the Co. were based: that test we have no purpose in here repeating. The quinquennial results were as follows:
No, out of which
After making various comparisons with other existing tables, Mr. Homans remarks: We here find that our adjusted experience gives a higher expectation of life at all ages than the co's. theoretical table, and also higher than any English table below the age of 70, except the "Friendly Sos.," according to Neison, which is uniformly higher at all ages. At and above the age of 70, the expectation by several English tables is higher than that shown by our own experience. The favourable results in the experience of this co. may be attributed to two principal causes, namely, the influence of selection, by which persons in sound health only are admitted as members of the Co., and the fact that among these members are to be found so many married men as heads of families.
MEECH'S TABLE (1860).—Mr. Meech, a well-known American actuary, has deduced a mort. table for white male life, from the results of the U.S. census of 1860, of which the following is an abstract :
Mr. Hine, contrasting the results of this table and the Pennsylvania Tables of 1814, with that of 1789, says:
It will be noticed from these figures that the most marked improvement has taken place in young life. The lapse of time has acclimated the American nation; has increased the efficiency of medical and sanitary science, and the volume of national wealth has secured greater protection and care to the American youth. From 40 onwards the last century go hand in hand, the sturdy forefathers living out their full term of years, sustained by their hardy constitutions; the descendants of the third and fourth generations preserving through the refinement of the age the longevity of the past.
In May, 1859, a general convention of life underwriters was held in N.Y. for the purpose of exchanging views and devising means for collecting statistical information. The committee on vital statistics, composed of some of the leading actuaries, was appointed, one of its purposes being the preparation of a table of mort. based on the experience of American Life Ins. This committee met just one year afterwards (May, 1860), and reported progress. Has anything since been done?
In 1865 Professor Elizur Wright prepared for the Ins. Commissioners of Massachusetts, of whom he was and had been one of the chief for many years, a table gathered from the experience of all the American offices up to that date. We have not that table at hand, but shall prob. have an opportunity of dealing with it hereafter. The table used for the State valuations in Massachusetts is the EXPERIENCE (Old) with int. at 4 p.c. The following note will furnish an outline of the results of Prof. Wright's table:
The mortuary experience of 14 cos., reported to the Commissioners of the State of Massachusetts, which was equal to about 50,000 years of life, prove that life in the U.S., at ages between 31 and 55, is not subject to as high rate of mortality as similar returns prove it to be in Gt. Brit. and Germany. It was also shown, by facts in possession of the cos., that at middle ages, life is subject to a lower rate of mort. in the States than in Europe; while, at the two extremes, at younger and older ages, it is the
In 1866 the experience of the Mutual Benefit of New Jersey was brought down to that date.
AMERICAN EXPERIENCE (1868)-HOMANS' TABLE (No. 2).—Soon after 1864 Mr. Homans again made an investigation into the affairs of the Mutual, which now covered an experience of 20 years. The results of this investigation, we believe, have not been made public in the same manner as the former ones. In the absence of such exact details, we gladly avail ourselves of some observations of Professor C. F. McCay (Spectator, July, 1870):
As this co. was growing rapidly, each new period of five years doubled its experience. The deaths were now more numerous than at Carlisle; they were reported with more accuracy; the numbers of the living and the dying were both more reliable; and the table obtained from the observations was constructed with more ability and skill by Mr. Homans than by Mr. Milne. It thus surpassed the Carlisle in every particular. Compared with Dr. Farr's, it was founded on fewer observations, but these were on ins. lives, and not on the general pop.; which made it more valuable for the uses of a life co. If we compare it with the Actuaries' [OLD EXPERIENCE], it had two advantages. It was founded on American experience, and not on policies, but on lives. The disadvantages were, however, serious; the observations were fewer, and the duration of the ins. shorter; and to weigh these properly, we will consider the two tables carefully, and examine particularly the grounds on which they may claim our confidence.
We shall have occasion to notice some of Mr. McCay's criticisms hereafter.
The table referred to by Professor McCay is the new table of Mr. Homans, first made public in 1868, after it had received all the careful consideration, adjustment, and elaboration, which its talented author had, alike by study and experience, qualified himself to bestow upon it. It will be seen that the expectation shown by this table differs somewhat materially from that shown by the table of 1859. Here is the new table: AMERICAN MORT. TABLE-HOMANS, Numbers Numbers Expectation
Numbers Numbers Expectation living. dying. of life. 718 36.73
85,441 720 35'33
723 33'92 83,277 726 33'21 82,551
81,090 737 31.07
80,353 742 30°35
We have said the table was made public in 1868. The manner in which that was accomplished was by the legislature of the State of N. Y. making it the standard for their State valuations. To this end the Ins. Law of 1853 and 1866 was amended by the add. of various provisions, of which the following only are material for our present purpose:
SEC. 13. It shall be the duty of the Superintendent of the Ins. Department to arrange the information contained in the statements required in the last section in a tabular form, or in abstracts, and to prepare the same for printing in his annual report to the legislature. It shall also be the duty of the said Superintendent, at least once in every five years, and ann. in his discretion, to make valuations of all the outstanding policies, additions thereto, unpaid dividends, and all other obligations of every American L. Ins. Co. transacting bus. in this State; and for the purposes of such valuations, and for making special examinations under the 17th section of this act, and for valuing registered life and other policies under chapter 708 of the laws of 1867, the rate of int. assumed shall be four and a half p.c. p.a. and the rate of mort, shall be that established by the American Experience Table, in which table the expectation of life and the numbers of living and dying at each age from ten to ninety-five out of one hundred thousand persons living at age ten, are as stated in the schedule hereto annexed. The Superintendent may, in his discretion, vary the above standards of int. and mort. in cases of cos. from foreign countries, and in particular cases of invalid lives or other extra hazards. The superintendent may also, in his discretion, value policies in groups, use approximate averages for fractions of a year and otherwise, and calculate values by the net, the actual, or the gross premiums or otherwise, deducting in cases of gross valuations, from the gross value of future premiums, one-sixth thereof for future expenses and contingencies.
The act was passed 6th May, 1868, and took effect immediately. The former standard of valuation in that State had been the ENGLISH TABLE, with 5 p.c. int. Since the adoption of the AMERICAN EXPERIENCE TABLE by the State of N. Y., the States of MICHIGAN and MISSOURI had adopted it, each in 1869.
The individual offices were not called upon by this act of the legislature, or any other, to alter their tables of rates, etc. They may still, for their own purposes, retain or adopt any table they please for their internal working; but all their operations will be from time to time measured by the AMERICAN EXPERIENCE TABLE. Some of the offices have already adopted the table. The Ins. Commissioners of Massachusetts have not yet adopted it; but it has been under their consideration. A belief is entertained in many quarters that it will ultimately become the National Standard Table of the U.S. The elaborate tables of the N. Y. Ins. Department [ACTUARIAL TABLES] are all based upon it. On the subject of these tables issued by the department, as also on the adoption of Mr. Homans' table by the State Legislature, the following passage from a circular from the then Superintendent, Hon. Wm. Barnes, should be preserved; the more so as it is the only utterance of his upon the latter question which we can recall :
To persons experienced in the construction of tables of this nature, it is entirely unnecessary to speak of the many elements of error which perpetually gravitate towards the printed page. More than ordinary precautions to insure accuracy have been taken in the preparation, printing and electrotyping of these tables. The herculean nature of the work must be apparent to any one, but more especially so to the mathematician and actuarial expert. The actual results in net values according to the new standard of int. and mort. adopted in this State are especially gratifying; and the "American Experience" at four and a half p.c. int. will commend itself to officers, act., cos., and legislatures, as the most practical and best standard yet compiled for the various purposes of life ins. Whenever American experience shall have been more thoroughly tested by our cos, having reached their maximum and decline, it may be advisable to recollate and tabulate the statistics of all of them; but this period is too remote for present purposes.
This table of Mr. Homans, as was to be expected, has been the subject of a great deal of criticism by American actuaries. Professor Wright has said of it:
I think the rate of mort. Mr. Homans proposes is somewhat better gradulated than Dr. Farr's table. It is free from some of the imperfections and anomalies of that scale; but the general effect of it is to reduce the initial value and to make the grade of the self-ins. a little steeper. The gradulation is hardly as perfect, I think, as that of the Actuaries' scale, though I have not examined it sufficiently to test that point. I have examined it for only one rate of interest-4 per cent., and with regard to a limited number of single prems., a few endows., and ordinary life prems. I think it preferable to the Carlisle table, and preferable to Dr. Farr's; but for some of the ages, I think it is affected with the peculiar experience of the Mutual Life Co. which we all know was estab. by a subs. of policies to be ins., including rather advanced lives and a good many not very good risks; so that the losses, for the first fifteen years, as compared with the expectation by the Carlisle Table, over the age of 54, were in excess; whereas, taking all the lives together, they did not lose so much as was to be expected by that
Again, and on a separate occasion :
That standard is what Mr. Homans calls "American mortality;" but I presume it is derived chiefly from the experience of the Mutual Life, so far as it is not purely hypothetical. That has been adopted as the standard for valuation in the State of N.Y., together with a rate of int. of four and a half p.c. The principal change affecting the value of policies is the half p.c. add. int. The mort. being a little more favourable, brings the values down somewhat more than the English; but as the rate of int. is reduced only one-half p.c. instead of one p.c., it does not produce so great a change as the former rule, Dr. Farr's rate of mort, with five p.c. int. It makes a pretty important change in paid-up insurance, reducing the value about ten p.c.
Mr. White says:
Unless I am misinformed, the table prepared by Mr. Homans covers really an experience from 20 to 74 only in his own co., and the numbers down to 10 and up to 95, are run out on what he calls "general principles," and not actual from experience; and inasmuch as they do not agree with Farr's or with the Carlisle table on these older ages, it would seem probable that the "general principles" have led him into an error in regard to those ages. One thing more. All these tables, so far as I have observed them, seem to run the duration of life up to 103 or 105, which is in accordance with the experience shown by our own census tables; but Mr. Homans' table runs them all out at 95, and leaves none living, making the duration of life apparently 8 or 10 years less.
Prof. McCay (Spectator, July, 1870) says:
Mr. Homans' American Table has been much esteemed. It has been adopted by many life cos. not only for the prems., but for the valuation of pol. and the distribution of surplus profits;. ... and as American experience is attractive, and the actual experience of large and ancient life offices form the very best basis for the future expectation of mort. among the lives ins. by our cos., this table presents great claims to the favour and approval of American actuaries. It is well adjusted, admirably constructed, and has added much to the reputation of the distinguished act. of the Mutual of N.Y. With all these recommendations in its favour, it is nevertheless the duty of every one to consider its claims with fairness and impartiality.
The learned Prof. then makes a series of comparisons between this and other tables, of which we shall avail ourselves in speaking of some of those tables. He says down to 1858 the oldest life ins. in the Mutual office was only in his 78th year; and the number over 65 was less than seventy.
Referring to this table running out at age 95, he says:
In almost every table there are some survivors after the age of 95. Out of 100,000 at 10, the Actuaries has 37; Dr. Farr has 155; and the Carlisle 356; but Mr. Homans has none. At the age of 77 his rate is below the Actuaries'; five years later it is 5 p.c. higher; ten years later the excess is 25 p.c.; and 15 years later it is 57 p.c. At 77 it is 3 p.c. below Dr. Farr's, at 82 it is 7 p.c. above; at 87 it is 37 p.c. higher, and at 92 more than a 100 p.c. higher. At 76 it is below the Carlisle; at 81 it is 14 p.c. higher; at 86 the excess is 37 p.c.; and at 91, 87 p.c. The rapid progress from the age of 75 upwards is even in excess of the Old Northampton, and of every other good table; and as the Mutual Life could not have had more than 4 or 5 deaths at these higher ages, it is difficult to guess where Mr. Homans could have obtained any authority for these high rates at the older periods of life. There is nothing in the mortuary reports of our cities that justify them; and the private experience of almost every one will tell him that some of his acquaintances have lived beyond 96, which is the extreme of human life according to this American table.
Mr. D. P. Fackler thus spoke of it at the National Ins. Convention, 1871 :
The differences between rates and valuations on the American Experience Table, and the same on the Actuaries', are quite trifling-about as much as the variations that would be due to a difference of one quarter p.c. in the rate of int. I think the American Experience Table is a better index of the mort. of American cos. than the Actuaries', and should be very glad to see it adopted [by the States generally). It should, however, I think, be better graduated so as to correspond to its author's orig ideas. If it were made on a basis of 10 millions of lives instead of 100,000, the graduation would be perfect. As it is now it is quite irregular. There are such differences in the successive results obtained from it that one is often led to think that he has made some considerable error. . . . A table for State purposes ought to be perfectly graduated, so that the numerous tables based on it would be susceptible of check by the method of differences.
Mr. Edwin W. Bryant, one of the most accomplished actuaries in the U.S., and now the actuary of the N.Y. Ins. Department, says:
The table shows a lower rate of mort. than the English (No. 1), or than the Experience table from 30 to 77. At the other ages a higher rate than the last-named table. In this particular it is confirmed by the experience of the Mutual since 1858, as well as before, and by that of several other American Life Cos.
The first complete valuation of all the policies existing on the books of the Life offices transacting bus. in the State of N.Y. was made by the Superintendent of the Ins. Depart ment in 1870. It embraced over 650,000 separate policies in existence at the close of 1869. This was made upon the AMERICAN EXPERIENCE TABLE.
There is one especial feature which remains to be noticed, and it is that in the State