PARREN'S LIFE - CONTINGENCY TABLES.* NiNE years ago, the growing interest in

the question of life-assu- rance induced us to give an historical précis of the subject, and a summary of its principles and practice as then generally esta- blished. Since that time, articles, popular compendiums, pamph- lets, and mathematical putlicatiens on the matter, have been rife enough. Two really new and important points, however, alone have been settled during the interval. One has been determined by Mr. Farr • -whose researches into the true value of life have shown the too low idea that insurance-offices have formed of the subject, and the consequently extravagant premiums they have been charging. 'The other refers. to the legal advantages the offices have taken-of the public, so as to be able to dispute -the validity of a policy pretty, much at their pleasure ; the public exposition of which is due to the founders of the,London Indisputable Com- pany, or their Actuary , Mr. Bebertson. A point of great import- ance is still-to be solved; and that is the real value or probable expectancy, of impaired lives---of persons who cannot be reported as sound or as being free;from the symptoms of some disease tend- ing to shorten life, though otherwise in a fair state of health. So far as we know, Mr. Burt, " Secretary to a Life Office," was one of the first to call attention broadly to this point. In the Life-Contingency Tables before us, Mr. Farren has investigated the subject in a systematic and scientific way, and thrown some cu- rious light upon, a portion of it.

The great point, indeed, is left directly untouched—what is the real expectancy of life in "unsound" lives? But this seems at pre- sent unattainable, from the absence of sufficient data : we shall have to buy experience in this as in other matters. Mr. Farren's tables, however, appear to show that the difference of risk between select and impaired lives is less than might be supposed, and very much less than the premiums of the insurance-offices would lead people to imagine. The opinions of Mr. Milne and Mr. Gompertz, given incident- ally on a cognate matter, seem to infer that the offices, notwith- standing all their care, are often imposed upon by bad lives; which gives an unfair increase to the average mortality as shown by the experience of insurance-companies. It is as just, perhaps, to ascribe it to the uncertain nature of human health, to the risk and exposure to which strong men often subject themselves, and to. their greater danger in attacks of acute disease. It is a remark of Celsus, that a man is never more obnoxious to illness than when in the highest health and condition ; for life is never at a stand-still, and as he cannot advance he must retrograde. The tables Mr. Farren has constructed exhibit life in several conditions. One shows the proportionate annual mortality of: select lives during the first year of their selection, before the life' has had time as it were to deteriorate ; the basis in this case being the deaths during the first year of insurance. Opposed to this table is the average mortality, taken from the malepulation

of the country without selection : and a very curious exhibition ' it is, well worth careful inspection and consideration. At twenty- one, the difference per cent in favour of selection is as .5891 to .8122. The proportion in favour of selection increases steadily up to forty, when it is .8588 against F3195 of the male population at large. This proportion then begins to diminish, till at fifty • Life-Contingency Tables. Part I. The Chances of Premature Death and the Value of Selection among Assured Lives. By Edwin James Farren. Published by Smith and Elder.

there is not much greater 'difference between the two classes of lives than there was at twenty-one. From fifty to sixty this difference further diminishes, being least towards the. middle of the decade. Above sixty, the greater stamina, of the select men. begins to tell considerably in them favour, though after all the. real difference in the mortality per thousand is less than it might appear to many from the decimal representation.

" From the comparison of the two tables it appears,. that although se- lection' has a decided influence in reducing the ensuing year's mortality, yet that premature deaths, even among; lives recently selected, must be ex- pected to occur at all ages, and to be irreducible below a certain minimum whenever large numbers are e,oneerned The value of selection during the first year has beenta ward. off from the life-assurance societies during that year about the Marring number of additional claims,;* that might have been expected to occur if the assurances had been issuectupen lives among the male population at larg,e,, without discrimination, as to hvalfh. Between ages Among 1,000 new assurances granted 21'and 30; about'3 claims, f (9,1347-6,2214 31 - 40 4. — (11,7947-7,4866

.......... 41..- 50 — 4 — 15,2106-11,4201 51- 60- — 2Z9492-21,3524

6L- 70 — 7— 4818662-41,4718 — 5,000 21 — 70 —20 — (107,9554-87,9523) Or-the- claims were less to the extent stated„orl.Oper 5,000-assurances during the. first year after. selection,. than if such ha& been issued at the rate of 1,000 to each decennial of age among the male population indiscriminately,. supposing that a fair average of such lives could have been secured."

Hitherto Mr. Yarren has been dealing with choice lives, in the first or most ficvourable years of their selection. In a third table he exhibits RAH insurances for twenty-five consecutive years ; showing the annual mortality the annual number of policies dis- continued, the mortality per Loveland, and other points. He also considers the general- mortality among insurers in further tables, always with a reference to the common mortality, and always with the same result-that the (Er-11110n, of select lives over the general ulation, is mueh less than would. a priori. be inferred:. The

erences vary according,to the particularmode of working, them.,, same being more, some less favourable.. Tn. the. following: sum- mary, which seems a fair mode of taking the case, since it embraces the ages within which perhaps. the_hulkof insurances range, the approximation is wonderful.

Ages

fallinwing each decennial of Equitable Society. Male Population. Difference.

admission. Average Duration'. Average Duration:

35 29'10 29*83 ..... ......... - 0'73

0•39

16*63 ..... 0,35

10-86- -I- 0'22

These results, and the numerous facts collected, by 3f.r; Yarren, seem to show the safety with which unsound lives- orpersons not free.from symptoms of disease "-tending to shorten' might be' insured at a moderate addition to. the present rates of premium. Still they do not establish. the average mortality among that class-. The experience of the Asylum and other offices which receive un-- select lives, is the only true datum ; for it is obvious that if this class-does not bear a pretty large proportion to the population of the. country, it would be a riskful thing to insure them as a body, 4.111/0. their. greater ratio of deaths would operate but little on the mortalityof the. country at. large. At the seine time, when the greaterhygienic and medical advantages, of the class who-insure, over those of the bulk of the community, are-taken into account; as well as the greater general! care and prudence of the- delicate, itis. probable-that this risk will tarn- outmatch less than it seems;, at all evaints, that ages. at twenty-sevenneed not Be. charged the premium of . sixty..

131

tot