the selection exercised by the employee would not be as severe as in Option II, the British offices life table (OM) was selected as the basis of this charge. Compound interest at 4 per cent, which is to be guaranteed by the Federal Government as an aid to the plan, was used in computing the figures for Options II, III, and IV. Census Bulletin No. 12, The executive civil service of the United States, giving statistics as of July 1, 1903, was used as a basis for determining the amount of annual appropriation by the Federal Government to provide annuities for old employees. As these data were not considered sufficiently recent, the Census Bureau has compiled new statistics as of July 1, 1907, which will be used this winter in revising the figures already obtained. The chief requirement in determining the annual appropriations was that the amounts be the maximum required to provide these annuities. For this reason, although the right of an employee to an annutiy from the Government is to be relinquished upon withdrawal or death, the withdrawals were disregarded and deaths alone considered in obtaining these figures. The method pursued is best shown by taking a particular age attained, as age 45 in Group I, the age of retirement being 60 years. The total of the annuity payments based upon 1 per cent for each year of service of the average salary for the past ten years was obtained for each age, and then discounted by the probability of dying before reaching the age of retirement. As the ages given in the census bulletin were for last birthday, and, consequently, the average age understated six 760 months, and, as the annuity at retirement is to be payable quarterly, the factor · was considered to be a safe approximation. To obtain the amount of annuity payments for employees now aged 45 during the second year and after retirement, the above result was multiplied by the factor 161 and so on each year to the end of the 1601 745 mortality table. When these calculations had been completed for all ages and groups, the results were arranged and the total appropriations necessary to provide annuity payments in 1908, 1909, and later years were obtained. The mortality table selected as the basis of these computations was the American experience table of mortality. This decision was reached after consideration of the effect of the conditions involved in this part of the proposed plan upon the amounts of appropriations and all other information obtainable which would throw light upon the character of the class under observation from the standpoint of probable longevity. As stated before, withdrawals, other than by death which take place before the retirement age is attained, will tend to decrease the amount of annuity payments each year. These amounts will also be diminished because of the fact that there are many employees at the older aegs, such as old soldiers, who have been in the service but a short time and who are receiving small salaries-virtually pensions--for unimportant work, who will be allowed to retain their positions rather than be retired upon an annutiy which would be but a small percentage of their annual salary. For these reasons, and because of the standard of the mortality table adopted, the appropriations determined would seem to be a safe outside estimate of the amount necessary to provide the annuity payments. To establish and maintain this savings system, although it might require considerable clerical work, should not prove a difficult task. From a table giving the amount of $1 per month at 4 per cent compound interest for certain terms of years, the amount of accumulated savings at retirement could be easily determined for any age at entrance. At each increase in salary the table could be entered for the amount of monthly increase for the term of years remaining before attaining the age at retirement, and the value thus obtained added to the amount of accumulation which was being provided for under the old salary. A card could be written for each employee upon which could be kept his account, and, in addition, other data which could be used in obtaining the mortality experience, rate of withdrawal, and other valuable information in regard to the members of the classified civil service. Because of the large number under observation the experience could be safely subdivided to show the mortality of employees in the District of Columbia, as distinguished from the other main division of the service, elsewhere; also the experience in certain branches of the service, such as postal railway clerks and city letter carriers, which are affected by conditions peculiar to these classes of employees. The experience thus obtained would not only prove of considerable statistical value, but would serve as a guide in future valuations or adjustments of the plan. Total.. APPENDIX E. Statement submitted by Jacob W. Starr, showing separations from competitive positions by branches of the service and by fiscal years from 1898 to 1905, inclusive. Branch of service. 75 327 Departmental service Post-office service. Customs service.. Internal-Revenue Service.. Government printing service. Indian service.. Total. 933 2, 717 334 58,991 1, 497 3, 479 383 64, 622 10, 255 17, 683 2, 169 30, 107 468 2,017 280 71,098 428 2,231 294 81,596 3,810 11, 428 2, 039 17,277 83 124 67 4,935 191 167 80 5,398 1,131 964 564 2,659 47 68 36 2,043 26 51 25 2,057 1,101 830 213 2,144 46 140 33 3,862 49 144 48 4,043 1.106 916 270 2,292 67 626 103 11,301 81 402 82 12,171 501 2,397 570 3,468 15 421 9 1,863 28 489 9 1,920 247 2,731 69 3,017 1,659 6,113 862 154, 093 2, 300 6,963 921 171, 807 18, 151 36,949 5, 894 60, 994 APPENDIX F. EXECUTIVE ORDER OF JANUARY 31, 1902, PROHIBITING ATTEMPTS OF GOVERNMENT EMPLOYEES TO INFLUENCE LEGISLATION. "All officers and employees of the United States of every description, serving in or under any of the Executive Departments or independent Government establishments, and whether so serving in or out of Washington, are hereby forbidden, either directly or indirectly, individually or through associations, to solicit an increase of pay or to influence or attempt to influence in their own interest any other legislation whatever, either before Congress or its committees, or in any way save through the heads of the Departments or independent Government establishments in or under which they serve, on penalty of dismissal from the Government service." In response to a letter from the United States Civil-Service Retirement Association, requesting permission for the formulation of a bill for the retirement of superannuated employees, the President, under date of February 21, 1902, replied, in part: "Permission is restricted to acts pertinent and necessary to the single object proposed, namely, to obtain information, upon which legislation for the retirement of superannuated employees may be based." APPENDIX G. ANALYSIS, BY MR. FRED BRACKETT, OF THE UNITED STATES TREASURY DEPARTMENT, OF THE ASSESSMENT SCHEME FOR THE RETIREMENT OF CLERKS, BASED UPON A 3 PER CENT ANNUAL ASSESSMENT ON SALARIES, A 25 PER CENT ASSESSMENT ON ALL INCREASE OF SALARY, AND AN ASSESSMENT OF ONE-TWELFTH OF SALARY ON ALL ORIGINAL APPOINTMENTS. In estimating the approximate cost of any retirement scheme there should be a definite basis as to number of clerks in the classified service to proceed upon. Census Bulletin No. 12, issued July 11, 1904, gives the following varying numbers: • Excluding those paid by the piece and those reported without salary, or for whom salaries were not reported. Page 252, Report of Civil Service Commission for 1904, gives a grand total of 154,093 "competitive positions." A recent report made by the Director of the Census to the retirement association (see p. 107) gives a total of 128,098 employees, with salaries aggregating $114,025,497. Reports received from various departments and offices (see p.) show that $133,913,363.84 is paid this year for salaries to classified employees. There is therefore an apparent difference in the two reports as to salaries of $19,887,866.84. As the report of the Director of the Census did not include 613 special agents of the Census Bureau receiving small pay (not more than an average of $200 per annum), nor 1,587 pieceworkers (see p. 20, Bull. 12, supra), this would account for quite a large sum of the difference, but taking— The maximum amount paid to Census special agents.. We have a total of..... Which is less the difference noted by. $122, 600 2, 380, 500 2,503, 100 17, 384, 767 19, 887, 867 The final difference noted may be the pay of rural free-delivery employees, but there is no report to that effect. It seemed to me impossible to harmonize in any satisfactory way these conflicting statements. I therefore made several computations, as follows: A. Based upon 22,273 clerks in the District of Columbia, serving twenty years and retiring at age 70, not deducting deaths before 70, but estimating them. 38257-08-9 B. Based upon 22,273 clerks in the District of Columbia, and deducting deaths before 70. C. Giving product of assessments, annual retired pay at $600 and $500, on basis of 6,705 times the number of clerks in Table B. D. Showing available fund, payments, balances, and increments from insurance, based upon foregoing tables. E. Based upon 149,333 clerks and employees in United States, retiring at age 70 without regard to term of service, giving survivors and years of salary (265,907) which, multiplied by $500, gives the cost for thirty years, but does not give the yearly balances and interest. F. Giving the probable actual cost of retiring all clerks now in service when they shall have reached age 70, without regard to term of service, giving yearly payments, balances, and compound interest on balances at 3 per cent per annum, covering the period from 1906 to 1935-thirty years. It may properly be asked why so many tables are presented, and in reply I have to say that they are the product of an evolution process, for which varying conditions are responsible. Then, too, it is desirable to try various methods in order to prove that the deductions are correct when comparison of the various tables are made. The first table covers only that portion of the clerks in the classified service who are employed in the District of Columbia, and the tables following up to and including Table D are predicated on the first two. The proportion of the clerks in the District of Columbia (22,273) to the entire force in the United States (149,333), given in Table 65 of Census Bulletin No. 12, is 6.705, and in various tables the estimates are increased that much. An examination of Table E, however, will show that the product of such a multiple is too great for certain ages (see columns 8 and 10), for the proportion of aged clerks is greater in the District of Columbia than elsewhere. The result, however, gives more than the greatest possible maximum of clerks ever likely to be retired of those now on the rolls. The total of years service of clerks on Table E to be paid for under the first tables is 310,984, but under the amended list, column 8, it is only 265,907, which covers the actual conditions as the ages are divided to-day. The average of deaths per year in the entire service, as given by reports of Civil Service Commission for 1903 and 1904, is 859, or 0.00575 per cent per annum. In thirty years, therefore, we should lose in entire service 25,770 by death, provided the conditions remain as they are to-day, but as clerks grow older the average loss by death will increase. In fixing the average of retired pay at $500 I am certain I have reached the maximum. The average salary of 149,333 clerks, whose salaries aggregate $122,000,000 would be $817 per annum. The average salary in statement of Director of the Census (Appendix H) is $890.14, while I have estimated the average salary at $1,000 and average retired pay at $500, thus giving a large increase, probably equal to one-tenth more than the maximum amount required. If the total amount paid is greater than $122,000,000 per annum, the product of a 3 per cent assessment would be greater than I have estimated it, while the average salary under any circumstances is not likely to exceed $1,000. In estimating the income to the retirement fund I have taken $122,000,000 as amount of annual expenditure for salaries, on which I have reckoned a 3 per cent assessment, producing $3,660,000. There is an average of 8,082 deaths, resignations, and removals per year, and I have allowed an equal number of new appointments at $720 each, from which we receive $60 each, or one-twelfth of the whole amount. This produces $484,920 per annum, but as probably 5 per cent of the original appointments would be of $900 grade, we must add $6,060, or 404 by 15, which would give us $490,981 from appointments. From promotions we should receive an average of $120, less 5 per cent for appointments to vacated grades, which might not involve promotions. Each promotion ought to average $114, as follows: $120 less 5 per cent ($6) $114. $114 by 8,082=$921,348. $490,981+$921,348=$1,412,329. $1,412,329+$3,660,000-$5,072,329, the amount of annual proceeds from three sources. The average "expectancy❞ of life for persons aged 70 is 8.48 years; that means that two persons of that age will probably live in the aggregate 16.96 years. One may live but two years, while the other will complete sixteen years (after reaching age 70). The terms lived by each may vary but the sum of the terms will be 16.96 years. This is the rate fixed by life insurance companies. I have allowed in the beginning (in order to facilitate computations) a term of nine years after age 70 and have subsequently deducted the difference between 8.48 and nine years, or 0.0577 per cent. Resignations from the service average 5,606 persons per annum, therefore, if any of the 168,180 persons that will resign from the service within the next thirty years, are of such age as will bring them to age 70 by the year 1935, they must be deducted from the number (36,827) listed for retirement up to and including that year. If but 1,000 of the 168,180 named are included among those whom we have treated as retired, and living in retirement for nine years, we will be able to reduce our retirement pay $4,500,000, less 0.0577 per cent, or $4,240,350 below the amount estimated on Table E. It is proper to suggest, in conclusion, that if those who can give but little time to the investigation of this subject do not care to traverse all the tables given they will find in Table F the essential facts governing the cost of retirement scheme outlined in the Brownlow bill. The estimates cover a period of thirty years from 1906 to 1935, inclusive. I am satisfied that the plan will be self-sustaining for forty years as well, and for all time under the same conditions, as the maximum number of clerks would be on retired list about the year 1947. If allowed 8.483 years, instead of 9 years life as average after retirement, would reduce amount 0.0577 per cent, or................ $43,478, 400 15,649,800 27,828,600 23,190,500 155,492,303 8,971,906 Probable actual cost (from 1906 to 1935). 146,520,397 |