From surveys of old-age insurance beneficiaries conducted by the Bureau of Old-Age and Survivors Insurance, it seems clear that the retirement test is not a major factor in most people's decisions to retire from full-time jobs. These surveys have consistently shown that a great majority of beneficiaries retire either because they are in ill health or because they lose their jobs. Only about 5 percent retire voluntarily while still able to work. Among beneficiaries drawing benefits at a given time, a very large proportion are disabled. In the 1951 national survey of aged beneficiaries, almost two-thirds of the old-age insurance beneficiaries drawing benefits as retired workers at the end of the year reported themselves unable to work. Four-fifths of the aged widows who are beneficiaries said they could not work.

At the same time, the retirement test in its present form does not fulfill its function in an entirely satisfactory way. For example, the present monthly test for wage earners probably tends to some extent to discourage retired persons from engaging in part-time employment. Beneficiaries do not want to accept jobs paying somewhat more than $75 a month if the amounts they would earn would not be as high as the benefits they would lose. Moreover, short-term or seasonal employment during the year can cause loss of benefits for some months even though the beneficiary is essentially in retirement status and his work earnings for the year are relatively low. In this respect the retirement test operates more satisfactorily for the self-employed than it does for the wage earner, since up to $900 of self-employment earnings are allowed in the course of a year before benefit payments are suspended.

The present retirement test also has anomalous results in some situations. At the present time it applies only to earnings in work covered by the program, and not to earnings outside the coverage. This makes it possible for some persons to receive social security benefits in addition to full-time employment earnings by entering noncovered employment. This situation would to a large extent be automatically resolved by the extension of the program to cover most types of employment. But short of completely universal coverage, the problem will not disappear altogether. Another anomaly in the retirement test is that some persons enjoy a "double exemption" of their work earnings-one for wage earnings and one for self-employment earnings. A person who has both wage earnings and self-employment income may earn up to $900 a year from self-employment income and up to $75 a month in wages without the loss of any benefits.

The Social Security Administration believes that the retirement test can and should be redesigned so as to remove anomalies and reinforce incentives to productive work. Methods of accomplishing this are now under study.

BENEFIT AMOUNTS

Similarly, the basic question of whether the benefit amounts provided under old-age and survivors insurance assure individuals and families an adequate measure of security needs to be carefully examined. Recommendations for program changes already made will result in increasing benefit amounts payable in many cases in the future. The extension of coverage, in addition to enabling many people to qualify for benefits who could not otherwise have done so, will also result in the payment of higher benefits to individuals whose working life is split between employment covered at present and that recommended for coverage.

Another provision, recommended in connection with the proposal for extension of coverage, will also increase benefits payable in the future. This recommended provision would permit the omission of the three or four years of lowest or no earnings-depending on the date of coverage extension-from the computation of the individual's average monthly wage under the program. This amendment is designed to remove the handicap in benefit computation which newly covered workers would otherwise suffer because of their late entrance into the system. Equally important, however, it will give to those already covered the advantage of some future protection against the lowering of the average monthly wage because of periods of unemployment, disability or low earnings.

Thus the measures already recommended for improving the program would increase its effectiveness in terms of benefits payable. The recommended extension of coverage, in particular, would in the long run greatly improve the effectiveness of the program in rural areas and thereby reduce the need for public assistance. Further evaluation of the benefit provisions is necessary, however. Study must be given to such questions as what income from other sources and savings aged individuals and survivor families may be expected to have; and what levels of benefits must therefore be paid under old-age and survivors insurance so that retired workers and families that have lost the earner through death will not need to apply for public assistance to supplement the insurance benefits. Analysis is also needed of the results of the present benefit provisions, including the benefit formula, the maximum on annual creditable wages, and the method of computing the average monthly wage.

Continuing study is being devoted to these problems, as well as to other aspects of the program. In a dynamic economy such as ours no social insurance program can be expected to maintain a reasonable relationship to the economy of the Nation without constant study, review, and change.

FINANCING THE PROGRAM

Congress has made clear its intent that the old-age and survivors insurance program be self-supporting, and has set the contribution rates of the program at levels calculated to attain that objective. Congressional committees, however, in determining the contribution rates, have recognized the difficulties involved in making exact predictions of the status and operations of the program that reach into the distant future. For example, the Committee on Ways and Means of the House of Representatives, in its report on H. R. 6000 dated August 22, 1949, page 31, stated:

. . . Your committee has recommended a tax schedule which it believes will make the system self-supporting (or in other words, actuarially sound) as nearly as can be foreseen under present circumstances. Future experience may differ from the estimates so that this tax schedule, at least in the distant future, may have to be modified slightly-either upward or downward. This may readily be determined by future Congresses after the revised program has been in operation a decade or two.

As economic and other conditions change it is essential that new cost estimates for the insurance program be prepared from time to time in order to take into account the latest operating experience and other newly available information. It is to be expected, of course, that any new estimate will differ somewhat from previous estimates.

Normally, several different estimates are prepared, combining different assumptions as to low and high employment, low and high cost factors, and various interest rates. No one of the estimates is considered to be the most probable for the long run. Each simply represents a reasonable set of assumptions tending in one direction or the other. The assumptions used are reviewed by an Actuarial Advisory Committee from outside of the Government. At present the members of the Committee are Reinhard A. Hohaus, Vice President and Chief Actuary, Metropolitan Life Insurance Company; Clarence A. Kulp, Professor of Insurance, University of Pennsylvania; Kermit Lang, Assistant Actuary, Equitable Life Insurance Company of Iowa; and W. R. Williamson, Consulting Actuary.

The result of preparing such a series of estimates is of course a range of possible costs, rather than a single figure. It is clearly impossible, however, to base a schedule of contribution rates on a range of possible costs. Congressional committees, in the past, therefore, in determining the contribution schedules, have adopted the practice of relating the contribution rates to the "intermediate" cost estimate based on the high employment assumptions. This "intermediate" estimate is merely the midpoint between the high and low cost estimates; it is not an attempt to predict the precise cost of the program and should not be regarded as such.

In the fiscal year 1953 the seventh set of actuarial estimates of the cost of the program was issued. These new estimates, based on recent

operating experience and current population data, show the cost of the program to be different from that shown in previous cost estimates. According to the last previous estimates made in 1952 (which were projected through the year 2000 and assumed to be level thereafter) the level-premium cost of the program as amended in 1952, on an intermediate cost basis, was 5.93 percent of covered payroll, which was very close to the level-premium equivalent of the graded contribution schedule. According to the new estimates, the level-premium cost of the program projected through the year 2000 on an intermediate basis is expected to be 6.09 percent; projected through the year 2050 it is expected to be 6.58 percent. Thus the new estimates show a higher cost than the last previous set of estimates.

It should be pointed out that, if somewhat different assumptions had been used in developing the estimates, different figures would have resulted. For example, assumption of a higher interest rate would show lower costs. The interest rate used in developing the level-premium cost figures indicated above was 214 percent; present interest rates are somewhat above this figure.

It should be remembered too, that there are a number of factors which could change the picture considerably-a further rise in interest rates, for example, or less improvement in mortality than has been assumed, or further increases in wage levels, any of which would result in lower costs. (It is worth noting that the experience in this country over the years has been that wage levels have risen, and the general view is that this trend will continue.)

Thus it must be emphasized that no single set of cost estimates is, or can be, "final." Experience over the last 15 years has been considerably more favorable than was anticipated in the cost estimates prepared when the system was established. Despite the fact that the program has been liberalized substantially since it was originally established, the cost of the program as a percentage of payroll for 1955, for example, is expected to be very close to, and perhaps somewhat below, the cost estimated for the year 1955 in the early days of the system.

Public Assistance

The number of persons receiving assistance continued to go down during the fiscal year, as it has since 1950. In June 1953, for the first time since the Social Security Act was passed, fewer people were receiving assistance than were receiving benefits under old-age and survivors insurance. For most children of incapacitated and separated parents, however, and for most disabled and blind persons, there is no insurance coverage, and public assistance is the primary support for needy persons in these groups.

Although more aged persons received insurance benefits than old-age assistance as early as 1951, the total number of recipients under all assistance programs at the beginning of 1953 still exceeded the total number of insurance beneficiaries by nearly a million. However, the continuing decline in the number of assistance recipients, coupled with a substantial increase in insurance beneficiaries, had reversed the position of the two programs by the end of the fiscal year 1953. The shift in programs for the aged was especially marked. In June 1953, 32 percent of the population aged 65 and over were drawing insurance benefits, and less than 20 percent of the aged were receiving assistance.

Thus, while old-age and survivors insurance is beginning to fulfill its long-term role as the basic income maintenance program, public assistance helps to provide a minimum level of living to needy persons who are not eligible for this social insurance program or whose insurance benefits and other resources are too low to meet their minimum needs. Today, more than 5 million persons, whose eligibility has been determined by State and local welfare agencies under the provisions of the States' own laws, are dependent on public assistance for the primary essentials of daily living.

As more retired persons and the dependents and survivors of insured workers become eligible for social insurance benefits, the proportion of persons receiving assistance because of disability, or because they are children deprived of parental care, has risen sharply. This shift in the composition of the assistance caseload emphasizes the need for further development of social services in public assistance agencies as well as the need for effective working relationships with other community agencies providing health and rehabilitative services. Federally, this shift of emphasis has been reflected this year in closer working relationships between the Bureau of Public Assistance and other programs of the Department, such as vocational rehabilitation, public health, and child welfare services, and also in more cooperative work with outside agencies, particularly national agencies carrying standard-setting responsibilities for institutions.

Case Load and Expenditures1

About 200,000 fewer persons received public assistance at the end of the 1953 fiscal year than a year earlier. From 5.5 million in June

1 For old-age assistance, aid to dependent children, aid to the blind, and aid to the permanently and totally disabled, data on payments now include payments to the suppliers of medical care, and data on recipients include persons who received no money payment but on whose behalf vendor payments for medical care were made. In previous issues of the annual report, data on average payments included only cases receiving money payments and the amount of such payments. Data for earlier years have been adjusted for comparison with 1953 data. For general assistance the data continue to exclude vendor payments for medical care and the number of persons on whose behalf only such payments are made.