The Retirement Test

As part of the 1958 Social Security Admendments, the Congress directed the Secretary of Health, Education and Welfare to conduct a study, with recommendations, to address the operations of the Social Security "retirement test." This report, issued in March 1960, is the response to that legislative mandate.

March 23, 1960

THE RETIREMENT TEST
UNDER OLD-AGE AND SURVIVORS INSURANCE

A Report on a Study Called for by the Committee on Ways and Means of the House of Representatives, 85th Congress, Second Session, in House Report No. 2288 on H.R. 13549, the Social Security Amendments of 1958


Department of Health, Education, and Welfare
Washington, D.C.: 1960

Hon. Wilbur D. Mills Chairman,
Committee on Ways and Means
House of Representatives
Washington 25, D.C.

Dear Mr. Chairman,

I have the honor to transmit to you a report on "The Retirement Test Under Old-Age and Survivors Insurance." This report was occasioned in part by the request of your Committee in its report accompanying H.R. 13549 the Social Security Amendments of 1958 (85th Congress,, 2nd Session, H. Rept. 2288). You will recall that the report asked the Department to study certain aspects of the retirement test.

Although in the opinion of the Department the retirement test is an essential part of the old-age and survivors insurance program, the test in its present form is subject to a number of valid criticisms. The Department has taken this occasion to make a study, not only of the particular problem the Committee expressed concern about, but also of other aspects of the retirement test. The Department presents in this report a proposal that it has developed to improve the retirement test so as to eliminate or greatly reduce the causes for criticism.

We are convinced that this is the best solution that can be devised to the problems involved; but we have not concluded upon whether it is desirable to provide for an increase in taxes in order to make enactment of the proposal possible. The Department plans to continue its evaluation of this question

Sincerely yours,
Secretary

THE RETIREMENT TEST
UNDER OLD-AGE AND SURVIVORS INSURANCE

The report of the Committee on Ways and Means of the House of Representatives on the Social Security Amendments of 1958 (House Report No. 2288) contains the following section calling for a study of the retirement test under the old-age and survivors insurance program:

"The committee has asked the Department of Health, Education, and Welfare to study certain aspects of the present test of retirement which seem to the committee to have questionable results. The present test is basically on an annual basis but under one of the provisions benefits are nevertheless paid for any month in which an individual earns $80 or less ($100 or less under the bill) and does not render substantial services in self-employment. Thus a person may have very high earnings in a single month and yet get benefits for the remaining 11 months in the year. We have asked the Department to consider possible changes in this provision."

In response to this request, the Department of Health, Education, and Welfare has conducted a comprehensive study of the retirement test, with particular attention to various changes in the retirement test designed to meet the problem the Committee had expressed interest in. This report sets forth the findings of that study.


WHY THERE IS A TEST OF RETIREMENT IN OLD-AGE AND SURVIVORS INSURANCE

The basic purpose of the old-age and survivors insurance program is to provide benefits for workers and their families when the worker's earnings can be presumed to have stopped or to have been substantially reduced as a result of his retirement, disability, or death. Since it is not reasonable to presume that all workers retire or suffer a significant reduction in earnings upon attainment of age 65, the program includes a "retirement test"--a provision intended to restrict the payment of benefits to those among the aged who can be presumed to have suffered such a loss.

If the retirement test had been removed from the program in June 1959, about 1.4 million people age 65 and over (working people and their dependents) who had not been getting benefits up until then could have immediately started to get benefits. Many of these people are working full time and earning as much as they ever have in their lives; the payment of full benefits to them would serve no socially useful purpose. And the removal of the test would not help the vast majority of beneficiaries now on the rolls, who are unable to work or to get jobs.

Payment of full benefits to all of the aged who are still working would be very costly, both in the immediate future and in the long run. Benefit costs in 1959 would have been increased by about $2 billion if the retirement test had not been in effect for that year. In terms of long-range costs, the removal of the test would increase the level-premium cost of old-age and survivors insurance by 1 percent of taxable earnings--an increase of 12 percent in the estimated level-premium cost of the old-age and survivors insurance provisions (8.38 percent of taxable earnings). [1]

[1] The disability insurance part of the program is estimated to cost an additional 0.35 percent of taxable earnings.

The Social Security Board, the Federal Security Agency (predecessor of the Department of Health, Education, and Welfare) and the Department have always recommended in the past, and the Department recommends now, that a test of retirement be retained in the old-age and survivors insurance program.


THE PROVISIONS OF THE PRESENT RETIREMENT TEST [1]

Under the present retirement test a beneficiary gets all of his benefits when his earnings are $1200 or less in a year (this is the concept referred to in the report as the "exempt amount"). Anyone making $1200 or less is, in effect, presumed to be retired. Beneficiaries may get benefits, therefore, even though they have a significant amount of part-time work on a regular basis or have relatively high earnings for part of the year. (In fact, there undoubtedly are a few people at even this relatively low level of $1200 a year who are working full time and earning as much as they did before age 65.)

Ordinarily, a beneficiary has a check withheld [2] for each $80, or part of $80, in excess of $1200 in earnings (the concept referred to throughout this report as the "unit of excess earnings" or the "excess unit"). This means that a beneficiary gets at least one benefit if his earnings are $2080 a year or less. [3]The reason why the number of benefit payments that can be made in a year varies as earnings vary between $1200 and $2080 is to avoid a sharp line below which all benefits would be payable for a year and above which none would be payable. If the test were not graduated, it would not be uncommon to have the payment of $2000 or more in benefits depend on a few dollars of earnings. The law also provides that no matter what his annual earnings, a beneficiary gets a benefit for any month in which he neither earns wages of more than $100 nor renders substantial services in self-employment (this provision is referred to as the "monthly measure of retirement").

[1] The limitation on the amount of earnings a beneficiary may haveand get benefits, although designed primarily as a test of retirement for the aged worker, also applies to beneficiaries receiving dependents' and survivors` benefits under the program.

[2] Where the dependents of a retired worker are getting benefits based on his earnings, those benefits are withheld for any month for which his benefit is withheld.

[3] The figure $2080 is the result of adding to the $1200 exempt amount 11 times $80. Thus at least one month's benefit is payable when earnings for a full year are $1200 plus $880 ($80 for each of the 11 months).

The retirement test does not apply to beneficiaries aged 72 or over; after that age, benefits are payable regardless of the beneficiary's earnings. (This provision was enacted in recognition of the fact that a few people--particularly the self-employed--continue working to a very advanced age. Without this provision these people might never get any benefits even though they had paid contributions longer than most other beneficiaries.) The test applies to the earnings of beneficiaries in covered and non-covered work in the United States and covered work outside the United States. (A special provision applies to beneficiaries working in non-covered work outside the United States so that levels of earnings in foreign countries need not be equated with those in the United States.)


HOW THE RETIREMENT TEST DEVELOPED

The reasons why the test has taken its present quite complicated form will be easier to understand if the considerations that led to the various changes that have been made in it through the years are reviewed.

From 1940 through 1950 the test of retirement applied only to earnings from covered employment. During those years work as an employee in commerce and industry was, generally speaking, the only employment covered by the program. The test was entirely on a monthly basis; the beneficiary got a benefit for any month in which he earned less than $15 in covered employment.

Effective in 1951, when the self-employed were brought under the program, the test of retirement for the self-employed was put, for the most part, on an annual basis. (This was necessary because it is practically impossible in most cases for a self-employed person to compute his earnings on a monthly basis.) Specifically, it was provided that a person with self-employment earnings of $600 or less for the year could get benefits for all months in the year no matter what his earnings were in any single month.

One part of the test, however, was placed on a monthly basis even though the earnings were figured over the whole year. No matter how high his annual earnings, a self-employed beneficiary could get a benefit for any month in which he did not render substantial services in his business. This latter provision served three purposes; First, it placed the self-employed beneficiary on a par with the wage earner in that he could receive a benefit for any month in which he did not work or in which he worked very little. Second, it allowed the payment of benefits to a self-employed beneficiary for months in which he did no work in the year in which he retired, even though his total earnings for the year were above the exempt amount by reason of work done before retirement. And third, the provision allowed payment of benefits to a person whose self-employment income came, not from work in operating the business, but rather from the investment he had in the business.

Two important criticisms of the test soon developed. First, there was criticism on the basis that the self-employed person could work, say, for three months, earn up to the annual exempt amount, and still get benefits for the whole year, while the wage earner who worked in three months and had the same total yearly earnings had three months' benefits withheld. (The 1950 amendments provided that a person could not get a benefit for any month in which he earned over $50 in covered wages.) Second, a beneficiary who had both self-employment income and wages was in an unwarrantedly favorable position because he could meet the two tests separately; that is, he could have earnings from self-employment for the year of as much as the annual exempt amount, and also have wages in every month amounting to as much as the monthly exempt amount, and still get all of his benefits. The 1954 amendments removed these two anomalies by providing that earnings from self-employment and wage employment would be combined for retirement test purposes and by providing a test with an annual exempt amount ($1200) for both the self-employed beneficiary and the wage-earner beneficiary.

The 1954 amendments also provided that the wage earner could get a benefit for any month in which he earned no more than $80 (this amount was changed to $100 beginning with 1959) regardless of his earnings for the year. This provision was included partly to avoid situations where a worker would not be able to get benefits under the 1954 amendments although he could have gotten them before. The provision also solved the problem of finding a way to pay benefits for the rest of a year when a worker retired in the middle of the year after his earnings were over the exempt amount. (Without the provision, a worker who retired in July, for example, after earning $2500 in that year could not get benefits for any part of the year.)


THE NUMBER OF PEOPLE AFFECTED BY THE TEST

At the end of June 1959, 11.1 million of the more than 15.5 million people age 65 and over in the United States were either getting old-age and survivors insurance benefits or could have gotten them if the breadwinner of the family had not been working. Of the 11.1 million, 4.9 million--44 percent--were age 72 or over and thus did not come under the operation of the retirement test. Of the remaining 6.2 million, an estimated 3.2 million did not have any earnings at all. Generally the retirement test does not affect the benefits these people get because most of them cannot work or cannot find work. [l] Another 1.0 million are expected to earn less than $900 in the year and also generally would not be affected by the test.[2] We have, then, a total of 9.1 million--about 4/5--of the 11.1 million eligible for benefits who in all likelihood are not directly or concretely affected by the retirement test.

There are, then, about 2 million of the aged who are directly affected by the retirement test. The 2 million fall into 3 groups. One group, numbering about 300,000, is the group of people who are getting full benefits and who are earning just under $1200 a year (between $900 and $1200). A sizable proportion of these can be assumed to be deliberately holding their earnings to $1200 or Just under that amount in order to get full benefits. For these people the test is clearly operating as a deterrent to work. The second group, also numbering about 300,000, is the group of people who are earning between $1200 and $2080 and under the present retirement test are getting some benefits for the year, the number of benefits being determined by the specific provisions of the retirement test. [3] The third group, numbering about 1.4 million, is composed of people who are making over $2080 in a year and therefore are generally not getting benefits. Some of these people are working full time and earning about as they did before they attained age 65; others may have suffered significant reductions in earning power, although not enough to reduce their income below $2080. In general, the existence of the retirement test in its present form prevents these people from getting any benefits for the year.


[l] In the last survey of beneficiaries conducted by the Bureau of Old-Age and Survivors Insurance, about 70 percent of those not working said that they were not able to work.

[2] This group may include a few who are deterred by the provisions of the test from earning more than they do.

[3] These estimates exclude those who came on the rolls during the year and earned amounts falling in the indicated range, in most instances before "retirement."

RET chart 1

THE COMMITTEE'S REQUEST

The Committee asked the Department to consider possible changes in the provision of law under which a person may have very high earnings in a single month and yet get benefits for the remaining 11 months of the year. The situation the Committee is concerned about grows out of the provision in the law setting up a monthly measure of retirement. This is the provision under which benefits are not withheld for any month in which the beneficiary neither earns wages of more than $100 nor renders substantial services in self-employment, regardless of what his total earnings are within the year.


Possible Elimination of the Monthly Measure of Retirement

In seeking a solution for the problem raised by the Committee the Department considered first whether the monthly measure of retirement should be eliminated, but came to the conclusion that it should not.

A major function of the monthly measure is to make it possible to pay benefits to a retired worker beginning with the first month of his retirement. Without a monthly measure of retirement, if a person retired from full-time work at the end of June, for example, after earning more than $2080 in a year, he could not get benefits until the following January. It would not seem reasonable for the program to require that a beneficiary go through the first several months of retirement without getting benefits. The benefits should start as soon as possible after earnings cease, when the need for the benefits arises; the monthly measure of retirement is the provision in the present law that makes this possible.

Moreover, since people move in and out of employment after reaching retirement age, the problem is not confined to the year of initial retirement for each beneficiary. For example, take a person who has been on the benefit rolls for a year, getting $100 a month, and now has a chance to take a job in January, and does so. He thinks he will be able to keep on working and he spends most or all of his earnings for current living expenses. In June of the following year his employer goes out of business and he is unable to get another job. Now, because in the first months of the year he has earned, say, $2100, he cannot get benefits for the next 6 months, and will not have earnings either.

And the problem exists not only for people who leave employment in the middle of a year, but also for those who return to work during the course of a year. Take for example a beneficiary who has been retired from a regular full-time job and has gotten benefits of $110 a month for a year or so. He is not satisfied to be idle and would like to work. An opportunity comes up for him in July of his second retirement year to take a full-time job paying $85 a week. If he takes it he will earn $2210, and therefore will have to return the $660 in benefits that he has already gotten or have his benefits withheld at a point when he is no longer earning. This will seem quite unfair to him, since he was not working during the months when he got the benefits. Moreover, the need to repay the benefits will be a significant barrier to his taking the job.

Removing the monthly measure of retirement would prevent the payment of benefits in cases where under present law some benefits are paid to the beneficiary in a year even though he may have had high earnings for a few months of the year. But removal of the monthly measure would make it impossible to pay benefits promptly upon retirement, and therefore would prevent the program from carrying out a major one of its objectives. Accordingly, the Department recommends that a monthly measure of retirement be retained in the program.


A Separate Retirement Test for People with Relatively High Earnings

The Department believes there is only one feasible proposal for preventing benefit payments where a person has relatively high earnings in a few months of the year. The Department is of the opinion, however, that the way in which the law operates at present is to be preferred, and it does not recommend adoption of this proposal. The proposal would add, on top of the present retirement test, a provision that no matter how little he worked, a person could not get full benefits if his earnings for the year were above some fairly high figure--for example, $4800 (the present maximum on taxable and creditable earnings). One form such a proposal might take would be to withhold one monthly benefit for, say, every $400 of earnings above $4800 in a year. To take account of the peculiar circumstances of self-employed people, it might be provided that the proposal would not apply to a self-employed person unless he actually had done some work during the year.

Here are some examples illustrating how the proposal would operate:

1. A farmer moves into town after turning the operation of his farm over to a paid manager. He gets a profit of $6000 from the operation of the farm. If he did absolutely no work in connection with operating the farm he would get benefits for the full year just as he does under present law. If, however, he helped with the work during the spring plantings he could get benefits for only nine months; three benefits would be withheld because of his income from the farm even though he worked in only one month.

2. A movie star works on a picture during 2 months of a year and earns $10000. He does not work in the other months of the year. Under present law he could get benefits for the 10 months in which he did not work. Under the proposal he would get no benefits.

3. An operator of a mail order business turns the management of the business over to his son and moves to Florida; his business yields him $7200 in earnings. Like the farmer in the first example, if he did not work at all he would get benefits for the whole year. If he did do any work at all in connection with the business he would get benefits for only six months; six benefits would be withheld.

As the examples show, a self-employed beneficiary with a business that produces a high income would, under the proposal, have to completely disassociate himself from the operation of his business in order to get full benefits. Under the proposal as much as a whole year's benefits would depend on whether the beneficiary did any work at all.

When a self-employed beneficiary has high earnings and yet works in his business very little during the year most of his earnings probably come from his investment in the business rather than from the work he performed. Thus the effect of the proposal in many cases would be to withhold benefits on the basis of investment income. But the wage-earner beneficiary can get his benefits even though his invested savings yield a large income. There is no good reason why the self-employed person should not be able to have income from his investment in his business without losing his rights to benefits. [1]

[1] Income from investments in real estate, stocks, bonds and the like is not taxed for social security nor credited toward benefits, nor does it count for purposes of the retirement test. If benefits were withheld from people who have income from investments and other forms of savings, the program would discourage personal savings.

More important than the undesirable effect just described is the fact that the proposal would add complexity to an already too complex provision in order to deny benefits to a very small group of beneficiaries. And the beneficiaries would be people who really are essentially retired. Generally speaking, any person who would be prevented from getting benefits solely by this provision is likely to be a person who has retired within the normal meaning of the word, and there is real justification for paying benefits to him for months in which he does not work. Because the proposal would deny benefits to only a very small group, it would not save any appreciable amount of money; there would be virtually no saving in the long-run costs of the program.

Because of the serious disadvantages associated with it, the Department does not favor the adoption of this proposal nor of any proposal that would have a similar effect.

Conclusion

The Department of Health, Education and Welfare recommends that no action be taken to remove the monthly measure of retirement or to put an additional earnings limitation on top of the present retirement test. Acceptance by the Congress of this recommendation would mean that the situation that the Committee has expressed interest in--the payment of benefits to a person even though he may have had very high earnings for a few months in a year--will not be changed. The Department believes that, for the very few retired people who come out of retirement for a short time and earn substantial amounts of income, the most appropriate action is to suspend their benefits for the months in which they actually worked. This is the approach taken in the present law and in fact the approach that has always been taken under the law. Ever since the program started paying monthly benefits the retirement test has been so framed that a person could get a benefit for any month in which he did not work, regardless of how much he worked or earned in any other month; and generally, over the years, this approach has been accepted without disfavor. Moreover, it is the only approach that is consistent with the treatment that should be, and is, accorded to a person in the year in which he first retires (that is, it is consistent with the payment of benefits to a person for months after retirement even if he has earned large sums in that same year before he retired).


THE RETIREMENT TEST AND THE QUESTION OF INCENTIVES

The Problem

Powerful incentives to work for people age 65 and over now exist. Generally, earnings from work make possible a higher standard of living than most people can manage to obtain for themselves in retirement, since earnings generally are much higher than benefits. There are many intangible satisfactions in work--interesting activity, and the feeling that the man has a contribution to make to the economy and is an important part of the community. Actually, most beneficiaries who are not working either are not well enough to work or cannot find jobs.

Nevertheless, it is unquestionably true that many older people would do more work than they do if the provisions of the retirement test did not operate so as to reduce the net addition to their income as a result of working. This is particularly true of people who are retired from their regular jobs and who would like to find same part-time or less demanding work to do. To the extent possible, retired people should be encouraged to accept jobs, earn money to improve their economic situations, and make a contribution to production and the national economy. Under present law it frequently happens that a beneficiary finds himself in a situation where, while he will be better off if he does a given amount of work than if he does no work at all, he would be still better off if he could have managed to restrict his work to a point where he would have earned somewhat less than he did. Thus the retirement test causes beneficiaries to restrict their earnings to lesser amounts than they could and would like to earn in order not to suffer a loss in total income.

An example or two may help to clarify the effect that the test has on incentives to work. Take the case of a beneficiary getting $1200 a year in benefits and faced with a choice between a job paying $1800 a year and one paying $1200. If he takes the $1800 job he will be only $1000 better off than if he does not do any work [1]; but if he takes the $1200 job his increase in income for the year will be $1200. Obviously, he would do better financially to take the $1200 job, although he might make more of a contribution to the economy, and feel better about his activities, if he could afford to, and did, take the $1800 job.

[1] The $600 of earnings in excess of $1200 causes the withholding of eight benefits of $100 each-$800. Therefore, the worker has $1800 in earnings and $400 in benefits, or a total of $2200 for the year--$1000 more than the $1200 in benefits he could have gotten if he had not worked at all.

Or take the case where a beneficiary has occasion to earn just over the $1200 exempt amount and lose a full month's benefit as a result. (Usually if he does earn just over 1200 it is through inadvertence or as a result of demands made upon him by his employer.) Whether the beneficiary actually does do the extra work and loses a month's benefit, or refrains from doing the extra work in order to get full benefits, the test is operating in an undesirable manner, since it either discourages him from work or penalizes him for working. And this situation can occur not only at the $1200 point, but at every one of the breaking points from $1200 to $2080. [1]

As a final example, take a man who with his wife has a benefit income of $180 a month (the maximum under present law) and is offered a job paying $3000 a year. In this situation it is impossible for the family to lose in income as the result of the man's work; but the addition to his income if he takes the job and does $3000 worth of work will be only $840. [2] If he takes the job he will be somewhat better off financially than if he does not. And if he is chiefly interested in maximizing his income, or if the job is particularly interesting or not too demanding, he may take it in spite of its not being very profitable. On the other hand, he may well think that the extra $840 in income does not make it worth his while to take the job. This is a situation in which it might be highly desirable, for the economy, the beneficiary, and the old-age and survivors insurance program, for the man to take the job and make whatever contribution he can. Yet the present law greatly reduces his incentive to do so.

[l] By "breaking point" is meant the point at which the beneficiary loses an additional month's benefit as a result of the operation of the $80 unit of excess earnings; that is, if he earns more than $1280 he loses 2 months' benefits, if he earns more than $1360 he loses 3 months' benefits, and so on.

[2] The beneficiary loses benefit income of $2160--12 months' benefits at 180 a month. Since he would have had $2160 had he not worked and since he has $3000 as a result of working, the net addition to his income is $840.

It is generally agreed that provisions of law that operate to discourage people from working as much as they can and want to work are, in that respect, undesirable. Even when a person has attained an age that is generally regarded as the time when retirement from work is taken for granted, it is probably better for him to continue active, so far as his health will permit; both the individual himself and the economy as a whole will benefit by his continuing in productive activity.

It would be highly desirable, then, to bring the provisions of the law into harmony with the general system of incentives; that is, to devise a retirement test that would result in a person's having increased income as a result of increased work that he does.

On the other hand, a point that must be kept in mind in connection with any proposal that would eliminate or reduce the disincentive effect of the retirement test is that any such change that can be devised has the result of increasing the earnings level at which some benefits are payable. Generally speaking, at present no benefits can be paid to anyone who works throughout the year and makes more than $2080. All of the proposals described in the following discussion would increase the level of earnings up to which some benefits can be paid.

The fact must be faced that the retirement test is the center of an insoluble dilemma. There is, on the one hand, the need to conserve the funds of the program by not paying benefits to people who have substantial work income, and on the other hand, the need to avoid interfering with incentives to work. Both of these objectives cannot be fully accomplished. The best that can be done is to accommodate the two, so that while the funds of the system are in large part directed to the most socially useful purposes, at the same time interference with incentives to work is kept at a reasonably low level.


Proposals to Improve Incentives


An increase in the exempt amount--The proposal for changing the retirement test that is most frequently advanced is to increase the exempt amount above the present $1200 level--for example, to $1500. This proposal has a great deal of popular appeal. It is the kind of change in the retirement test that people usually think of first--in some cases, perhaps, because the $1200 exempt amount is the only part of the retirement test that they are familiar with. And an increase in the exempt amount would result in increased income for many beneficiaries. People who are able to control their earnings and who now limit them to $1200 in a year would be encouraged to increase their work to the point where they earned $1500 (if that were the new exempt amount), and all those who earn between $1200 and $2380 [1] would get more benefits than they can under present law.

[1] The $2380 figure is $1500 plus $880 (i.e., 11 x $80).


Increasing the exempt amount would not, however, have much effect on the problem of improving incentives to work, except for amounts of earnings up to the new exempt amount, nor would it remove any of the problems and inequities of the present test; it would merely change the point at which they occur. If the new amount were $1500, a man who had a choice between a job paying $1800 and a job paying $1500 would generally do better financially to take the lower-paying job; and the person who planned to earn exactly $1500 and inadvertently went just over that amount would have the same problem of losing more in benefits than his earnings above the exempt amount.

If the exempt amount were raised an increase in the other elements of the test--the unit of excess earnings (now $80) and the monthly measure of retirement (now $100)--might seem to be called for. Setting the excess unit and the monthly measure at the same amount, and both at one-twelfth of the exempt amount, has the merit of simplicity, but it is not essential that all three elements correspond. It is quite important for the sake of public understanding that the monthly measure of retirement be one-twelfth of. the exempt amount. People interpret $1200 a year to mean $100 a month. Before the 1958 amendments, when the exempt amount was $1200 and the monthly measure $80, many people did not understand that they could not get benefits for a month in which they made over $80 but less than $100, and many incurred losses on that account. If in addition to an increase in the exempt amount to $1500 the monthly measure of retirement were increased to $125, the increase in the cost of the program would be 0.11 percent of payroll. If the exempt amount were increased to $1800 and the monthly measure were set at $150, the increase in the cost of the program would be 0.24 percent of payroll. Estimates of the cost effects of other possible changes are given in the Appendix.

An increase in the unit of excess earnings-another way of reducing the effect of the retirement test as a deterrent to work at certain levels would be to increase the unit of excess earnings--the amount (now $80) by which earnings in excess of $1200 are divided to determine the number of benefits that must be withheld because of earnings. Since a month's benefit is withheld for every $80 in excess earnings, anyone whose benefits amount to less than $80 has some incentive to work and earn more than $1200 now, since in general he loses less in benefits than the amount of his excess earnings. Increasing the $80 unit would provide a positive incentive to earn above $1200 for all those whose benefit amounts were less than the amount of the new excess unit, and for all other beneficiaries it would in general reduce the loss in total income because of earnings in excess of $1200.

In order to eliminate reductions in income as a result of work for the great majority of the beneficiaries, a substantial increase in the unit of excess earnings would be necessary. An increase to $125 would mean that a million retired worker beneficiary families--15 percent of all such families--would still be losing more in benefits than the unit of excess earnings that caused the loss. Actually an increase to $175 or $200 would be necessary to approach a complete solution to the problem. With an excess unit of $175 all but six-tenths of one percent of the retired worker beneficiary families would have benefits lower than the excess unit and hence would stand to lose less in benefits than the amount of their excess earnings. At $200 the figure would be four-tenths of one percent.

Generally, the families that would still be at a disadvantage with a $175 or $200 excess unit would be those consisting of a retired worker, wife and child, or a retired worker with two or more children, getting benefits at the higher amounts. Families of this composition are, of course, rare.

An example may be helpful to show how the proposal would work. Take the case of a beneficiary with a benefit of $100 a month and suppose he were to earn $1760 in a year. Under present law, 7 months' benefits ($700) would be withheld for his $560 of excess earrings, so that in comparison with the situation in which he could earn exactly $1200, he would lose $140 ($700 minus $560 of excess earnings) in total income for the year. He therefore would not earn $1760 if he understood the law and had any control over how much he could work and earn. Under a proposal to increase the unit of excess earnings to, say, $175, this same beneficiary, because of his $560 in excess earnings, would have 4 months' benefits ($400) withheld. He would thus have gained $160 in total income from his earnings of $560 above $1200.

A peculiarity of this proposal may be brought out by changing the benefit amount in the foregoing example. Suppose a man's benefit were $80 a month instead of $100. The beneficiary would still have 4 months' benefits withheld, but the amount withheld would be only $320 instead of $400. Thus the second beneficiary would have gained $240 rather than $160 as a result of the same amount of work.

The effect of this proposal on incentives to work is quite capricious; the net addition to the beneficiary's income as a result of work is not related at all to the amount earned by doing the work.

An increase in the excess unit, moreover, does not completely solve the problem of benefit losses as a result of earnings, even for the beneficiary whose family benefit amount is smaller than the excess unit. In any situation where a beneficiary makes just over the exempt amount, or just over that amount plus one or more excess units, and consequently loses a month's benefit as a result of having excess earnings amounting to a fractional part of the unit, he can lose in total income. Thus if a beneficiary made the mistake of making 1201 in a year, no matter what the amount of the excess unit, he would lose a whole month's benefit for the extra dollar in earning:.

Still another example may be helpful at this point. Assume that the excess unit were increased to $175. A beneficiary who has benefits amounting to $100 a month has an opportunity to take a job at $1900 a year. If he does, he will lose 4 months' benefits--one for each $175 in excess of $1200. He expects, then, that his total income will be $1900 in earnings and $800 in benefits, for a total of $2700. His employer is caught with a rush job and asks him to do extra work. He does so, and is paid $25 extra for the work. For the $25 additional earnings he loses a whole month's benefit of $100, so that he is actually $75 worse off as a result of doing the extra work.

It is clear, then, that an increase in the unit of excess earnings could not of itself solve the problem of benefit losses as a result of work even if the unit were increased as high as $175 or $200. And an increase in the excess unit would of course increase the long-range cost of the program, the amount of the increase depending upon the size of the increase in the unit. An increase in the unit to $175 would cost 0.15 percent; to $200, 0.19 percent.

Since increasing the excess unit does not completely solve the problem, since under it the amount of additional net income to the beneficiary as a result of work has no relation to the amount he earned by doing the work, and since it would involve significant increases in cost, the Department does not recommend its adoption.

A delayed retirement credit--Another proposal that has been urged on the grounds that it will improve incentives for older workers to get and keep jobs is one that would give a higher benefit amount at retirement to the worker who delays retirement beyond age 65. A proposal of this sort is actually independent of any change in the retirement test; but it is often advanced as an alternative to changes in the test as a way to encourage older people to continue working.

Actually it is doubtful that a delayed retirement credit of this sort--a slight increase in the benefit upon eventual retirement--would be very effective in increasing the number of people who work after 65. For those who can continue in their regular jobs, their regular full-time earnings are a much stronger incentive than a delayed retirement credit could provide. For those who must retire from their regular jobs but might be induced to take on less demanding work, a more effective incentive than a delayed retirement credit would be a revision in the retirement test that would make it possible for them to gain more in income from working than they can under present law.

Aside from the question of incentives a case can be made for the idea that a man who does not retire until some years after reaching age 65, and therefore does not get benefits when he is first eligible and continues to make contributions to the program, should get a higher benefit amount than a man with the same average earnings who retired at 65. Many people argue that the first man's additional contributions and the fact that he will get fewer benefits than he would have drawn if he had retired at age 65 should be recognized by a higher benefit when he does retire.

Many of the sponsors of proposals for a delayed retirement credit either imply or state explicitly that the cost of such a proposal would be very little or nothing because people would be encouraged to go on working beyond age 65 in order to increase their benefits when they do retire and would therefore claim their benefits at older ages than they do now. Actually, a delayed retirement credit would increase program costs because it would give extra benefits to the many people who now work past retirement age. The number of people who would receive extra benefits is suggested by the fact that more than two-thirds of the men coming on the rolls in the last few years have been over 65. The average age at which men become entitled to benefits is about 68.

A substantial increase in benefits for persons who continue to work after age 65 would be costly. For example, a credit large enough to provide, on the earnings record of a person who delays retirement, benefits actuarially equivalent to the amounts that would have been paid to him and his dependents if he had retired at age 65 would amount to an increase of 7.2 percent, in all benefits payable to the family, for each year of delayed retirement, and would raise the level-premium cost of the program by about one percent of taxable earnings--the same as the cost of eliminating the retirement test. Most of the proposals for delayed retirement credits that have received serious consideration have been much more modest--1 or 2 percent a year--and correspondingly less costly (by the same token, it is likely that they would be correspondingly less effective in encouraging people to work who otherwise would not work.) A credit of 1 percent would increase program costs by 0.14 percent of taxable earnings; a credit of 2 percent, by 0.28 percent of taxable earnings.

The increased cost of a delayed retirement credit would be incurred, of course, only for the benefit of those who are able to continue their employment beyond age 65. The delayed retirement credit would not provide any increase in benefits for those who are unable, because of health, labor market conditions, or the retirement policies of employers, to continue in employment beyond that age. There is nothing to suggest that the worker who retires later than 65 needs higher benefits than the worker who had to retire at 65.

Primarily because of the high cost of even a modest delayed retirement credit, and the real doubt whether it would prove to be a significant incentive to continuing work, the Department does not recommend that such a credit be provided under the program.

A reduction in the age at which benefits are paid without regard to earnings--Still another approach to the problem of incentives involved in the test would be to reduce from 72 to 68 the age at which benefits are paid without regard to earnings.

There would be increased incentive to work under the proposal both for workers under 68 and for those 68 and over. For a worker under 68 there would be an incentive to stay on in his job because when he reached age 68--only three years after the time when he could first get benefits--he would have the advantage of getting all of his benefits on top of his earnings. Under present law he cannot get both until age 72, and to most people who would not otherwise continue to work it probably does not seem worthwhile to hang onto a job and forego benefits for seven years in order to have both benefits and earnings at the end of the 7-year period. For a period as short as three years it is not unreasonable to expect that many workers among those who have some control over the decision on whether or not they will retire would stay in their jobs so as to be assured of having both benefits and earnings from a regular job after that period. For workers 68 and over there would obviously be the usual incentive to work--full earnings--because their earnings could not reduce their benefit income. And there are real advantages, in the form of lessened irritation and increased public acceptance, in a provision under which many more beneficiaries could feel free to earn what they could without keeping records or worrying about whether they might earn too much.

The basic approach of the law to the problem of retirement, if this proposal were enacted, could be described in these terms: Age 68 would be looked upon as the normal age of retirement, and benefits would be payable at that age regardless of earnings. (Actually, about half of all workers claim their benefits before they reach age 68.) Because in fact many workers retire before age 68, either because they are no longer able to work or because they are retired by their employers, benefits would be payable before age 68 to anyone who met a fairly strict test of retirement (the present test). Prior to age 65 (62 for women) benefits would be payable only if the person demonstrated that he could not work--that is, if he met the requirements for disability insurance benefits.

There are advantages in moving in the direction of establishing in the minds of people the idea that a higher age than 65 is the normal retirement age. Such a change in attitude toward retirement age (which might in time induce changes in the retirement policies and hiring policies of employers), plus improvement in incentives to work might help to meet the criticism that the Nation is wasting its human resources by policies that force older workers out of jobs they are still capable of doing and anxious to have.

On the other hand, there are disadvantages to the proposal, too. Many people do continue after age 68 to work full time at their regular jobs--many more than do so after age 72--and there is no reason related to the basic objectives of old-age and survivors insurance why benefits should be paid to them on top of their full-time earnings, no matter how high.

Enactment of the proposal, if no other changes were made in the retirement test, would mean that all of the shortcomings of the present test described earlier in the report would remain--though a much smaller number of people would be affected--and, by the same token, dissatisfaction with those shortcomings would remain, though it would probably be lessened since each worker would be under the test for only three years.

The proposal would cost 0.35 percent of payroll, over one-third of the cost of complete elimination of the test. The Department does not recommend the proposal, in spite of its advantages, because it is expensive, because it does not correct any of the anomalies and problems that arise in the present test, and because the additional money that would be spent would go to many who are working full time at high earnings.

A proposal to withhold $1 in benefits for each $2 in excess of $1200 and up to $2400, and withhold $1 in benefits for each $1 in earnings in excess of $2400--The Department has also considered a proposal to eliminate the $80 unit of excess earnings and substitute for it a provision for reducing the benefit amount payable by $1 for each $2 of earnings in excess of the $1200 exempt amount. The Department believes that this proposal is as good an approach as can be devised to the problem of how to preserve incentives for older people to work at a moderate cost to the social insurance system.

An example will help to make clear how the proposal would work. Take the case of a man who, with his wife, is eligible for benefits amounting to $150 a month. He has a chance to take a job paying $1500 a year. Under present laws for the $300 in earnings over $1200 he would lose 4 benefits or $600. Instead of gaining by doing extra works he would actually lose twice as much in benefits as he gained in earnings. With a one-for-two-reduction proposal for his $300 in excess earnings he would lose, of course, only $150 in benefits.

Let us take another example, that of a man who is eligible for benefits of $90 a month and has an opportunity to make $1600 in the year. Under present law he would lose 5 monthly benefits--$450--for his $400 of earnings in excess of $1200. Under the one-for-two reduction proposal he would lose $200 and would therefore be $200 better off, rather than $50 worse off.

Under this proposal there would no longer be any reason for beneficiaries to seek out jobs at $1200 and to limit their work activity. The effect of the provision would be to support rather than interfere with the desire of older people to continue to work to the extent they are able to do so. The proposed provision would furnish an incentive to work throughout the entire range of benefits and would avoid the anomalies that arise at the various breaking points in the present annual test.

Elimination of the $80 unit of excess earnings, moreover, would be a significant simplification of the earnings test. The $80 excess unit is a completely arbitrary element in the test included to avoid the sharp borderline that would occur if there were not some graded reduction of benefits to take account of earnings over the exempt amount. (It would be unreasonable to withhold a year's benefits for a dollar of excess earnings and the $80 provision was included in the law to prevent that result.) Since the amount of $80 is arbitrary it is difficult for beneficiaries to understand.

Although most beneficiaries would find the new test to their advantage some would, at least theoretically, be somewhat worse off than under present law. They would be those beneficiaries who get less than $40 in benefits and yet earn in excess of $1200 in a year. Even for people at these low benefit levels it is likely that the disadvantage would be largely theoretical, since probably few people at these benefit levels work for earnings of as much as $1200 in a year. And those who do would of course gain as a result of their work, although they would not gain as much as they do under present law. (Now, as a result of the $80 excess unit, a beneficiary getting benefits of, say, $35 loses only that $35 for every $80 of excess earnings, whereas under the proposal he would lose $40.)

The proposal for a $1 reduction in benefits for $2 of earnings, if applied to all earnings, would result in the payment of some benefits to people earning relatively high amounts. A man and wife getting the present maximum benefits of $180, for example, would get $100 in benefits for the year with earnings of $5320. To reduce costs somewhat by preventing these partial payments to people who have relatively high earnings it could be provided that earnings above $2400 a year would reduce benefits dollar for dollar. Without this modification the change in the test would cost 0.19 percent of payroll. With this modification, the change would cost 0.15 percent.[ l]

[1] Other types of reduction proposals were also considered by the Department, including a straight dollar-for-dollar reduction and a proposal to add on top of the present retirement test, with all of its complexities, a provision to make a dollar-for-dollar adjustment at the end of the year for beneficiaries who had more withheld in benefits than the amount of their excess earnings over $1200. Neither of these proposals furnishes a positive incentive to the beneficiary to work; and the second proposal does not correct the situation where a beneficiary can lose by earning just over a breaking point as compared with earning at the breaking point. And the dollar-for-dollar reduction, with the elimination of the $80 unit of excess earnings, would mean such a significant de-liberalization for people with benefit amounts less than $80 that it would need to be accompanied by an increase in the exempt amount in order to be acceptable, and would therefore mean a rather heavy increase in cost.

A proposal that would increase costs to this degree would require at this time that provision be made for additional financing, in view of the current close balance of the fund. We have not in this report tried to conclude on whether additional cost should be incurred for this purpose, but rather have tried only to develop the best solution for the problems involved in the retirement test. We believe the solution that we have developed is the best.

SUMMARY OF FINDINGS

The findings and recommendations of the Department with respect to the retirement test under the old-age and survivors insurance program can be summarized as follows:

I. The retirement test in the old-age and survivors insurance program is necessary in order to assure that the funds of the program will be employed for socially useful purposes. Elimination of the retirement test would substantially increase the cost of the program, and the additional cost would be incurred chiefly as a result of paying full benefits to people who are fully employed at relatively high earnings. The Department therefore recommends that a test of retirement be retained in the program.

II. The Department recommends that the monthly measure of retirement be retained in the test, since to remove it would prevent the program from attaining its objective of making benefits available to people immediately upon retirement and during other periods when they do not have income from work.

III. The Department has developed but does not recommends a proposal that would eliminate the payment of benefits in the sort of case the Committee asked the Department to study--the case of a person who is retired throughout most of the year but comes back into employment for a month or two and has high earnings. Under present law, because of the monthly test, he gets benefits for the months in which he did not work. The Department believes that it is desirable to withhold benefits only for the months in which the person works, as is done under present law.

IV. The Department has developed a proposal designed to improve incentives for older people to work, to remove certain inequities that now arise in the operation of the retirement test, and to make the test more understandable and acceptable. That proposal is to eliminate the present provision requiring that a whole month's benefit be withheld for every $80, or fraction of $80, of earnings in excess of $1200, and to substitute for it a provision for withholding $l in benefits for each $2 in earnings in excess of $1200 and up to $2400, and for withholding $1 in benefits for each $l in earnings in excess of $2400. We are convinced that this is the best solution that can be devised to the problems involved; but we have not concluded upon whether it is desirable to provide for an increase in taxes in order to make enactment of the proposal possible. The Department plans to continue its evaluation of this question.

APPENDIX

COST OF CERTAIN COMBINATIONS OF PROPOSALS
FOR CHANGES IN THE RETIREMENT TEST

Exempt Amount [1]

Monthly Measure of Retirement [2]

Unit of Excess Earnings [3]

Other [4]

Cost [5]

$1,200

$100

$ 80

$1 for $1

.05

1,200

100

None

$1 for $1

.04

1,200

100

None

$1 for $2

.19

1,200

100

None

[6]

.15

1,200

100

100

---

.03

1,200

100

125

---

.07

1,200

100

150

---

.12

1,200

100

175

---

.15

1,200

100

200

---

.19

1,500

125

80

---

.11

1,500

125

None

$1 for $1

.23

1,500

125

125

---

.20

1,800

150

80

---

.24

1,800

150

None

$1 for $1

.40

1,800

150

150

---

.38

Lower exempt age to 68

.35

Complete elimination of the test

1.00

[1] The amount of earnings a beneficiary may have and get all his benefits.

[2] The amount of wages a beneficiary can earn for a month without losing benefits for that month, regardless of his total annual earnings.

[3] The amount by which the total yearly earnings in excess of the exempt amount are divided to determine the maximum number of monthly benefits that may be withheld because of earnings.

[4] The "$1-for-$1" adjustment means that, after taking into account the other provisions of the retirement test, $1 of benefits would be withheld for each $1 of earnings in excess of the exempt amount. The "$1-for-$2" adjustment means that, after taking into account the other provisions of the retirement test, $1 of benefits would be withheld for each $2 of earnings in excess of the exempt amount.

[5] Percent of payroll--intermediate level-premium estimate.

[6] A reduction of $1 in benefits for every $2 of earnings in excess of $2,200 and up to $2,400 and a reduction of $1 in benefits for every $1 of earnings in excess of $2,400.