ACTUARIAL NOTE
 
 
SOCIAL SECURITY ADMINISTRATION
Number 103
May 1981
Office of the Chief Actuary
Baltimore, Maryland
AVERAGE WAGES FOR INDEXING UNDER THE SOCIAL SECURITY ACT
AND THE AUTOMATIC DETERMINATIONS FOR 1979-81
by Eli N. Donkar, Ph.D., A.S.A.
INTRODUCTION
CONSTRUCTION OF THE 1951-77 AVERAGE WAGE SERIES
1. Historical precedents for wage indexing and average wage
2. The preliminary series of average wages
3. Limitations on other historical sources of data available from SSA
4. The construction of the final average wage series for 1951-77
5. The final series—indications of reliability

AVERAGE WAGES FOR 1978 AND 1979

1. Sources of data for years after 1977

Beginning with 1978, quarterly reports of wages for individual employees are no longer available. Instead, SSA is receiving annual reports of all wages, including wages in noncovered employment and covered wages in excess of the maximum annual amount taxable. These data are obtained by SSA in its processing of employer copies of W-2 Forms which are now submitted directly to SSA. Thus, in determining the average wage figures for years after 1977, it was not (and will not be) possible to rely on the same source of data that had been used for indexing under the law for years before 1978. Nevertheless, as was mentioned previously, it was assumed in the deliberations of the congressional committees preceding the enactment of the 1977 Amendments that the basic principles underlying the construction of the historical indexing series were to be preserved in the development of the indexing figures for years after 1977. In particular, this implied that wage data, rather than wages plus SEI, were to be used as a basis for calculating the indexing figure. This was true even though the wage data available for use after 1977 are tabulated on an annual basis because there continued to be compelling reasons for not including SEI data along with wage data in the determination of the indexing figure.

The first reason concerns the timing of the processing of SEI data. While wage data for a given year now become available from Forms 1040 by September of the following year, on a fairly complete basis, SEI data are still not very complete by that time. Second, even if the self-employment reports were processed in a timely enough fashion, the desire to obtain a consistent year-to-year measure would be thwarted by the evident unreliability of the self-employment reports. Recent data published in the "Survey of Current Business" (July 1979 issue, p. 69, Table 8.6) show that the reported totals of self-employment income are substantially modified by "posttabulation amendments and revisions including allowance for audit." These revisions have represented as much as 20 to 22 percent of the total reported SEI in two recent years. This variation in the reliability of the reported data makes the inclusion of such data impracticable as part of a measure that must be consistent from year to year.

The committees were also especially aware of the problems that were to arise in making the transition from first-quarter wage data to annual wage data that was necessitated by the implementation of annual reporting. For this reason, the provisions of the law were written to allow the Secretary flexibility in making this transition. The committees understood that, ultimately, the basis for calculating the indexing average wage was to be the data tabulated by SSA in the processing of all W-2 Forms submitted to SSA by employers. However, in the transition to this data source, the committees recognized that comparable data for 1977 could not be obtained from all W-2 Forms. Therefore, the committees provided for the use of data from Forms 1040 tabulated by the Internal Revenue Service (IRS) for both years 1977 and 1978, so that the increase in average wages from 1977 to 1978 would be based on comparable data for both years. Explicit reference to the use of Form 1040 data is made in the report of the Senate Finance Committee on H.R. 5322. These considerations provided the basis for the methodology underlying the calculation of the 1978 indexing figure.

The congressional committees anticipated that beginning with the 1979 wage indexing figure, the calculation of such indexing figures would be based on data tabulated in the processing of all W-2 Forms. However, it became clear, soon after the initiation of SSA's annual reporting procedures, that this massive job of data processing would not be able to provide the requisite data for making the 1979 average wage determination in time to meet the operational and legal deadlines imposed on this calculation. Therefore, it was necessary to continue the use of the Form 1040 data obtained from IRS for the determination of the 1979 average wage. These IRS data and their use in the calculations of the 1978 and 1979 indexing figures are summarized in the following section. It is currently anticipated that such use of the IRS data will continue for the next few years.

2. IRS wage data for 1977-79

Following the enactment of the annual reporting legislation, SSA executed a contract for reimbursable services with IRS under which IRS would provide the underlying data to be used for indexing according to increases in average wages. As mentioned above, the original negotiations with IRS envisioned the use of IRS data for tax years 1977 and 1978 only. However, the contract was extended to include data for tax year 1979 as well. Prior to the existence of these requests for data by SSA, IRS collected data annually on total wages reported in tax returns, but did not regularly obtain comparable information on the number of wage earners who earned those wages. This was true, in part, because such data are not readily available on the tax returns themselves, since a given joint return does not contain an explicit entry which indicates whether the wages reported in the return were earned by one or both of the persons listed in the return. Such information could only be deduced by examining the W-2 Forms attached to the given joint return, and that tedious procedure was not included in the normal processing of returns before 1977.

Beginning with the returns for tax year 1977, the SSA contract with IRS provided for the collection of data on numbers of wage earners by including an examination of the W-2 Forms attached to returns as part of the normal processing of joint returns. A summary of the data supplied to SSA by IRS for tax years 1977-79 is contained in Table 9. The tabulation of total wages shown was made directly from the wages reported on the Forms 1040. The number of "wage earners" earning those wages was derived as follows:

  1. For "non-joint" returns it was assumed that all wages reported on the return were earned by one wage earner.

  2. For "joint" returns there were two possibilities:
    a. There were no W-2 Forms or W-2 Forms for only one wage earner attached to the Form 1040. In this case, it was assumed that all the wages reported on the return were earned by one wage earner.
    b. There were W-2 Forms for more than one wage earner attached to the Form 1040. In this case, it was assumed that the wages reported on the return were earned by two wage earners.

The data for 1977 were based on 1977 tax returns processed during the period January 1, 1978 through December 31, 1978. The corresponding processing periods for the 1978 and 1979 data were January 1, 1979 through September 15, 1979 and January 1, 1980 through August 11, 1980, respectively. The cutoff dates in each case were dictated by the need to have the required average wage determinations for a given year made by early October of the preceding year. This is necessary because benefits must be computed for those persons reaching age 62 in January of the given year who may begin filing claims the preceding October. In each year 1977-79, IRS estimates that the tabulations reported to SSA represent over 97 percent of the individual returns that would be processed for comparability of the data from year to year. The average wages per wage earner calculated directly from the IRS data are $10,043.15, $10,840.68, and $11,789.01 for 1977, 1978, and 1979 respectively.

3. Average wages for indexing, 1977-79

In order to construct average wage figures for 1978 and 1979 for use in indexing that were consistent with the published series for 1951-77, the annual percentage increases measured by the IRS averages were used to increase the published SSA average wage for 1977 of $9,779.44. Thus, the 1978 indexing average wage was determined by multiplying the published average wage for 1977 by the ratio of the average annual wage for 1978, from IRS data, to the average annual wage for 1977, from IRS data, and rounding the result to the nearest cent. The calculation is as follows:

Average wage for 1978 = $9,779.44 x ($10,840.68/$10,043.15) = $10,556.03
The 1978 average wage figure of $10,556.03 was announced in the Federal Register on November 1, 1979.

Similarly, the 1979 indexing average wage was determined by multiplying the published average wage for 1978 by the ratio of the average annual wage for 1979, from IRS data, to the average annual wage for 1978, from IRS data, and rounding the result to the nearest cent. The calculation is as follows:

Average wage for 1979 = $10,556.03 x ($11,789.01/$10,840.68) = $11,479.46
The 1979 average wage figure of $11,479.46 was announced in the Federal Register on November 18, 1980.

DETERMINATION OF WAGE-INDEXED PROGRAM AMOUNTS FOR 1979-81
1. The retirement test exempt amounts
2. Amount of earnings required for a quarter of coverage
3. Bend points in PIA formula and maximum family benefit formula
4. Contribution and benefit base
5. Contribution and benefit base under provisions in "old law"
Relevant Federal Register Publications
Tables
1. Average wage series for indexing earnings under the Social Security Act, calendar years 1951-79
2. Social Security program amounts determined under the automatic provisions which depend on increases in average wages, calendar years 1978-81
3. First-quarter averages used in constructing the preliminary indexing series, calendar years 1951-77
4. First-quarter averages used in constructing the final indexing series, calendar years 1951-77
5. Average wages, percentage increases and indexing factors resulting from the preliminary average wage series, the final average wage series, and an average wage series derived from Department of Commerce (DoC) estimates, calendar years 1951-77
6. Number of "wage items" per employee tabulated from the 1-percent sample LEED file, calendar years 1957-73
7. Average wages for the first quarter for private industry population covered by unemployment insurance, calendar years 1951-75
8. Estimated amounts of Average Indexed Monthly Earnings for maximum wage earners retiring at age 62 in each year 1979-90, calculated using three different wage series for indexing
9. Summary of wage data tabulated by the Internal Revenue Service from individual tax returns filed for tax years 1977-79

Notes published in the 1980s