Because there is no pressing administrative need for it, much potentially useful information about tax-deferred saving is not recorded on either Form 1040 or any of the tax information returns (such as Forms W-2 and 5498). Much of that information--for example, 401(k) balances, employer contributions to 401(k)s, and expected benefits from defined-benefit plans--is available from the Survey of Consumer Finances (SCF), conducted by the Federal Reserve Board every three years. One such survey was conducted in 1998, making it comparable with income tax returns filed for tax year 1997. Although the SCF collects more-complete data on retirement plan contributions and balances than the Internal Revenue Service (IRS) does, it is based on a much smaller sample (fewer than 7,000 households) and depends largely on the memories and recordkeeping practices of voluntary participants. Furthermore, the SCF does not concern itself with details of the tax code, such as whether a plan is "qualified." Nevertheless, the SCF data are considered to be of high quality and are widely used in studies of the distribution of wealth in the United States. For this analysis, the Congressional Budget Office (CBO) used the survey to answer questions for which tax-return data were inadequate.
 

Reconciliation of Data

In cases in which items from the IRS's Statistics of Income (SOI) files overlapped with those from the SCF, CBO gave precedence to the SOI data. Four important categories of data, however, could be derived only from the SCF:

• Employer contributions to defined-contribution plans;
  
  
• The number of workers participating in defined-benefit plans;
  
  
• The number of workers participating in noncontributory defined-contribution plans; and
  
  
• The number of workers expecting to collect retirement benefits from former employers in the future.

To ensure consistency with the SOI data, the relationships between variables from the SCF were maintained, but values were adjusted to conform to control totals from the SOI. For example, CBO estimated employer contributions to defined-contribution plans by calculating the ratio of employer contributions to employee contributions from the SCF by income, age, or marital status/earner role, and then applied those ratios to employee contributions in the same categories as tabulated from the SOI data.

Estimating the number of active participants in defined-benefit plans was more complicated. Both the SCF and SOI files were initially split between those participating in 401(k)-type plans and those participating only in noncontributory plans. The percentage of workers in each of those groups who were participating in a defined-benefit plan was calculated from the SCF, and those percentages were applied to the corresponding group in the SOI file. For workers participating only in a noncontributory plan, the percentage participating in a defined-contribution plan was calculated from the SCF and was applied to the same group in the SOI file.

The estimate of the number of inactive participants in employment-based plans has three parts: those who owned an individual retirement account (IRA) to which they did not contribute in the current year; those who collected benefits from former employers; and those who expected to collect benefits from former employers in the future. The first two groups can be tabulated from the SOI file. To estimate the third group, CBO identified in the SCF workers who were not participating in an employment-based plan and calculated the percentage expecting benefits from a former employer in the future. CBO then applied that percentage to workers in the SOI file who were not participating in employment-based plans.
 

Discussion of Tables C-1 Through C-3

Table C-1 provides a finer breakdown of worker participation in employment-based plans than that presented in Table 3. The SCF makes it possible to tabulate all participants in noncontributory plans, not just those who do not also participate in 401(k)-type plans. Furthermore, it allows noncontributory plans to be divided into defined-benefit and defined-contribution plans.

The SCF also makes it possible to tabulate inactive as well as active participants, as detailed in Table C-2. As noted above, inactive participants are those who did not contribute to a retirement savings plan in 1997 but were either already collecting benefits, expected to receive benefits from a former employer's plan, or owned an IRA from which they could draw funds after their retirement. The resulting patterns are quite different from those revealed through active participation alone, particularly among the various age groups.

Table C-3 supplements the information on employee contributions to 401(k)-type plans, which is provided in Table 5, by presenting estimates of employer contributions as well. Those contributions are, on average, smaller than those of employees but follow the same general income and age patterns.