However, an issue brief from the Center for Retirement Research at Boston College notes that the four major household data sets that provide information about pension coverage use different measures of coverage and participation and are subject to individual reporting error. For example, the Current Population Survey (CPS), produced jointly by the U.S. Census Bureau and the U.S. Bureau of Labor Statistics, asks individuals two questions. First, did the employer you worked for have a pension or other type of retirement plan for any of its employees? And second, were you included in the plan? The survey does not gather any information about whether the plan was defined benefit or defined contribution.

In the CPS, the percentage of workers covered depends on two factors. These include 1) the definition of coverage—does the individual’s employer have a plan or does the individual participate in a plan?; and 2) the population group under consideration. Restricting the population to full-time wage and salary workers ages 25 to 64, including both public and private sector workers, and using employer sponsorship as the criterion, the CPS shows 63% of workers have the potential for participation in a retirement plan. At the other extreme, focusing on private sector workers only, using participation as the criterion, including both part-time and full-time workers, and eliminating the age constraint, the CPS shows only 38% are covered by a plan.

According to the report, while pension participation has risen and fallen for short intervals, all four data sets show relatively stable participation rates over the period 1991 to 2012. The surveys suggest, over the last two decades, between 40% and 55% of private sector employees participated in some kind of retirement plan in any given year, with each data set suggesting slightly different levels of participation.

The issue brief points out that a look at the National Compensation Survey (NCS), an employer survey sponsored by the U.S. Bureau of Labor Statistics, indicates employers report a higher level of retirement plan coverage and participation than levels reported by employees. The NCS suggests nearly 80% of full-time workers have an employer that sponsors a retirement plan. The figure refers to full-time workers in both the private and state/local sectors. Virtually all (99%) full-time state or local government workers are offered a pension. Eliminate state and local employers and the coverage figure for full-time workers in the private sector drops to 74%. Add in part-time workers and the 74% drops to 64%.

In addition, only three-quarters of private sector workers who are offered a plan choose to participate. Thus, the NCS reports only 48% of private sector workers (including both full-time and part-time workers) participate in a retirement plan.

According to the issue brief, a 2011 study merging data from the Survey of Income and Program Participation (SIPP) from the Census Bureau with W-2 tax records shows participants may not understand their own situations. About 85% of SIPP households were matched with tax records, and the study found SIPP respondents in total underreported being offered a defined contribution plan by three percentage points and participating in a plan by five percentage points.

For example, with respect to participation, about 30% of respondents actively participated in a defined contribution plan and reported correctly, as verified by the W-2. About 14% self-reported incorrectly that they did not participate when the W-2 records showed they did. On the other hand, 9% of workers self-reported incorrectly that they did participate in a defined contribution plan when their W-2 record showed no tax-deferred contributions. The researchers assumed that the 9% reporting “false positives” were participating in a defined benefit plan.

Despite the discrepancies in data, retirement plan coverage is a problem in the private sector, the paper concludes.

The issue brief may be downloaded from here.