Pension Plan Participation Among Married Couples

Introduction

Employer-provided pensions and retirement plans constitute one of the pillars of income at retirement in the United States, while Social Security, own savings or assets, and earnings are the other pillars. Evidence shows that among people aged 65 or older, accounting for all sources of family income, about 45 percent receive income from a pension or an annuity (SSA 2012, Table 2.B1), and two-thirds of that share receive income from a private pension. Among people aged 65 or older, income from pensions constituted one-fifth of their retirement income in 2010 (SSA 2012, Table 10.1). Thus, investigation of workers' participation in employer-provided pensions or retirement plans is important in understanding retirement income of future retirees.

Defined benefit (DB) pensions are funded by employers and traditionally provide a guaranteed monthly payment at retirement for the worker and if elected, for their surviving spouse. Thus, the employer assumes the funding and investment risk during the accumulation phase and the longevity risk during the distribution phase. Defined contribution (DC) plans, on the other hand, are investment accounts funded mainly by employees (often with matching contributions from employers) during their working lives who are responsible for investing their contributions. After retirement, workers are also responsible for managing and drawing down their account balances (Mackenzie 2010). Thus, the type of retirement plan available is important given plan differences in risks assumed during the accumulation and distribution phases.

This article presents descriptive statistics on the overall participation in employer-provided pension plans by plan type among married couples over a decade, from 1998 to 2009. The private sector's pension environment during this period was characterized by a continued shift from traditional DB pensions to DC retirement accounts. This decade saw a rising prevalence of employers “freezing” their DB plans while also establishing new DC plans or increasing the employer match to current DC retirement plans and thus shifting the risks and responsibilities for retirement from employers to the employees (GAO 2008; Mackenzie 2010; Anguelov, Iams, and Purcell 2012). Furthermore, the 2006 Pension Protection Act, which permitted employers to automatically enroll their employees into DC plans, is likely to have had an impact on this trend. In addition, during the Great Recession of 2007–2009, the decline in the financial markets led to sharp declines in retirement account balances, whereas the drastic increase in unemployment led to decreases in participation and contributions to DC plans and increases in loan activities from those accounts. These changes revealed more clearly some of the causes that may derail employees in DC plans from accumulating sufficient funds for retirement (Gorton and Metrick 2012; Utkus and Bapat 2011, Figure 1; Butrica 2012; GAO 2012; Haaga and Johnson 2012; Johnson and Butrica 2012). Finally, the decade was a period when women's labor force participation declined particularly among unmarried women, those with no children, and women with more than 16 years of education (Macunovich 2010), which is likely to translate into lower participation in pension plans.

Evidence from previous research indicates that almost half of wage and salary workers participate in an employer-provided pension or retirement plan (Munnell and Sunden 2004; Copeland 2009, 2010; Purcell 2009; Gustman, Steinmeier, and Tabatabai 2010, Table 5.1). The common findings are that participation is higher among full-time workers, those with higher earnings, and those who work for larger firms. In addition, participation is higher among older workers, men, married people, whites, and those with higher levels of education.

Few studies estimate participation in retirement plans among couples. Those that do find that the participation rate of couples as a unit is higher than when looking at married men and women separately (Iams 1995, Table 5; Mackenzie and Wu 2008). Using data from the pension modules of the 1984 and 1993 Panels of the Survey of Income and Program Participation (SIPP), Iams (1995, Table 5) estimated that about 75 percent of married women aged 40–54 and their husbands participated in a pension plan through their current or previous job. The participation rate was lower when looking at married women's own pensions (33 percent and 40 percent, respectively in 1984 and 1993) and married men's own pensions (66 percent and 60 percent, respectively in 1984 and 1993). Using data from the 2004 Survey of Consumer Finance, Mackenzie and Wu (2008, Table A.6) found that 60 percent of working couples had pension coverage through their current job and 66 percent had coverage from their current or previous job (the authors do not provide separate estimates for married workers). In sum, previous research on trends in pension participation and pension type in general underestimates pension coverage of married workers by omitting coverage available through the spouse.

Data

The data for this analysis come from the 1996 and 2008 Panels of SIPP, which are matched with information from Social Security administrative W-2 records. We restrict the sample to respondents and spouses with matching administrative records. Retirement plan characteristics are collected in years 1998 and 2009, respectively, for the 1996 and 2008 SIPP Panels. The samples for this analysis consist of full-time wage and salary workers aged 25–60 who were either not married or married with a spouse present.¹ We create a couples unit comprised of married men working full time and merge their wives' information to their records.

We identify pension participation and the type of plan by combining respondent-reported information in the SIPP with information in the W-2 records regarding tax-deferred contributions for the same year. Using information from W-2 records, previous research has shown that respondents misreport either whether they participate in a plan or the type of plan they participated in (see Dushi and Honig (2008); Dushi and Iams (2010); Dushi, Iams, and Lichtenstein (2011)). Following Dushi and Iams (2010), we adjust respondents' reports and plan types as appropriate. Thus, respondents who according to W-2 records have positive tax-deferred contributions are defined as participating in a DC plan. In cases where a respondent reported not participating in a pension plan but W-2 records indicate that the respondent made a tax-deferred contribution, we assign him or her as participating in a DC plan. In cases where a respondent reported participating in a DB plan and W-2 records indicate that the respondent made a tax deferred contribution, we assign him or her as participating in both types of plans (DB and DC).² In cases where a respondent reported participating in a DC plan but the W-2 records indicate that he or she did not make tax-deferred contributions in the last 3 years, then we assume the respondent misreported his or her plan type and thus we assign that person as participating in a DB plan (see Dushi and Honig (2008); Dushi and Iams (2010); Dushi, Iams, and Lichtenstein (2011)).³

Results

Participation rates in employer-provided pension plans were relatively constant between 1998 and 2009 (Table 1). Among all full-time workers, about two-thirds participated in a pension plan in both years. Nonmarried workers, both men and women, were less likely to participate in a pension plan than were their married counterparts. About 72 percent of married men (with spouse present) working full time participated in a retirement plan in 1998 and 2009. Similarly, 72 percent of married women (with spouse present) working full time participated in a pension plan in 2009, an increase of about 5 percentage points from the 1998 participation level. As expected, the participation rate is higher when looking at couples as a unit of analysis instead of just looking at married workers. Both in 1998 and 2009, about 80 percent of couples had at least one of the spouses participating in a pension plan, an increase of about 10 percentage points compared with looking at married men and married women separately. This suggests that the typical analysis that focuses on married workers without considering coverage of their spouses is likely to underestimate the participation rate in pension or retirement plans from which the couples are expectedly going to draw their retirement income.

An important aspect of retirement security is the type of plans from which workers are expecting to draw their income. About 20 percent of nonmarried full-time workers in 2009 participated in a DB-only plan, compared with about 25 percent in 1998. About 25 percent participated in a DC-only plan and around 16 percent participated in both types of plans in 2009, compared with about 21 percent in a DC-only plan and 14 percent in both types of plans in 1998 (Table 1). Among married men, the proportion participating in a DB-only plan decreased from 26 percent in 1998 to 21 percent in 2009, while the proportion participating in a DC-only plan increased from 24 percent to 29 percent, respectively. Similar trends to those for married men are observed for married women, although their level of participation in DB-only plans stayed more constant at about 25 percent in both 1998 and 2009.

About 20 percent of married workers were covered by both a DB and a DC plan in 1998 and 2009. However, looking at couples as a unit, the proportion participating in both types of plans increased to about 35 percent in 1998 and 2009. Thus, it is clear that when considering couples' joint participation in pension plans, more than one-third are expected to draw their retirement income from both a DB plan (at least potentially in the form of an annuity, if elected) and from a DC account (which is more liquid and can be drawn down depending on individual choices).

Among couples in both 1998 and 2009, 20 percent did not participate in a pension through either spouse, whereas 80 percent participated in a pension through either the husband or the wife (Table 2). About one-third of couples had both spouses participating in a DB pension or a DC retirement plan. In 2009, 8 percent of couples had the wife as the sole participating spouse, whereas 37 percent had the husband as the sole participant in the household (10 percent and 39 percent, respectively in 1998).

Among couples, 55 percent in 1998 and 60 percent in 2009 had at least one of the spouses contributing to a DC plan (Table 3), although the prevalence of contributing was higher among husbands. Thus, of the 60 percent with at least one spouse contributing in 2009, only in 18 percent of couples were both spouses contributing to a DC plan, whereas in 32 percent the husband was the only contributor to a plan, and in the remaining 10 percent the wife was the only contributor to a plan. A similar pattern is evident for 1998, although the proportion of couples where both the husband and wife contributed to a DC plan was 4 percentage points lower than in 2009.

Looking at the annual W-2 contribution amounts (expressed in 2009 dollars, Table 4), it is evident that both the mean and median tax-deferred contributions have increased between 1998 and 2009 for all workers, while contribution rates were relatively similar.⁴ As expected, couple contributions were higher than husband or wife contributions separately, and they have increased substantially over the period from a mean of $3,891 in 1998 to $6,655 in 2009 and a median of $2,637 in 1998 to $3,861 in 2009. This increase is not only because husbands' contributions increased over the period but also because the proportion of wives contributing to a DC plan increased (Table 3), and conditional on contributing, wives' annual contribution amounts increased both at the mean and the median (Table 4).

When we look at the distribution of the wife's contributions as a proportion of total family contributions (Table 5), among couples where at least one of the spouses contributes to a plan, we notice that 56 percent of wives in 1998 (and 53 percent in 2009) did not contribute to a plan and the husband was the sole contributor. Conversely, 18 percent of husbands in 1998 (and 17 percent in 2009) did not contribute to a DC plan, and the wife was the sole contributor in the family. Among couples where both spouses contributed to a plan, for a majority of wives, their contribution share ranges from 26 to 75 percent of total family contributions. Among couples where both spouses contributed to a plan, both at the mean and the median, the wife's contribution comprised about 40 percent of the total family contribution in 1998 (42 percent in 2009; authors' calculations, not reported in the table).

Conclusion

The decade between 1998 and 2009 saw many changes related to retirement plans, including expanded access to DC plans, DB-plan freezes, expanded autoenrollment into DC plans after the implementation of the 2006 Pension Protection Act, and the Great Recession of 2007–2009. Using Census Bureau SIPP data matched with Social Security administrative records, we examine participation in employer-provided retirement plans by plan type among couples where both spouses are present and the husband is a full-time wage and salary worker aged 25–60. We focus particularly on measuring participation by specific plan type for married couples rather than married workers separately because couples share their retirement income, regardless of whether those contributions are through the husband or the wife.

Our findings show that the participation rate is much higher among couples as a unit than when looking at married men and married women separately. While about 30 percent of married men with a spouse present did not participate in a plan both in 1998 and 2009, when we accounted for their spouses' participation we found that in only 20 percent of couples did neither spouse participate in a pension plan. In about 10 percent of couples in 2009, the wife was the only one participating in a pension plan compared with about 37 percent of couples where the husband was the only one participating. Regarding participation and contributions to a DC plan, 60 percent of couples in 2009 had at least one of the spouses contributing to a DC plan. In half of those couples, the husband was the only one contributing. Among couples where both spouses contributed to a plan, the wife's contribution comprised around 42 percent of the total family contribution both in 1998 and 2009. In sum, we find little change in participation rates and in the prevalence of plan types between 1998 and 2009. Given the major changes affecting retirement plan participation and types of plans offered by employers, further analysis is needed to investigate the reasons behind the stability in participation rates over this decade.