How Effective Is the Social Security Statement? Informing Younger Workers about Social Security

Introduction

In 1995, the Social Security Administration (SSA) began large-scale mailings of earnings and benefit statements to workers.¹ One of the statement's primary purposes was to provide workers with information on their Social Security benefits and to help them plan their financial futures. The Social Security Statement has been widely acknowledged as one of the most important of federal government communications with the public (Jackson 2005).² It stands as the largest customized mailing ever undertaken by a federal agency (SSA n.d.). Developing and distributing the Statement each year required a massive effort in terms of resources and work hours.

To date, research assessing the Statement's effect on public knowledge about Social Security has focused on older workers approaching retirement (Mastrobuoni 2009; Biggs 2010; Greenwald and others 2010; Liebman and Luttmer 2010). Even SSA-commissioned research, based on surveys conducted between 1998 and 2004 to measure the Statement's effect on public understanding of Social Security programs and benefits, emphasized the impact on older workers.

We decided to focus instead on younger workers, in part because of the changing nature of retirement income. Younger workers are less likely than older workers to be covered by defined benefit pension plans and more likely to be responsible for their own retirement security. For them, Social Security benefits are bound to be increasingly significant. Thus, it is essential that younger workers understand how different factors might affect the Social Security benefits they can expect to receive. Younger workers are also of interest because, although in a position to benefit most from additional information, they are less likely than older groups to seek it.³ For these reasons, we assessed how much younger workers know about Social Security, identified their knowledge gaps, and considered ways to provide them with additional information.

In addition to its focus on younger workers, this article contributes to the literature on the impact of the Statement by looking at changes over time in the public's understanding of Social Security. Most other studies focus on a single year and compare individuals who received a Statement that year with those who did not. Using surveys commissioned by SSA, we were able to look at a group of younger workers who had not received the Statement at the time of the first survey and contrast them with workers of the same age who had received a Statement at the time of a later survey. This enabled us to observe the increase in knowledge associated with receipt of the Statement. It also enabled us to compare groups of younger workers who had not received the Statement in either year and observe whether knowledge about Social Security changed even in its absence. In addition, we used a unique source of information in the analysis: the surveys conducted by the Gallup Organization in 1998 and 2001.

We assessed how younger workers' understanding of Social Security, in both broader program-level aspects and narrower benefit-specific aspects, changed across time and with receipt of the Statement; and we found that Statement receipt is associated with large and statistically significant increases in knowledge. Younger workers who received the Statement were very knowledgeable about the programs that Social Security offers.⁴ Particularly wide majorities understood that payroll taxes finance benefits and that benefit levels depend on lifetime earnings. However, far fewer young people knew that the full retirement age (FRA)—the age at which one becomes eligible for full retirement benefits—would begin rising incrementally in 2003, and even fewer were aware that benefits are inflation-indexed. In the article's Conclusion, we discuss how being unaware of those facts might negatively affect the retirement security of younger workers and suggest approaches SSA might take to address that knowledge gap.

Social Security Statement Background

This section presents a brief overview of the Statement. First, we discuss the history of its implementation. Next, we describe its content and appearance, to indicate how the Statement conveys various types of information to recipients. Then, by discussing the Gallup surveys, we summarize the efforts to measure the Statement's effectiveness.

Implementation

The Omnibus Budget and Reconciliation Act of 1989 amended the Social Security Act to require SSA to issue estimated benefit and earnings statements beginning in 1995.⁵ The legislation mandated that SSA mail these statements to workers aged 60 or older in fiscal year (FY) 1995 and, in FYs 1996 through 1999, to workers turning 60 during those years.⁶ The legislation also required SSA to send estimated benefit and earnings statements to all eligible workers aged 25 or older beginning in FY 2000. SSA modified the mailing schedule specified in the legislation to include increasingly younger recipients during FYs 1996 through 1999. As shown in the implementation schedule below, this modification enabled the agency to increase the volume of mailings gradually over time.

SSA staggered the Statement mailings throughout each year, with workers receiving their Statements about 3 months before their birthdays. Note that from this implementation schedule, we can identify, by their age, which workers would be unlikely to have received a mailing by a specific date. For example, workers aged 25 did not begin receiving mailings until FY 2000.

Content and Appearance

Legislation determines the basic content of the earnings and benefit statements. The Omnibus Budget and Reconciliation Act of 1989 specified that Statements must contain the worker's earnings history and Social Security and Medicare taxes paid; estimated retirement benefits payable at early retirement age (62), FRA, and age 70; estimated disability, survivor, and auxiliary benefits payable on the worker's account; and a description of benefits payable under Medicare. The Social Security Protection Act of 2004 further mandated that the Statement include sections on the Windfall Elimination Provision and the Government Pension Offset beginning in 2007.⁷

Over time, the content and the placement of information in the Statement have undergone slight changes. In FY 1995, the first page of the original Statement contained a message from SSA's commissioner. The second page provided the worker's earnings history as well as Social Security and Medicare taxes paid. The third page contained the worker's estimated retirement, disability, and survivor benefit amounts, as well as a description of Medicare benefits. The fourth, fifth, and sixth pages provided additional information (including data sources and estimation methods) on the worker's earnings record, Social Security taxes paid, Social Security and Medicare credits, estimated benefits, types of benefits, and the effect of working while receiving benefits. SSA used this format for the Statement through FY 1999.

The FY 2000 version of the Statement reflected both content changes and significant focus-group-tested design changes. It eliminated information on taxes paid in each earnings year and provided only cumulative lifetime Social Security and Medicare taxes paid. A new paragraph encouraged recipients to think about the advantages and disadvantages of retiring early. A list of publications on topics related to retirement benefits also appeared.

The design changes included shortening the Statement from six to four pages and rearranging the order of presentation so that information on retirement, disability, and survivor benefits preceded the worker's earnings history. The revised Statement showed only two columns of numbers representing taxed Social Security earnings and taxed Medicare earnings for each year in a worker's earnings history. The Statement was modified so that numbers filled only one-half of a page, instead of an entire page as before. More white space and greater use of different font sizes and styles made the Statement easier to read.

Effect on Public Awareness

As the agency was implementing the Statement, it sought to measure the Statement's effect on public awareness of and knowledge about Social Security. SSA identified this objective in its strategic plans and commissioned surveys to assess the Statement's impact.

Strengthening public understanding of Social Security programs was one of the five goals of SSA's Strategic Plan 1997–2002: Keeping the Promise, issued in September 1997.⁸ In 1998, as part of that strategic plan, SSA established the Public Understanding and Management System, under which it commissioned the Gallup Organization to conduct six surveys between 1998 and 2004 to evaluate SSA's outreach efforts, including the Statement.

The first survey, conducted in 1998, provided the baseline for public knowledge about Social Security. It found that Americans aged 18 or older were relatively well-informed about basic program facts. Eighty percent of respondents knew that Social Security provides survivor benefits, 83 percent knew that Social Security provides disability benefits, 87 percent knew that a tax on workers and employers finances Social Security benefits, and 89 percent knew that benefit amounts depend on earnings histories. However, fewer respondents knew certain facts affecting their future benefits: Only 65 percent knew that the FRA was going to rise, and only 59 percent knew that benefits increase with the cost of living. Respondents who stated they had received a Statement knew more about Social Security's programs and benefits than did those who did not report receiving a Statement.

The 2001 survey results revealed a significant increase from 1998 in the percentage of respondents who knew that Social Security provides survivor benefits (88 percent versus 80 percent), that Social Security provides disability benefits (89 percent versus 83 percent), that Social Security is financed by a tax on workers and employers (93 percent versus 87 percent), and that benefit levels depend on earnings histories (93 percent versus 89 percent). Although higher percentages of respondents in 2001 knew that the FRA would rise (70 percent versus 65 percent) and that benefits increase with the cost of living (67 percent versus 59 percent), knowledge in 2001 about their own benefits still lagged behind their Social Security program knowledge. Nonetheless, these results indicate a positive impact of SSA's public information campaign.

As required by its implementation schedule, SSA had sent Statements to workers aged 47 or older by the time of the first Gallup Organization survey in 1998. In 2001, about one-half (46.8 percent) of respondents aged 18 or older reported receiving a Statement in the year prior to the survey. Thus, simple contrasts of the percentages of correct responses for all workers surveyed in 1998 with those surveyed in 2001 do not cleanly capture gains in knowledge associated with Statement mailings because in 1998, many older workers had already received a Statement.

To establish a more meaningful comparison, we focused on workers aged 46 or younger who, because of their age, probably would not have received a Statement at the time of the initial survey. Then, we contrasted their knowledge with that of workers aged 46 or younger several years later, after full implementation of the Statement. Using that approach, we were able to highlight more clearly how much additional information the Statements conveyed to younger workers. Although it would have been useful to provide similar contrasts for older workers, the dating of the surveys and the implementation schedule of the mailings do not align in a way that would have allowed us to perform that analysis. Nonetheless, we think researchers and policymakers can learn about the effectiveness of the Social Security Statements by examining only the younger individuals.

Data and Methodology

We used data from the first Gallup survey, conducted in October and November of 1998, and data from the fourth survey, conducted between August and December of 2001.⁹ Both were nationally representative surveys of adults aged 18 or older.¹⁰ Gallup interviewed 4,000 respondents in 1998 and 20,700 respondents in 2001.¹¹ Each survey included both broad-scope questions about Social Security programs and narrower questions about benefits that would more likely apply directly to a respondent's own situation. We focused on adults aged 46 or younger, relatively few of whom would have received a Statement at the time of the 1998 survey. To that end, we excluded from our calculations anyone who reported in that survey that he or she had ever received a Statement.¹² The 1998 survey thus serves as the baseline, measuring the level of understanding of Social Security programs and benefits among younger workers prior to receipt of the Statement.

Following the implementation schedule that sent Statements to progressively younger workers each year, SSA sent at least two and possibly three Statements to individuals aged 46 or younger by the time of the 2001 survey.¹³ By using the responses of persons aged 46 or younger to the 2001 Gallup survey and contrasting them with the responses to the same questions in 1998, when those aged 46 or younger would not have received a Statement, we were able to evaluate the Statement's effectiveness over time.

To assess the impact of the Statement on younger workers' knowledge, we looked at the percentages of respondents that correctly answered each of six questions: three about Social Security benefits and three about program aspects. Following is a list of the questions, presented verbatim.¹⁴

Benefits-knowledge questions

1. Social Security benefits go up automatically with the cost of living. (a) agree, (b) disagree, (c) don't know.
2. Is the youngest age you can retire and collect FULL Social Security retirement benefits fixed or will it rise in the future? (a) fixed, (b) will rise in the future, (c) don't know, (d) (refused).
3. Do all people who receive Social Security retirement benefits receive the same amount, or does it depend on how much people earned when they were working? (a) people receive the same amount, (b) it depends on how much people earned when they were working, (c) don't know, (d) (refused).

Program-knowledge questions

4. Social Security provides benefits to families of workers who die.¹⁵ (a) agree, (b) disagree, (c) don't know.
5. Social Security pays benefits to workers who become disabled.¹⁶ (a) agree, (b) disagree, (c) don't know.
6. Social Security is paid for by a tax placed on both workers and employers. (a) agree, (b) disagree, (c) don't know.

In all instances, the percentages of respondents answering “don't know” or refusing to answer were very low.¹⁷ For each question, we calculated the percentage of correct responses for all workers in our sample in the two survey years.¹⁸ We then disaggregated the results by education, income, race/ethnicity, and sex to see if the impact of the Statement differed across or within demographic groups.

For each of the following three pairings, we examined the difference in the percentage of correct responses:¹⁹

• Individuals who did not receive a Statement in 1998 and individuals who did not receive a Statement in 2001;
• Individuals who did not receive a Statement in 1998 and individuals who received a Statement in 2001; and
• Individuals who did not receive a Statement in 2001 and individuals who received a Statement in 2001.

The first two pairings allowed us to distinguish the effect of Statement receipt from general changes in knowledge about SSA's programs and benefits that might have resulted from information obtained through other channels. The third pairing enabled us to assess the impact of the Statement within a single year. We have also provided breakdowns by education, income, race/ethnicity, and sex in order to determine if initial and subsequent levels of knowledge are associated with any of those characteristics.

We expected to see an increase in the percentage of correct responses between 1998 and 2001, even for respondents who did not receive a Statement in either year. One reason we did so is that SSA embarked on an extensive public information campaign during the rollout of the Statement.²⁰ Another reason is that younger workers who did not receive a Statement in 2001 might have received (and read and remembered) a Statement in 2000; or, they might have had contact with one or more Statement recipients in 2001. However, we expected to see an even greater increase in the percentage of correct responses between the younger workers in 1998 who did not receive a Statement and the younger workers in 2001 who did. To illustrate, we expected the following general trend in the percentages of questions answered correctly in each respondent group:

After calculating the actual percentages of correct responses, we calculated standard errors for the differences in the percentage of correct answers that we observed within each of the three pairings. We then used a two-tailed t-test to determine whether the changes in the percentage of correct answers between the groups in each of the three pairings were statistically significant.

Results

Table 1 shows that, for all workers aged 46 or younger in 2001, the percentage of correct responses for five of the six questions examined was greater among respondents who reported receiving a Statement than among those who did not. That outcome is consistent with our expectations. The table also shows that, for most questions, a greater percentage of respondents answered correctly in 2001 than did so in 1998, regardless of whether they received a Statement in 2001.²¹

However, the increase in the percentage of correct answers between the two years was noticeably larger for those who received a Statement in 2001. For example, for individuals who did not receive a Statement in either year, correct answers increased by 1.2 percentage points (from 78.5 percent to 79.7 percent) between 1998 and 2001 for the program-knowledge question about Social Security providing disability benefits. Statement recipients in 2001 were 6.7 percentage points more likely than 1998 respondents to answer that question correctly (from 78.5 percent to 85.2 percent). Similarly, for individuals who did not receive a Statement in either year, correct answers increased by 2.4 percentage points (from 62.4 percent to 64.8 percent) for the benefits-knowledge question about the future rise in the FRA. Statement recipients in 2001 were 9.3 percentage points more likely than their 1998 counterparts to answer that question correctly (from 62.4 percent to 71.7 percent). That outcome suggests that the Statement is effective in informing the public about the programs and about benefits.

In significance tests, we found that the percentage of correct answers for respondents who received a Statement always differed significantly from the percentage of correct answers for respondents who did not. In all but one case, the percentage of correct answers was significantly higher for respondents who had received a Statement than it was for those who had not. In the anomaly, Statement recipients were less likely to provide the correct answer to the question about benefits increasing with the cost of living than were 2001 respondents who had not received a Statement.

Knowledge By Topic: Benefits Versus Programs

Table 1 also shows that younger workers were more knowledgeable about the types of programs SSA administers (and program financing) than they were about the details of the benefits they could expect to receive. In 1998, before most respondents aged 46 or younger had received a Statement, 76.1 percent knew that Social Security provides survivor benefits, 78.5 percent knew that Social Security provides disability benefits, 88.0 percent knew that Social Security is financed by a payroll tax, and 81.4 percent knew that benefit levels depend on earnings. By contrast, only 62.4 percent of respondents knew that the FRA was scheduled to rise, and only 43.8 percent knew that Social Security benefits are inflation-indexed.

Among 2001 respondents who had received the Statement, 85 percent or more correctly answered all three of the program-knowledge questions as well as the question about benefit levels depending on earnings. However, even after receipt of the Statement, only 71.7 percent knew that the FRA was going to rise, and only 49.8 percent knew that benefits increase with the cost of living. Those lower percentages hint at potential vulnerability in the retirement security of those younger individuals.

Also of interest is the correlation between the frequency of correct responses and the placement of information in the Social Security Statement. High percentages of respondents correctly answered program-knowledge questions about survivor and disability benefits. The Statement displays information on those topics at the top of the second page, in a section titled “Your Estimated Benefits,” along with the estimated benefit amounts the individual can expect to receive. The Statement presents information on the third aspect of program knowledge, payroll taxes, at the top of the third page in a column of numbers labeled “Your Taxed Social Security Earnings,” under the section heading “Your Earnings Record.” By contrast, the Statement does not present information underlying two aspects of benefits knowledge so prominently. It contains a single sentence on the cost-of-living adjustment of benefits at the bottom of the second page, in the middle of a description of the benefit calculation. An individual who glances over this section might not find that information. The Statement also does not mention the increase in the FRA directly; rather, that information is noted indirectly, on the line that provides the benefit estimate at FRA, by including the individual's own FRA in parentheses. The percentages of correct responses to those two benefits-knowledge questions are lower than the percentages for the program-knowledge questions. However, the Statement notes the relationship between benefits and earnings—the subject of the third benefits-knowledge question—in two prominent locations (the first paragraph of the benefit-estimation section on the second page and the “Your Earnings Record” section at the top of the third page). Respondents were highly knowledgeable about the dependence of retirement benefit levels on lifetime earnings.

We do not argue that the placement of information in the Statement has a cause-and-effect relationship with the percentage of correct answers about Social Security's programs and benefits. Other SSA informational publications may also play a role. We reviewed seven publications (five available in both print and online formats, and two online only) that provide information on retirement benefits.²² Of the five publications available in both formats, four furnish the program knowledge that Social Security provides survivor benefits and the benefits knowledge that the FRA is rising and that benefit levels depend on earnings histories.²³ Three of those publications note that Social Security provides disability benefits (program knowledge). Only two of them note that Social Security is financed by a payroll tax (program knowledge), although that information is also widely communicated on pay slips and W-2 forms. Interestingly, only one publication notes that Social Security benefits increase with the cost of living (benefits knowledge)—a fact that is not readily available elsewhere. Our review found that the Statement and other relevant SSA publications place more emphasis on providing program information than on benefits information. One of the consequences of that emphasis is that survey respondents are less likely to be knowledgeable about the specifics of their benefits than they are about the programs generally.

We recognize that younger workers could have learned about Social Security, especially its programs, from other sources. For example, high school and college courses in history, social studies, economics, and civics often contain references to and discussions of Social Security. Further, younger workers could have family or friends who are Social Security retired-worker, survivor, or disabled-worker beneficiaries.

Program and Benefits Knowledge by Demographic Characteristics

Statement receipt increases the percentage of correct answers, regardless of education level, income level, race/ethnicity, or sex. Table 2 shows the percentage of correct responses across demographic groups for one sample program-knowledge survey question and one sample benefits-knowledge survey question. (The appendix tables show the percentages of correct responses to all six questions for all demographic groups.)²⁴ We found no strong pattern across education levels in the percentage of correct responses, although respondents in the lowest levels tended to have lower percentages. For example, the percentage of correct responses to the benefits-knowledge question in 1998 ranged between 41.7 percent and 60.1 percent for those with no more than a high school diploma and between 66.1 percent and 67.0 percent for those with some college or a college degree. However, we note anomalies, such as the lower percentage of correct answers among those with some postgraduate education than that for respondents with only a high school diploma. Likewise, for the program-knowledge question, the percentage of correct answers in 1998 generally increased with the level of education, albeit not as sharply, from 73.8 percent for those without a high school diploma to about 78 percent for respondents with at least some college (overlooking the anomalous 64.0 percent correct for those with trade, technical, or vocational training). The results for the other benefits- and program-knowledge questions follow a roughly similar pattern (see Table A-1).

In many cases, respondents in the education levels that had the lowest percentages of correct responses in the first survey year registered a statistically significant increase in correct responses after receiving the Statement. For example, Table 2 shows that respondents in the three education levels with the lowest percentage of correct responses to the benefits-knowledge question in 1998 (less than a high-school diploma, high school diploma only, and postgraduate work or degree) exhibited the greatest percentage-point increases after Statement receipt in 2001.²⁵ We noticed similar results for the program-knowledge question. In 1998, respondents with no more than a high school diploma and those with trade, technical, or vocational training had the lowest percentages of correct responses. Those three groups exhibited the largest increases following receipt of the Statement. In effect, the Statement compressed the range of the percentage of correct responses across the education levels. In 1998, the difference between the lowest and highest percentages of correct responses to the benefits-knowledge question across the education levels was 26.3 percentage points (from 41.7 percent to 68.0 percent). Receipt of the Statement reduced that range to 13.0 percentage points (from 64.3 percent to 77.3 percent). For the program-knowledge question, receipt of the Statement reduced the range of correct responses from 14.7 to 4.2 percentage points. Consequently, the Statement appears to provide the most help to those with the least knowledge about Social Security.

In comparing results by income level, we note that respondents with incomes below $50,000 tended to have lower percentages of correct answers in 1998. For the benefits-knowledge question, the percentages of correct answers among respondents in the three income groups below $50,000 ranged from 54.7 percent to 61.5 percent. By contrast, the percentages of correct answers among the income groups above $50,000 ranged from 65.2 percent to 69.5 percent. However, results for the program-knowledge question reveal a less consistent pattern. Broadly, the percentage of correct responses for those with incomes under $50,000 was around 73 percent, while for those with incomes of $50,000 or more, it was closer to 80 percent. However, the percentage of correct responses to the program-knowledge question spiked to 77.4 percent among those with incomes from $20,000 to $34,999; and for those with incomes of $100,000 or more, it dipped to 72.5 percent. Similarly, no consistent pattern emerges across income levels for the other survey questions (see Table A-2).

As with the educational levels, the income groups with the lowest percentages of correct responses in the first survey tended to register the largest increases in the percentage of correct responses after receiving the Statement. Table 2 shows that the increases in the percentage of correct answers to the benefits-knowledge question were larger and statistically significant for those with incomes under $50,000. Results for the program-knowledge question reveal a similar pattern, although the highest income group ($100,000 or more) had one of the highest increases in the percentage of correct responses (14.3 percentage points). Again, receipt of the Statement reduced the range in the percentage of correct responses across the income levels. In 1998, the difference between the lowest and the highest percentages of correct answers to the benefits-knowledge question by income group was 14.8 percentage points (from 54.7 percent to 69.5 percent). Receipt of the Statement reduced that range to 5.4 percentage points (from 69.0 percent to 74.4 percent). For the program-knowledge question, Statement receipt reduced the range of correct responses from 12.9 to 5.3 percentage points. That outcome further indicates that the Statement most helps the people with the least knowledge about Social Security's benefits and programs.

Among racial/ethnic groups, Table 2 shows that the percentage of correct responses to the benefits-knowledge question in 1998 was highest for white respondents, followed by blacks, Asians, and Hispanics. However, the ranking of the racial groups by percentage of correct answers differed for 2001 respondents with no Statement, and differed yet again for 2001 respondents who received a Statement. That general pattern of changing rankings was repeated for the other benefits-knowledge questions and the program-knowledge questions (see Table A-3). An increase in the percentage of correct answers after receipt of the Statement was statistically significant only for white respondents (for five of the six of the questions) and for black and Asian respondents (for one of the benefits-knowledge questions). However, as was seen with the education and income groups, receipt of the Statement narrowed the difference between the lowest and highest percentages of correct answers across the racial/ethnic groups.

We found slight and essentially offsetting differences by sex in the percentage of correct responses to the benefits- and program-knowledge questions, as each sex scored higher than the other on three of the six questions (see Table A-4). Men were generally more likely to have higher scores on the benefits-knowledge questions, while women tended to score higher on the program-knowledge questions. After Statement receipt, the differences between men and women in the percentages of correct answers narrowed for all questions.

Conclusion

We found that younger workers were better informed about Social Security in 2001 than they were in 1998. Before the introduction of the Statement, more than 76 percent of individuals aged 46 or younger knew that Social Security provides survivor and disability benefits and that a payroll tax finances those benefits. Eighty percent knew that the Social Security benefit amount depends on a worker's earnings history. Those high percentages resulted in part from the extensive outreach campaign SSA initiated in early 1998. In 2001, after the distribution of the Statement, close to 90 percent of younger workers knew about SSA's programs, the financing of its benefits, and the relationship between benefits and earnings.

However, even after Statement receipt, only about 70 percent of respondents knew that the FRA was going to rise, and less than 50 percent knew that benefits are inflation-indexed. The gap in knowledge about those program aspects could affect the retirement security of some younger workers. For example, an individual who mistakenly equates Social Security benefits with noninflation-indexed retirement savings might see no difference between claiming benefits early and drawing down his or her retirement savings. Additionally, as the age of eligibility for full benefits rises, benefits claimed at earlier ages will be subject to greater reductions. Individuals who do not know that benefits are inflation-indexed and that the FRA is rising are at risk of claiming benefits before their optimal time. It is thus important for workers to know those facts and to understand the implications for the benefits they will receive.

We also noticed a correlation between the type of information SSA provides in its publications, including the Statement, and the percentages of correct answers to the benefits- and program-knowledge questions. SSA provides program information in most of the publications we reviewed. The agency is less likely to provide information on benefits, especially information on the inflation-indexing of benefits, which is rarely mentioned in its publications.

Our results suggest that further SSA outreach efforts should provide more detail on benefits rather than overemphasizing program knowledge. Further, the lack of consistent patterns across demographic categories suggests that SSA should distribute additional information widely among the general population rather than targeting it to particular groups.

SSA might consider developing and testing some format and content changes for future versions of the Statement (whether mailed or online) to emphasize the adjustment of benefits for the cost of living and the rise in the FRA. The surveys cited in this article demonstrate that the Statement is effective in increasing recipients' knowledge of SSA's programs and benefits. Providing more emphasis on benefits information in the Statement and in SSA's other publications seems likely to increase younger workers' knowledge about their benefits, as future surveys could verify.

SSA should revise its publications to add information about the adjustment of benefits for the cost of living and its implications for retirement security because few of the publications that younger workers are likely to use currently mention inflation indexing. Similarly, the agency might highlight benefit information in e-mails explaining how to access the online Statement. SSA might also develop and test informative messages using social media such as Twitter, Facebook, and YouTube to deliver information on benefits and to encourage younger workers to access the Statement online.

Appendix