Statement on Retirement Readiness in America Approaches for Retirement Security in the United States Before the ERISA Advisory Council

Transcription

1 Statement on Retirement Readiness in America Approaches for Retirement Security in the United States Before the ERISA Advisory Council Stephen P. Utkus, Principal Vanguard Center for Retirement Research September 17, 2009 Chair [Elizabeth J.] Dill and Vice-Chair [Marc] LeBlanc Thank you for the opportunity to address the ERISA Advisory Council on the question of retirement readiness in America. As you know, Vanguard is one of the leading retirement services providers in the U.S. We provide administrative and investment services for both defined contribution and defined benefit plans. We are also known for our investment services provided to individual investors and financial advisors. As of June 30, 2009, Vanguard managed assets of $1.1 trillion on behalf of our clients and we estimate that more than three-quarters of these assets are in some way directly or indirectly linked to the goal of retirement saving. At Vanguard, I head up our Center for Retirement Research. The Center is part of our Strategic Retirement Consulting organization, which is responsible for retirement policy, research and consulting activities at Vanguard. For several decades, I have been an avid student of the research on the question of how well prepared Americans are for retirement. My goal in this testimony is to provide you with a synopsis of my interpretation of this research. In so doing, I hope to assist you in your deliberations on the broader question of the evolution of retirement systems in the United States. Who is ready for retirement? At first blush, you might imagine that it would be a relatively easy task to understand who in America is and isn t prepared for retirement. Unfortunately, that is not the case. Because of different data sources and methodologies for calculating retirement preparedness, and because of an ever-changing retirement landscape, there are somewhat divergent views about how well prepared Americans are for retirement today and how well prepared they will be in the future. 1 One look at readiness is provided by three studies in Figure 1. (I emphasize that throughout this testimony, the interpretation of the studies I cite is my own, not necessarily that of the authors.) From this set of studies, it appears that about four in 10 Americans are inadequately prepared for retirement. The size of the well-prepared group ranges from 30% to nearly 60%. And as two studies demonstrate, there is also an intermediate or potentially prepared group. These are households that need to take additional but not insurmountable steps, like working a few years longer or saving a few more percentage points of income, to attain prepared status. 1 Skinner (2007) provides an excellent and accessible overview of the challenges of estimating retirement readiness over the lifecycle. 1

2 Figure 1. Studies of U.S. Retirement Readiness Degree of retirement preparedness Potentially Not Study Prepared Prepared Prepared 1. Moore and Mitchell (2000) 30% 30% 40% 2. Ameriks and Utkus (2006) 40% 17% 43% 3. Munnell et al. (2006) 57% 43% Source: Author s computations from cited studies. This first set of studies underscores why I believe it is not useful to describe retirement readiness in either/or terms. These studies clearly indicate that it is a mistake to claim that all or most Americans are well prepared for retirement. But it would also be a mistake to conclude that all or most Americans are poorly prepared. A better approach in thinking about the policy issue is to imagine (at least) three groups: the well-prepared, the inadequately prepared, and the individuals who are part way there and need to take additional action. The scientific debate is about how large each of these groups is and whether the size of these groups is changing over time as the nature of retirement and retirement systems changes. Another fact about retirement readiness is to recognize that small changes in behavior today can produce meaningful improvements in readiness outcomes, at least for part of the U.S. population. As shown in Figure 2, one study (Munnell et al., 2006) initially found that 57% of American households were prepared for retirement and 43% were not. But if American households saved three percentage points more of their income, or delayed retirement until age 67, the proportion judged to be retirement-ready increased to as much as two-thirds (depending on the specific change). It s not a large jump, but it is still meaningful. Another study (Moore and Mitchell, 2000) found that saving shortfalls can be cut in half by delaying retirement three years, from age 62 to 65. Figure 2. Changing Savings Rates and Retirement Ages Degree of retirement preparedness Potentially Not Study Prepared Prepared Prepared 3. Munnell et al. (2006) Initial estimate 57% 43% If households save 3% more 64% 36% If households retire at age 67 68% 32% Source: Author s computations from cited studies. These studies point to an important lesson namely, that there is a group of Americans who, with relatively small changes in behavior, could improve their outlook for retirement in a meaningful way. They also remind us that there is a sizeable group, the unprepared, for whom small changes are not enough. 2

3 So far, the studies I ve cited suggest a somewhat large minority of Americans is inadequately prepared for retirement, although the typical American is on track for an adequate retirement or can reasonably get on track. Another recent set of studies, summarized in Figure 3, paints a more optimistic view. These studies suggest that the problem of inadequate preparation is relatively small, and is limited mainly to the one-fifth of poorest households in the U.S. These researchers argue that 80% or more of American households are well prepared for retirement. Figure 3. Three Additional Assessments of Retirement Readiness Degree of retirement preparedness Potentially Not Study Prepared Prepared Prepared 4. Engen, Gale and Uccello (2004) Many or Perhaps one-fifth* most* 5. Scholz, Seshadri and Khitatrakun (2006) 80% 20% 6. Love, Smith and McNair (2008) 87% 13% * Given the nature of this study, it isn t really possible to assign definitive percentages to these categories. Source: Author s computations from cited studies. Taken together, these studies come to one of two conclusions: either the problem of poor retirement preparation is rather widespread, affecting four in 10 (or more) of American households, or the problem is about half that, affecting the poorest two in 10 households. Another observation I draw is that there is a group of households which is partially prepared, but which could become retirement-ready by taking manageable steps. The critical role of assumptions What accounts for these two distinct assessments of retirement readiness in the US? The biggest explanation is differences in assumptions. On a personal level, to determine whether you are prepared for retirement, you must go through a fairly arduous financial planning process gathering all asset and liabilities for your household as well as benefit statements for all current and prior employers, providing all of this information accurately to a qualified planner, and then engaging in a discussion of income goals and retirement needs. Next, you or your planner must make a series of complex assumptions about how the future will unfold projecting future incomes, expected retirement dates, lifespans, inflation rates and market returns, taxes, and estimates of retirement wealth in personal accounts, in retirement plans and from government programs. To compute retirement readiness at the national level, economists in academia, industry, and the government undertake similar calculations for various demographic groups. Yet they are hampered by important data limitations. One major source of data on retirement readiness, the Survey of Consumer Finances (produced by the Federal Reserve), is well known for its high-quality data on American wealth holdings. But it has limited information on the features of employer retirement plans that many Americans participate in and knowledge of those features is critical to understanding retirement readiness. Another excellent survey, the Health and Retirement Study, is much richer in terms of data on retirement plans and household finances, but it only includes households age 50 and above. There are other data sets that are helpful, including a popular March supplement from the U.S. Census Bureau s Current Population Survey, but these too have limitations. It s as if you, as an individual, went to a financial planner, but provided him or her with only part of your financial picture, or only a summary of your financial situation rather than the details. 3

4 There are other differences between the studies, most importantly their varying assumptions and methodologies. Probably the most important assumption differentiating them is the definition of what constitutes an adequate retirement. The various studies usually tackle this in one of two ways an income replacement ratio, the practitioner s view you re probably familiar with, or a consumption smoothing model, an academic view of how much in spending an individual would prefer over a lifetime. In the end, regardless of which approach they take, the studies with more optimistic results tend to set lower targets for what constitutes an adequate income or an adequate level of resources. Consider a calculation I undertook using the results from one study (Love et al., 2008). Their top-line result, using a replacement rate yardstick, was that only 13% of Americans are not retirement ready. 2 But to arrive at this conclusion, the researchers assumed that all Americans would take their entire stock of liquid and illiquid wealth, including all financial assets and housing equity, and convert it to an inflation-adjusted annuity. They further defined as adequate an inflationadjusted annuity (with Social Security and any DB pensions) that would replace 50% of pre-retirement earnings. By contrast, suppose we were to use a higher replacement standard, such as the two-thirds income replacement rate proposed by Mitchell and Moore (2000). Then, as shown in Figure 4, the prevalence of poorly prepared households jumps from 13% to 29% (based on my calculations from the study s published results). If we further introduce a more realistic treatment of housing equity, and assume that only 50% of that amount can be accessed and converted to an income stream (itself a somewhat unrealistic assumption, as few households actually do this today), the level of poorly prepared households jumps to 34%. Figure 4. The Impact of Assumptions Degree of retirement preparedness Potentially Not Study Prepared Prepared Prepared 6. Love, Smith and McNair (2008) Initial results 87% 13% Higher replacement rates (66%) 69% 29% With only half of housing 66% 34% Source: Author s interpolations from Table 7 of Love et al. (2008). This illustration gets to the heart of an important policy question. Should the aim of policy be to ensure that Americans reach some minimal level of retirement adequacy, such as a modest replacement rate or perhaps a low multiple of the federal poverty level? Or should the goal be to reach a much higher standard, something like the conventional practitioner s view that households should replace 70% to 80% of their pre-retirement income? As you can see, the former goal (a minimum standard) leads you to a more optimistic view of retirement preparation than the latter (a more expansive standard). 2 They also reported that 18% of Americans were unprepared for retirement using 150% of the federal poverty level as a measure. 4

5 Overall I believe that the more optimistic studies set the bar too low. It seems likely to me that more than one-fifth of American households are poorly prepared for retirement. Moreover, there are other reasons to believe that retirement estimates may be skewed too positively. Many studies assume that Americans convert all of their retirement savings into an inflation-adjusted annuity, and use their home equity for retirement income. Yet few actually do. 3 Another common but untenable assumption is that scheduled Social Security benefits can be paid without reducing benefits or raising payroll taxes, or that Medicare premiums will remain constant. A strength of this research, I should emphasize, is that these studies take a comprehensive view of retirement readiness, examining behavior at the household, not individual level, and including all assets (and liabilities) accumulated by households, including Social Security, employer benefit plans, personal assets, and, as I ve noted, the value of their primary residences. I note this comprehensive view of wealth for the Advisory Council for a reason. The Council is typically focused on improvements to the qualified plan system, which covers approximately half of private-sector workers in the U.S. Many sponsors, consultants and lawyers working in the qualified plan system tend to think about retirement readiness exclusively in terms of assets or benefits in a given employer s qualified plan. But qualified plans only represent one component of wealth accumulated by U.S. households. 4 Any analysis of readiness must do more than consider the benefits accumulated in the retirement program of one s current employer. From my readings, I draw several conclusions about retirement readiness. First, it is possible to take a somewhat pessimistic view of readiness, what I sometimes call the 1/3 rule namely, that 1/3 of Americans are clearly prepared, 1/3 are unprepared, and 1/3 are possibly prepared depending on what actions they take. Second, it is possible to argue for a more optimistic interpretation, namely that two-thirds of Americans are prepared or could be (if they make some reasonable changes), leaving the focus on the one-third who are poorly prepared. Finally, it is wrong to conclude that today, most Americans aren t prepared for retirement. Rather it s more accurate to say that a large majority is prepared or could be prepared by taking certain manageable steps. Finally, it s important to emphasize that studies I ve reviewed here are only the tip of the iceberg when it comes to understanding retirement preparedness. In our offices at Vanguard, we have several foot-high stacks of retirement readiness studies, and we add regularly to those stacks. A recent Wharton Pension Research Council conference was devoted to the topic and produced an important reference volume in this area (Madrian, Mitchell, and Soldo, 2007). Additional studies appeared at this year s Retirement Research Consortium conference in Washington. As you hear additional views about the state of retirement readiness, I hope that my analysis affords you a benchmark for interpreting other studies you may read. 3 A related policy question is whether retirement adequacy should be measured based on what households actually do, or what they could do. Although they choose not to do so, Americans could annuitize savings and convert home equity into income in order to boost retirement outcomes. 4 Moore and Mitchell (2000) find that for the median American, approximately 40% of retirement wealth is from Social Security, and a fifth each from employer-sponsored plans, personal savings and home equity. 5

6 Is readiness getting worse? Many economic studies of retirement readiness strive to give a current snapshot of retirement readiness how the situation looks at a particular point in time. Researchers have more recently turned to analyzing readiness over time, and whether it is improving or deteriorating. The conventional wisdom is, of course, that readiness must be deteriorating. The list of reasons to worry is long, and includes the following developments: The well-known shift from defined benefit (DB) to defined contribution (DC) plans, with the questionable but nonetheless common assumption that current DC plans will be less generous than earlier DB plans; The aging of the U.S. population and the need to slow the growth of Social Security and Medicare benefits in order to make the programs fiscally sustainable; A lack of thrift among the baby boom generation, accompanied by rising debt levels (both mortgage and credit card debt) across society; Two recent large declines in global stock markets, in and , which have led to one of the worst ten-year periods for equity investments; and, Increased longevity combined with a trend toward earlier retirement, meaning that individuals must finance a longer period of retirement with fewer years of work. Research on the dynamic aspects of retirement readiness has arrived at several conclusions. 5 The first, on the good news side of the ledger, is that the typical aging baby boomer has accumulated retirement wealth at a reasonable pace and is likely to have a standard of living similar to (or possibly slightly better) than his or her parents. The second, on the bad news side, is that boomers have had higher working-age incomes and have been better educated than their parents generation; as a generation, they should have saved and accumulated more wealth than their parents. Unfortunately, relative to incomes and education, they have not. As a result, their retirement replacement rates (retirement incomes as a percent of working age incomes) will be lower, and so boomers may feel relatively more economically constrained than their parents, even while in absolute terms they may be similarly or better off. And third, the trend toward growing income inequality is now predicted to carry over to retirement. Thus the median boomer may do as well or better than the generation before. Yet higher income retired households will do even better, and lower income retirees even worse, than in the past. What about the developments reshaping the retirement landscape? With reasonable confidence, we can say that the long-term financing problems facing Social Security and Medicare are likely to put pressure on retirement readiness levels (just as, for prior generations, the increasing generosity of Social Security and Medicare improved readiness). As an illustration of this point, Figure 5 adds a scenario (from Love et al. based on my calculations) where Social Security benefits are reduced by 25% and households must pay for projected increases in out-ofpocket medical costs. 6 In this scenario, the unprepared category rises to 40% of Americans. 5 See Madrian, Mitchell, and Soldo (2007) for a comprehensive set of papers in this area, including in that volume, Butrica, Iams and Smith. Also see Love et al. (2008). 6 This chart also includes my two earlier assumptions: a 66% replacement rate with an inflation-adjusted annuity, and the use of only half of housing equity to finance retirement. 6

7 Figure 5. The Impact of Assumptions and Future Changes Degree of retirement preparedness Potentially Not Study Prepared Prepared Prepared 6. Love, Smith and McNair (2008) Initial results 87% 13% Higher replacement rates (66%) 69% 29% With only half of housing 66% 34% With Social Security, health costs shock 60% 40% Other factors, such as the trend from DB to DC and recent equity market losses have more complex effects. It is true that generous DB pension benefits at large companies are being reduced or eliminated; a significant group of Americans have been affected. But it is not at all clear how this affects the broader population, since generous large-company DB pensions only covered a small part of the American working population. DC plans, which are becoming a substantial source of wealth for a more mobile workforce, are expected to generate benefits in excess of those available from DB plans. 7 So for a variety of reasons, the shift from DB to DC plans, I believe, is likely to enhance retirement well-being for the typical American worker. In terms of market risk, the impact of the recent market declines is likely to have its greatest impact on Americans approaching or in retirement. The weak equity market returns over the past decade we sometimes refer to as equity cohort risk the risk of retiring after a period of weak capital market returns. Possible responses to the downturn include working longer, saving more, or choosing to live at a lower standard of living in retirement. But whether the market decline has an enduring effect will depend on the period over which markets recover, and how many of these offsetting steps households actually take. It s also important to recognize that many households approaching retirement and experiencing equity market losses have also benefited from prior periods of strong equity market returns. The impact of cohort risk cannot be judged over a single market cycle, but must be evaluated over a working and saving career. On balance, it seems to me that major factors influencing retirement readiness in the future will be the baby boom generation s lower saving rate, plus the shortfalls in Social Security and Medicare. But before concluding that the future consists mostly of financial headwinds, it s important to recall several other features about the next generation of retirees. First, baby boomers have had higher incomes and higher levels of education than their parents. Second, their general health status is better, and disability levels are lower. These facts suggest that future generations could work longer, either for financial or psychological reasons, and these work earnings could play an important role in financing retirement. Labor force participation of older households has been rising. In 2007, 21% of Americans age were in the workforce, versus 15% two decades earlier (Gendell, 2008). Additional years of work could significantly address part of the saving s shortfall likely to arise from a shifting retirement landscape. 7 See Poterba, Rauh, Venti, and Wise (2007), who find that DC benefits will outpace those from DB, but also note that some DC plans can generate very low levels of retirement saving. 7

8 Who is not ready? One of the major concerns in the policy world is with the population not ready for retirement, and so it is worth spending some time considering the factors leading people to be poorly prepared. Most of the studies on retirement readiness provide some insight into this group of Americans. While each study draws its own conclusions, I identify three sets of factors that appear linked to inadequate retirement preparation. The first is human capital factors. People with low labor force attachment (meaning little or erratic attachment to the full-time workforce), with low lifetime incomes, and with low levels of education, are very likely to be unprepared. I have sometimes said that the best way to improve retirement readiness is not by improving savings rates; instead, it would be to send everyone to college or to get everyone steady full-time work with a great employer. Retirement readiness is clearly linked to the human capital you accumulate in your lifetime, not simply the narrower question of how much you are currently saving. Regular participation in the workforce, along with the accumulation of labor market skills through education and work experience, leads to higher incomes, an increased ability to save, and better retirement outcomes. A second set of factors contributing to retirement readiness are the so-called traditional demographic factors. We know that across the economy, minority households generally are less prepared than white households although whether this is due strictly to race and ethnicity, or to differences in income, education and other socio-economic factors, is an unresolved question. Couples benefit from pooling of household results; thus, being married or partnered helps with retirement adequacy, while being single raises the risk. Because of these pooling effects, an important factor that reduces retirement preparedness is divorce. Widowed women are also particularly vulnerable because of the loss of benefits or assets often associated with their spouse. Further, there are uncontrollable life shocks that can occur, including disability or job loss that can shift someone suddenly from the prepared to the unprepared category. A final set of factors relates to issues that are less tangible but I believe equally important, namely the inability to plan and financial illiteracy. 8 It s hard to know whether the grasshopper or ant mentality (to use Aesop s analogy) is hard-wired or learned in human beings. But what we do know is that, independent of income, education, or other factors, some people seem to lack the skill or motivation to plan for the future, while others are good at it. The lack of planning skills is also linked to a broader set of what are known as behavioral biases, drawn from the field of behavioral finance, such as procrastination and inertia in saving and investing for the future. Another aspect of this problem has to do with low levels of financial knowledge and experience the financial literacy problem. Households have varying skills when it comes to saving, investing budgeting, and debt management, and they have varying degrees of interest in financial matters. Both behavioral biases and financial literacy influence retirement preparation. As we know from recent developments in qualified DC plans, exploiting the tendency toward inertia with automatic enrollment is a valuable mechanism for increasing saving among low-wage or young workers. And as one researcher in the financial literacy field has said to me, one of the best ways to improve retirement readiness might be to teach Americans how not to misuse their credit cards. Policy solutions As members of the Advisory Council are aware, there have been a number of policy initiatives in recent years designed to improve retirement readiness. Although my intention is not to endorse any specific policy proposals, I would like to suggest several ideas the Council may want to keep in mind. 8 See Mitchell and Utkus (2004) for a summary of behavioral biases influencing retirement plan participants, and Lusardi and Mitchell (2007) for an analysis of the links between financial planning, financial literacy, and retirement wealth. 8

9 First, in the Pension Protection Act of 2006, the Congress endorsed a number of initiatives to improve outcomes in qualified DC retirement plans. These included automatic enrollment incentives, better default investment options and strategies for offering investment advice to participants. Continuing these initiatives, and encouraging employers to adopt these strategies, makes sense. A second dimension for promoting retirement readiness has focused on the uncovered population: those who lack access to a qualified plan. Probably the critical benefit that qualified DC plans provide for participants is the availability of a payroll deduction saving system. The auto IRA proposals are very much in this spirit. If implemented in a cost-effective manner, they have the potential to reach a large group of lower-wage workers (wages are generally lower in small firms with no qualified plan benefits), and improve retirement outcomes among such households. A third issue is the question of Social Security. Most reform plans, whether seeking to retain the system s defined benefit character or seeking to add a defined contribution element, have also included the proposal to boost minimum benefits paid by the system. In some ways, this step, in the context of an overall plan for Social Security s financial stability, could be one of the most efficient ways of improving adequacy among many households in the unprepared category. Finally, as I ve mentioned earlier, retirement readiness is broadly linked to the overall condition of the economy. Stable macroeconomic growth, sensible fiscal and monetary policies, productivity growth over time, and improvements in the job market and education systems: all play a role and influence readiness, albeit an indirect one and obviously one not directly within the purview of the Council s discussion today. I d like to conclude by thanking the Advisory Council for focusing its attention on the question of retirement readiness, and for inviting me to offer testimony today. From time to time, it s important to step back from our day-to-day work in the retirement system, and consider our overall progress toward the goal of providing retirement security for all Americans. The good news is that for a large group of Americans, retirement readiness is an attainable goal. The challenge is to develop strategies for assisting those who are poorly prepared, and to think longer-term about the headwinds facing the retirement system in the years ahead. 9

10 References Ameriks, John A. and Stephen P. Utkus Vanguard Retirement Outlook Vanguard Center for Retirement Research, Malvern, PA. Gendell, Murray Older workers: increasing their labor force participation and hours of work. Monthly Labor Review. January Lusardi, Annamaria and Olivia S. Mitchell Baby Boomer Retirement Security: The Roles of Planning, Financial Literacy, and Housing Wealth. Journal of Monetary Economics. 54(1) January: Madrian, Brigitte, Olivia S. Mitchell and Beth J. Soldo Redefining Retirement: How will the Boomers Fare? Oxford University Press, Oxford, UK. Mitchell, Olivia S. and Stephen P. Utkus Pension Design and Structure: New Lessons from Behavioral Finance. Oxford University Press, Oxford, UK. Moore, James F. and Olivia S. Mitchell Projected Retirement Wealth and Saving Adequacy. In Forecasting Retirement Needs and Retirement Wealth, Olivia S. Mitchell, P. Brett Hammond and Anna M. Rappaport (eds.). University of Pennsylvania Press, Philadelphia, PA Munnell, Alicia, Anthony Webb and Luke Delorme A New National Retirement Risk Index. Issue Brief No. 48. Center for Retirement Research at Boston College. June. Engen, Eric M., William G. Gale and Cori E. Uccello Lifetime Earnings, Social Security Benefits, and the Adequacy of Retirement Wealth Accumulation. CRR Working Paper Center for Retirement Research at Boston College. Poterba, James, Joshua Rauh, Steven Venti and David Wise, Defined Contribution Plans, Defined Benefit Plans, and the Accumulation of Retirement Wealth. Journal of Public Economics. 91(10) Scholz, John Karl, Ananth Seshadri and S. Khitatrakun Are Americans Saving Optimally for Retirement? Journal of Political Economy Skinner, Jonathan Are You Sure You re Saving Enough for Retirement? Journal of Economic Perspectives. Summer All investing is subject to risk The Vanguard Group, Inc. All rights reserved. WDCT