Institute Letter to OMB Supporting Department of Labor Proposal on Default Investments (pdf)

May 31, 2007 Ms. Susan Dudley Administrator Office of Information and Regulatory Affairs Office of Management and Budget 725 17th Street, N.W. Washington, DC 20503 Re: Department of Labor Regulation on Qualified Default Investment Alternatives Dear Ms. Dudley: The Investment Company Institute is writing in regard to the Department of Labor’s proposed regulation on qualified default investment alternatives (QDIAs), which will enable employers to select default investments for plan participants that are appropriate for long-term investing. Recently, the American Council of Life Insurers (ACLI) urged the Office of Management and Budget to return the Department’s forthcoming final regulation for further consideration if the QDIA safe harbor does not include capital preservation products sold by the insurance industry, such as stable value funds. This suggestion seeks to maintain a status quo that Congress intended to change for the benefit of participants and beneficiaries. We believe that, as proposed, the Department’s regulation would significantly advance the interests of plan participants and beneficiaries by allowing employers to select default investments that are more likely to generate adequate retirement assets. For this reason, the Institute strongly supports the range of default investments identified in the proposed regulation.1 The range of options outlined in the Department’s safe harbor proposal will achieve the goals of automatic enrollment. Including stable value funds would be inconsistent with the purpose of measures enacted in the Pension Protection Act of 2006 (PPA) to facilitate automatic enrollment and enhance the utility of 401(k) plans. Research cited in ACLI’s letter in support of using stable value funds as default investments is incomplete or misleading and ignores important policy considerations, as described below. 1 The Institute is not the only supporter of the Department’s approach. For example, the U.S. Chamber of Commerce Commission on the Regulation of U.S. Capital Markets in the 21st Century recently issued a bipartisan report recommending that all pre-retirement default investments be in lifecycle, target retirement, asset allocation, or balanced funds. The Commission’s report is available here: http://www.capitalmarketscommission.com/portal/capmarkets/default.htm. Office of Regulations and Interpretations Revision of Form 5500 (RIN 1210-AB06) Page 2 of 9 Importance of Automatic Enrollment and Default Settings The Department’s proposal reflects the goals of Congress and the Administration in facilitating automatic enrollment in 401(k) plans and is consistent with the statutory language of the PPA. Research shows that automatic enrollment into employer-sponsored retirement plans boosts participation rates and savings.2 With this in mind, Congress carefully crafted measures in the PPA to encourage employers to automatically enroll their employees, providing a safe harbor from nondiscrimination testing and preemption of state laws that might otherwise interfere with automatic payroll deduction. Congress recognized that the success of automatic enrollment depends on selecting appropriate contribution rates and investments on behalf of participants who give no instruction. Research shows that higher contribution rates and default investments with equity exposure make a positive difference in generating adequate retirement savings.3 The default contribution rate and default investment option send a signal to participants generally, not just to those who are defaulted.4 Accordingly, the PPA requires automatic escalation of contribution rates under the nondiscrimination safe harbor and includes fiduciary relief for default contributions invested in accordance with Department of Labor regulations. Sponsors of defined contribution plans long have faced the necessity of investing participant accounts absent specific instruction. Until now, the majority of plan sponsors favored low-yield, low- risk default options – which save money rather than invest it – largely to shield themselves from litigation by a defaulted participant whose account might have lost value. Both Congress and the Department of Labor recognized that the status quo needed to change. In fact, the Department began work on its default investment regulation before the PPA was enacted. The proposal is designed to protect sponsors who want to invest participant money in capital appreciation products that are more appropriate for long-term retirement investing but involve principal risk. In today’s litigious environment, employers may gravitate toward the most conservative option available. Including stable 2 See James J. Choi, David Laibson, Brigitte C. Madrian, and Andrew Metrick (2004), “Saving For Retirement on the Path of Least Resistance,” originally prepared for Tax Policy and the Economy 2001, updated draft: July 19, 2004; James J. Choi, David Laibson, Brigitte Madrian, and Andrew Metrick (2002), "For Better or For Worse: Default Effects and 401(k) Savings Behavior," Pension Research Council Working Paper, PRC WP 2002-2; and Brigitte C. Madrian and Dennis F. Shea (2001), The Power of Suggestion: Inertia in 401(k) Participation and Savings Behavior,” The Quarterly Journal of Economics, Vol. CXVI, Issue 4, November 2001: pp. 1149–1187. 3 See Sarah Holden and Jack VanDerhei (2005) “The Influence of Automatic Enrollment, Catch-Up, and IRA Contributions on 401(k) Accumulations at Retirement,” ICI Perspective, Vol. 11, No. 2 and EBRI Issue Brief, No. 283, Washington, DC: Investment Company Institute and Employee Benefit Research Institute, July 2005. 4 Evidence suggests that instituting automatic enrollment affects the contribution rates of those who otherwise would have participated on their own. See Choi, et al. (2002) and Madrian and Shea (2001). For example, in Madrian and Shea (2001), approximately one-third of participants who likely would have participated absent automatic enrollment, with a contribution rate other than the default rate, ended up at the default rate. Office of Regulations and Interpretations Revision of Form 5500 (RIN 1210-AB06) Page 3 of 9 value funds in the safe harbor likely would result in retention of the status quo and render the PPA’s default investment provision ineffective in enhancing retirement savings for participants. In addition, including stable value funds is inconsistent with the statutory language and purpose. The PPA mandates that default investments include a “mix of asset classes” to receive fiduciary relief. Stable value funds include only one asset class – fixed income investment contracts. Consistent with the safe harbor requirement that a default include a mix of asset classes, the Department’s proposal also leaves out money market funds, a mutual fund product often used as a default investment. Approximately $83 billion of 401(k) assets is invested in money market mutual funds, some portion of which was invested by default. We believe money market funds are equally inappropriate as default investments for retirement savings, and support the Department’s determination to exclude them from the safe harbor. Regulatory Impact Analysis The ACLI letter challenges the Department’s regulatory impact analysis using inaccurate and inconsistent data. Its arguments also are contrary to the policies espoused by Congress and the Department of Labor. In short, ACLI’s letter states that some employees change jobs frequently and cash out their retirement plan balances. Because, the letter argues, these and other defaulted employees may be risk averse, employers should have the protection of the safe harbor to choose a principal- protected default option such as a stable value fund or guaranteed investment product. Job-Changers ACLI’s letter cites data on job turnover and average tenure to argue that public policy should accommodate individuals who do not intend to use their retirement plan contributions for retirement. This argument runs counter to the fundamental goals of automatic enrollment – encouraging the accumulation of retirement savings. Public policy should advance this goal, not hinder it. The purpose of a 401(k) plan is not to encourage workers to withdraw and spend their savings when they change jobs. In fact, Congress has spoken in regard to job changers – 401(k) money is meant to stay invested. For example, Congress imposed a 10 percent excise tax on premature distributions (including distributions upon separation from service prior to age 55) and mandatory 20 percent income tax withholding on distributions eligible for rollover but not directly rolled over. More recently, Congress enacted a requirement to automatically roll over small account balances (greater than $1,000) into IRAs when an employee separates from service. Not only is the letter’s policy premise unsound, it relies on data that does not tell the whole story. Office of Regulations and Interpretations Revision of Form 5500 (RIN 1210-AB06) Page 4 of 9 • By citing job turnover rates, ACLI’s letter understates typical employee job tenure. The Department of Labor measures job turnover by totaling the number of workers who separate from jobs in a given year as a percentage of the labor force. The separation numbers count each time an individual separates from a job in a given year, including temporary and seasonal workers who may enter and exit multiple jobs in a year. Using these turnover rates leaves the impression that “many American workers have short job tenures”5 instead of the correct inference that some workers enter and exit multiple jobs in a given year, but the length of time that the typical worker is at an employer is much longer than implied by the turnover rates. In fact, the median number of years in 2006 that workers age 25 or older have been at their current employer is 4.9 years.6 The median job tenure has remained around 5 years since 1983.7 • Although some defaulted employees leave their employers after short tenures, studies show that of those who stay, most remain invested in the default option.8 The Department’s regulation should not ignore the interests of these workers. • ACLI’s letter ignores important facts in stating that “many plan participants … cash out their account after leaving employment. These short-term investors would benefit significantly from an investment option insulated from equity market volatility.” First, a plurality of investors who withdraw their account balances when they leave a job roll over their balances into a tax- qualified retirement account, and so by definition are not short-term investors. 9 Second, various researchers have found that 40 to 43 percent of 401(k) participants leave their account balances in the plan when they change jobs.10 Taking together employees who leave their balances in the plan and employees who roll them into a new qualified plan or IRA, two-thirds of employees do not behave like “short-term” investors. 5 See page 5 of the ACLI letter. 6 See Craig Copeland, Employee Tenure, 2006, EBRI Notes, April 2007, Vol. 28, No. 4 for summary of trends of employee tenure based on the U.S. Census Bureau’s Current Population Survey. 7 See Copeland (2006). 8 See Choi, et al. (2001), Choi, et al. (2002), and Madrian and Shea (2001). 9 Craig Copeland, “Retirement Plan Participation and Retirees’ Perception of Their Standard of Living,” EBRI Issue Brief No. 289, January 2006, p. 22. 10 See Michael Hurd, Lee Lillard, and Constantinj Panis “An Analysis of the Choice to Cash Out Pension Rights at Job Change or Retirement,” October 1998 DRU-1979-DOL; Fidelity Investments, Building Futures Volume II: Opportunities and Challenges for Workplace Savings in America, A Report on Corporate Defined Contribution Plans. Fidelity Investments (2001); Mike McCarthy and LIz McWhirter, "Are Employees Missing the Big Picture? Study Shows Need for Ongoing Financial Education." Benefits Quarterly, Volume 16, No. 1, First Quarter 2000: pp. 25-31; James Porterba, Steven Venti, and David Wise, “New Estimates of the Future Path of 401(k) Assets, NBER Working Paper 13083, May 2007. Office of Regulations and Interpretations Revision of Form 5500 (RIN 1210-AB06) Page 5 of 9 In essence, the ACLI letter implies that perpetually short-tenured employees should be invested in a capital preservation vehicle for the duration of their careers. We think this result does a disservice to the mobile workforce. 401(k) plans have taken on an increased role in generating retirement income, as the role of defined benefit plans has declined, precisely because they are portable. ACLI’s letter states that certain participants use their retirement accounts for current consumption, not for retirement income. Is this the participant behavior that we want public policy to endorse? Making Default Choices for the Risk-Averse The ACLI letter contends that capital preservation vehicles such as stable value funds should be available so that employers can take into account a participant’s risk tolerance when selecting a default investment. We disagree with this assumption. In the absence of investment direction from the participant, the employer’s objective should be to make a prudent choice on his or her behalf, not to guess what he or she would have picked. The Department’s proposal recognizes that the plan sponsor in most cases will not be able to identify the risk tolerances of its individual participants. In cases where an employer believes its workers are particularly risk averse and determines that a stable value fund would be the most prudent default investment, the employer is not precluded from going outside the safe harbor. The Department stated in the preamble to its proposal that capital preservation products may be prudent default investments for some participants or beneficiaries. Safe harbors are designed to protect participants and provide the greatest good for the greatest number. The Department's QDIA safe harbor requires that qualified default investments include a mix of asset classes to combine asset appreciation and asset preservation. The default can be either a balanced portfolio with a set asset allocation appropriate for participants as a whole, or a balanced portfolio that changes over time based on the age, target retirement date, or life expectancy of the participant. An age varying asset allocation is based on research showing that in many cases it is appropriate to have greater exposure to equity investments when young, but to shift the portfolio toward bond investments over time. 11 As people age and their potential working years shorten, lifecycle and target date funds are designed to increase holdings of fixed income securities to buffer the retirement assets of aging workers. Nothing in the Department's regulation precludes asset managers from using stable value funds as an asset class within a QDIA. Indeed, it is possible for asset managers to use stable value funds as an important component of these products and increase their share of the 11 For a discussion of the appropriateness of age-based asset allocation, see ZviBodie, Robert C. Merton, and William Samuelson (1991), “Labor Supply Flexibility and Portfolio Choice in a Lifecycle Model,” Journal of Economic Dynamics and Control 16, 427-449; John Y. Campbell, Joao Cocco, Francisco Gomes, and Pascal Maenhout (2001), “Investing Retirement Wealth: A Life-Cycle Model,” in John Y. Campbell and Martin Feldstein (eds), Risk Aspects of Investment-Based Social Security Reform (Chicago: University of Chicago Press); Joao Cocco, Francisco Gomes, and Pascal Maenhout (2005), “Consumption and Portfolio Choice over the Life Cycle,” Review of Financial Studies, 18(2), 491-533; Christian Gollier and Richard J. Zeckhauser (2002), “Horizon Length and Portfolio Risk,” Journal of Risk and Uncertainty 24, 195-212. Office of Regulations and Interpretations Revision of Form 5500 (RIN 1210-AB06) Page 6 of 9 portfolio as the target date nears. Furthermore, the Department’s proposal does not prevent asset managers from creating a lifecycle fund with a more conservative mix of assets throughout the life cycle of the fund or a more conservative balanced fund, to be used for workforces that may, as a group, be more risk averse. On the other hand, the proposal would prevent employers from using a single asset class for all of its workers, recognizing that such a default investment in principle would not provide the greatest good to the greatest number of workers. Importance of Equity Exposure ACLI’s letter also criticizes the Department for overstating the impact of asset allocation on retirement savings, stating that “participant contribution rates and plan participation rates have much more of an impact on retirement savings than investment allocations.” Participation in a 401(k) plan certainly is a necessary condition for accumulating 401(k) assets, and automatically enrolling participants in a 401(k) plan will certainly increase the balances that workers accumulate in these plans. Contribution rates also are important determinants of asset accumulation.12 Nevertheless, how the money is invested also plays an important role in the accumulation of assets. The ACLI letter misrepresents a study by Putnam Investments by stating that investment returns play a minor role in asset accumulation.13 The point of the Putnam study was not to compare stable value with lifecyle funds; it was to show that the performance of individual mutual funds was much less important than other factors in accumulating retirement assets. The study found that contribution rates, asset allocation/diversification, rebalancing the portfolio over time, and age-appropriate investments were the most important drivers. Rather than supporting the points made by the ACLI letter, the Putnam study provides evidence supporting the Department’s proposal. It can be difficult to compare alternative investments, as the return earned on any investment differs depending on the time period analyzed. Over time, investments in stocks tend to have higher average returns and more variable returns than investments in stable value funds. Attached are results from stochastic simulations comparing the returns from a lifecycle fund to returns from a stable value fund, which strikingly show what a difference the allocation of a worker’s retirement assets can make. These so called “Monte Carlo” simulations assume that a worker invests in a 401(k) plan over his or her working career and in each year experiences investment returns that are drawn randomly from the distribution of returns experienced over history. 14 Five thousand simulations are run to represent five 12 See Sarah Holden and Jack VanDerhei (2005) 13 DC Plans Missing the Forest for the Trees, Putnam Investments, August 2006. 14 The Monte Carlo simulations used historical data to model investment returns. Specifically, returns are assumed to be distributed normally with the mean and standard deviation of returns set equal to the arithmetic mean and standard deviation from history. Lifecycle funds were assumed to be invested in a combination of large U.S. company stocks and long-term corporate bonds. After inflation and fees, stocks were assumed to have a geometric average return of 5.5 percent per year. Bonds were assumed to have a geometric average return of 2.1 percent per year after inflation and fees. Because we Office of Regulations and Interpretations Revision of Form 5500 (RIN 1210-AB06) Page 7 of 9 thousand possible combinations of annual investment returns a worker might experience over his or her working career. The results show that workers can expect to be much better off in retirement if they invest in a lifecycle fund. For example, a worker who begins investing at age 30 could expect, on average, to have more than twice the retirement assets at retirement by investing in a lifecycle fund compared to investing in a stable value fund (Figure 1). Although the returns over any given time period will vary, the simulations show that this worker investing in a lifecycle fund would experience investment returns greater than the average stable value fund return nearly 90 percent of the time. Indeed, lifecycle funds performed better than stable value funds in the vast majority of cases, even for those investors who began to make contributions later in life, when the lifecycle fund would be more conservatively invested. We believe the approach the Department followed in its analysis was sound. For purposes of our simulations, however, we make minor adjustments that address many of the concerns ACLI’s letter raised. The simulations use historical stable value returns rather than money market fund returns. Because the return to stable value funds over history is very close to the return on intermediate-term government bonds, it is typically higher than the return to short-term Treasury bills. The simulations also account for investment expenses. Total expenses for equity funds are assumed to be 120 basis points, for bond funds total expenses are assumed to be 70 basis points, and for stable value funds the account level expenses are assumed to be 45 basis points.15 After adjusting for inflation and investment expenses, the average historical return on equities is 5.5 percent, and the average return on corporate bonds and stable value funds is 2.1 percent, leaving an implicit premium of equities over stable value funds in the simulations of 3.4 percent. The “equity premium” assumed in the simulations is in line with the premium Jeremy Siegel has recently argued equities should earn.16,17 do not have access to a time series on stable value fund returns, we use a proxy for stable value fund returns by taking a 5-year moving average of 10-year Treasury yields. This series corresponds well to the Hueler Stable Value Index. For example, for the ten-year period ending in 2006, stable value funds, gross of account level expenses, returned 5.5 percent before inflation. Over the same period, the geometric mean return for a 5-year moving average of the 10-year Treasury yield was also 5.5 percent. Stable value funds are assumed to have a mean geometric return of 2.1 percent after inflation and fees. All returns are based on the period from 1958 to 2006 – the period for which we have data on 10-year Treasury bond yields. The glide path used to simulate the returns on lifecycle funds was drawn from Craig L. Israelsen, “Target-Date Funds: Cookie Cutter Methodology,” Financial Planning Magazine, April 2007. 15 The Hueler Stable Value Index reports returns net of contract expenses but gross of account level expenses. 16 See Jeremy Siegel, “Perspectives on the Equity Risk Premium,” Financial Analysts Journal, Volume 61, Number 6, pp. 61- 73. Siegel argues that the spread between the gross real return (that is, excluding expenses) on stocks and Treasury bonds should be in the range of 3.5 percent to 4.0 percent. Excluding investment expenses, our simulations assume an equity premium of 3.9 percent. 17 More detailed information regarding the assumptions, methods, and computer programs used in these simulations are available upon request. Office of Regulations and Interpretations Revision of Form 5500 (RIN 1210-AB06) Page 8 of 9 The ACLI letter also cites the volatility of the equity markets as a reason certain participants should invest in stable value funds. It may indeed benefit long-term investors to invest a portion of their portfolio in a stable value fund or other capital preservation product for purposes of diversification. But investing an entire portfolio in a capital preservation product over the long term would cause participants to miss out on significant equity market gains. Comparing Returns over Short Periods Can be Misleading ACLI’s letter notes that the equity premium relative to stable value funds over the past 10 years is just over 2 percent.18 The average total return on large company stocks from 1997 to 2006 was 8.4 percent, according to Morningstar, and the average total return to intermediate-term government bonds was 5.8 percent, which was about in line with stable value fund returns.19 This results in an equity premium relative to intermediate government bonds of 2.6 percent over this period, and an equity premium relative to stable value funds of 2.9 percent. This decade was a period of lower-than- average returns to stocks due to the prolonged bear market, and higher-than-average returns to bonds because of falling interest rates. By comparison, during the period of 1926 to 2006, large company stocks averaged an annual return of 7.1 percent real (10.4 percent nominal) and intermediate-term government bonds averaged an annual return of 2.1 percent real (5.3 percent nominal). During the period of 1958 to 2006 (which is the time period used in the simulations presented above),20 the spread was tighter, but still exceeded the spread from the last ten years significantly. Over this period, large company stocks averaged an annual return of 6.6 percent real (11.1 percent nominal) and intermediate- term government bonds averaged an annual return of 2.7 percent real (6.8 percent nominal). There is no reason to believe that the most recent decade, in which the equity premium was low by historical standards, will be predictive of future returns. Costs Associated with Change 18 See pages 6-7 of the ACLI letter. 19 Although yields differ in the short run, the yield on stable value funds is very similar to intermediate-term government bond funds over longer periods of time. The difference is primarily due to differences in accounting: intermediate-term government bond funds mark-to-market and stable value funds essentially use book accounting. In a falling interest rate environment (as we have experienced over the past 10 years) a stable value fund maintains its principal value and the rate of return is generally higher than the market rate of return, but declines over time. Similarly, in an increasing rate environment, a stable value fund maintains its principal value and the rate of return is generally lower than the market rate of return but increases over time. Thus, the value remains stable by using book value for the assets and allowing the rate of return to lag the market rate of return, on both the way up and the way down. In contrast, bond funds mark-to-market, resulting in the value of the fund fluctuating over time and the current yield on the fund matching the market yield. 20 See footnote 13 above. The time period was chosen based on the availability of data on Treasury yields. Office of Regulations and Interpretations Revision of Form 5500 (RIN 1210-AB06) Page 9 of 9 ACLI’s letter states that the absence of stable value funds in the QDIA safe harbor will lead to large-scale migrations out of these products over a short period of time, which in turn will negatively affect the returns of those products. We think this does not justify inclusion. Rather than continuing to use a bad policy because of potential transitional problems, responsible public policy would develop methods for addressing transitional problems so that the good policy can take hold. We believe the Department could and should develop transitional rules to address any potential issues relating to migration out of stable value funds. Proposed Regulation Does Not Exclude Guaranteed Products We also note that many guaranteed products may fit within the framework of the Department’s proposed QDIA list. Contrary to the ACLI letter and statements made in the popular press, the categories of default investments set forth in the proposed regulation do not require use of a mutual fund. Any product that invests in a mix of equity and fixed income securities, with either an allocation that changes over time to become more conservative with increasing age or an allocation that is appropriate for participants in the plan as a whole, could be a QDIA. The fact that the product also has a guarantee would not make the product ineligible. * * * The Department of Labor is charged with protecting the interests of employees and their retirement benefits. With this mission in mind, the Department used its expertise to develop a reasoned and substantiated regulatory analysis that demonstrates the cost-effective benefits required of a federal agency regulation. It should not be overturned. The Department’s proposed safe harbor will allow employers to act free of the threat of litigation in making default investment choices that are appropriate for long-term investing and provide greater likelihood of ensuring retirement income security. Sincerely, /s/ Brian Reid /s/ Elena Barone Brian Reid Elena Barone Chief Economist Assistant Counsel – Pension Regulation Figure 1: Results of 5,000 Stochastic Simulations, Comparison of a Lifecycle Fund and a Stable Value Fund1 Age at Which Contributions Begin2 30 40 50 Balance at the end of age 66 (2006$) Prototype Lifecycle Fund Average $398,097 $215,052 $101,259 By Decile Top 915,019 417,945 165,986 Middle 343,421 196,610 96,417 Bottom 148,026 99,129 57,614 Stable Value Fund Average $195,470 $134,789 $78,092 By Decile Top 238,933 159,786 89,219 Middle 194,239 134,390 77,834 Bottom 158,224 112,732 68,054 Percent of Cases where Lifecycle Fund Performs Better Than Average Stable Value Fund 87% 82% 77% 1 Lifecycle fund is modeled as changing from 90 percent equity and 10 percent corporate bonds at age 40 and below, moving to 40 percent equity and 60 percent corporate bonds at age 70. Returns based on historical returns from 1958 to 2006. This time period was chosen because of the availability of data on 10-year Treasury bond yields. Real equity returns less assumed investment expenses equal to 120 basis points have a geometric average of 5.5 percent. Real corporate bond returns less investment expenses of 70 basis points have a geometric average of 2.1 percent. Returns to stable value funds proxied by 5-year moving average of 10-year Treasury bond yields. Real stable value fund returns less account expenses equal to 45 basis points have a geometric average of 2.1 percent. Inflation used is historical prior to 2006 and equal to the Social Security Intermediate Projection thereafter (generally 2.8 percent per year). 2 Assumes contributions of 6 percent of earnings, with earnings in 2006 at age 40 of $55,000, and average approximately $58,300, in real 2006 dollars, between age 30 and age 66. The level and path of earnings is not a significant factor in the relative performance of the investments. Sources: ICI simulations using data from Morningstar, Bureau of Labor Statistics, and Federal Reserve Board