ICI Letter to California Officials Urging Delay on ‘Secure Choice’ Retirement Plan (pdf)

March 24, 2016 Delivered Electronically The Honorable John Chiang California State Treasurer Chairman, California Secure Choice Retirement Savings Investment Board 915 Capitol Mall, Room 110 Sacramento, CA 95814 Re: Overture Financial Final Report to the California Secure Choice Retirement Savings Investment Board Dear Treasurer Chiang: The Investment Company Institute1 appreciates the opportunity to provide comments on the prepared by Overture Financial LLC to the California Secure Choice Retirement Savings Investment Board .2 The Report reflects the market analysis, financial feasibility study, and program design recommendations of Overture Financial and its subcontractors with respect to the . The Program is contemplated as a state-run retirement savings plan for private-sector workers in California, pursuant to the California Secure Choice Retirement Savings Trust Act (SB 1234), enacted in 2012. The Institute strongly supports efforts to promote retirement security for American workers. We understand and appreciate the interest shown by the state of California in ensuring that its residents have sufficient resources for retirement and share the goal of increasing workplace retirement plan access. Our member companies devote considerable effort to helping Americans prepare for and achieve a financially secure retirement. Due in part to the innovation that has taken place over the last few decades in the private sector, Americans currently have $24.0 trillion saved for retirement, with more than half of that amount in defined contribution ( DC ) plans and individual retirement 1 The Investment Company Institute (ICI) is a leading global association of regulated funds, including mutual funds, exchange-traded funds (ETFs), closed-end funds, and unit investment trusts (UITs) in the United States, and similar funds offered to investors in jurisdictions worldwide. ICI seeks to encourage adherence to high ethical standards, promote public members manage total assets of $16.9 trillion and serve more than 90 million U.S. shareholders. 2 The Report is dated March 17, 2016 and is available at www.treasurer.ca.gov/scib/report.pdf. The Honorable John Chiang, State Treasurer March 24, 2016 Page 2 of 31 accounts ( IRAs ).3 About half of DC plan and IRA assets are invested in mutual funds, which makes the mutual fund industry especially attuned to the needs of retirement savers. The Institute has 36 member companies located in California with about 16,000 employees in the state and $3.5 trillion in assets under management. These California-based companies, as well as mutual fund companies based outside of California, provide investments and other services to retirement plans and individual retirement savers in California. Our members are eager to serve this marketplace with increasingly competitive product and service offerings. We appreciate that the Board faces a significant challenge in making an informed, sound assessment of the Program. Unfortunately, the Report does not provide adequate support to meet that challenge. While the Report contains a great deal of survey data on the characteristics and views of workers without employer-provided retirement plan coverage, it fails to provide an adequate analysis of the financial feasibility of the Program. We are concerned that Program participants or California taxpayers or most likely both will find themselves bearing unanticipated costs as a result of the Program. We urge the Board to conduct further analysis before moving forward with the Program. Many of the assumptions and conclusions in the Report, which are , appear unrealistic or incomplete. This is the case even under the Report. Without additional information about the financial feasibility of the Program, we question how the Board can truly assess the Program, much less recommend it for further action by the California State Legislature. The Program also raises important legal questions, as described in our November 15, 2013 letter to Mr. Grant Boyken responding to the California Secure Choice Retirement Savings Program Request for Information.4 These legal questions including the application of the Employee Retirement Income Security Act of 1974 ERISA and federal securities laws have yet to be sufficiently answered. In particular, SB 1234 requires that prior to implementation, the Board must find that the Program accounts will qualify for the favorable federal income tax treatment accorded to IRAs under the Internal Revenue Code, and that the Program is not an employee benefit plan under ERISA. As you know, the ERISA status of state-based programs is the subject of a pending rulemaking project at the U.S. Department of Labor .5 As the Report acknowledges, it is unclear whether 3 See Fourth Mar. 2016); available at www.ici.org/info/ret_15_q4_data.xls. 4 The letter is available at www.treasurer.ca.gov/scib/rfi/ici.pdf. 5 DOL has proposed a regulatory safe harbor from coverage under ERISA for certain payroll-deduction IRA arrangements established and maintained by state governments. Savings Arrangements Established by States for Non-Governmental Employees, 80 Fed. Reg. 72006 (November 18, 2015). Under the safe harbor, these state arrangements would not be treated as employee benefit plans under ERISA, as long as specified conditions are met, including that state law requires certain employers to make the program available to employees. We do not believe that this proposal settles the ERISA status of any The Honorable John Chiang, State Treasurer March 24, 2016 Page 3 of 31 certain investment structures would be permissible in a state-run retirement savings program and we understand that the Board may be seeking guidance from the Securities and Exchange Commission on those matters. These matters must be resolved before the Board considers recommending the Program. More broadly, we are concerned that initiatives like that under consideration in California will ultimately lead to the creation of a fragmented, state-by-state system of retirement savings for private- sector workers.6 A patchwork of state-run programs, each with its own unique rules, has the potential to harm the voluntary system for retirement savings that is helping millions of American private-sector workers achieve retirement security. In our view, the research and data suggest that these state initiatives are misplaced and that there are other more targeted changes at the national level that will be more effective at increasing access to payroll-deduction savings opportunities.7 We discuss our views of the Report below. First, we explain our concerns about the adequacy of the financial feasibility study in the Report, noting that the Report fails to adequately consider probable that call into question the financial feasibility of the Program. The most significant of these are -out rates, contribution rates, and withdrawal and turnover activity . We also failure to adequately account for all likely costs in implementing and operating the Program. Second, we describe our broader concern about a fundamental assumption on which the perceived need for mandatory state-run retirement plans is based that workers currently not covered particular state-run program, as it has not been finalized and, even if finalized in its current form, application of the safe harbor to any particular program would be a facts and circumstances determination that ultimately would rest in the hands of a court of law. ultimately litigated, the courts would conclude that state payroll deduction savings arrangements are preempted by ERISA 80 Fed. Reg. 72009. ements of the safe harbor added). 6 See letter from Investment Company Institute to U.S. Department of Labor, dated January 19, 2016; available at www.dol.gov/ebsa/pdf/1210-AB71-00062.pdf. Section 514 of the Employee Retirement Income Security Act of 1974 tatute. The intent of ERISA preemption is to avoid subjecting employers to a patchwork of different and likely conflicting requirements under potentially 50 state laws. With state laws such as SB 1234, employers that operate in multiple states or employ workers residing in more than one state will face significant burdens complying with differing requirements regarding covered employees, the type of retirement plan that will exempt an llment features, among others. 7 For a description of such targeted changes, see letter from Investment Company Institute to U.S. Department of Labor, dated January 19, 2016; available at www.dol.gov/ebsa/pdf/1210-AB71-00062.pdf. The Honorable John Chiang, State Treasurer March 24, 2016 Page 4 of 31 by employer-sponsored retirement plans will be best served by being automatically enrolled in such programs. I. The Report Fails to Adequately Consider Probable Scenarios That Call into Question the Financial Feasibility of the Program The Report - sustaining even under adverse conditions with poor investment returns and high opt-out rates. 8 The Report rests this conclusion on a long list of assumptions, many of which are not likely to hold true. As a result, we fear that the Program is not nearly as financially secure and self-sustaining as it is portrayed to be. We recognize that attempting to model costs of any proposed initiative is a difficult and uncertain exercise, but we strongly caution that the Report does not appear to have fully considered many scenarios that are entirely possible, if not probable. Any combination of these alternative scenarios could result in much lower assets and much higher costs than suggested by the baseline scenario or the one-off changes studied. In addition, there appear to be costs that are not considered in the analysis even under the -year payoff period and $186 million funding gap.9 In this respect, there are several risks that could significantly affect the asset growth and cost estimates. The most significant of these risks opt-out rates, contribution rates, and withdrawal activity and turnover, and account for all likely costs. We discuss each of these risks in more detail below. The Report also appears to rely on averages, which as a methodology does not recognize the range of workers and range of account balances and participation rates that may occur. This variability can have a material impact on the assets and costs of the Program. A. The opt-out rate is a key risk to the ability to build assets and manage costs It is important for the Board to be aware that it is quite likely that automatic enrollment in the Program may not have results close to those produced by automatic enrollment in voluntary private- sector retirement plans. In this respect, the opt-out rate for the Program may be higher on average than is projected in the Report, or could be distributed across employers in a way that is more costly than projected. Plan design and workforce demographics affect opt-out rates and both are quite different between the private-sector retirement plans voluntarily implementing automatic enrollment and the proposed state-mandated Program to be applied to employers without plans. 8 See page 110 in the Report. 9See page 119 in the Report. The Honorable John Chiang, State Treasurer March 24, 2016 Page 5 of 31 1. Plan design affects opt-out rates The Report baseline opt-out rate assumption for the Program likely is overly optimistic. Automatic enrollment may not be anywhere nearly as successful in increasing participation in as it is among employers who voluntarily adopt it for the plans that they offer their employees. First, automatic enrollment has been adopted more widely in the private sector by larger employers. 10 Such employers often combine automatic enrollment with other participation incentives such as employer contributions (which provide an immediate and positive incentive to save) and the availability of participant loans (which provides flexible access to the savings).11 The Board would be ill-advised to assume that participation and opt-out experience in the Program will be close to the private-sector experience, because it is difficult to disentangle the impact of one plan feature in isolation and some of the results achieved with automatic enrollment may also reflect the influence of other plan features. For example, BrightScope and ICI analyzed a sample of nearly 54,000 401(k) plans with 100 participants or more and at least $1 million in plan assets and found that 401(k) plans tend to have combinations of plan features.12 The most common combination of plan features offered to workers includes employer contributions, which provides immediate growth in the 401(k) balance, and participant loans, which provides flexibility that in turn promotes larger contributions.13 While the Program presumably would offer access to the accounts through withdrawals, which provides some flexibility, it would not provide employer or state contributions.14 10 See Utkus and Young, How America Saves, 2015: A report on Vanguard 2014 defined contribution plan data, Valley Forge, PA: The Vanguard Group (2015); available at https://institutional.vanguard.com/iam/pdf/HAS15.pdf. 11 In the case of state-sponsored retirement plans that are IRAs, individuals could access the accounts through withdrawals. However, amounts withdrawn may be subject to penalties and/or income tax. 12 Private-sector 401(k) plans with automatic enrollment are more likely to have both employer contributions and participant loans outstanding than plans without automatic enrollment. In 2013, 74 percent of 401(k) plans with automatic enrollment had employer contributions and outstanding participant loans. Nearly nine in 10 401(k) plans with automatic enrollment had employer contributions. See BrightScope and Investment Company Institute, The BrightScope/ICI Defined Contribution Plan Profile: A Close Look at 401(k) Plans, 2013, San Diego, CA: BrightScope and Washington, DC: Investment Company Institute (December 2015); available at www.ici.org/pdf/ppr_15_dcplan_profile_401k.pdf. 13 Id. See also NBER Retirement Research Center Paper, no. NB 09-05, Cambridge, MA: National Bureau of Economic Research (September 2010); available at www.nber.org/aging/rrc/papers/orrc09-05.pdf. 14 Id. In addition, Beshears et al. (2007) studied savings plan participation at nine firms with automatic enrollment and variation in their match structures. Although they caution that the potential existence of firm-level omitted variables means al matching structure a match of 50 [percent] up to 6 [percent] of pay contributed to no match would reduce participation under See Beshears, Choi, Laibson, and Madria NBER The Honorable John Chiang, State Treasurer March 24, 2016 Page 6 of 31 Second, sponsors and administrators of private-sector plans provide extensive participant education on the importance of saving and investing, through materials and website tools, which plays a key supporting role in increasing participation (Figure 1). The depth and breadth of these educational efforts help inform employees of the benefits of the 401(k) plan and the importance of saving for retirement. All of these features contribute to the success of automatic enrollment in the voluntary private retirement system, but it is far from clear that they will be present in the context of state- mandated payroll-deduction IRAs. Academic research has examined the effectiveness of automatic enrollment in the context of private-sector 401(k) plans that added the feature to already existing plans.15 These plans typically have extensive educational programs in place, including materials to promote the importance of saving for retirement, explanations of investment types and the trade-off between risk and return, and the features of their plans.16 Household survey results highlight that about nine in 10 households with DC plan accounts agreed that their employer-sponsored retirement plan helped them to think about the long- term, not just their current needs.17 Third, private-sector 401(k) plans offer an array of investment options, typically covering a range of investment risks and returns (Figure 1). Household surveys find that DC-owning households generally appreciate the investment choice and control and agree that their DC plan offers a good line- up of investment options.18 All of these factors suggest that the Program will find it difficult to replicate the success of the private sector. A final factor that may depress participation rates for the Program compared with private- sector 401(k) participation rates is the complexity around IRA contribution rules. It is not clear how the educational materials and enrollment process will help workers make sure that their contributions to the Program are within the legal requirements surrounding IRAs, which may result in some workers Retirement Research Center Paper, no. NB 07-09, Cambridge, MA: National Bureau of Economic Research (August 2007); available at www.nber.org/aging/rrc/papers/orrc07-09.pdf. 15 For example, see The Quarterly Journal of Economics 116, no. 4 (2001): 1149-1187; available at http://qje.oxfordjournals.org/content/116/4/1149.abstract. 16 See Plan Sponsor Council of America, 58th Annual Survey of Profit Sharing and 401(k) Plans: Reflecting 2014 Plan Experience, Chicago: Plan Sponsor Council of America (2015), which reports on the educational materials and activities used in 401(k) plans and the goals of the educational programs. See also, Figure 1 in this letter. 17 See Figure 2 in H ICI Research Report, Washington, DC: Investment Company Institute (February 2016); available at www.ici.org/pdf/ppr_16_dc_plan_saving.pdf. Results are based on a survey of more than 3,000 U.S. adults from mid- November 2015 to mid-December 2015, of which more than half owned DC plan accounts. 18 Id. The Honorable John Chiang, State Treasurer March 24, 2016 Page 7 of 31 opting out due to confusion,19 and others having difficulties when they ultimately file their taxes. The instructions for determining IRA contribution eligibility may themselves have the effect of putting off workers.20 19 The historical data indicate that when traditional IRA contributions were universally allowed from 1982 to 1986, many low-income workers joined the ranks of traditional IRA contributors. When income limits and restrictions based on employer-sponsored retirement plan coverage were placed on traditional IRA contribution eligibility, the data indicate that many lower-income taxpayers stopped contributing, even if they were eligible to make tax-deferred contributions suggesting confusion around the rules governing contributions. See discussion in Holden, Ireland, Leonard-Chambers. and Bogdan Investment Company Institute Perspective 11, no. 1 (February 2005); available at www.ici.org/pdf/per11-01.pdf. 20 For the instructions on determining IRA contribution eligibility, see U.S. Department of Treasury, Internal Revenue Publication 590-A; available at www.irs.gov/pub/irs-pdf/p590a.pdf. The Honorable John Chiang, State Treasurer March 24, 2016 Page 8 of 31 Figure 1 Private-Sector Retirement Plans Provide a Great Deal of Support Private-sector 401(k) plans Proposed state program Employer contributions 76% of 401(k) plans covering 88% of 401(k) participants have employer contributions1 None Participant education 68% of 401(k) plans email plan communications2 Plan materials for distribution by employer 57% websites for the plan 45% individually targeted communications 53% seminars or workshops 40% newsletters 36% retirement gap calculators 29% retirement income projections 17% mobile apps Employee contributions 58% of 401(k) plans allow Roth 401(k) contributions2 Roth IRA; unless not eligible All 401(k) plans allow pre-tax contributions Deductible traditional IRA; unless not eligible After-tax traditional IRA Account access 87% of 401(k) participants are in plans that offer loans3 Hardship withdrawals Hardship withdrawals In-service withdrawals (age 59½ or older) Penalty-free withdrawals (age 59½ or older) Investments Average 27 investment options4 Unclear number of options Default, typically a target date fund5 Default, under review 1 BrightScope and Investment Company Institute, The BrightScope/ICI Defined Contribution Plan Profile: A Close Look at 401(k) Plans, 2013 (December 2015). 2 Plan Sponsor Council of America, 58th Annual Survey of Profit Sharing and 401(k) Plans: Reflecting 2014 Plan Experience, Chicago, IL: Plan Sponsor Council of America (2015). Plans provide multiple types of educational materials. Figures reported are percentage of 401(k) plans using the educational material indicated. 3 Holden et al., "401(k) Plan Asset Allocation, Account Balances, and Loan Activity, 2013," ICI Research Perspective and EBRI Issue Brief (December 2014). 4 BrightScope and Investment Company Institute, The BrightScope/ICI Defined Contribution Plan Profile: A Close Look at 401(k) Plans, 2013 (December 2015). Sources: See notes above. The Honorable John Chiang, State Treasurer March 24, 2016 Page 9 of 31 2. Workforce demographics also affect opt-out rates Workforces of employers without retirement plans differ from those with retirement plans.21 As the Program is intended to serve workforces at employers that have not voluntarily adopted retirement plans, workforce demographics may adversely affect participation and opt-out rates for the Program relative to those for voluntary private-sector plans. As is the case nationwide, workforces at California employers without retirement plans tend to be younger, lower-income, and not working full-time, or full-year (Figure 2).22 In this respect, younger workers tend to be focused on other savings goals and paying down debt, and lower-income workers tend to be focused on saving for emergencies and meeting current needs. Lower income workers are also more likely to receive high replacement rates from Social Security.23 In California, 35 percent of workers at employers without retirement plans are younger than 30, compared with 22 percent of workers at employers with plans. Fifty-eight percent of workers at employers without retirement plans earn less than $27,000, compared with only 24 percent of workers at employers with plans. Only 57 percent of workers at employers without retirement plans work full-time, full-year, compared with 77 percent of workers at employers with plans. 21 Differences in workforce composition appear to be a primary cause for the lower rate at which small employers sponsor retirement plans. See nationwide analysis in ICI Research Perspective 20, no. 6 (October 2014); available at www.ici.org/pdf/per20-06.pdf. 22 ICI used the description of the CPS sample analyzed in the Report to do additional analysis of the pool of eligible California workers (see page 26 in the Report). Although, we caution that the latest CPS resulted in a change to the survey that understates retirement plan coverage (see note 42). Additionally, there is the difficulty that the weights in the CPS are for national calculations not regional or state-specific. We were able to closely replicate the wage and salary distribution of the eligible California workers, but found different results for employment status. Figure C-6 on page 31 of the Report indicates that 83 percent of eligible California workers are full-time, while our analysis of the CPS data finds that 73 percent are full-time, with the key difference in the percentage that are full-time, full-year, which is 66 percent in the Report, and 57 percent in our analysis of the data (Figure 2 in this letter). It appears that the feasibility study assumes 75 percent of eligible workers are full-time and 25 percent are part-time (see page 112 of the Report), even though we believe that it makes more sense to group full-time and not full-year workers with the other workers also less connected to the workforce. 23 The Congressional Budget Office reports estimated replacement rates from scheduled Social Security payments, and Social Security replaces a higher percentage of pre-retirement earnings for workers in lower-income households than it does for workers in higher-income households. See Congressional Budget Office, -Term Projections for Social Security: Additional Information (December 2015); available at www.cbo.gov/sites/default/files/114th-congress-2015- 2016/reports/51047-SSUpdate.pdf. The Honorable John Chiang, State Treasurer March 24, 2016 Page 10 of 31 Figure 2 Workforces at Employers with Retirement Plans Differ from Those Without Plans Percentage of private-sector wage and salary workers age 18 to 64 in California, 2012 2014 Workforces at employers without retirement plans tend to be younger Workforces at employers without retirement plans tend to be lower-income Workers at employers without retirement plans tend to be less connected to the employer Note: Sample is California private-sector wage and salary workers, age 18 to 64. The data are from CPS conducted in the years 2013 2015, reflecting retirement plan coverage in the prior year (2012 2014). Source: ICI tabulation of Current Population Survey data 2 7 20 28 37 34 41 31 Plan No plan 45 to 64 30 to 44 21 to 29 18 to 20 Age (years) 24 5816 16 38 20 22 7 Plan No plan $90,000 or more $40,000 to $89,999 $27,000 to $39,999 Under $27,000 Annual earnings 77 57 8 14 11 16 4 13 Plan No plan Part-year, part-time Part-year, full-time Full-year, part-time Full-time, full-year Work status The Honorable John Chiang, State Treasurer March 24, 2016 Page 11 of 31 Opt-out rates may impose significant costs on workers, as well. Participants who opt out after accounts have been created may face tax penalties or incur additional consumer debt, perhaps suffering avoidable financial stress. The DOL has expressed concern that certain workers who fail to opt out of state programs may be very economically vulnerable.24 As explained above, workers at employers without retirement plans often are lower-income (58 percent of California workers without retirement plans at their current jobs have annual earnings of less than $27,000; Figure 2). While analysis of this potential outcome is beyond the scope of the financial feasibility study for the Program, the Board should be mindful of these risks. 3. Opt-out rates will affect the financial feasibility of the Program As discussed above, research regarding the impact of automatic enrollment cannot predict the opt-out rates, which may prove significantly higher than those predicted by the experience of private-sector automatic enrollment in 401(k) plans. The Report shows that higher opt-out rates will impact the funding and breakeven period for the Program. Its analysis indicates that the breakeven period will increase from 6 to 10 years and the financing needs will increase by 30 percent (from $89 million to $116 million) if the opt-out rate rises from 10 percent to 70 percent (Figure 3). Given sunk costs, fixed-costs, and per-employer costs, it is perplexing that the Report finds that the opt-out rate must rise to 70 percent before that rate has dramatically different effects on Program financing. Although the Report identifies this risk, it does not consider high opt- 24 For research supporting this concern, see -Income Social Security Administration Retirement Research Consortium Paper (October 11, 2012); available at http://gsm.ucdavis.edu/sites/main/files/file-attachments/lowincomedefaults.pdf. The Honorable John Chiang, State Treasurer March 24, 2016 Page 12 of 31 Figure 3 Impact of Opt-Out Rates on Financing Needs to Consider More Scenarios Opt-out rate changes financing and payoff year little except in the extreme scenario Note: See page 118 of the Report. Source: Overture Financial Final Report The Board should analyze more opt-out scenarios and ensure that the underlying calculations fully take into account the impact of sunk costs, fixed costs, and per-employer costs. Figure 4 presents simple numerical examples that highlight the impact of higher opt-out rates on the variable recordkeeping costs. In the first panel, the variable recordkeeping costs from the direct servicing model are calculated for the scenario where there are 100,000 employers offering the Program to their workers. On average, 10 employees per employer participate in the Program, resulting in 1 million participants with IRA balances. The average variable cost of recordkeeping is $35 per participant in this scenario (Example 1). But, what if the opt-out rate is higher and fewer employees decide to open an IRA under the Program? If on average, there were only one participant per employer, the average $89 $100 $116 6 7 10 10% 50% 70% Opt-out rate Required financing ($millions) Payoff year The Honorable John Chiang, State Treasurer March 24, 2016 Page 13 of 31 variable cost of recordkeeping climbs to $170 per participant (Example 2).25 This is not to suggest that only one worker per employer will participate, but to highlight that the number of participants as well as the number of employers are important to determining the costs of the Program. These calculations do not even take into account sunk or fixed recordkeeping costs or any of the other fixed costs of running the program; the lower the number of accounts, the greater these expenses will be on a per- account basis. Understanding the likely opt-out scenarios is critical to being able to determine whether the Program is economically viable. Figure 5 provides additional numerical examples that highlight how variable recordkeeping costs could be affected depending on the pattern of the opt-out rates across firms even when the same 25 Examples 3 and 4 in the lower panel of Figure 4 repeat the variable recordkeeping cost exercise using the EDD (State of California Employment Development Department) servicing model. Figure 4 Higher Opt-Out Rates Increase the Variable Costs of Recordkeeping Direct servicing model Number Per unit cost Total cost Cost per participant account Example #1 Employers 100,000 $150 $15,000,000 Participants 1,000,000 $20 $20,000,000 $35 Example #2 Employers 100,000 $150 $15,000,000 Participants 100,000 $20 $2,000,000 $170 EDD servicing del Number Per unit cost Total cost Cost per participant account Example #3 Employers 100,000 $120 $12,000,000 Participants 1,000,000 $17 $17,000,000 $29 Example #4 Employers 100,000 $120 $12,000,000 Participants 100,000 $17 $1,700,000 $137 Note: Cost data are for existing employers and participants. Costs would be higher for new employers in the direct servicing model. See page 117 in the Report. Sources: ICI tabulation and Overture Financial Final Report The Honorable John Chiang, State Treasurer March 24, 2016 Page 14 of 31 number of workers open IRAs in the Program (and therefore the aggregate opt-out rate is constant). In the top panel, the variable recordkeeping costs from the direct servicing model are calculated for the scenario where 500,000 workers participate in the Program and open IRAs. These 500,000 workers are employed by 100,000 different employers. The average variable cost of recordkeeping is $50 per participant in this scenario (Example 1). But, what if the 500,000 workers are employed by 300,000 different employers? In this scenario, the average variable cost of recordkeeping rises to $110 per participant (Example 2).26 This exercise highlights that the number of employers as well as the number of participants are important to determining the costs of the Program. 26 Examples 3 and 4 in the lower panel of Figure 5 repeat the variable recordkeeping cost exercise using the EDD servicing model. Figure 5 Different Opt-Out Rates Across Employers Impact the Variable Costs of Recordkeeping Direct servicing model Number Per unit cost Total cost Cost per participant account Example #1 Employers 100,000 $150 $15,000,000 Participants 500,000 $20 $10,000,000 $50 Example #2 Employers 300,000 $150 $45,000,000 Participants 500,000 $20 $10,000,000 $110 EDD servicing od l Number Per unit cost Total cost Cost per participant account Example #3 Employers 100,000 $120 $12,000,000 Participants 500,000 $17 $8,500,000 $41 Example #4 Employers 300,000 $120 $36,000,000 Participants 500,000 $17 $8,500,000 $89 Note: Cost data are for existing employers and participants. Costs would be higher for new employers in the direct servicing model. See page 117 in the Report. Sources: ICI tabulation and Overture Financial Final Report The Honorable John Chiang, State Treasurer March 24, 2016 Page 15 of 31 B. If contribution rates are lower than projected, the economic viability to the Program is at risk The Report finds that contribution rates have a significant impact on Program expenses, the required financing for the Program, and the payoff year. The baseline scenario assumes a 5 percent contribution rate on an average full-time annual salary of $45,000 or an average annual part-time salary of $20,000.27 Changing the contribution rate to 3 percent raises required financing by $81 million, or 91 percent, extends the payoff year by 3 years, or 50 percent, and increases Program expenses in the first year by 1.61 percentage points, or 51 percent (Figure 6). Although the Report acknowledges that the contribution rate will have a significant impact on the ultimate financing required, it uses a rate that is higher than the survey results suggest may occur. The Board should consider additional scenarios that include a 2 percent contribution rate, as well as a variety of combinations of contribution rates and opt-out rates. Figure 6 Financial Feasibility of Program Greatly Impacted by the Contribution Rate Required financing ($millions) Payoff year Program expenses as a percent of assets Year 1 Year 5 Year 10 (5% contribution rate; 25% opt- out) $89 6 3.17 0.76 0.45 3% contribution rate $170 9 4.78 1.07 0.59 Memo: Difference $81 3 1.61 0.31 0.14 % difference 91% 50% 51% 41% 31% Note: See page 118 in the Report. Sources: Investment Company Institute calculations and Overture Financial Final Report In contrast to the assumed 5 percent contribution rate, the survey results presented in the Report indicate that the majority of eligible workers say they would likely contribute very small amounts into IRAs in the Program. The median maximum expected monthly amount respondents indicated they would likely contribute falls in the $50 to $99 category (Figure 7). Indeed, most eligible workers surveyed said they could contribute to such a program, but 64 percent indicated that the 27 See page 112 in the Report. The Honorable John Chiang, State Treasurer March 24, 2016 Page 16 of 31 maximum they could contribute would be less than $100 a month, including one-third who said the most they could contribute was between $25 and $49 a month. Such contribution amounts would generate much smaller accounts than those estimated in the baseline scenario in the feasibility study in the Report. Based on the maximum likely contributions, the average annual contribution would range from about $1,000 to about $1,700, which is well below the average annual contribution assumed in the feasibility study. The Report projected 1.6 million participants with $3.2 billion in assets after one year, which is an average account of about $2,000.28 Corroborating what the survey found with regard to contribution amounts, the wage composition of the uncovered California workers also suggests that many accounts will have only modest contributions. One-fifth of eligible California workers have annual salaries less than $10,000 (Figure 8). Even if the Board were to focus on eligible California workers who are employed full-time and full-year, one-fifth of those workers have annual salaries of $20,151 or less. Given that 57 percent of eligible workers are full-time, full-year (Figure 2), 43 percent are less connected to the workforce and more likely to be lower-income and experiencing financial stresses. 28 See page 114 in the Report. Figure 7 Eligible Workers Report Possible Contribution Amounts That Are Modest Percentage of respondents, 2015 Note: Sample is 1,000 respondents; workers not offered retirement plans at work. Source: Greenwald & Associates Online Survey 4% 11% 33% 20% 17% 6% 3% 6% Could not save anything Less than $25 $25 to $49 $50 to $99 $100 to $199 $200 to $299 $300 to $399 $400 or more Maximum expected monthly contribution amount The Honorable John Chiang, State Treasurer March 24, 2016 Page 17 of 31 C. Withdrawal activity and turnover may be higher than assumed in the Report Because the Program represents an entirely new set of enrollment and participation experiences, it is not possible to extrapolate from private-sector experience with withdrawals and turnover. Additional scenarios should be stress-tested to determine the impact of withdrawals and turnover on the asset growth and costs of the Program, and hence on its financial feasibility. There are several variables that would affect withdrawal activity and turnover: (1) access to account balances for self- certified hardship withdrawals; (2) behavior at job change of participants in a mandatory automatic enrollment program; (3) IRA rules which permit individuals to change financial services providers at any time; and (4) realization by participants that a private-sector IRA may offer a more attractive investment opportunity. Research on withdrawal activity and rollover and cash-out behaviors at job change or retirement has largely been based on the behavior of participants in retirement plans with voluntary enrollment. Even for plans with automatic enrollment, the employer had the choice to set up automatic enrollment, not the requirement to do so. Moreover, research shows that the availability of plan loans in 401(k) or other DC plans helps contribute to low withdrawal rates while working, because typically a plan participant must take a loan before seeking a withdrawal. Thus, withdrawal rates in employer- Figure 8 Wage Distribution of Eligible California Employees Percentiles and average of annual earnings, 2012 2014 Note: Data for all eligible employees are from Figure C-4 on page 30 of the Report. Data for full-time, full-year, versus part-year or part-time are from Investment Company Institute tabulations of Current Population Survey data. ICI tabulations indicate that 57 percent of eligible California workers are full-time, full-year, and 43 percent are part-time or part-year (see Figure 2 above). Sources: Overture Financial Final Report and Investment Company Institute tabulations of the Current Population Survey $4,642 $16,000 $2,000 $9,624 $20,151 $4,000 $23,009 $33,749 $10,227 $48,487 $65,000 $22,166 $74,278 $92,043 $32,726$35,146 $48,305 $17,161 All eligible employees Full-year, full-time Part-year or part-time 10th 20th Median 80th 90th Average The Honorable John Chiang, State Treasurer March 24, 2016 Page 18 of 31 sponsored DC plans will not provide insight into the withdrawal activity that might occur in the Program. One might then consider withdrawal activity among IRA investors, but again, the withdrawal activity among IRA investors who have voluntarily created their IRAs may not provide a good measure of withdrawal activity in the Program. The Report assumes that worker turnover is the sole factor in determining withdrawals. Linking withdrawal activity solely to worker turnover, however, ignores the fact that participants in the program will also have the ability to change service providers and therefore likely understates the extent of potential participant withdrawals from the Program. The Report indicates that participants will shoulder the start-up and fixed costs of the plan in addition to a 0.18 percent investment management fee. Total participant fees will be capped at 1.00 percent per year, with excess revenue from this fee used to pay back the costs of the Program in the initial years.29 Effectively, the Program places the burden of start-up costs on participants in the Program, with the promise of eventual lower fees when those start- up costs have been recovered. It is far from certain these participants will stick with the Program to eventually experience the lower fees. If the Program creates true IRAs, participants may change service providers, or transfer their IRA balances from one service provider to another, at any time.30 Workers not covered by retirement plans at their current employers have access to the vibrant IRA market, which is served by a range of financial services firms offering a wide variety of investment options.31 Indeed, 61 percent of traditional IRA owning household with rollovers indicated that one of the reasons they rolled over the assets from their employer-sponsored retirement plans was to get access to more investment options (with 21 percent saying that was the primary reason they rolled over) and 48 percent rolled over to use a different financial services firm.32 Program participants also may realize that they could find a far more attractive deal in the private sector, with additional investment choice, flexibility, and lower-cost options. Private-sector IRA investors are able to obtain lower-cost fund investing and, indeed have concentrated their assets in 29 See pages 115 and 117 in the Report. 30 is not a rollover. This includes the situation where the current trustee issues a check to the new trustee but gives it to you to deposit. Because there is no distribution to you, the transfer is tax free. Because it is not a rollover, it is not affected by the 1- year wai see U.S. Department of Treasury, Internal Publication 590-A; available at www.irs.gov/pub/irs-pdf/p590a.pdf. 31 See Holden and Schrass, ICI Research Perspective 22, no. 1 (February 2016); available at www.ici.org/pdf/per22-01.pdf. 32 Id. The Honorable John Chiang, State Treasurer March 24, 2016 Page 19 of 31 lower-cost mutual funds. For example, at year-end 2014, more than 90 percent of IRA equity mutual fund assets were in equity mutual funds with operating expenses of less than 1.0 percent, including 40 percent with operating expenses less than 0.50 percent (Figure 9).33 All of these factors could be expected to lead to higher turnover and withdrawal activity than the Board may anticipate. D. Program costs may be higher than projected, which will burden Program participants, California taxpayers, or both The Report does not appear to have contemplated all the costs of the Program. The Report 33 If one analyzes the distribution of IRA mutual fund assets by total fund expense ratio, it also is clear that IRA investors concentrate their assets in lower-cost mutual funds. See Statement of the Investment Company Institute, Brian Reid, Chief Economist Restricting Access to Financial Advice: Evaluating the Costs and Consequences for Working Subcommittee on Health, Employment, Labor, and Pensions Committee on Education and the Workforce, United States House of Representatives June 17, 2015); available at www.ici.org/pdf/15_house_advice.pdf. Figure 9 IRA Investors Concentrate Their Assets in Lower-Cost Mutual Funds Percentage of equity mutual fund assets held in IRAs, 2014 Note: This figure reports the distribution of equity mutual fund assets held in IRAs by mutual fund operating expenses so as to focus on investment management expenses (rather than total mutual fund expense ratios which would include the 12b-1 fees investors pay through the fund for some or all of the services they receive from financial professionals and other financial intermediaries). The operating expense ratio is reported as a percent of assets. Components do not add to 100 percent because of rounding. Sources: Investment Company Institute and Lipper The Honorable John Chiang, State Treasurer March 24, 2016 Page 20 of 31 largest cost item and can be the primary determinant of its financial feasibility 34 But the analysis does not include enforcement costs,35 and it is not clear the extent to which it fully considers compliance costs.36 Some assumptions on patterns of participation in the Program could have significant impact on the cost of administering the Program. The feasibility exercise, for example, appears to focus on the number of IRAs actually created, but the number of IRAs attempted but not created could also drive costs higher. The Report estimates that the 6.3 million eligible California participants] 37 Opting those non- participants out of the system could incur processing costs, including the cost of resolving their Social Security number issues. As explained on page 14 in the Report, in the EDD (State of California Employment Development Department) servicing model, the recordkeeper may also need to provide refunds if payroll deduction commences before the recordkeeper resolves the Social Security number issues. In either scenario, the recordkeeper will incur costs related to sorting out the situation with these workers (Figure 10). 34 See page 91 of the Report. 35 See page 113 of the Report. 36 It is difficult to figure out where this cost is accounted for on page 117 of the Report, which provides a detailed breakdown of expense drivers. Enforcement, compliance, and audit costs could prove substantial, and their role in the current analysis should be clarified. 37 See page 112 in the Report. The Honorable John Chiang, State Treasurer March 24, 2016 Page 21 of 31 Another problem with the Report is it assumes that workers who are between jobs do not have an impact on Program costs. The Report states that full-time workers and 25 [percent] of part-time workers would be in between jobs and not represented 38 Even if those estimates are accurate, they do not appear to account for part- year and seasonal workers who will not make periodic contributions throughout the entire year and could end up with smaller balances, which will result in higher per-account costs in the Program. 38 See page 112 in the Report. Figure 10 Sorting Out Social Security Issues May Result in Costly Recordkeeping Note: See page 14 in the Report. Source: Overture Financial Final Report The Honorable John Chiang, State Treasurer March 24, 2016 Page 22 of 31 The Report also does not appear to consider the full extent of the costs to the Program of developing and delivering participant education or communications to report account activity, account balances, and other matters. Employers provide extensive educational materials about private-sector 401(k) plans through multiple touch points. In the IRA market, financial services firms provide extensive materials about opening and investing in IRAs, and have systems in place to comply with Form 5498 and 1099-R reporting requirements. As an entirely new program involving mandatory automatic enrollment, the Program will require extensive educational materials and multiple communication channels to explain the Program to employers and workers. The cost of such materials, initially and on an ongoing basis, needs to be fully incorporated into the feasibility analysis. The Report indicates that education and communication materials will be developed by EDD and the recordkeeper, 39 but the Report does not clearly spell out the extent to which costs associated with education and communication materials for both employers and employees have been incorporated into the feasibility study.40 Any incorrect estimate of the cost of creating and maintaining these materials, communication channels, and reporting systems would impact the financial feasibility of the Program. II. Many Workers Not Covered by Employer-Sponsored Retirement Plans Have Other, More Pressing Financial Needs In addition t Program, we believe that it is important for the Board to appreciate that in contrast to a fundamental assumption underlying the perceived need for mandatory state-run retirement plans workers currently not covered by employer-sponsored retirement plans may not be best served by automatic enrollment into such programs. In this respect, analysis of the data on retirement plan coverage suggests that workers not currently covered by retirement plans tend to have other, more pressing financial needs or savings goals. The Board should be thoughtful about potentially causing inadvertent harm to workers who fail to opt out but really cannot afford to contribute to the plan.41 Analysis of household 39 See page 100 in the Report. 40 See how the estimates were arrived at or the extent to which the development and maintenance of educational and communication materials (for both employers and employees) has been fully captured. 41 Significantly, in the notice accompanying its proposed safe harbor regarding state plan programs, DOL mentions that such inadvertent savings could cause damage to the overall household balance sheet if, for example, debt were incurred or not paid down. DOL mentions the possibility that a college student might reasonably focus on paying down student loans and a young family might focus on saving for education. 80 Fed. Reg. 72012. Household survey data from the Survey of Consumer Finances provide evidence that there is a life cycle of saving: Households tend to focus on building education, a family, or money to purchase a home earlier in life, before focusing on saving for retirement later in life; see Figure 1 in Brady and ICI Research Perspective 20, no. 6 (October 2014), available at www.ici.org/pdf/per20-06.pdf; see also Figure 7.2 in Investment Company Institute, 2015 Investment Company Fact Book: A Review of Trends and Activities in the U.S. Investment Company Industry (2015), available at www.icifactbook.org. In addition, the Federal Reserve Bank of New York Consumer Credit Panel data indicate that in 2015:Q4, student loan debt The Honorable John Chiang, State Treasurer March 24, 2016 Page 23 of 31 balance sheet data indicates that households without retirement accumulations tend to face significant and immediate pressing financial stresses, which would only be heightened if they are automatically enrolled into these plans and a portion of their wage income is set aside into a retirement savings account. A. The Report does not reflect the complexity of factors associated with retirement plan coverage or the potential for economic harm to workers Discussions about retirement plan coverage often rely on misleading or incomplete coverage statistics. The Institute has published extensive research on the difficulties that arise in determining the scope of retirement plan coverage. The most commonly used data understate retirement plan coverage,42 and the most commonly used measure a snapshot of coverage at a single point in time across workers of all ages, incomes, and degrees of attachment to the workforce is not a good was $1.2 trillion, which is larger than the $0.7 trillion in credit card debt and the $1.1 trillion in auto loan debt. See Federal Reserve Bank of New York, The Center for Microeconomic Data, Consumer Credit Panel, Household Debt & Credit, 4 www.newyorkfed.org/microeconomics/data.html. Federal Reserve Board researchers note that see page 26 in Federal Reserve Bulletin 100, no. 4 (September 2014); available at www.federalreserve.gov/pubs/bulletin/2014/pdf/scf14.pdf. They also analyze how education debt burden varies across households. 42 The most commonly used data to analyze retirement plan coverage is the Current Population Survey (CPS), which is a household survey. The CPS typically shows lower rates of pension coverage than surveys of business establishments, such as the National Compensation Survey (NCS). For example, the CPS data show that 59 percent of all full-time, full-year private-sector wage and salary workers had pension coverage in 2013 (pension coverage includes DB and/or DC plans; ICI tabulations of 2014 CPS data). The March 2014 NCS, on the other hand, shows that 65 percent of all private-industry workers and 74 percent of all full-time private-industry workers had access to a pension. See Table 1 in U.S. Department of News Release USDL-14-1348 (July 25, 2014); available at www.bls.gov/ncs/ebs/sp/ebnr0020.pdf. The March 2015 NCS reports that 66 percent of all private-industry workers and 76 percent of all full-time private-industry workers had access to a pension. See U.S. d States News Release USDL-15-1432 (July 24, 2015); available at www.bls.gov/news.release/pdf/ebs2.pdf. The analysis in Figure 11 uses the March 2014 CPS data which provide insight into benefits available in 2013. The CPS, which tends to understate retirement plan coverage, changed the survey in March 2015 and the survey changes inadvertently impacted the retirement plan coverage question responses. The March 2015 CPS data for 2014 find that retirement plan coverage dropped in 2014, particularly among the groups of workers most likely to have retirement plans at work. Copeland conflicting time series of the participation levels in CPS relative to other surveys raise doubts about the use of CPS data to assess future See ent- EBRI Notes 36, no. 12, Washington, DC: Employee Benefit Research Institute (December 2015): 1 11; available at www.ebri.org/pdf/notespdf/EBRI_Notes_12_Dec15_CPS-WBS.pdf. The Honorable John Chiang, State Treasurer March 24, 2016 Page 24 of 31 indicator.43 It is important to understand the typical characteristics of the workers at employers that do not offer plans in order to formulate effective solutions to increasing coverage among the minority of workers who are without access. As explained below, the majority of private-sector workers without employer-sponsored retirement plan coverage are younger, lower-income, or less connected to the workforce. This is the case whether the data are examined for the nation as a whole, or for the State of California. As a result, many of these workers may face financial stresses and savings priorities more pressing than retirement saving. B. Workers not currently participating in retirement plans at work may have other, more pressing financial priorities Without a doubt, inadequate retirement savings can affect o meet basic needs, such as needs for food, housing, health care, and transportation. For many workers not covered by employer-sponsored retirement plans, the difficulty in meeting these basic needs does not begin in retirement, but occurs during their working years as well. Workers not currently covered by employer- sponsored retirement plans who tend to be younger, lower-income, or less connected to the workforce may have other, more immediate savings priorities. Part-time employment in particular may be a signal of financial stress. In its proposed ERISA safe harbor regulation, DOL notes that the state initiatives might have some unintended consequences for such workers, explaining: Workers who would not benefit from increased retirement savings could opt out, but some might fail to do so. Such workers might increase their savings too much, unduly sacrificing current economic needs. Consequently they might be more likely to cash out early and suffer tax losses, and/or to take on more expensive debt. Similarly, state initiatives directed at workers who do not currently participate in workplace savings arrangements may be imperfectly targeted to address gaps in retirement security. For example, a college student might be better advised to take less in student loans rather 43 employers across the entire private-sector workforce. This measure is a poor indicator of whether households will have retirement plan coverage at some point over their lifetimes and approach retirement with retirement accumulations. If this snapshot measure is refined to take into consideration the lifecycle of saving, to recognize the role that Social Security plays in replacing lifetime wage income for lower-income households, and to account for the degree of connection to the workforce it is clear that the majority of private-sector workers most likely to contribute to an employer-sponsored retirement plan have pension plan coverage as part of their compensation. See ICI Research Perspective 20, no. 6 (October 2014); available at www.ici.org/pdf/per20-06.pdf. Put another way, the number of private-sector workers who are likely to be focused on saving for retirement but do not have access to an employer-sponsored retirement plan is lower than suggested by a cursory look at the aggregate data. The Honorable John Chiang, State Treasurer March 24, 2016 Page 25 of 31 than open an IRA, and a young family might do well to save more first for their children's education and later for their own retirement.44 As important as retirement savings is, DOL is correct to point out that these workers may have other priorities for take-home pay. The data suggest that about three-quarters of private-sector workers without retirement plan coverage may be focused on other savings goals or experiencing other financial stresses. The policy rationale underlying the state initiatives does not give adequate consideration to the fact that lack of retirement savings is not the beginning of the financial difficulties for many of these individuals. It also does not give due regard to the important resource that Social Security plays in replacing earnings for U.S. retirees, particularly lower-income workers, who get high earnings replacement rates from Social Security.45 A substantial portion of private-sector workers not currently covered by retirement plans at work may face immediate financial stresses. Among the 50.6 million private-sector wage and salary workers aged 21 to 64 who work for employers that do not sponsor retirement plans, nearly four in 10 (39 percent) work only part-time or part-year (Figure 11). Part-time or part-year work in a given year may be an indicator of financial stress, whether it is a long-term or temporary situation. If these workers usually work part-time or part-year, they are less likely to have additional disposable income to reduce their current consumption to save for retirement, because the vast majority of part-time, part-year workers have low earnings.46 As low lifetime earners, these workers likely will receive a high earnings replacement rate from Social Security.47 If some of these workers who are currently working part-time or part-year usually work full-time or for a full year, then earnings in the current year likely are below their typical earnings, and these individuals are unlikely to want to reduce current consumption further by saving for retirement or for any reason. In either case, part-time, part-year workers are unlikely to be focused on saving for retirement in the current year. 44 80 Fed. Reg. 72012. 45 The Congressional Budget Office reports estimated replacement rates from scheduled Social Security payments, and Social Security replaces a higher percentage of pre-retirement earnings for workers in lower-income households than it does for workers in higher-income households. See Congressional Budget Office, -Term Projections for Social Security: Additional Information (December 2015); available at www.cbo.gov/sites/default/files/114th-congress-2015- 2016/reports/51047-SSUpdate.pdf. 46 See available at www.ici.org/info/per20-06_data.xls. 47 See note 45. The Honorable John Chiang, State Treasurer March 24, 2016 Page 26 of 31 Another 35 percent of private-sector workers without retirement plan coverage at work are very young or lower earners (Figure 11), which suggests they may well have other savings goals, have less need to supplement Social Security benefits, or have other financial stresses. Of this 35 percent, the 14 percent who are full-time, full-year but aged 21 to 29 are likely to be saving for other goals, such as a home, for the family, or education.48 The primary concern for the 13 percent of full-time, full-year 48 According to 2013 Survey of Consumer Finances data, 32 percent of households with head of household aged 21 to 29 indicate that saving for home purchase, the family, or education is their primary savings goal, while only 13 percent of such young households report that retirement is their primary savings goal. See Figure 7.2 in Investment Company Institute, 2015 Investment Company Fact Book: A Review of Trends and Activities in the U.S. Investment Company Industry (2015); available at www.icifactbook.org. Household education loan debt has grown in recent years; see discussion in note 41. Figure 11 About Three-Quarters of Workers Without Retirement Plan Coverage Likely Have Other Financial Priorities Percentage of private-sector wage and salary workers, aged 21 to 64, whose employers do not sponsor a retirement plan, 2013 Note: Components do not add to 100 percent because of rounding. See Figure 6 in the research paper for additional detail. Source: Investment Company Institute tabulations of March 2014 Current Population Survey; see Brady and Bogdan, "Who Gets Retirement Plans and Why, 2013," ICI Research Perspective 20, no. 6 (October 2014) 17 8 8 13 14 39 Percentage of total Part-time or part-year Full-time, full-year, aged 21 to 29 Full-time, full-year, aged 30 to 64, earn less than $25,000 Full-time, full-year, aged 30 to 44, earn $25,000 to $44,999 Full-time, full-year, aged 45 to 64, earn $25,000 to $44,999 Full-time, full-year, aged 30 to 64, earn $45,000 or more (50.6 million private-sector workers without coverage) 39% of workers without retirement plan coverage are not employed full-time, full-year, and likely are financially stressed 35% of workers without retirement plan coverage work full-time, full-year, but are very young or in the lower earnings quintiles, likely with financial stresses or other savings goals The Honorable John Chiang, State Treasurer March 24, 2016 Page 27 of 31 private-sector workers aged 30 to 64 earning less than $25,000 per year more likely will be that they do not have enough to spend on such immediate needs as food, clothing, and shelter. In fact, many are eligible for government income assistance so that they will be able to spend more than what they earn on these items. If these workers consistently have low earnings throughout their careers, Social Security will replace a high percentage of their lifetime earnings, allowing these workers to use more of their wage income to meet current needs and allowing them to delay additional saving for retirement.49 The remaining 8 percent of private-sector workers age 30 to 44 who earn between $25,000 and $44,999 a year may have the ability to save, but may have other saving priorities, such as starting a household and providing for the needs of their children. Given that they get a substantial replacement rate from Social Security, they are likely to delay saving for retirement until later in life.50 Analysis of household balance sheet data indicates that households without retirement accumulations are more likely to face significant and immediate pressing financial stresses compared with those with retirement accumulations. Focusing on older households who have had much of a lifetime to address retirement savings needs, the data show that those without retirement accumulations tend to have indicators of financial stress. Figure 12 examines older households those with a head aged 55 to 64, whether working or not by their retirement accumulation status. Retirement accumulations can be in the form of DC plans, IRAs, or defined benefit (DB) plan benefits. Older households without retirement accumulations are more likely to report that they received income from public assistance: 35 percent of households without retirement accumulations, compared with 4 percent with retirement accumulations. Older households without retirement accumulations are more likely to be lower income: 52 percent are in the lowest per capita household income quintile, compared with 8 percent of households with retirement accumulations. More than one-quarter (27 percent) of older households without retirement accumulations have no health insurance and almost one-quarter (23 percent) do not have checking accounts. All told, 76 percent of older households without retirement accumulations face at least one of these financial stresses, compared with only 20 percent of households with retirement accumulations. 49 For a simulation exercise that explores the relationship and timing of 401(k) plan saving taking into account the role that Social Security plays for American workers in preparing for retirement, see ICI Research Perspective 21, no. 7 (November 2015); available at www.ici.org/pdf/per21-07.pdf. 50 See ICI Research Perspective 20, no. 6 (October 2014); available at www.ici.org/pdf/per20-06.pdf www.ici.org/info/per20-06_data.xls. The Honorable John Chiang, State Treasurer March 24, 2016 Page 28 of 31 C. Workers have access to many tax-advantaged retirement savings opportunities It is certainly essential that workers have easy access to tax-advantaged retirement savings opportunities to supplement the broad-base of the Social Security system. This premise has served as the foundation for the strong voluntary U.S. retirement system. Millions of workers in California have such access already through employer-sponsored retirement plans. For those without employer- sponsored retirement plans, access is available through traditional IRAs (since 1974), Roth IRAs (since Figure 12 Older Households Without Retirement Accumulations Tend to Have Financial Stresses Percentage of U.S. households aged 55 to 64 by retirement accumulation status, 2013 1 Retirement accumulations include retirement assets and DB benefits. Retirement assets include DC plan assets (401(k), 403(b), 457, thrift, and other DC plans) and IRAs (traditional, Roth, SEP, SAR-SEP, and SIMPLE), whether from private-sector or government employers. DB benefits include households currently receiving DB benefits and households with the promise of future DB benefits, whether from private-sector or government employers. 2 Income from public assistance includes TANF, SNAP, and other forms of welfare or assistance such as SSI. 3 Households with a head aged 55 to 64 at the time of the survey were ranked by per capita household income before taxes in 2012. 4 No health insurance indicates that no individual in the household had public or private health insurance. 5 Households may fall into multiple categories. Note: The sample represents 23.0 million households with head of household aged 55 to 64 in 2013; 73 percent had retirement accumulations and 27 percent did not. Source: Investment Company Institute tabulations of 2013 Survey of Consumer Finances 4 8 8 7 20 35 52 27 23 76 Income from public assistance Per capita household income in the lowest quintile No health insurance No checking account At least one indicator of economic stress With retirement accumulations Without retirement accumulations1 2 3 5 4 Indicator of economic stress 1 1 The Honorable John Chiang, State Treasurer March 24, 2016 Page 29 of 31 1998), and myRAs (since November 2015). Traditional and Roth IRAs can easily be opened through a variety of avenues whether through investment professionals or directly with a mutual fund company or discount broker and myRA is available online.51 As federal options, these different IRAs are available to workers across the country, regardless of state of residence or changes in state of residence. Thus, before the State of California embarks on creating another plan, it should remember that such access exists for the more than 6 million California workers who do not have employer-sponsored retirement plans at their current jobs. Indeed, the Report indicates that 71 percent of eligible workers were saving for retirement already, with 45 percent indicating they were saving 5 percent or more for retirement (Figure 13). Many even may be saving in an employer-sponsored retirement plan, as Current Population Survey (CPS) data indicate that about 10 percent of uncovered California workers have a spouse whose employer offers a plan.52 51 Traditional IRA owning households surveyed in mid-2015 report that 80 percent of traditional IRA owning households hold traditional IRAs through investment professionals such as full-service brokerages, independent financial planning firms, banks or savings institutions, or insurance companies, and 33 percent hold traditional IRAs directly through mutual fund companies or discount brokerages (households may have multiple traditional IRAs). See Holden and Schrass, ICI Research Perspective 22, no. 1A (February 2016); available at www.ici.org/pdf/per22-01a.pdf. In addition, firms and individuals interested in the myRA can learn more and set up accounts at https://myra.gov. 52 ICI tabulation of the same sample studied in the Report, which represents 6.8 million California workers without employer-sponsored retirement plan coverage at their current jobs, 2012 2014. The Honorable John Chiang, State Treasurer March 24, 2016 Page 30 of 31 * * * * * The Board must engage in further study before taking any action to recommend moving forward with the Program. Analysis of the data provides reasons to believe that the Program will not be as effective at increasing retirement plan participation and savings in California as the Report assumes. The Program is dramatically different than the voluntary retirement plan system in which automatic enrollment has been so successful. The Report likely underestimates the true costs of setting up and running the Program and would benefit from deeper analysis of scenarios that consider the impact of higher opt-out rates, variation in opt-out rates across employers, lower contribution rates, higher levels of withdrawal and turnover activity, and more comprehensive cost estimates. In addition, the Institute believes strongly that policies that cooperate with, rather than coerce, employers, who best know the demographics and needs of their workers, present far more efficient and effective solutions for expanding coverage. We hope you find the foregoing comments helpful to your consideration of the Report. If you need additional information or you have questions regarding our comments, please feel free to contact Figure 13 Majority of Uncovered Workers Already Are Saving for Retirement Percentage of respondents, 2015 Percentage of household income currently being saved for retirement Note: Sample is 1,000 respondents; workers not offered retirement plans at work. Source: Greenwald & Associates Online Survey 29% 26% 26% 19% Not saving Saving 1% to 4% Saving 5% to 9% Saving 10% or more The Honorable John Chiang, State Treasurer March 24, 2016 Page 31 of 31 me at (202) 326-5815 or david.blass@ici.org; Sarah Holden, Senior Director, Retirement and Investor Research, at (202) 326-5915 or sholden@ici.org; or David Abbey, Deputy General Counsel Retirement Policy, at (202) 326-5920 or david.abbey@ici.org. We welcome the opportunity to discuss these comments further or to provide additional information to you and your staff as you work on this important issue. Sincerely, /s/ David W. Blass David W. Blass General Counsel