Most state and local governments provide traditional defined benefit (DB) pensions, enabling their employees to retire with secure, predictable, and adequate retirement incomes. However, in the past decade, historically low market returns, along with many governments’ failure to keep up with required contributions, have caused many state and local pension plans to be underfunded.
Rhode Island’s pension fund was no exception to this trend. It faced an additional problem in that Rhode Island had been slower than most states to move to an advance-funded pension system, and faced a shortfall even before the 2008 market crash. The fund was further weakened by the state’s failure in some years to keep up with required contributions and pay for retroactive benefit increases, and by adopting arguably unrealistic actuarial assumptions that, when adjusted, significantly weakened its financial outlook (GRS 2011a; GRS 2011b; GRS 2012; Raimondo 2011).1 Consequently, as of fiscal 2010, Rhode Island’s state pension funds reported a combined funding ratio of just 49 percent, the second-lowest of all states (Pew Center on the States 2012).
Ostensibly in response to this low funding ratio, Rhode Island passed the Rhode Island Retirement Security Act (RIRSA) in 2011, which took the radical step of cutting pension benefits for both future and current state employees and teachers through a variety of means, including the replacement of their DB pensions with a new hybrid system composed of a smaller DB pension and a new defined contribution (DC) plan.
This issue brief begins by comparing RIRSA with the pension plan it replaced. It then analyzes RIRSA’s impacts on Rhode Island state employees, as well as on the state’s finances. Key findings include:
- The shortfall in Rhode Island’s pension plan for public employees is largely due not to overly generous benefits, but to the failure of state and local government employers to pay their required share of pensions’ cost.
- The savings from the Rhode Island Retirement Security Act (RIRSA) of 2011 are due to its higher retirement age and lowering or suspending the cost-of-living adjustment.
- RIRSA also cut the defined benefit (DB) pension accrual rate and introduced a new defined contribution (DC) plan. The new DC plan doesn’t save the state money, but will cost retirees.
- RIRSA will result in an average benefit cut of 14 percent for future full-career employees. Furthermore, due to the market risk introduced by the DC plan, many future employees will likely do even worse than this average: For the quarter of future employees who are in the lowest quartile of investment returns on their DC plan, the cuts will be 22 percent or higher.
- These cuts to retiree incomes stemming from the hybrid DB+DC plan are not projected to translate into savings for the state, and will do little, if anything, to improve the health of Rhode Island’s pension funds. The changes will actually increase the average annual cost for taxpayers.
- Rhode Island can and should make its pension funds solvent without exposing future retirees to the risks and higher costs of DC plans.
Comparing Rhode Island’s old and new retirement systems
Rhode Island’s retirement system prior to the 2011 RIRSA reforms was not particularly generous. Consequently, RIRSA’s provisions—including suspending COLA payments, raising the normal retirement age, and introducing additional risk into employees’ retirement via a supplemental defined contribution plan—erode already-lackluster benefits.
Old system (Schedule B)
Before RIRSA, under the benefit scheme known as Schedule B, Rhode Island public employees enjoyed a standard defined benefit pension, where benefits at retirement were based on years of service and final average salary, providing a predictable, constant stream of income in retirement. Benefits increased along with years of service, such that state employees, including teachers, earned a pension of 16 percent of their final average salary after 10 years of service (meaning they earned a 1.6 percent benefit per year), but received 55.25 percent after 30 years (averaging 1.84 percent per year of service). This is not particularly generous in comparison with other states, where the most common accrual rate is 2 percent per year of service (author’s analysis of CRR and CSLGE 2009).
The normal retirement age was 65 after 10 years of service and 62 after 29 years of service. Retirees received a cost of living adjustment (COLA) tied to inflation and capped at 3 percent on the first $35,000 of their pensions; in other words, up to $1,050 per year. Again, neither the retirement age nor the COLA was particularly generous in comparison with other public plans (author’s analysis of CRR and CSLGE 2009).
In 2010, state employees2 approaching retirement who had worked for the state for most or all of their careers (30 years of service) had a final (five-year) average salary of around $60,000 (author’s calculations based on GRS 2011a and assuming 4 percent salary growth). Based on a replacement rate of 55.25 percent after 30 years, the average pension for a full-career state employee would have been approximately $33,500. However, because many workers do not work for the state their entire careers, the actual average annual benefit paid to state employees was less than half that amount—about $16,400. Nearly all state employees and roughly half of teachers are part of the Social Security system, so most employees’ retirement income is supplemented by Social Security benefits (Nuschler, Shelton, and Topoleski 2011; Raimondo 2011).
The normal cost3 of providing benefits under the old ERSRI system was, on average, 11.4 percent of employees’ salaries (GRS 2011a), of which employees paid a flat rate of 8.75 percent and state and local governments together paid 2.64 percent. However, because of the large hole in the system’s finances, Rhode Island state and local governments combined contributed over $300 million in 2010 in total to the teachers’ and state employees’ pension funds, which is equal to 19.5 percent of employees’ total salaries that year.
Taken together, these findings suggest that the shortfall in Rhode Island’s pension plan for public employees is largely due not to overly generous benefits, but to the failure of state and local government employers to pay their required share of pensions’ cost.
New system (RIRSA)
Most immediately, RIRSA, which was passed in 2011, suspends COLA payments until the aggregate funding ratio for all state pensions reaches 80 percent (except for an interim COLA calculated at five-year intervals).4 Afterwards, a COLA of between 0 and 4 percent (based on the funds’ average investment returns) will be awarded on the first $25,000 of an employee’s pension. While this is the most immediate cut, in what follows we will focus more on the other cuts introduced by RIRSA.
Additionally, RIRSA raises the normal retirement age for calculating DB benefits from the current 62 after 29 years of service and 65 after 10 years of service to match the normal retirement age for Social Security (currently 66 but gradually rising to 67). This increase in the normal retirement age constitutes a significant benefit cut on its own.
Finally, RIRSA changes the accrual rate under the current DB plan while also instituting a new DC plan for state employees. Under RIRSA, state employees receive a pension of 1 percent of final average salary per year of service. After factoring in the increased retirement age, this leaves lifetime DB benefits at just around half what employees would have received under Schedule B. Using the previous example, a state employee who retired with a final average salary of $60,000 after 30 years of service would receive DB benefits of $18,000 per year instead of $33,500 per year, and would receive these benefits for four fewer years. While employees’ total contribution rates remain the same under RIRSA, more than half of their total contribution—5 percent of salary—will be mandatorily deposited into a newly created defined contribution account, administered by TIAA-CREF, in which all state employees are automatically enrolled. The state will also contribute 1 percent of salary into these accounts. Each employee can choose to invest their savings in a limited menu of stock and bond mutual funds. Most funds are index funds or lifecycle funds, and all are very low-fee (TIAA-CREF 2012). Benefits at retirement are based on an account’s balance and can either be taken as a lump-sum or annuitized through TIAA-CREF’s traditional annuity.
In a very real sense, RIRSA’s changes are directly analogous to plans floated in the mid-2000s to introduce private accounts to Social Security. Very large cuts to the guaranteed portion of the plan (either the traditional Social Security benefit or, in the case of Rhode Island, the defined benefit plan) were paired with a new defined contribution component. This new defined contribution component was often presented so as to convey the perception that it effectively neutralized the cuts to guaranteed benefits (even if more risk was being introduced). But as with plans to introduce private accounts to Social Security, the expected benefits stemming from the introduction of defined contribution elements is nowhere near large enough to offset the cuts to the guaranteed benefits.
Despite containing many good features (particularly low-fee investment options) most private-sector defined contribution plans lack, RIRSA actually lowers benefits for state employees and introduces more risk. Further, these benefit cuts do not translate into lower employer contributions; under RIRSA, the state will actually pay a higher average normal cost, even in the unlikely event that the state is able to implement these benefit cuts without any effect on employee recruitment or retention.
Impact on state employees
Rhode Island’s new pension system will significantly reduce the retirement income of state employees in several ways. First, the increase in the full retirement age from 62 to 67 is in effect a large benefit cut, as it reduces the average duration of workers’ retirement and therefore their total retirement income. According to the Social Security Administration, the average life expectancies at ages 62 and 67 are 20.9 years and 16.3 years, respectively (SSA 2012). Thus, raising the retirement age from 62 to 67 reduces the average retirement by 4.6 years, or 22 percent.5
Besides raising the normal retirement age (an unambiguous and large benefit cut), RIRSA also combines reductions to the accrual rate of the DB portion of employee retirement plans with significant contributions to a new defined contribution plan. Despite the addition of new contributions to a DC plan, the net effect of all of these changes is, on average, a benefit cut.
To estimate the size of the cut experienced by a future employee, we used the average salary and salary increase data from ERSRI’s annual report (GRS 2011a) to estimate the career salary path of a typical full-career employee. Using the mandatory contribution rate into and portfolio choices offered by the TIAA-CREF–administered DC plan (TIAA-CREF 2012), we performed Monte Carlo simulations6 to project returns on a balanced portfolio of funds in the plan, split evenly between stocks and bonds. Using the simulation results, we estimate the balance at retirement in an employee’s DC account under three scenarios, reflecting average, low (25th percentile), and high (75th percentile) returns over a working lifetime of 7.35 percent, 5.95 percent, and 8.75 percent, respectively. We then annuitize the projected account balance at retirement through TIAA-CREF’s Traditional Annuity (TIAA-CREF 2012) to convert the nest egg into a lifetime stream of income, allowing us to compare directly the new system with the old.
Figure A summarizes the results of these calculations. On average, a typical full-career future employee (i.e., one with 30 years of service) will have nearly $8,000 less per year in retirement income, a 14 percent cut. Many future state workers, however, will fare significantly worse: The quarter of employees who earn low returns on their defined contribution portfolio over their working lifetimes will experience a cut of at least $12,500 per year in their retirement incomes, which translates to at least a 22 percent cut. Though the quarter of employees who are lucky enough to work during a lifetime “bull market” will enjoy retirement incomes nearly equal to what they would have had under the old system, on balance the new system introduces significant risk that retirees will retire with less—risk not present under the old system.
Estimated yearly retirement income of a typical full-career Rhode Island state employee, old system (Schedule B) versus new system (RIRSA)
|Defined contribution||Defined benefit|
|RIRSA, mean returns (7.35%)||$19,277||$28,963|
|RIRSA, low returns (5.95%)||$14,651||$28,963|
|RIRSA, high returns (8.75%)||$25,724||$28,963|
Source: Author's analysis of GRS (2011c, 2012)
Impact on state finances
Over time, RIRSA will likely lead to a gradual improvement in the Rhode Island pension funds’ funding ratio. However, this improvement can, on net, be entirely attributed to the increase in the retirement age and suspension and reduction of COLA benefits. The change in DB accrual rates combined with the introduction of the DC plan actually increases costs to state and local governments and taxpayers while making retirement incomes less secure and failing to make up for the cuts to the DB portion of employees’ pensions. Under the old system, the most recent estimate of the normal cost to the state per employee was 2.64 percent of salary (GRS 2011a). The projected long-term normal cost to the state for the DB portion of the new system will be around 2.43 percent of payroll. Adding the state’s 1 percent contribution for the new DC plan means the state’s total contribution per employee will rise by about 30 percent (author’s analysis of GRS 2011c). To put this in perspective, in 2013, the total payroll of all state employees, teachers, and municipal employees is projected to be just over $2 billion (GRS 2011c). Over the long term, RIRSA may cost the state upwards of $15 million a year in additional contributions while providing a smaller benefit for the average full-career worker.
The failure of the introduction of a new DC plan to offset cuts to the guaranteed portion of Rhode Island employees’ pensions is therefore not due to this change leading to lower state contributions. Rather, it is due to the inherent inefficiencies of defined contribution plans, inefficiencies that Rhode Island’s TIAA-CREF plan shares. On net, even as the state will contribute a higher amount to ERSRI than before, benefits received by retirees will be lower because defined contribution accounts’ average returns and annuitization rates are lower (DC portfolios are necessarily less diversified than pension funds’, and TIAA-CREF’s traditional annuity, while one of the better annuities offered in the private market, is still more expensive than an annuity provided through a state pension fund). Further, the accounts’ exposure to market risk creates the possibility that many individuals’ retirement income will be significantly lower than average.
Rhode Island’s pension cuts are not only bad for its employees and the state, they promote the false notion that providing retirement benefits through DC plans is a sensible or viable option for pension plans facing shortfalls. Defined benefit pensions remain the most efficient means to provide retirement income for all workers, public and private, and states looking to solve shortfalls in their plans should not turn to DC plans, which simply shift risks without lowering costs. Instead, they should find ways to preserve retirement benefits through their existing systems. However, states can still learn from Rhode Island’s example by recognizing that they must preserve the pensions that have provided retirement security for generations of Americans if they wish to continue to provide the public services that taxpayers rely on.
About the author
Robert Hiltonsmith is a policy analyst at Demos. He joined Demos in March 2010 to provide research and analysis on issues surrounding retirement security in the United States. Since then, he’s written on a wide variety of topics, including tax policy, fiscal policy, education policy, health care, and the labor market. Robert’s research has been widely covered in the press, including the Washington Post, Time, The New York Times, and The Wall Street Journal. He has appeared on national and regional television and radio, including Fox Business News, PBS Frontline, and Fresh Air. Robert’s writing has appeared in a variety of news outlets including POLITICO, Newsday, and The American Prospect. He has a M.S. in economics from the New School for Social Research and a B.S. in mathematics and philosophy from Guilford College.
1. Among other things, ERSRI went from using a relatively high (8.25 percent nominal, 5.25 percent real) rate-of-return assumption compared with other public pensions to a relatively low assumption (7.5 percent nominal, 4.75 percent real) (GRS 2011a; GRS 2011b; NASRA 2013). The state also selectively switched from a smoothed to a market valuation of assets during the technology bubble in order to lower employer contributions (Raimondo 2011).
2. For simplicity, the numbers in this brief are for “general” (i.e., administrative/clerical) state employees, who represent 34 percent of all Rhode Island state and municipal employees. The figures for teachers and municipal employees, the other two major groups of public employees in Rhode Island, are very similar.
3. The normal cost is the present value of projected future benefits attributed to the current year of service.
4. While the regular COLA is suspended, an interim COLA will be awarded based on the plan’s five-year average investment rate of return minus 5.5 percent, and will range from 0 to 4 percent. The interim COLA will be awarded only on the first $25,000 (indexed to inflation) of a member’s pension benefit, and thus will range from $0 to $1,000.
5. In present value terms, the benefit reduction is even greater than 22 percent because initial pension benefits are valued more than those received later in retirement. In other words, the real value of the benefit cut for employees is actually greater than 22 percent taking into account expected investment earnings on the forgone benefits.
6. A Monte Carlo simulation is a statistical method for modeling situations that present uncertainty, such as future investment returns. The simulation samples from a given distribution with a set mean and standard deviation (in this case, historical returns on stocks and bonds), and draws a random sequence (of set length) from that distribution. To ensure statistical accuracy, the process is often repeated 10,000 times, and the average of all the sequences is then taken to obtain an estimate of, in this case, future returns.
Center for Retirement Research (CRR) at Boston College and Center for State and Local Government Excellence (CRR and CSLGE). 2001–2009. Public Plans Database.
Gabriel Roeder Smith & Company (GRS). 2011a. Employees’ Retirement System of Rhode Island Actuarial Valuation, June 30, 2010. https://www.ersri.org/public/actuarialValuations/ERS_VAL10.pdf
Gabriel Roeder Smith & Company (GRS). 2011b. Employees’ Retirement System of Rhode Island Actuarial Experience Study for the Six–Year Period Ending June 30, 2010. https://www.ersri.org/public/actuarialValuations/ERS_VAL10.pdf
Gabriel Roeder Smith & Company (GRS). 2011c. Actuarial Analysis of the Rhode Island Retirement Security Act of 2011, as Described in S1111A and H6319A. http://www.treasury.ri.gov/secure-path-ri/legis/RetirementSecurityAnalysisAmendNov11Final.pdf
Gabriel Roeder Smith & Company (GRS). 2012. Employees’ Retirement System of Rhode Island Actuarial Valuation, June 30, 2011. https://www.ersri.org/public/actuarialValuations/ERS_VAL11.pdf
National Association of State Retirement Administrators (NASRA). 2013. Public Pension Plan Investment Return Assumptions. NASRA Issue Brief. http://www.nasra.org/resources/issuebrief120626.pdf
Nuschler, Dawn, Alison M. Shelton, and John J. Topoleski. 2011. Social Security: Mandatory Coverage of New State and Local Government Employees. Congressional Research Service. http://www.nasra.org/resources/CRS%202011%20Report.pdf
Pew Center on the States. 2012. The Widening Gap Update. http://www.pewstates.org/research/reports/the-widening-gap-update-85899398241
Raimondo, Gina. 2011. Truth in Numbers: The Security and Sustainability of Rhode Island’s Retirement System. Rhode Island Office of the General Treasurer. http://www.law.yale.edu/documents/pdf/cbl/RI_TIN-WEB-06-1-11.pdf
Social Security Administration. 2012. “Period Life Table 2007” [last reviewed or modified April 10, 2012]. http://www.ssa.gov/oact/STATS/table4c6.html
Teachers Insurance and Annuity Association of America (TIAA-CREF). 2012. “Retirement Investments: Current Performance” [data tables, accessed online December 2012]. http://www1.tiaa-cref.org/public/performance/retirement/267.html