A Stress Test of Philadelphia’s Retirement System

Overview

The Pew Charitable Trusts conducted a stress test analysis of Philadelphia’s city retirement plan in late 2018 to help policymakers evaluate how the plan would weather various economic conditions, given recent reforms.

Philadelphia sponsors a retirement plan as part of the compensation package for over 28,000 municipal employees.¹ As of 2017, the most recent year for which data were available for Philadelphia and the other cities cited in this brief—Baltimore, Chicago, Houston, and Pittsburgh—the retirement system was 43 percent funded, a higher percentage than in Chicago but lower than in the other three cities. The annual contribution by city government was over $700 million; the contribution rose to $782 million in 2018.² Philadelphia has increased contributions to the pension system substantially since 2014, when they were $553 million, to make up for previous years in which the city’s contributions had not kept pace with the system’s needs. In addition, reforms that started in 2016 increased employee contributions and established a stacked hybrid structure for new, nonuniformed employees—combining elements of traditional defined benefit plans and 401(k)-style defined contribution plans—to reduce the cost of future pensions.

This brief provides summary results of the stress test analysis for Philadelphia, which Pew conducted in partnership with the actuarial firm The Terry Group. Specifically, the brief assesses whether recent reforms will allow the city’s pension promises to be kept in a fiscally sustainable way, whether the plan will more effectively manage the risk of market volatility and underperformance as a result, and how future retirement system costs will affect Philadelphia’s budget.³ To illustrate the range of outcomes that cities face as a result of their different fiscal situations, plan designs, and funding policies, the brief also presents summary data for the four peer cities cited above.

This research focuses on three key issues: 1) the possibility that investments could not perform as well as expected, 2) the prospect that contributions could fall below the rate required to meet funding objectives, and  3) the impact these risks can have on the plan’s fiscal health and the city’s budget. The analysis takes as a starting point the plan’s own assumptions and actuarial methods—including long-term financial projections released by the Philadelphia city retirement plan each year—before applying risk scenarios. The result is intended to be accessible to all stakeholders and was designed to inform planning and decision-making.⁴

Following are some of the analysis’s key findings for Philadelphia:

• If maintained in the years ahead, Philadelphia’s current funding policy is projected to sustainably pay down the city’s pension debt even if returns fall short of expectations.
• Although market volatility will affect the city government’s pension costs, these costs are not projected to be materially higher than they are now—in terms of the percentage of the city’s overall revenue to be deposited into the pension system—under any of the scenarios tested.
• The cost to the city is likely to remain high for years to come. This means that contributions to the pension fund will continue to crowd out spending on other budget priorities.
• The stacked hybrid plan design, now in effect for some newly hired workers, will lower the city’s exposure to investment risk if its current form is maintained; however, any future increases in the defined benefit portion  of the hybrid will add to the cost of benefits.

Philadelphia findings

As part of the stress test analysis, three economic scenarios were simulated for the Philadelphia pension system through 2037. One modeled the baseline projections under current plan policy and actuarial assumptions, in which the plan meets its 7.65 percent expected annual return on investments. Another modeled a likely lowreturn scenario in which the return on assets is 5 percent, reflecting financial experts’ consensus that future returns will perform below historical averages.⁵ The third scenario, an asset price shock simulating a recession, showed market returns initially declining by approximately 26 percent in year one, followed by a three-year recovery and then low, 5 percent equity returns over the long term.⁶

The analysis takes into account several changes that have been made to the city’s main retirement system in recent years, including the pension board’s adoption of a revenue recognition policy that has resulted in increased contributions by city government; the dedication, by state law, of a portion of the local sales tax to the pension system; higher employee contributions; and the hybrid plan for new workers.

As shown in Figure 1, if the plan meets its 7.65 percent assumed rate of return annually, the government’s contributions will drop from 13 percent of revenue in 2017 to about 9 percent in 2034 and then to below  3 percent.⁷ Absent the Philadelphia Water Department’s and Philadelphia International Airport’s funds, which were included in the definition of revenue based on feedback from city budget officials, the 2017 contribution  as a share of revenue was 17 percent. Philadelphia’s 2017 estimated employer contribution as a share of revenue is currently among the highest of the cities assessed, but it is projected to be the lowest at the end of the  forecast period if all assumptions are met.

In a low-return scenario in which the plan earns 5 percent annually, the data project a more gradual decline  in employer contributions, which would remain above 10 percent of revenue throughout the forecast period.  And in an asset shock scenario, projected contributions based on current policy would rise during the initial market downturn, then start gradually declining in a pattern similar to that projected in the low-return scenario.

Likewise, in each of these scenarios, the system’s funded ratio—the amount of funds on hand divided by benefits owed—is expected to improve, although it would fall initially in the asset shock scenario, as shown in Figure 2. The ratio reaches at least 80 percent by 2035, even under the two lower-return scenarios.

These results are largely a function of the recent changes in the system, particularly the relatively high level of current contributions from the city and employees combined—equal to 95 percent of benefit payments in 2017. By way of comparison, the next highest contribution-benefits ratio among the four comparison cities was

68 percent, in Pittsburgh; the highest contribution ratio among state systems in 2017 was 89 percent, in North Dakota. Another contributing factor is the hybrid tier introduced in the 2016 reforms, which will lower plan costs over time by providing some new employees with a defined benefit plan for only a portion of their salaries.

Long-term market underperformance is not the only source of investment risk to retirement systems. Annual fluctuations in market returns can cause volatility in required employer contributions or result in decreased pension plan funding even if returns match plan actuaries’ assumptions over the long term. For this reason, the analysis also estimates financial outcomes using stochastic analysis, a simulation tool that generates thousands of possible forward-looking trials to examine the probable impact of market uncertainty on financial outcomes.

The next two graphics illustrate how future market volatility may affect Philadelphia’s pension costs. Figure 3 depicts three trial runs with varying sequences of returns that average 7.65 percent over the 20-year forecast period—and compares them with what would happen if the returns were 7.65 percent each year.⁸

Figure 4 shows the impact of this volatility on employer contributions, with the city’s annual contribution to the pension fund—expressed as a percentage of payroll—projected over the forecast period for each of the trials illustrated in Figure 3. The city’s contribution in 2017 was 42 percent of payroll; however, the subsequent trajectory of costs depends on the timing of returns in addition to the long-term investment performance. In other words, volatile market returns could potentially lead to longer-term high costs and effectively delay full funding of the system even if long-term plan assumptions are correct, as shown most clearly in Trial 3 of Figure 4.

Philadelphia’s high contribution rate is a primary driver behind the findings of projected cost stability and funding improvements. In particular, Philadelphia’s funding policy pays the full actuarial determined contribution and additionally sets aside a portion of sales tax revenue and employee contributions as pension plan payments above and beyond the actuarial contribution.⁹ The adoption of a stacked hybrid plan design for new, nonuniformed city employees also plays a role. The plan provides a defined benefit to all workers based on salaries up to $65,000. Employees with higher salaries may also participate in a defined contribution plan, with employees contributing up to 3 percent of their salaries above $65,000 and employers matching 50 percent of those contributions. Because city funding of the defined contribution portion of the stacked hybrid remains fixed regardless of market performance, this plan reduces Philadelphia’s exposure to investment risk.

The benefit design, based on a collective bargaining agreement, does not provide for indexing the $65,000 salary threshold to any measure of inflation. As a result, as salaries rise over time, a declining share of total salary is subject to the defined benefit plan, thereby reducing employer cost and risk. But if the threshold were to increase in future years, it would lessen the reduction of employer cost and risk.

How Philadelphia compares with its peers

Because it can be helpful to compare the fiscal resilience and plan design of one public retirement system with those of similar systems, the researchers performed a stress test analysis on municipal and public safety employees’ pension plans in four of Philadelphia’s peer cities: Baltimore, Chicago, Houston, and Pittsburgh.

Figure 5, which provides the key fiscal metrics of this comparative analysis, shows that Philadelphia is unique in the level of total contributions it makes to the municipal retirement system—95 percent as a fraction of current benefit payments, which means it had nearly as much coming into the funds as was going out. In essence, annual contributions from government and employees are paying for retiree pension checks, leaving investment returns on the plan’s assets to pay down the system’s legacy debt. As described above, Philadelphia’s high contribution ratio helps to insulate its retirement system and budget from adverse scenarios. However, this is not the case for other cities, highlighting the importance of maintaining the funding commitments that Philadelphia has made in recent reforms, including the revenue recognition policy.

Projected funded ratios under a 5 percent return scenario using each city’s current contribution policy yield striking differences. Philadelphia’s funding policy is more robust than those of the comparison cities: It is the sole system to generate significant improvements in funded status even under this low-return scenario. Comparisons of Philadelphia’s forward-looking stress test results with those of the other cities will be explored in greater detail in a full report later this year.

Conclusion

Philadelphia’s recent reforms demonstrate that improved funding of a municipal pension system is attainable if the city strictly adheres to scheduled contributions. The city’s high contribution rate, although challenging from  a budgeting perspective, provides protection from future investment underperformance. Even under adverse return scenarios, required contributions are not projected to increase appreciably over time as a percentage of budget resources. And perhaps most significantly, if Philadelphia adheres to its new, higher contribution levels, the pension system’s funded ratio will increase gradually and substantially under any economic scenario.

The projected improvement in Philadelphia’s retirement system funded level is also due, in part, to the stacked hybrid plan design for new hires, which exposes the city to lower levels of investment risk over time as the stacked hybrid becomes a larger portion of total pension liability.

But even under the best scenarios, high pension costs are likely to persist in Philadelphia for years, making that money unavailable to address other budget priorities. And if policies are changed at a later date—by reducing contributions or increasing benefits without funding them—these findings are likely to change.

Philadelphia’s pension reforms have set the city retirement system on a path to sustainably deliver on pension promises as long as policymakers remain committed to their current contribution policy in the years to come.

Maintaining those policies will have real costs for taxpayers and those who depend on city services.

Endnotes

1. Additionally, Philadelphia sponsors the Gas Works plan, which accounts for 7 percent of the city’s total pension liability. Due to its relatively small size, the Gas Works plan was not included in the analysis. This analysis does not include teachers in the School District of Philadelphia, who are part of the statewide Public School Employees’ Retirement System (PSERS). Contributions to PSERS are made by the school district and partially reimbursed by the commonwealth. Baseline projections were estimated using the city retirement plan’s own assumptions and actuarial information, including an assumed rate of return of 7.65 percent. The plan board has scheduled an additional reduction in the discount rate to 7.6 percent that is not incorporated in this analysis.
2. See Appendix A for methodological details. The 43 percent funded ratio is based on the market value of assets; using the actuarial value of assets, as Philadelphia’s Board of Pensions and Retirement does, the funded ratio is 45 percent.
3. The analysis is informed by the Actuarial Standards Board’s recent guidance on risk reporting and was created using Pew’s stress test methodology as described in Greg Mennis, Susan Banta, and David Draine, “Assessing the Risk of Fiscal Distress for Public Pensions: State Stress Test Analysis,” Harvard Kennedy School Mossavar-Rahmani Center for Business and Government Associate Working Paper No. 92 (2018), https://www.hks.harvard.edu/centers/mrcbg/publications/awp/awp92.
4. Stress test analysis can also be used to estimate the impact of longevity risk, mortality risk, changes in workforce behavior, and other actuarial assumptions. These factors are held constant in this analysis to examine investment and contribution risk.
5. Pew’s capital market assumptions yield a 50th percentile expected return of 6.4 percent and a 25th percentile return of 4.8 percent when applied to a typical public pension asset allocation. When applied specifically to the Philadelphia city retirement plan’s asset mix, they yield 50th and 25th percentile returns of 6.4 percent and 4.7 percent, respectively.
6. Appendix A contains more detail on investment risk scenarios, including additional information on how they were developed. The asset shock scenario is based on the Federal Reserve’s “2017 Supervisory Scenarios for Annual Stress Tests Required Under the Dodd-Frank Act Stress Testing Rules and the Capital Plan Rule.” Applying the economic and financial data points from that scenario, a drop in equity prices of 39 percent is assumed in the first year of the shock, followed by equity returns of 18.75 percent, 20.6 percent, and 16.24 percent, respectively, in the three years of the subsequent recovery. In addition to the four-year forecast from the Dodd-Frank stress test, an assumption of 5 percent return on equity is applied in years following the recovery.
7. Pew’s Philadelphia projections were calculated assuming contribution levels as prescribed in the Revenue Recognition Policy and revenue growth as provided in the City of Philadelphia’s FY 2018-22 Five Year Financial Plan. Revenue here is defined as own source revenue, calculated as the city’s total revenue from governmental funds, plus revenue from the Water Department and airport funds (which participate in the city retirement system), less revenue from other governments as reported on the city’s Statement of Revenues, Expenditures and Changes in Fund Balances. See Appendix A for additional methodological details.
8. See Mennis, Banta, and Draine, “Assessing the Risk,” for a more detailed discussion of stochastic methods as they apply to stress testing public pensions. See also Yimeng Yin and Don Boyd, “Analyzing the Interplay Between Public-Pension Finances and Governmental Finances: Lessons From Linking an Economic Model to a Pension Fund Model” (2018), https://www.brookings.edu/wp-content/uploads/2018/04/Boyd-Yin.pdf.
9. Projected cost stability and funding improvements are therefore subject in part to the accuracy of sales tax revenue as forecast and reported in the plan actuarial valuation.