Following are questions submitted to the Congressional Budget Office and its responses, following a recent Senate hearing that focused in part on the CBO’s recent report comparing federal pay and benefits to that of the private sector.
Question. You mentioned at the hearing that converting the defined benefit portion of the federal pension plans into a defined contribution would likely entail an up-front cost in the budget window, with savings achieved only beyond the budget window.
Do you believe that having the federal defined benefit pension plans (under both FERS and CSRS) scored on a cash-flow basis—instead of a long-term accrual basis—is the best way to capture the budgetary impact of these pension programs?
Do you think the current scoring methodology dissuades policy makers from appropriately considering the long-term impact of changes to these pension programs?
Answer. There is no single best way to capture the budgetary effects of federal pension plans. As required by law, the Congressional Budget Office estimates the budgetary effects of federal defined benefit pension plans on a cash basis—that is, transactions are recorded when annuity payments are made or revenues are received. Although that method recognizes the cost of outlays when they occur, it does not reflect the cost of expected future outlays when the commitment to make those outlays occurs.
Relative to cash estimates over a 10-year projection period, an accrual measure would offer three advantages:
The cost of providing defined benefit annuities in the future would be recognized as employees earned those benefits. By contrast, cash-based measures do not recognize such costs until after workers leave federal service, making it difficult to implement new policies that would change those costs without altering prior commitments.
All forms of current and deferred compensation would be measured on a consistent basis, making the trade-offs between defined benefit pension plans and defined contribution pension plans clearer to policymakers.
By summarizing long-term budgetary effects up front, an accrual measure would give policymakers a more accurate sense of whether proposed changes to deferred compensation would increase or decrease the deficit. That is particularly important when considering modifications to defined benefit pension plans because such plans involve commitments over long periods.
However, an accrual measure would also have some disadvantages:
Cost estimates can vary substantially depending on the discount rate used for the accrual measure, which may make them less transparent than a cash-based measure.
Accrual accounting also depends on decades of projections for future wages and inflation—which are highly uncertain. The farther into the future such projections extend, the greater the uncertainty.
It generally requires more time and additional resources to provide the initial estimate of future costs and the periodic reestimates that would be required to reconcile actual cash flows with the initial estimate.
Question. In CBO’s April 25, 2017, report Comparing the Compensation of Federal and Private-Sector Employees, 2011 to 2015, the increase in employees’ contributions to their defined benefit pensions does not factor into the comparisons of benefits presented in this report because workers first hired after 2012 had not yet accumulated the five years of service needed to receive the defined benefit pension. Could you tell us how much smaller the difference between the sectors would have been if all federal workers over the 2011 to 2015 period were subject to the higher contribution rate of 4.4 percent?
Answer. The government’s cost of total compensation—the sum of wages and benefits—for federal employees would have been about 2 percent lower if, during the 2011–2015 period, all of those workers had contributed 4.4 percent of their salaries to the Federal Employees Retirement System (FERS) annuity. Under the laws in place during that period, most employees contributed 0.8 percent of their salaries to the annuity, so the higher contribution rate would have reduced their compensation by an amount equal to about 3.6 percent of their salary. Because such wages represent about 60 percent of total compensation, on average, the government’s cost would have fallen by about 2 percent. The percentage reduction in federal employees’ total compensation would be slightly larger for more educated workers than for less educated workers because salaries are typically a larger percentage of total compensation for more educated workers.
The difference between the compensation of federal employees and that of similar private-sector workers varies widely depending on workers’ educational attainment. If federal workers had been subject to higher contribution rates, the difference between their compensation and that of private-sector employees would have been about 2 percentage points less than indicated in CBO’s April 2017 report. In particular:
Among workers whose education culminated in a bachelor’s degree, the cost of total compensation would have averaged 19 percent more for federal workers than for similar workers in the private sector.
Among workers with a high school diploma or less education, total compensation costs would have averaged 51 percent more for federal employees than for their private-sector counterparts.
Total compensation costs among workers with a professional degree or doctorate, by contrast, would have been 20 percent lower for federal employees than for similar private-sector employees, on average.
Overall, the federal government would have paid 15 percent more in total compensation than it would have if average compensation had been comparable with that in the private sector, after accounting for certain observable characteristics of workers.
CBO’s estimate of the effect of higher employee contributions on compensation is uncertain for at least two reasons. First, if federal workers had been subject to the higher contribution rate between 2011 and 2015, employees who stopped working for the federal government might have been more likely to have their contributions refunded to them instead of eventually receiving an annuity. For some of those workers, the forgone annuity payments would have been less than the additional contributions they made. Other workers who highly value having more money to spend before retirement might have decided to forgo larger annuity payments that would have cost the government more than the refunded contributions.
Second, if federal workers had been subject to the higher contribution rate, some of them might have left federal employment, which might have increased or decreased total compensation, on average. Most of those workers would probably have left earlier because the higher contributions would have made federal employment less attractive when compared with nonfederal jobs, retirement, or other alternatives. Other federal employees might have extended their service because the higher contributions would make an earlier retirement harder to afford. The way in which those shifts in retention changed average compensation would depend on the characteristics of the affected employees because the cost of the FERS annuity varies considerably with the worker’s age and years of service.
The increase in contributions made by employees hired after December 31, 2013, will eventually lead to a reduction in compensation for federal employees, holding all else equal, but it is unclear how their compensation will compare with that of their private-sector counterparts in the future. The full reduction in federal compensation will not be realized for several decades (that is, until the entire federal workforce consists of people hired after December 31, 2013). Forecasts of differences between federal and private-sector compensation that far into the future are subject to considerable uncertainty.