Business School research suggests better oversight needed for defined benefit public pensions
A series of studies by two University of Colorado Denver Business School professors indicate a large level of surplus deferred compensation and near-retirement salary growth among both K-12 and university employees.
Overall, the three studies point to potential conflicts between agents managing employee compensation and taxpayers providing the benefits. At the very minimum, the studies suggest, improved oversight is needed for these pension plans. With state and local public employee defined benefit public pension plans facing underfunding problems, many state and local governments are considering ways to reduce large deficits.
The first two studies used data sets of retiree characteristics and salary histories from two groups — non-faculty university employees (278) at the University of Colorado Denver, University of Colorado-Boulder and Metropolitan State University of Denver and K-12 Denver Public Schools retirees, all of whom retired between 2001-06. The K-12 retirees were members of the Denver Public School Retirees System, which was a separate retirement plan for all K-12 retirees. Now it is a division of the Colorado Public Employees’ Retirement Association (PERA) defined benefit pension plans.
Prompting the studies was the fact that deferred retirement compensation is a major component of defined benefit pension plans, but hadn’t received much study, not even by the PERA Commission, said Michael Mannino, PhD, associate professor of information systems in the Business School. Surplus deferred compensation is defined, simply, as a retiree getting more out of a pension than was put in, he said.
“I wanted to get some idea of the private-sector valuation of the pension benefits, and compare that to what was really put into the plans by both the employee and employer,” he said. “I think an understanding of that comparison is essential to construct reasonable policy.”
The studies by Mannino and Elizabeth Cooperman, professor of finance, showed a large gap between the Colorado Department of Personnel and Administration’s value of a retiree’s defined retirement benefits and a private-sector valuation. “I came up with a method of doing a private-sector evaluation of the pension value. In other words, how much it would cost to purchase that pension in the private sector at retirement.”
In the data sets, he said, the average surplus deferred compensation — the additional amount needed to purchase an annuity above what was available at retirement from employer and employee contributions invested at historical rates — was more than $500,000, or generally 25 to 35 percent in additional compensation.
Taxpayers are essentially on the hook for the additional amount. “The employer’s contribution in the plan valuation was historically around 10 to 12 percent through PERA, but what the study showed is that you should actually add another 25 to 35 percent,” Mannino said. “So, essentially, you’re looking at retirement contributions by the taxpayer of about 35 to 45 percent, if you count the 10 percent that is already there.”
Data sets used in the three studies are unique, Mannino said, as other researchers and the Colorado State Treasurer have not been able to obtain comparable data. The first two studies were published in the the Journal of Pension Economics and Finance.
In the latest study, currently under journal review, Mannino and Cooperman focused on benefit enhancement — various ways an employee increases retirement benefits near his or her retirement date. Methods of spiking pensionable income include salary growth near retirement, excessive overtime, adding unused vacation and sick time, adding uniform allowances, and enhancing education, such as an advanced degree.
Mannino and Cooperman used diverse sources of evidence, including reports on pension benefit enhancement in the financial press; a survey of managers of state and local public employee pension plans; and an empirical evaluation of two unique data sets of retiree characteristics and salary histories. Read the study abstract here.
Using the unique data sets already procured on retirees and salary histories, they compared wage growth in the several-year period used to determine final average pay with the Consumer Price Index (CPI) and other wage indexes. “I compared the actual geometric growth of the compensation — the salaries — in that period compared to these indexes,” Mannino said. “I found a pretty good surplus there, indicating there was some level of, you might say, pension spiking or enhancement of benefits. But the employer wasn’t paying attention.”
For administrators, the average salary growth rate in the five-year period (2001-06) was about 7 percent, Mannino said. For professionals and non-professionals in the data sets, growth rates were about 4.7 and 4.4 percent, respectively. Meanwhile, the CPI growth rate in the same period was about 2.3 percent. “So, for administrators (the growth) was almost three times the CPI and for the other groups it was about 2 times.”
The most surprising part of the study was the dominance of professional and non-professional employees with extreme salary growth in the five-year period before retirement. According to the study, “The top 10 outliers (0.196 to 0.520) contained only professional and non-professional retirees. In addition, the non-professional class contributed a reasonable number of extreme and mild outliers. Thus, the results from the distribution of salary growth and outliers indicated widespread principal-agent conflict, not limited to administrative employees.”
Mannino said this principal-agent conflict can circumvent a primary advantage of defined benefit pension plans — the ability to retain personnel with low costs during the near-retirement period. “Special pension enhancement provisions that allow employees to boost salaries near retirement to enhance future pension benefits provide incentives for opportunistic employees to game the system, often resulting in higher future employee and employer contributions to cover future underfunding,” Cooperman noted.
The state pension bill passed in 2009 addressed some of the issues by stretching the salary period and placing more limitations on salary increases in the pension calculation period, Mannino said. However, the study concludes, there should be better oversight associated with large salary growth that occurs just outside the pension benefit calculation period. An even better policy is to consider defined contribution pension plan now provided for many exempt employees at CU Denver, the study suggests.
Abstracts of the first two studies published in the Journal of Pension Economics and Finance:
- Surplus deferred pension compensation for long-term K-12 employees: an empirical analysis for the Denver Public School Retirement System and four state plans.
- Deferred compensation for career employees in public defined benefit pension plans: evidence from Colorado PERA.