Later this fall, Governor Terry Branstad will begin negotiating new two-year contracts with several labor unions that represent state employees. He will come to the table armed with research that purportedly shows Iowa has been over-compensating its public workers.

Branstad has been beating the drum about excessively generous state employee benefits for months, but few unionized workers took up his call to voluntarily contribute to their health insurance premiums. In September, the Iowa Department of Administrative Services released a report supporting the governor’s position:

A new report concludes the benefit package for state government employees is better than what’s normally offered to employees in 13 other states and a small group of businesses considered to be competitors for the same workers.

Iowa Department of Administrative Services director Mike Carroll says the analysis found state worker health care benefits were “well above the norm.”

Michelle Minnehan Golightly,  the chief operating officer for the Iowa Department of Administrative Services, says Iowa is spending about $116 million more than its peers do on benefits and the main difference is the state now pays the entire health insurance premium for most of its workers.

“The market shows that 20 percent is in line,” Golightly says.

It’s no secret that state workers receive generous benefits. A 2011 report by the Iowa Policy Project found,

When education, work experience, annual hours worked, race, sex, disability status, and firm size are accounted for, male public-sector workers earn nearly 12 percent less and female public-sector workers earn over 16 percent less than private-sector workers. Male state government workers earn 9 percent less than comparable workers in private industry, while for local government the public-sector wage penalty was 14 percent. Among women, the earnings penalty was over 13 percent for state workers and 19 percent for local government workers.

Many critics have argued that it is not public-sector pay that is so out-of-line, but rather public-sector benefits, such as health insurance and pension contributions. It is true that such benefits comprise a larger share of public employees’ overall compensation than for most private-sector workers. However, even after adding these benefits into the mix, total compensation for Iowa’s male and female public employees are 7.9 percent and 10.8 percent less, respectively, than for their private-sector counterparts. The gap between private and public compensation narrows to 6 percent and 8 percent among male and female state government workers, and 9 percent and nearly 13 percent for male and female local government employees.

I recommend reading the whole report (pdf), which includes lots of supporting data. Pages 3 through 6 discuss the differences between education levels in Iowa’s public sector and private sector.

Research has demonstrated time and time again that educational attainment is the critical predictor of income. Further, there is a wide divergence in educational attainment between Iowa’s public-sector workers and its private-sector workers. […]

he contrast in educational attainment between the two sectors is stark. While a third of private-sector workers have just a high school degree, for example, only 16.8 percent of public-sector workers have this level of education. Over half of Iowa’s state and local government employees have a four-year degree or higher; just a quarter of private-sector workers in Iowa have a four-year degree. […]

Nevertheless, accusations of “overpaid” public employees often neglect this reality and lump all public and all private-sector employees together without regard for educational attainment differences. However, when the earnings between the two sectors are compared across education levels, it is apparent that there is no public sector advantage.

A new report hailed by the Branstad administration contradicts those findings. The human resources consulting firm Aon Hewitt concluded that Iowa is paying substantially “above market” base salaries to many state workers. The Des Moines Register gave the governor the news lead he wanted:

The base pay for 18,500 state government employees in Iowa is 17.9 percent higher than the market pay of comparable employees working for private and public employers, according to a consultant’s report released today.

The study by Aon Hewitt, a nationally known human resources consulting firm, said that 83 percent of Iowa state government positions studied had average base pay that was 20 percent  or higher than comparable states. But when considering the minimum and maximum for base pay, State of Iowa employees are competitive with the market.

The Register’s William Petroski didn’t delve into the details of the report, which you can download here (pdf). I found it difficult to follow, because Aon Hewitt didn’t provide details about which private-sector jobs were deemed “comparable” to various state positions.

I didn’t notice a huge methodological problem flagged by Peter Fisher of the Iowa Policy Project.

First, you lose an apples-to-apples comparison when you compare a mean, or average, with a median. Yet that is what this study does, comparing an average or mean base salary in state government with the median pay in the “market.” Income and wages are usually distributed in such a way that the mean is higher than the median. There is not enough information provided to show those distributions in this study, but the danger is that it creates a built-in bias to make the public-sector wages appear higher than the market. (Suppose seven workers earn $40,000 and three earn $60,000, in both the public and the private sector. The average is $46,000, the median is $40,000 – a wide disparity.)

Aon Hewitt buried the fact that it compared average (mean) state base pay to median private-sector pay in the tables that begin on page 24 of the report. Any competent high school math student knows that when you’re analyzing a group of numbers, the mean and the median can be very different. A small number of higher-paid supervisors can make the average Iowa public employee’s salary seem higher than it is in reality.

Fisher raises other important questions about Aon Hewitt’s study, so I recommend reading his whole post at the Iowa Policy Points blog. Any journalist who will cover the upcoming labor negotiations needs to understand the potential flaws in this analysis.

Meanwhile, consider one of Aon Hewitt’s policy recommendations.

For incumbents who are paid well above the estimated market value, Aon Hewitt does not generally recommend reducing pay. Rather, Aon Hewitt recommends reducing or eliminating future increases to base pay to allow the labor market to “catch up” to State of Iowa compensation levels over time. As a best practice, Aon Hewitt recommends weighing individual performance against market to determine if pay is appropriate, as high-performing employees may warrant pay higher than the estimated market value.

Expect the Branstad administration to take this position into the labor negotiations. State employees will not only be asked to contribute 20 percent toward their health insurance premiums (amounting to a significant pay cut for some), but also to forgo pay raises for who knows how many years until the private sector “catches up” to their wages.

What labor union representative would agree to a contract on those terms? My money’s on the talks going to arbitration sometime in 2013.