Rankings: What they can — and can’t — tell you

Last week, we issued a report through Govistics – a project of CGR – ranking U.S. states by average 2010 state worker salaries. New Jersey and New York topped the list, followed by California, Alaska, Maryland and Connecticut. All had average state worker earnings of over $50,000. Indiana, Missouri, West Virginia and the Dakotas rounded out the bottom of the list, with average salaries of less than $35,000. Of the six top-paying states, all but Connecticut saw an increase in state worker pay from 2009 to 2010, with New York state workers seeing a 3.4% increase in their paychecks. Of the bottom five, all but Indiana saw increases in state worker pay.

As states grapple with rising costs and tighter budgets, states at the top of the list may see the rankings as evidence that rising employee costs are crippling. On the other hand, states at the bottom of the list may feel defensive about the comparison. More than 230 news outlets in 42 states ran pieces on the rankings. And some found the report troubling. In an article in the Plains Daily, Ken Purdy, Classification and Compensation Manager with the State of North Dakota’s Human Resources Management Services—the state with the lowest average state payroll—said:

“An overall comparison of average salary is pretty baseless. It’s not very valid. It depends on the makeup of the workforce, and the employees at various levels. Throwing out a blanket number like that, in a business sense, isn’t very helpful. The true measure in trying to price your employees is a more direct job-to-job comparison of salaries to competing employees.”

The question for us at CGR is this: Should we be issuing straight rankings from our Govistics database without applying an additional analytical “screen”? Or are we contributing to a “dumbing down” of complex statistics, stats that should only be explored in the context of careful review and discussion?

Many take offense at rankings. US News and World Report has turned the ratings game into a business model. There isn’t a college president in the country that doesn’t roll his or her eyes at the latest release of USN&WR’s “top colleges” issue, even schools that score well. While they accept parts of the approach, they also know that much of what makes a college or university desirable and effective can’t be easily reduced to a numbered list.

A simple ranking is a crude indicator—in the case of public sector payrolls, for example, a robust study of variable pay would include more than just gross per person. Certainly cost of living is a factor—a dollar goes much further in Fargo than it does in Manhattan. Non-cash compensation also makes a difference. Pensions vary dramatically from state to state—how many years are required before the benefits are assured, even if the individual leaves public employment (called “vesting”)? What share of final salary is assured after, say, 20 years of service?  “Other than pension employment benefits” (referred to as “OPEB”), particularly health insurance, also influences total compensation. And consider job security—states with strong unions like New York or California have achieved much greater job security than in states with less influential public employee unions. Dollar for dollar, a safer job is worth more than a job that is more subject to the state’s fiscal condition or changing priorities of the state’s leaders.

But you have to start somewhere. State workers in North Dakota (again, the last ranked state) earned $34,000 on average—about 60% of earnings in New Jersey, the first ranked state. If the difference can be explained by differences in labor markets and cost of living—then let’s have that discussion. What is it about New Jersey and New York that have driven up average salaries? Why is it more expensive to live in New York than North Dakota? High salaries for public sector workers don’t emerge from a vacuum. While high average salaries for the public sector drives up cost to taxpayers—a bad thing—high average salaries in the private sector probably reflects an economic base in high value-added industries—a good thing.

So CGR’s simple Govistics rankings—just like the college rankings—can only be a starting point for deeper study and further discussion. A small difference in rank—consider Arizona and Alabama—is certainly meaningless. Big differences between similar states are another matter. But we are reminded that the lion’s share of cost for state government is salaries. We intend to continue to explore the information on the public sector contained in CGR’s Govistics database and to add to it additional information that can inform and empower taxpayers and community leaders.