Risher | The IT pay conundrum

In our recently released report, “Managing for Better Performance: Enhancing Federal Performance Management Practices” by IBM Center for the Business of Government, Charles Fay and I set forth a framework of steps to create a results-based, high-performance environment. This is not easy to achieve. It involves changes that are far-reaching. There is obviously resistance, but even opponents know change is necessary.One of the focal issues is pay for performance, which is effectively a universal policy outside of government and a global trend. The concept has been on the table for a long time. It was discussed in the meetings that preceded the Federal Employees Pay Comparability Act and pay reform in 1990. However, it was not then as important as the changes to make federal salaries more competitive. A few years later, I had a conversation with John Sturdivant, then president of the American Federation of Government Employees, who said his union would have to learn to live with merit pay. He saw the writing on the wall.In our IBM report, we chose not to focus on the problems with the federal pay system. At this point, there is a widely shared understanding that the government pay system is broken and that it is an impediment to meaningful change.For the information technology community, the General Schedule pay system is particularly problematic. Its inadequate sensitivity to market trends and governing regulations means that it is badly out of sync with best-practice thinking for managing technology specialists.This is not a new conclusion. In 2001, the CIO Council asked the National Academy of Public Administration (NAPA) to study the prospects for developing new human resources and management policies for IT professionals. That report, “The Transforming Power of Information Technology,” recommended a comprehensive set of human resources practices, including a new pay system that would be market sensitive and performance based. I was the author of the section on pay. The report, unfortunately, was released shortly before the 2001 terrorist attacks, and it was largely forgotten.The problems are, for the most part, widely acknowledged. Commercial salary surveys highlight one of the most obvious problems. The most widely used surveys report data for more than 200 positions and 160 metropolitan areas. Technology labor markets are not as hot as they were a decade ago, but specialized skills such as those associated with machine learning or wireless networking can command a premium in local markets. And those premiums can change rapidly. Any pay system should be sufficiently flexible to stay abreast of trends and help rather than impede recruiting.A market-sensitive pay philosophy also means the specialists with skills that are no longer in demand cannot expect automatic pay increases.I learned in CIO Council meetings that federal agencies often have large systems that in another context would be seen as antiques. The employees who maintain those systems may not be able to find jobs outside of government, and that obviously affects what the market is willing to pay for their skills.Another problem is the way career ladders must be force-fitted into the GS grade structure. The bottom and the top of technology career ladders don’t fit neatly into the standard GS grades. That was one of the reasons NAPA recommended the transition to salary bands aligned with the natural rungs in career ladders. The concept of knowledge workers could have been conceived to refer to technology specialists. Their value and expected contribution depends on what they know and are able to do. As individuals develop their capabilities, they should expect to realize career progress, and that does not mean automatic promotions or step increases.Research from several years ago showed that career ladders for knowledge job families have four natural levels – entry/trainee, developmental, full performance and expert. Salary surveys commonly report pay data for these levels, and bands defined around market pay at each level. Within a band, salaries can be set and managed to reflect the market and the premiums for hot skill jobs.The idea is that individuals can earn salary increases and promotions to higher level bands based on their skills. That philosophy is well-suited to technology careers but at odds with the GS system.The market for the best technology specialists demands they be treated as world-class experts and paid accordingly. That level of expertise justifies special status and pay as high as Senior Executive Service levels. Technology companies often have highly selective policies that make becoming an expert the capstone on a career.Planning the banded structure and determining market rates are straightforward tasks. Actually planning the pay-for-performance policy is also straightforward – the basic model and options are well-documented. There are numerous textbooks that discuss the design issues.The problem that has confounded government agencies is the process or system for managing employee performance. Over the years, I can only recall two or three quotes from leaders of technology companies stating that the best specialists are “four or five times more productive” — at least one statement claimed the gap is significantly larger — than an average employee working in an IT field. The issue is not the accuracy of their estimates but rather the implications for performance management. The stars stand out – that’s true in every occupation. The care and feeding of high performers should be a priority for every employer.The Government Accountability Office may have created a few disgruntled employees by moving to a pay-for-performance system, but it is still viewed as a success story. GAO’s performance system uses the phrase role model to describe specific behaviors that reflect the highest level of performance. More than a few employees are going to work to reach that level. In our report, we emphasized the importance of defining success.The notion of challenging employees and defining success is central to another important issue. Performance expectations should be job-specific. The expertise needed, for example, to plan the human/computer interface is different than that required to create a secure operating environment. Those skills are different compared with those needed in finance or human resources. It takes time for a supervisor to work with his or her staff members to define job-specific expectations, but research shows it is vital to high performance.Performance management is — or should be — more than completing a rating form. No one would contend that task serves much of a purpose. Performance management should be seen as a process that involves an ongoing dialogue between a manager and his or her workforce. They must discuss what is to be accomplished, and then have periodic feedback and coaching. The year-end evaluation should never be a surprise.When most federal employees start their careers, they aspire to accomplish great things. The performance management process should be an important tool for managers who want to support those aspirations. Somehow federal practices too often dampen an employee’s commitment. Surveys have shown repeatedly that federal workers see reward and recognition practices as the least effective of all federal workforce practices, a weakness that diminishes employee commitment.Every employer needs a rationale for increasing salaries over time. Pay for performance is not perfect. The transition is going to hit a few bumps, but it is decidedly better than the anachronism of automatic step increases that satisfy no one. Pay for performance has its critics, but a study by a National Academy of Sciences committee some years ago concluded the advantages surpass any problems. The experience in every demo project proves that federal managers and employees will accept pay for performance.