I have written two posts on the issue of General Schedule reform. They addressed some common fallacies and some facts about the General Schedule and how it is working today. My plan was to wrap up the series with one more post on my recommendations for replacing today’s General Schedule with something that is better suited to the knowledge workforce we now have.
After hearing from several folks about their concerns about pay for performance and why it is a bad idea, I decided to include one more post than I had intended. The next post on this issue will address my recommendations. This one is focused entirely on pay for performance.
What really surprised me is how many smart, well-informed and thoughtful people assume that the only alternative to the General Schedule is pay for performance. I do not believe that is true. The General Schedule is a pay and job classification system. It includes modest measures to deal with pay differences in different parts of the country, a small performance-based component (bonuses and Quality Step Increases), and longevity-based within grade increases to incentivize retention. Retaining employees, pay that recognizes labor costs are not the same in every locale and performance incentives are all good.
Replacing the General Schedule does not have to mean going to a system that is pay for performance (P4P). Here are some of the reasons why I believe that to be true:
- Performance management. I doubt anyone can find an agency where HR, managers, employees and unions will agree the performance appraisal process is working. My four posts on performance (on judging others, neuroscience and ratings, why the government does ratings and what can be done within existing law and regulations) outline the many problems with current appraisal processes. IF, and that is a big if, we could get to the point where we actually believe we know how to evaluate performance effectively, I would support an approach with a larger performance-based pay component. The only way to do that and make certain it works is to implement a substantially improved performance management process, test it for at least 2 or 3 years, assess it, fix problems that are uncovered, try it again, assess it again, and make certain it works. We can use it as the basis for pay if, and only if, it is proven to be effective. If that isn’t done, we are making decisions that affect employees for years to come based on unproven systems.
- Retention. It is common to see P4P systems that include no pay increases for longevity. Many proponents of P4P argue that pay raises should be based entirely on performance and longevity-based increases such as the current within grade increases are bad. If hiring people had no cost, training new employees had no cost, and vacancies had no cost, I might agree with them. But – they all have a cost. The lost productivity of vacancies affects agency performance and morale. Hiring new employees can be very costly. Training new employees is also expensive, both in terms of dollars and productivity. Given all of the costs associated with turnover, incentivizing retention is a good thing. Incentives that vary from year to year based on performance negate much of the benefit. A GS replacement should include, but not be limited to, incentives to encourage employees to stay with the government.
- Unintentional outcomes. My experience with P4P in government caused me to have concerns that P4P systems that are not fully tested prior to being used for pay decisions may have unintended adverse impact on some groups of employees. In 2008, a Federal Times article on DoD’s National Security Personnel System (NSPS) raised concerns about race, ethnicity and gender inequities in the system. Some other P4P systems have also had similar issues. Although those are the more commonly discussed issues, some unintended consequences are the result of types of jobs. For example, a pay system designed for an organization primarily comprising auditors, attorneys, or medical professionals may not recognize the contributions of other staff. Bias issues might actually be easier to deal with if they appeared to be the result of intentional discrimination. My experience in DoD was that the issues were not evident during the appraisal process or in pay pool deliberations, but surfaced when large amounts of data across multiple pay pools were collected and analyzed. Many pay pool managers I talked with were surprised and troubled by the results. They clearly had no intent to discriminate, but the numbers showed outcomes that no one intended. Some people (including the Government Accountability Office) recommended DoD implement a third-party review of NSPS ratings prior to them becoming final. DoD objected, citing grievance procedures and other protections. There is another reason that is even more fundamental. In a large-scale P4P system that includes third party review, how do you correct problems that are uncovered during a rating cycle? Redo everyone’s ratings? Overrule the managers ratings to correct disparities? Short of a complete do-over, correcting a flawed rating and pay process after the fact is incredibly hard to do. As I said above, the only realistic solution would be a tested and throughly validated rating process that works the first time. That isn’t impossible, but it is not easy and getting it right the first time is not the norm.
I believe performance should be a factor in pay decisions. So should retention and differences in the cost of labor throughout the country. I also believe a pure P4P system is too hard to implement quickly or fairly, and is not worth the numerous problems it would create. In my next post I will outline how we can reform the General Schedule in a way that meets all three needs, but does not rely on P4P for all pay decisions.