Controlling ratings inflation is a good idea...but a forced distribution isn’t

Performance MUST matter

According to Government Executive’s reporting, the Office of Personnel Management is currently circulating draft regulations that would (a) eliminate the “minimally successful” performance rating for virtually all rank-and-file federal civil servants, and (b) limit the number of top ratings—that is, a rating of 4 or 5 on a potential 5-point scale, or a rating of 4 if there’s no longer a “minimally successful” option—for those same feds, the latter by imposing limits on the number of such ratings that can be issued each year.

I want to focus on the latter; that is, controls on top performance ratings. As a career member of the Senior Executive Service for 20-plus years—I’ve also been Chief Human Capital Officer or equivalent at three agencies employing well over 1.3 million civil servants and was once in charge of Human Resources strategy at OPM—I applaud that goal. It’s critically necessary and a long time coming. Indeed, one wonders why it took OPM so long to address this, but that’s another matter. However, if the goal is to make performance really matter, those ratings controls cannot be arbitrary, and thus, I have some reservations about how they are to be imposed across government.

First, let me say the unsayable: That ratings inflation has been a perennial problem for the U.S. government. Indeed, it’s a problem for all organizations, but while some have solved it, few of my peers want to talk about it on the record. However, I’m encouraged, indeed ecstatic(!), that OPM Director Scott Kupor and his team are not afraid to. We do NOT live in Lake Woebegone, where every federal civil servant is above average. Our “intake system” (that is, the way we hire those civil servants) is just not that precise. It cannot be. That’s what a probationary period is for—and gist for another commentary—but realistically, the vast majority of our employees are NOT above average.

Lake Woebegone they aren’t...but civil servants aren’t chattel either! 

For years it’s been too easy to just give someone a too-high performance rating, NOT because they deserved it (those few truly outstanding performers do, by the way), but rather just to avoid the candid-but-difficult conversations that would occur if their annual performance were just “average” compared to their peers, or even below.  In other words, high ratings were too often given just to mollify subordinates and make them feel good...especially since their inflated ratings haven’t really meant anything for years. 

That may change with another reform that Kupor and company have proposed: that is, changing the RIF rules to give more weight to performance ratings in retention—another positive step that should and could have been taken years ago—by placing far more emphasis on one’s performance rating, vs. one’s veteran status or seniority.

But without imposing some limit, some brake, on the number of top performance ratings, those ratings will continue to be grossly inflated, and everything stemming from those ratings will remain meaningless. Or distorted. And the federal government’s supervisors and managers, those relative few who are responsible for the achieving their agency’s mission, thus must look for other ways (some extralegal) to hold their employees accountable. 

However, as noted, that’s easier said than done, especially given the archaic rules dealing with adverse actions for performance or conduct. It is simply too hard today to get rid of those who are demonstrably performing or behaving below standard. And I’m NOT talking about doing so rashly or precipitously or without good cause or evidence, nor am I talking about those new civil servants who may just not be a good “fit” for their jobs, because of a skills mismatch, etc. Rather, I’m talking about those who just don’t perform, and everyone (including their co-workers) knows who they are.

And making most civil servants “at will” is not the answer either. At least not anymore, as that term has been so overused as to become meaningless in and of itself. Traditionally, that term means just that...that someone can be fired “at will” by his or her boss, suddenly and for any (or no) reason whatsoever. But I’ve not met anyone in the Trump OPM that argues for a civil servant’s termination for arbitrary and capricious reasons; that is, for no reason and with no notice whatsoever. Rather, they all say “of course” when I suggest that everyone (not just civil servants) who’s fired from a job deserves to know why, and they also deserve the right to challenge those reasons with evidence.  

But that’s not the issue. Indeed, everyone I talk to says due process is fine, just so long as there are NO interminable external appeals (something that’s never been part of that due process in any case), except under the narrowest of circumstances. But today, it’s the years-long, overly legalistic, lawyered-up, forum-shopped, defensed-to-death appeals to third parties like MSPB, EEOC and the courts, appeals that can sometimes take a decade or more to adjudicate. That’s the issue. That’s what those officials (and I) would get rid of. 

Our own experience undermines arbitrary controls

But we’re talking about OPM’s proposed controls on the number of top performance ratings that frontline employees can receive. Trying to bring some realism—and hence, some meaning—to the annual performance evaluation process (and everything that that rating drives) is a worthy goal. And while clearly making those top ratings more realistic is laudable, one cannot do so in a way that demoralizes those affected by the controls. 

That has been the experience in the U.S. military, which experimented several decades ago with forced ratings limitations, by literally placing an arbitrary cap on the number of “top block” performance ratings that could be given in any given rating cycle. 

But what happens when everyone in a military unit, like groups of test pilots or astronauts or Green Berets or SEALs, already “walks on water?”  That was the dilemma faced by our military when confronted with the same ratings inflation problem that the federal government faces today, and they abandoned it after a couple of ratings cycles, in part because they risked demoralizing their very best troops by imposing an arbitrary limit on the number that could be realistically rated as outstanding. And the same would be true for today’s civil servants. 

Bottom line: We simply cannot risk demoralizing our very best federal employees by telling them that their performance notwithstanding, their annual performance rating won’t (and can’t) reflect it because of some arbitrary, externally imposed limit. That, or those top ratings will be rotated. They’ll be told “It’s not your turn this year, so just be patient.” The point is that while forced distributions and ratings quotas will quickly change the numbers and are thus appealing at first blush, they will not improve overall agency performance, and THAT should be the goal. 

Should we then just throw our hands up and surrender to the inevitable...that is, just live with a world in which almost every federal civil servant is wonderful and their annual appraisals are thus meaningless (unless you’re less than wonderful, of course)? No! There are plenty of viable alternatives, and they’re not very complicated. 

Give the ‘best and brightest’ the performance rating they deserve

One of them is to focus on agency performance first, above all else, and let it drive everything, regardless of what it shows. That agency performance is exactly what we tried to evaluate when I was Associate Director of OPM, under the tutelage of then-Director Kay Coles James, and with the willing help of career executives like myself, we led the effort to make ALL performance ratings more realistic by clearly linking them to the actual performance of their departments and agencies.   

Who evaluated that overall agency performance? In our case, we asked the President’s Management Council, comprised (then as now) of cabinet department deputy secretaries and executive agency equivalents, to do so. As a practical matter, the PMC’s members serve as the Chief Operating Officers of their respective organizations, and as such, we asked them to make those judgments based on the agency-wide performance measures required by the Government Performance and Results Act , then almost a decade old and now even older.

Another option is to rank—not judge but rank—everything and everybody in government by performance against those same metrics...first agencies, then their units, and then employees within those units. First by the PMC and then by agency heads or their Performance Review Boards. And if any of them want to exempt an agency or a unit or a group of employees for any reason (because they’re all good. Or because they tried hard but failed in the face of undue external interference. Or insufficient funding, whatever the reason), let them make the case to their peers. 

And what about DOD and its Laboratory and Acquisition Demonstration Projects, as well as its other pay-for-performance arrangements across government, now covering many tens of thousands of employees? Those now-permanent experiments have figured out a way to limit top ratings by using “crowd-sourced” evaluations to achieve a semblance of objectivity, in part because they’ve tied them to budget-limited pay raises and bonuses. What about them? At the very least, OPM’s new rules need to exempt them...and anyone else who has figured out how to achieve realistic ratings.

Bottom line: There are plenty of ways to do make performance matter WITHOUT imposing arbitrary caps AND still keep ratings inflation under control.    

What’s critical is that an agency’s top-to-bottom performance ratings should be generally consistent with that agency’s overall performance. Thus, if an agency met or exceeded its GPRA performance measures, one would expect its overall (and/or unit-wide) performance ratings distribution to be high...and vice versa. With exceptions permitted by those making those judgements. But that’s exactly what those making those judgments are appointed to do.   

Realistic performance ratings are worth pursuing, but easier said than done

Note in that regard that those judgements, whether it’s by GPRA measures or by ranking or something else, are not formulaic, nor should they be—after all, what happens if say, the Congress interferes in a way that undermines an agency’s performance? Was that a good enough excuse to justify high individual ratings, despite substandard overall agency performance? Perhaps, but that’s for “the judges” to decide.  

After all, an agency’s overall performance is a team effort, the result of political appointees, career executives, frontline supervisors and employees, and (yes) even unions working together, so that means that all of them need to be evaluated on agency performance, just as it should be!  

To be sure, these are all difficult issues, But as I said, the underlying premise—that is, the performance of government agencies, and the results they achieve—is still what it’s all about, right? And controlling rampant ratings inflation to reveal that bottom line can and should be addressed. That means that however it’s done, performance ratings realism remains a laudable and achievable goal, one worth striving for. So, have at it, OPM!

Ron Sanders is a fellow of the National Academy of Public Administration and was a federal civil servant for almost 40 years, including over 20 as a member of the Senior Executive Service. In that capacity, he served as director of civilian personnel for the Defense Department, chief human resources officer for IRS, associate director for HR strategy at OPM and associate director of National Intelligence for human capital, as well as the chairman of the Federal Salary Council.