There’s still time for Congress and the Administration to hit pause on its improvisational restructuring of the federal workforce and invest in long-term state capacity before it’s too late.
In the midst of impending mass layoffs, “voluntary” resignations, and early retirements driven by the Trump Administration, Americans are once again becoming familiar with this chart:
Many are surprised to learn that the federal workforce has remained remarkably stable, fluctuating between 2 and 2.3 million employees since the late 1960s. As a result, the size of the civil service has dramatically decreased in real terms as a percentage of the population—despite the fact that it does a lot more today than it did in the Johnson Administration.
With agencies reportedly reopening the “deferred resignation program” and preparing to implement significant reductions in force across the entire federal workforce, there has also been renewed interest in examining the last major effort to reshape the federal workforce.
Thirty years before Elon Musk declared his intention to cut the federal workforce by 5-10% to reduce spending by hundreds of billions of dollars, Bill Clinton began his first term by assigning Al Gore to lead an initiative aimed at “reinventing government.” This multi-year effort focused on shrinking the federal workforce, improving efficiency, and positioning the government for the post-Cold War era. From 1993 to 1998, the National Performance Review (NPR) and the National Partnership for Reinventing Government (the latter incorporating the former) pursued this vision with notable success—by 2000, the government looked and functioned very differently compared to 1992.
As the Trump Administration has embarked on its own ‘efficiency’ effort, the NPR has been a convenient foil for both Musk’s critics and his defenders. Critics from the left point out that Clinton’s effort proceeded more methodically, involved Congress, and tried to treat federal workers with dignity—all values that the DOGE appears to eschew in favor of working in private and just seeing what breaks. On the other hand, defenders from the right point out that the Clinton Administration successfully reduced the federal workforce by 400,000 people, or about 20% of the civilian workforce, through many of the same mechanisms that the Trump Administration is now employing, such as Reductions in Force or layoffs. The politics motivating these views may differ but both narratives start with the assumption that NPR successfully addressed a major issue: the size of the federal government. However, with 20 years of hindsight, it’s not clear that this was truly the case. In fact, it appears that the Trump Administration, like the Clinton Administration before it, is heading straight into the same trap that the NPR fell into. Both administrations became overly focused on headcount as its measure of success and (in the process) started making many choices that undermine the government’s long-term state capacity.
The Clinton administration had good ideas about reform
In the classic retelling of the NPR, the federal government of 1993 was procedurally bloated and full of middle management that did little to advance agencies’ missions. Writing in 1993, the National Performance Review diagnosed the problem this way: “Is government inherently incompetent? Absolutely not. Are federal agencies filled with incompetent people? No. The problem is much deeper: Washington is filled with organizations designed for an environment that no longer exists,” and argued that “[t]he federal government is filled with good people trapped in bad systems: budget systems, personnel systems, procurement systems, financial management systems, information systems.”
The idea was to improve the government’s ability to deliver by reducing red tape, providing federal employees with more flexibility, partnering with the private sector in places where they have expertise, and moving to more ‘performance-based’ approaches to carrying out the business of government. If this sounds like it rhymes with the work of contemporary critics like Ezra Klein or my colleague Jen Pahlka, that’s because it does. Much of the Clinton Administration’s diagnosis still rings true today. One could imagine this line from the NPR’s report being pulled from today’s New York Times op-ed page in a discussion of the Biden Administration’s rural broadband program: “It is almost as if federal programs were designed not to work. In truth, few are ‘designed’ at all; the legislative process simply churns them out, one after another, year after year. It’s little wonder that when asked if ‘government always manages to mess things up,’ two-thirds of Americans say ‘yes.’”This critique was basically correct in 1993 and it remains so in 2025. One might ask, then: If the federal government figured this out 30 years ago, why are we still tangling with these same issues?
“The era of big government is over”
Underlying the NPR’s argument was a very dim view of the (at the time) 2.1-million-person federal workforce and its role in hindering efficient government. Bob Stone, who directed the effort, expressed its view in stark terms:
“Roughly one of three federal employees had the job of interfering with work of another two. We called them the forces of micromanagement and distrust, and we wanted to reduce the number of inspectors general, controllers, procurement officers and personnel specialists.”
If this were true, the thinking went, it was worth it to try to shrink the size of the federal workforce by shedding these types of jobs. Congress, agreeing with this view, granted the Clinton Administration the authority to do this legally and expeditiously. By the end of the decade, according to their own count, the effort had managed to reduce the federal workforce by 426,000 jobs. Taken at face value, this approach is completely reasonable: if a third of the federal workforce is doing nothing to add to mission effectiveness, and if you could rethink the way programs operated to be successful without them, then reducing the size of the federal workforce by 20% seems like a great first step in the right direction. To their credit, the Administration set about doing just that.
However, political and practical realities asserted themselves. The President, wanting to promise something tangible and measurable to show progress, walked into the Capitol for the 1996 State of the Union and declared that “[t]he era of big government is over” and promised to cut the size of government by several hundred thousand people. In making that commitment, however, that useful measure of transformation (savable headcount) quickly morphed into a target (roles cut) to be managed against when the larger programmtic reforms failed to materialize. From there, the old management heuristic that “when a measure becomes a target it ceases to be a good measure” insisted on its rightness again.
Some of this is Congress’ fault. Leaders on the Hill could get behind the administration’s effort to cull the workforce, but could not uphold their end of the bargain. The personnel cuts that were supposed to be downstream of program reform ended up coming first. As John Kamensky, the Deputy Director of the NPR, wrote in 2020:
“The ‘thoughtful’ cuts did not happen as envisioned. Congress intervened, mandating cuts at a faster pace, without providing the flexibility envisioned by streamlining personnel or acquisition requirements.”
This meant that agencies were left scrambling, trying to figure out how to reduce headcount without the ability to actually de-proceduralize the work. To minimize the damage, agencies were forced to eliminate positions that they felt they could make do without or easily outsource, which in practice meant either eliminating entire functions, trimming large numbers of more-junior, clerical and administrative roles that could be replaced by contractors or were primarily developmental for new hires.
In terms of the process improvements they did manage to make, most of the signature changes to the process were cosmetic at best or operationally-challenging at worst. As Jeff Neal, a former HR leader at DOD and DHS argued, the reforms:
“…gave recruiting authority to agencies, but made them follow a 350-page OPM Handbook.…It eliminated the [Federal Personnel Manual], but kept the complex regulations and eliminated many of the OPM staff who were the real experts who could explain all those regulations. Not to worry – agencies had HR experts who could help. At least they did until the NPR cut them by half, leaving federal HR offices unable to do anything but the most essential work. The initial ideas of the NPR may have been sound, although the original report had a lot of snarky anecdotes. But, at least with respect to federal HR, the NPR was a half-baked set of reforms that broke the mold and did not put something functional in its place.”
By only going half-way and adopting a cut-first approach without a plan to meaningfully transform the entire system, the HR enterprise was simply hobbled rather than reformed.
At the same time, the Administration made several logical but shortsighted human capital management choices. To hit these types of aggressive reduction targets, it used blunt instruments like voluntary buyout offerings (VSIP), early retirement (VERA), and layoffs (RIFs).
As we are currently witnessing in the Trump Administration, these tools are poorly targeted. In the case of voluntary offerings like VERA and VSIP, the best employees (who consider themselves as well-suited to compete for private sector jobs) are often more likely to self-select into the programs. In the case of RIFs, because of the rigid statutorily-mandated procedure required to execute them, they disproportionately led to the most junior and least-tenured staff being laid off. Taken together, this meant that the government walked both its best talent and its future leaders out the door together, all at once, to hit a target.
By the end of the year 2000, the result of these choices on the size and character of the federal workforce was obvious, although the impact wouldn’t be felt for several years: Agencies’ mission support functions were hobbled by cuts that were concentrated on lower-graded staff and that severed the talent pipeline of younger federal employees.
The former effect was intended: between 1992 and 2000, the total number of human resources (HR) staff fell by 24%, including a 43% drop in HR assistants; the number of procurement specialists dipped by 16%, while the number of procurement support staff slid by 59%; the number of budget assistants dropped by 27%; the number of financial management/accounting jobs by 27%; and so on.
The latter effect may have not been intended but was no less transformative. The share of federal workers under the age of 35 shrunk from 26% in 1992 to under 17% in 2000 as the government shed over 300,000 younger workers, while the share over the age of 50 shot up from 25% to over 36%.
Due to the primarily administrative, white-collar nature of government work, the federal workforce had long been somewhat older and greyer than the broader national workforce. However, these reforms supercharged this trend, taking the federal government even further out of sync with the rest of the labor market. By 2000, in many sectors of the government, there were simply no junior people learning the ropes.
Government didn’t actually shrink, it just got harder to see and manage
In retrospect, by focusing on what they could control–namely, headcount reductions– the Clinton team succeeded in making good on the President’s promises but at the expense of long-term state capacity. They succeeded in convincing the public that government’s size was the core issue, yet failed to meaningfully reduce the scope of the government’s responsibilities. In doing so, they constrained the management options available to future administrations.
As the country entered the 21st century, the Clinton-era mantra of “do more with less” became “do the same with more, but hide the headcount.” This approach contributed directly to many of the present challenges we now face with effectively implementing policy.
With agencies under intense political pressure to keep these ‘official’ headcounts steady, over time they turned to contractors to fill in the gaps. This led to an explosion in the ‘blended workforce’ that, by 2005, had more people than it did in 1994 when accounting for contractors and grantees, according to analysis by a team at Brookings. Outlays to contractors shot up, especially at civilian agencies, reaching historic highs year after year after year. By 2023, there were nearly 2.1 contractors for every federal employee. Over time, various analyses have generally found that these contractors are varyingly more expensive than their federal counterparts, meaning that the government is likely paying a premium for this hidden workforce.
All of this also came after the contracting workforce had been decimated in the 1990s, meaning that fewer personnel had to manage more and increasingly complex contracts. The size of the procurement workforce didn’t rebound in nominal terms until 2009 and, when accounting for inflation-adjusted growth in the number of contracting dollars, has never really recovered. In 1992, the government spent $13.8 million (in 2024 dollars) for every procurement professional. In 2024, that number was nearly $17 million. At the same time, their mechanical workload grew significantly: the average contracting specialist processed 622 discrete contract actions in 1992 and by 2024 that number had grown to 2,718. Some of this can be attributed to higher productivity, but it’s also the case that contracts are more complex now than they were 30 years ago.
These challenges are directly upstream of some of the most visible failures in federal program delivery over the past few decades–from Healthcare.gov in 2013 or FAFSA in 2023. But they also manifest daily in less-publicized ways, creating persistent challenges for federal managers across the government.
Contractors, while often essential, may not share the government’s long-term incentive to build internal capacity; in some cases, they may even become embedded in operations in ways that hinder flexibility. Moreover, contract structures can create friction between federal managers and their “blended workforce,” as time and energy are spent negotiating scope and deliverables rather than advancing mission goals.
This is not to suggest that contractors are inherently problematic. On the contrary, when deployed wisely, they can bring valuable expertise and efficiency. But doing so requires skilled, experienced federal management–a capability that weakened significantly after the Clinton-era cuts. Like any muscle, this one atrophied when it wasn’t used.
Not everything can be outsourced. And when the government does hire directly, the lingering effects of the NPR era are still deeply felt. For one, despite staging a funeral for the 10,000 page Federal Personnel Manual in 1993 in the lobby of the OPM Headquarters and firing a quarter of the HR professionals in government, the fundamental hiring process that governed the federal government didn’t change much during the 1990s.
There were some attempts at reform. The Administration experimented with workforce innovations, including converting the U.S. Patent and Trademark Office, the Federal Aviation Commission, and Federal Student Aid into ‘performance based organizations’ that had exemptions from many civil service rules. It also launched personnel demonstration projects like AcqDemo at the DOD.
While these initiatives showed some promise, broader reform has remained out of reach. Many of these pilots have simply lingered in “trial” status for more than two decades. Meanwhile, meaningful updates to the foundational framework–the 1978 Civil Service Reform Act, which still governs most of the federal civil service–has stalled due to Congress’s persistent lack of interest in comprehensive reform.
Additionally, due to the artificial pressure to keep headcounts low amidst growing mission requirements, the cuts to junior roles during the 1990s were never fully reversed. Headcounts returned to ‘normal’ in the late 2000s as the War on Terror and the expansion of the security state elongated the executive branch’s to-do list but the number of early-career federal employees never fully rebounded.In 1992, the median pay grade for federal workers on the General Schedule–which is the main pay table for the majority of federal workers and has been since the 1940s–was GS-09, which roughly corresponds to the entry level for those with Master’s degrees or a bachelor’s degree holder with 2-3 years of experience. By 2000, the median job was a GS-11, which roughly corresponds to the entry level for a doctoral degree holder. In 2024, the median federal employee was a GS-12, which typically requires years of specialized experience on top of a bachelors and graduate degree. For someone just starting out in their career, there simply aren’t many jobs you’re qualified for in government anymore, and the ones that remain get more scarce every year–this problem isn’t unique to government, but it is uniquely rigid in government given the legally-mandated qualification standards from the 1940s and 1950s. Recent efforts to address these challenges via skills-based hiring have shown some promise but will not radically change the shape of the federal workforce to something more sustainable without rethinking the underlying classification system entirely.
This means that while other large employers are re-embracing modern versions of the apprenticeship model, where they train their workforce from novice to expert and into management, the federal government has increasingly closed itself off to early-career talent. Instead, the government continued to fill in only the middle and senior levels of the career ladder, relegating those more junior jobs to contractors, in an effort to manage to the ‘invisible ceiling’ of 2.3 million federal employees that had persisted for decades.
The administration needs to do better this time
With the government hurtling headfirst into another round of poorly-targeted reshaping, there is still time to avert a rerun of these mistakes.
Rather than barreling forward with cuts and hoping that Congress will trim the government’s to-do list down to match, the Administration should adopt a more thoughtful approach that optimizes for good governance instead of chaos and carelessness.
If the DOGE and OMB are serious about their stated goal of imposing fiscal responsibility, they should start with engaging with the source of the government’s spending commitments: Congress. That means working together to make deliberate choices about how much revenue to raise, which programs to scale back or eliminate, and which to retain. Only then should decisions be made about how to align the federal workforce with the effort required to administer what remains. That’s what genuine zero-based budgeting looks like.
Taking the ‘cut first, reengineer later’ approach will lead to the same results we saw in 1996: painful, but largely ineffective. It won’t fundamentally change how the government operates or what it does—it will just leave fewer people trying to do the same work. And when the next crisis hits, we’ll find ourselves scrambling to rebuild capacity. Without real structural reform to federal hiring, we won’t be able to rebuild fast enough for it to matter.
Once cuts are duly approved by Congress, the Administration must take a more strategic, deliberate approach in implementing them.. Early retirement, buyouts, the ‘deferred resignation program’ and layoffs are exactly the wrong types of tools to use if you want to build a better, more merit-based federal workforce.
As in the 1990s, these methods will disproportionately push out the most marketable and youngest employees— either by incentivizing their departure or removing them outright. While the full impact of this reshaping may not be immediately visible, the long-term consequences are already in motion. Inertia may carry the current workforce through the remainder of this term, but the executive branch is, once again, cutting off its pipeline of future leaders—the very people we’ll need to navigate the crises of 2035 and 2045.
Additionally, it’s important to recognize a key distinction between the Clinton-era NPR effort and the current moment: while the Clinton Administration may have held a skeptical view of the federal workforce, it still made an effort to treat public servants with a measure of dignity and respect. Today, there is little evidence of that same regard. Instead, the current approach seems to center on creating a hostile work environment–one that demoralizes civil servants, many of whom have served in the military, and drives them out of public service altogether.
Careers in government are being made increasingly unattractive to the people we need most: public-spirited, high-achieving, middle-class professionals who want to make a difference. Aspiring public servants may now approach a career in the federal government with skepticism, viewing it as being one election away from an arbitrary firing, and they are unsurprisingly hesitant to stake their family’s financial future on such uncertainty. Only the independently wealthy or those with no better prospects will elect to serve, and we will all suffer for it.
This Administration and its allies claim that present ruptures are necessary to change the long-term fiscal course of the nation and deal with rising deficits. This is a plausible argument–radical reform is perhaps needed to secure the fiscal future of our children and grandchildren in the long run. Yawning deficits may endanger that future, particularly as interest rates rise, though it is unclear that their plan will solve the problem. What’s jarring is that the same team’s management approach is the opposite: short-sighted, focused on scoring political wins today at the expense of delivery and organizational health in the future. We should place as much value on our children and grandchildren as future recipients of government services as the Administration places on them as future taxpayers.
No organization that aims to exist indefinitely would operate this way–they instead take the long view. Our government should be managed with the same level of good judgment and foresight. Congress and the Administration should pause their hasty workforce restructuring, carefully reconsider their approach, and execute their agenda with a focus not only on today’s efficiency but also on the impact it will have on future generations of civil servants and citizens.
We cannot afford to fall into the headcount trap again—we’ve been down that road before, and we’re still repairing the damage.