Elon Musk and his team of software engineers at the Department of Government Efficiency are rampaging through the government in search of fraud, waste, and abuse. The government, they claim, is misusing taxpayer money. They are willing to shut down entire agencies to prove their point. By treating the civil service as a business, as well as importing Musk’s ruthless managerial style into public agencies, the Silicon Valley disrupters believe that the country can produce a leaner and more efficient government.

Put aside for now the fact that these actions are likely illegal and unconstitutional. The central problem is the false assumption of this approach: that problems of inefficiency in federal agencies stem exclusively from the public administration. Much of the waste, inefficiency, and indeed fraud result from the government’s overreliance on private-sector organizations to conduct its work. This byzantine system of outsourcing to nonprofit and for-profit organizations adds costs and creates waste because each nongovernmental contractor exacts fees or imposes a profit margin. In addition, government officials have in recent years exposed multiple cases of embezzlement by USAID contractors, leading to millions of dollars in fines paid out by major companies and their executives.

If the country is serious about improving government efficiency, the emphasis should be on curtailing contracting and making the civil service more accountable—to American voters and to their elected representatives, not to freelance tech bros.

[Peter Wehner: The cruel attack on USAID]

The contracting swamp we have today emerged through previous attempts to streamline government by cutting the federal workforce. All such efforts ended the same way: a modest number of layoffs among public workers combined with a dramatic expansion of private-sector contracting. In fact, our postwar history shows that federal spending has ballooned while the size of the federal workforce has remained almost static: In 1946, the government employed about 2.5 million workers and ran a budget of $628 billion (in today’s dollars); in 2023, it had only half a million more workers, but its budget stood at $4.6 trillion.

How does the government disburse this colossal budget? Through contracting. A 2017 tally estimated that more than 5.2 million contract and grant employees worked for the federal government. On the largely unexamined assumption that the private sector is more efficient, past administrations of both parties have eagerly embraced contracting to make up for limited administrative capacity.

President after president has sought to cut the federal workforce—without reducing the responsibilities and demands on government agencies. That has meant ever greater reliance on contracting. President Ronald Reagan—who famously asserted that government was the problem, not the solution, to the country’s ills—oversaw shrinking the federal civil service by more than 130,000 employees in his first two years in office. Meanwhile, an equivalent number of civil servants was working solely on contract and grant administration. And over Reagan’s first term, spending on contract consultants increased from $1.1 billion to more than $1.5 billion. The Clinton administration echoed Reagan’s rhetoric and sought to win favor for the Democrats with a “reinventing government” program that cut the federal workforce by nearly 400,000 direct hires—even as it expanded the shadow government of contractor workers by nearly 300,000.

USAID, the foreign-aid agency that Musk has been busy feeding “into the wood chipper,” offers a prime example of how extensively privatized so much government work already is. The federal government first made overseas economic development a significant feature of U.S. foreign policy in the years following World War II. From foreign aid’s inception, the government contracted out much of this activity, largely because, after wartime state involvement in industry, a strident anti-statism took hold in Congress. As Cold War anti-Soviet sentiment rose, conservative politicians inveighed against the New Deal federal bureaucracy as evidence of a collectivist power grab. Congress responded by capping the growth of the civil service, but that did not limit the government’s scope or aspirations; foreign aid became a particular Cold War priority as an exercise of soft power. Expansive government aims instead swelled the number of private-sector contractors.

“The entire effort that the government agency carries out here is really carried out through private organizations,” Secretary of State Dean Acheson said in 1952, referring to the forerunner of USAID. “We do not have in the Government sufficient people to staff these operations, sufficient people to give us all the ideas, to give us all the working groups which are necessary.” So private actors would be drafted in to do the work.

Subsequent changes to foreign aid reinforced contracting as the core of U.S. foreign policy, including in the legislative text that originally established USAID in 1961, and in amendments a year later that mandated USAID to “utilize wherever practicable the services of United States private enterprise,” such as “the services of experts and consultants.” As early as 1964, USAID had committed more than $400 million through some 1,200 contracts for technical assistance from universities, for-profit firms, and nonprofits.

This contracting meant that most foreign-aid money went straight to domestic organizations. William Gaud, who led USAID during the Johnson administration, was blunt about this. “The biggest simple misconception about the foreign-aid program is that we send money abroad,” he said in a 1968 speech. “We don’t.” At the time, about 95 percent of the agency’s $1 billion–plus budget was spent directly in the United States.

USAID continued to be an important tool of U.S. foreign policy after the end of the Cold War, overseeing many of the “shock therapy” programs designed to help the former Soviet Union and Eastern Bloc countries transition to Western-style capitalism. But, as an early experiment in Bill Clinton’s “reinvention laboratory,” the agency suffered from the growing reliance on outsourcing. A 1993 Government Accountability Office report noted that expanding USAID work into newly independent countries “further burdened its operating expense budget, resulting in greater dependence on contractors and a greater potential that programs will be vulnerable to fraud, waste, or abuse for lack of adequate oversight.”

The way the contracting system favored U.S. firms was all too apparent. In the shock-therapy programs, U.S. aid was “composed for the most part of financial intangibles and technical assistance,” one U.S. diplomat complained. “The result is that very few Russians have seen anything at all of the vaunted billions … most simply do not believe the money ever existed.” Rather, money flowed from the U.S. government directly to “domestic contractors,” the diplomat added. For obvious reasons, this undermined U.S. foreign-policy goals, generating more frustration and cynicism than gratitude.

By 2001, more than 80 percent of USAID contracts went to domestic for-profit and nonprofit companies. Although Musk has singled out nonprofits that receive grants for projects he finds risible, in practice USAID has long given its most lucrative and important contracts to American corporations. By 1996, consulting and accounting firms such as Abt Associates, Booz Allen Hamilton, KPMG, and Chemonics International all held tens of millions of dollars’ worth of contracts. For such companies, this government spending was virtually their entire business: When Chemonics, for example, reported $85 million in revenue in the last nine months of 1999, 90 percent of its business came from USAID contracts. In 2009, Senator Patrick Leahy of Vermont, a Democrat, lamented that the agency had been reduced to a “check-writing agency.”

At least government-transparency mechanisms made the problems of over-relying on private-sector partners visible and public. But even that degree of accountability is disappearing in Musk’s bonfire of the agencies. USAID’s database of past research projects, for instance, provided scholars like me with valuable insights into foreign aid—and its flaws. Now that resource has been yanked offline. Likewise, inspectors general and the GAO produced reports that highlighted wasteful spending and outright fraud. Recent federal investigations have turned up staggering cases of graft and corruption. Contractors were indicted for a range of fraudulent charges while doing reconstruction and humanitarian work in Iraq and Afghanistan, including classic bribery and wire-fraud cases and engagement in illegal activities to help other contractors secure contracts. As recently as 2023, Booz Allen Hamilton agreed to pay $377 million—one of the largest procurement-fraud settlements in the country’s history—because the for-profit company was caught overbilling the government to cover its private-sector losses.

[David A. Graham: The world’s most powerful unelected bureaucrat]

Silicon Valley tech moguls were not necessary to root out waste; the government was already doing that. But last month, the Trump administration took a hammer to such oversight when the president fired more than a dozen agency watchdogs. Congress could certainly do a better job of oversight, with a greater emphasis on exposing corruption and conflicts of interest. It could also empower inspectors general as its enforcers against graft and waste. One obstacle to smarter reform is that the person charged with overhauling the government has himself been a huge beneficiary of its largesse. Musk’s companies have enjoyed more than $15 billion in contracts over the past decade (including some from USAID), a spigot that continues to pour hundreds of millions of dollars into his pocket every year.  

The contracting model has bloated government costs while hindering public accountability and insulating policy making from citizens. Rather than inviting billionaires to demonize the civil service, the true solution to government inefficiency is to reverse outsourcing. Severing private contractors and hiring more public servants would allow the state to cut time spent managing contractors and devote more resources to the actual work of government. Relaxing hiring regulations could help agencies hire bright and enthusiastic talent, as could raising pay for federal employees. Instead of dismantling the federal workforce, we should put the contracting system in the wood chipper.

The lawsuit filed last week to halt the Trump administration’s dismantling of the U.S. Agency for International Development stands on a bedrock constitutional principle: “Congress, not the President or the U.S. Constitution, creates and organizes the offices and departments” of the government—as a 2017 Heritage Foundation report accurately stated.

Good-faith arguments exist both for and against America having an independent USAID, or—to name another Donald Trump target—a stand-alone federal Department of Education. Over the decades, Congress has changed its mind about both. Constitutionally, however, that’s the point: The decision is up to Congress. Unilateral moves to dismantle USAID, to mothball the Consumer Financial Protection Bureau, or, if Trump’s advisers have their way, to disassemble the Education Department are beyond the president’s constitutional authority.

Since the Kennedy administration, foreign-assistance functions have been lodged in different agency homes. With authority granted him by the Foreign Assistance Act of 1961, President John F. Kennedy established USAID as a division of the State Department. Using powers delegated to him by statutes enacted in 1979, President Jimmy Carter moved USAID’s functions to the United States International Development Cooperation Agency. In 1998, Congress gave President Bill Clinton authority to either return USAID to the State Department or allow it to become an independent establishment within the executive branch; Clinton did the latter. Although presidential judgment thus informed the shape of USAID at every stage of its evolution, everything that presidents pre-Trump did with regard to the structure of USAID or the allocation of its functions was done pursuant to laws that Congress had enacted. No president asserted authority independent of Congress to create, reshape, or eliminate USAID.

This history reflects the Framers’ decision to give Congress, not the president, the authority to generate the executive-organization chart. The Constitution’s executive-branch charter, Article II, envisions what we now call the federal bureaucracy. The president is given explicit authority to “require the opinion, in writing, of the principal officer in each of the executive departments, upon any subject relating to the duties of their respective offices.” But Article II says nothing else about those “departments.” Instead, Article I of the Constitution, the charter for the legislative branch, assigns to Congress the responsibility to “make all laws which shall be necessary and proper for carrying into execution … all … powers vested by this Constitution in the government of the United States, or in any department or officer thereof.” The president’s job is to faithfully execute the law, but law—including law that establishes and structures executive offices and agencies—gets made by Congress.

[Read: The other fear of the founders]

Since the very first Congress, the legislative branch has jealously guarded its power over organization. When the first House bill creating the Department of Foreign Affairs was introduced in the Senate, Senator William Maclay of Pennsylvania suggested that the organization of the executive branch might be left to the president, as the holder of executive power. His scheme would have given to the president the power of a British monarch to create offices. The Senate rejected his position, and the First Congress enacted a round of statutes organizing the new departments—Foreign Affairs, War, and Treasury. The statutory duties of the secretaries heading Foreign Affairs and War were largely to carry out presidential instructions; Congress recognized that Article II envisioned significant discretionary roles in foreign and military affairs for the president. The Treasury, however, was organized in detail. Not only did Congress assign the Treasury Secretary a significant number of specific legal duties, but it also created additional offices within the department—all requiring Senate advice and consent. These additional offices, as explained by the administrative-law scholar Jerry L. Mashaw, “were meant to provide checks on the Secretary and each other in the crucial matter of safeguarding the integrity of the fiscal and monetary affairs of the nation.” Congress went on to create a variety of other agencies, including the Mint, the Post Office, a Customs Service, and a national bank, tailoring the structure of each according to its sense of how best to fit structure to mission. No one doubted that this was Congress’s prerogative to decide.

Supreme Court jurisprudence recognized Congress’s role. In Myers v. United States, the 1926 Supreme Court decision most protective of broad presidential power over administration, Chief Justice (and former president) William Howard Taft acknowledged: “To Congress under its legislative power is given the establishment of offices, the determination of their functions and jurisdiction, the prescribing of reasonable and relevant qualifications and rules of eligibility of appointees, and the fixing of the term for which they are to be appointed.” This proposition has never been open to serious question.

Congress has recognized, of course, that presidents may have valuable ideas regarding administrative organization. Beginning in 1939, Congress enacted a series of so-called Reorganization Acts, which gave presidents significant (but not unlimited) discretion to create, abolish, or restructure administrative agencies, subject to an important caveat. Presidential reorganization plans were subject to a “legislative veto”—that is, a resolution disapproving the plan enacted by both Houses of Congress, which could keep it from going into effect. This would be a concurrent resolution of the House and the Senate that the president could not veto and did not have to sign in order to make it binding. Through the threat of legislative vetoes, Congress kept control over what got created, abolished, or restructured.

In 1983, however, the Supreme Court held that legislative vetoes were an unconstitutional form of legislation. As a result, Congress took away presidential authority to implement reorganizations unilaterally. If presidential reorganization plans could not easily be blocked, Congress would no longer authorize them. Since 1984, presidents have been allowed only to propose reorganizations, which Congress could enact or reject through the ordinary legislative process. (A suggestion in 2023 by Vivek Ramaswamy that a 1977 Reorganization Act continues to empower presidents to abolish agencies despite the statutory changes Congress enacted in 1984 is an appallingly fanciful statutory interpretation.)

[Read: The Constitutional crisis is here]

In light of this legal background, the question is why Trump thinks a president can legally disassemble agencies on his own—assuming, that is, that he cares if it is legal. The likely answer would involve an especially ambitious version of an Article II interpretation called the “unitary executive theory.” The baseline premise of the unitary executive theory is that Article II guarantees presidents complete removal authority over every subordinate member of the executive branch. Bolder versions contend that he or she can also directly command how every function of the executive branch be performed—or even perform them personally.

The Supreme Court has never fully embraced the unitary executive theory. However, a broad reading of the Myers decision mentioned earlier—a reading the Court unanimously rejected seven years later—would invalidate any attempt by Congress to create independent administrators protected from presidential at-will removal. The Roberts Court has gone nearly all in on the broad reading of Myers, treating Humphrey’s Executor v. U.S., the 1935 opinion upholding the Federal Trade Commission, as a mere exception to Myers. (In the intervening decades, the Supreme Court had repeatedly reaffirmed Humphrey’s Executor as the controlling authority, most famously in its 1988 decision upholding the constitutionality of post-Watergate independent counsels.) As a result, the constitutionality of agency structures such as the Federal Trade Commission and the National Labor Relations Board now hangs by a thread; the Court could conceivably uphold the firing of the NLRB member Gwynne Wilcox.

Of course, even a presidential power to fire an individual agency head would not necessarily translate into authority to shut down entire government departments. However, in its 2024 opinion granting former presidents all-but-blanket immunity from prosecution for crimes committed while in office, the Court seemed to signal something far more ominous. The majority described the president’s authority to supervise the executive branch as a power that Congress may not touch—a conclusion that flies in the face of constitutional text. As explained by the Harvard law professor Jack Goldsmith, who had headed the Justice Department’s Office of Legal Counsel during part of George W. Bush’s second administration: “The ruling about the exclusivity of presidential enforcement discretion, especially vis-à-vis Congress, is entirely novel … And it has potentially massive implications, depending on its scope.” What the opinion now apparently implies to Trump is that the president, constitutionally speaking, is the entirety of the executive branch, and he can configure it however he wants.

That said, Trump’s record of legal success in the Supreme Court is a mixed one. But he presumably thinks it a good bet either that the legal challenges to his scorched-earth tactics will be too slow to stop him or that, if they reach the Supreme Court, that body’s right-wing supermajority will continue to improvise on behalf of de facto executive supremacy. Eyeing the latter possibility, the newly confirmed Office of Management and Budget Director Russell Vought has affirmed the administration’s position that Congress lacks authority to force the spending of appropriated funds—a position the Supreme Court has never endorsed, and which is constitutionally unfounded. But a majority that would proceed as vigorously and creatively as it did to protect Trump from prosecution might be willing to improvise some more.

[Read: Trump signals he might ignore the courts]

A government agency’s structure and location are not just abstract; they matter to the work the agency does on the ground. When Congress extracted a Department of Education from what was formerly the Department of Health, Education, and Welfare, it was to give federal support for education greater emphasis. When Congress moved the Coast Guard from Transportation to Homeland Security, it was presumably to prioritize the Coast Guard’s role in security rather than safety. The reason proposals to merge the Bureau of Land Management and the U.S. Forest Service have always failed is that the organizational DNA of the Interior Department, which houses BLM, favors conservation, whereas the reflexive policy mood of the Agriculture Department, which owns the Forest Service, is pro-development.

Perhaps the most worrying development is that the administration’s commitment to obeying court orders may not prove any more reliable than its dedication to following statutes. On Sunday, with a soupçon of Trumpian deniability in his precise wording, Vice President J. D. Vance posted on X: “Judges aren’t allowed to control the executive’s legitimate power.” Taken literally, Vance’s statement is accurate; what it fails to acknowledge is that the judicial power includes authority to state just how far the executive’s legitimate power extends. In rejecting President Richard Nixon’s claim of entitlement to withhold the Watergate tapes, the Court held in a unanimous opinion: “Many decisions of this Court … have unequivocally reaffirmed the [1803] holding of Marbury v. Madison that ‘[i]t is emphatically the province and duty of the judicial department to say what the law is.’” Should Trump ignore any court order to halt his demolition of the executive branch, he will have dismantled not just an agency, but the Constitution itself.

Elon Musk has brazenly dismantled government agencies because he can feel assured of his insulation from the law. By the end of Donald Trump’s presidency, he may well receive a pardon. That’s what many recent pardons (Paul Manafort, the Biden clan, the January 6 insurrectionists) suggest: Presidential loyalists and family members are, in effect, immune from prosecution. On the most disturbing scale, they have become like diplomats who can park wherever they want.

The dawn of this age of impunity can be dated to any number of administrations. In his new book, The Pardon, Jeffrey Toobin makes a compelling case that a primary culprit is the 38th president, Gerald Ford. Toobin’s thesis is brashly revisionist; Ford’s pardon of Richard Nixon has gone down in history as a great act of beneficence. According to conventional wisdom, by immunizing Nixon from prosecution, Ford short-circuited years of polarizing legal proceedings against the former president that would have torn the nation asunder. But Toobin argues that this overpraised act of catharsis established a precedent of lawlessness. The road to Trump begins, in some moral sense, with the absolution of Nixon.

At a glance, the amiable Ford, a college football star and World War II veteran, seems impossible to villainize. Compared with Trump or Nixon, he was the picture of humble decency. On the day he became president, he lumbered out of his suburban-Virginia house in a bathrobe to pick up the paper. In the White House, he toasted his own English muffins. He told dad jokes, played in celebrity golf tournaments, and had a reputation for basically wanting to do the ethical thing.

Having stumbled into the Nixon presidency, as the replacement for the venal vice president Spiro Agnew, he stumbled into the presidency after Watergate. As Chevy Chase portrayed him on Saturday Night Live, dooming him in popular memory, he was always stumbling. The shtick drew on President Lyndon B. Johnson’s famous aperçu, “Jerry Ford is so dumb that he can’t fart and chew gum at the same time.” (Johnson also declared, “There’s nothing wrong with Jerry Ford, except that he played football too long without his helmet.”)

As the Watergate scandal unfolded, Ford made it his mission to learn as little about it as possible. He defended Nixon in the vaguest terms, and essentially ran in the other direction when Nixon asked him to examine evidence in the scandal. Ford stubbornly, and somewhat inexplicably, refused to prepare for the possibility that he might become president. He had initially accepted the vice presidency in the hope that it would be a capstone to his long political career. Indeed, that was the reason Nixon picked him: He knew that Ford had so little appetite for the big job, and so little political guile, that he was unlikely to conspire to oust him.

[Read: Trump’s pardons are sending a crystal-clear message]

In the days leading up to his ignominious departure, Nixon hatched a very Nixonian plot to exploit Ford’s goodwill and naivete. He wanted to pressure the future president into pardoning him without ever making a direct ask—a strategy he conceived with the White House counsel Fred Buzhardt, under the cover of attorney-client privilege.

On August 1, 1974, Nixon told Alexander Haig, his chief of staff, that he wanted him to begin preparing Ford to assume the job. “Tell him what’s coming,” he instructed. Nixon knew that Haig would check in with Buzhardt before sitting down with Ford. This was the twist in his scheme: Buzhardt had prepared a memo for Haig, listing six “endgame” scenarios for Ford to consider. In classic Washington style, he arrayed the possibilities so that every plan entailed a messy, prolonged handoff except for one: “Nixon could resign and then Ford could pardon him.” This was the elegant solution, but it had the whiff of corrupt horse-trading.

The pardon wasn’t something that Ford had ever considered, so he peppered Haig with questions about it. Although they didn’t agree to anything in the course of conversation, Ford’s interest had been ignited. He came to believe that a pardon genuinely served his own interests. When he finally assumed the job, he wanted to be more than a pleasant placeholder, and he could never be his own man without first disposing of the looming presence of Richard Nixon.

And so Ford talked himself into the pardon. He read a 1915 Supreme Court decision, which ruled that the acceptance of a presidential pardon is tantamount to admitting guilt, and convinced himself that the public would accept that legal logic. He would tell aides that he felt sorry for poor old Nixon, who he worried was in physical decline.

Ford pushed the process forward without really debating the merits of a pardon with his staff. His poorly argued, nervously delivered speech announcing the decision to the nation was so rushed that aides didn’t have time to prepare a teleprompter. Ford barely gave congressional leaders a heads-up, and none of them could quite grasp his reasons for haste. Tip O’Neill, the majority leader in the House, asked Ford, “Then why the hell are you doing it?” He posed that question minutes before Ford went on national television.

In the most outrageous passage of the speech, Ford declared the fate of Richard Nixon “an American tragedy in which we all played a part.” The public, having been accused of complicity, took its revenge. In a single week, Ford’s popularity plummeted 21 percentage points. His party suffered catastrophic collapse in that year’s midterm election.

[Jeffrey Crouch: O]ur Founders didn’t intend for pardons to work like this

With the benefit of time, however, Washington revised its opinion of the decision. Bob Woodward, of all people, eventually concluded, “Ford was wise to act. What at first and for many years looked like a decision to protect Nixon was instead designed to protect the nation.” Ford slowly remerged with the reputation of a healer, a man of grace.

That revisionism is nostalgic gloss. Toobin makes a damning, nuanced case against Ford. Nixon had, at that point, committed the worst crimes in the history of the presidency, vividly and irrefutably captured on tape, and he escaped without any punishment. He received absolution without displaying remorse. “The pardon was just a free pass handed from one powerful man to another,” Toobin writes.

Despite his earnest desire to undo Nixon’s legacy, Ford’s pardon was itself an assertion of the imperial presidency. That’s because the pardon is an inherently Caesarean implement. In every other facet of the American system, carefully installed safeguards are designed to limit the authoritarian exercise of power. But there is no curb on the pardon other than the conscience of the executive issuing one. Presidents tend to tacitly admit that they are misusing this authority when they sheepishly hoard pardons for the final hours of their administration, waiting for the moment when there’s no political price to pay and hoping that their shabby behavior is drowned out by the inaugural hoopla.

By absolving his former patron, Ford helped create a new Washington ritual: the moment when presidents release their cronies, friends, and family from the bonds of justice. George H. W. Bush sprinkled his magic forgiveness dust over Casper Weinberger, Robert McFarlane, and Elliot Abrams, among others, letting them off the hook for the Iran-Contra scandal. Bill Clinton bailed out the financier Marc Rich, whose alleged crimes included buying oil from Iran in defiance of an embargo. (Rich’s wife was a generous donor to Clinton.) And then Joe Biden had the temerity to pronounce himself a defender of the rule of law before he used his presidential powers to insulate his own family from potential prosecution.

Over the past few weeks, Donald Trump has exposed the flimsiness of American institutions. Pressure-tested by his audacious assault on the civil services, those institutions instantly folded. But when a bridge tumbles into a river, the rivets and bolts don’t suddenly fail. They erode over generations. This is what happened in Washington: The unfettered power of the president kept expanding, Congress entered a state of sclerosis, the parties became apologists for their leaders, and courts fell into the hands of ideologues. As Toobin depressingly shows, even upstanding nice guys like Gerald Ford played their part in the collapse.