What Works: Removing Unfit Leaders
Part 3 of “The Tree of Liberty’s Roots: Repairing What Is Rotten”
The American constitutional system has a removal mechanism: impeachment. A president can be removed from office for “treason, bribery, or other high crimes and misdemeanors”[^1]—criminal misconduct by the official holder of the office.
It is an extraordinary mechanism. It is also a broken one.
In the 245 years since ratification, the House of Representatives has impeached four presidents: Andrew Johnson in 1868, Bill Clinton in 1998, and Donald Trump twice, in 2019 and 2021.[^2] The Senate has never convicted.[^3] The mechanism designed to remove incompetent or corrupt executives has become a partisan weapon—threatened periodically but never successfully used for removal.
Parliamentary systems have a different approach: votes of no confidence. A prime minister can be removed mid-term by a simple majority vote in parliament—and it happens, regularly, in democratic systems around the world.
The difference matters. When an executive becomes a liability—to competence, to reputation, to governing effectiveness—parliamentary systems can remove them. The United States has no reliable mechanism for non-criminal removal.
But the comparison requires nuance. Parliamentary removal has its own limitations. And the United States has one mechanism the article has not yet discussed: the Twenty-Fifth Amendment.
The Existing Mechanism: The Twenty-Fifth Amendment
Before examining what the United States lacks, it is worth examining what it has.
Section 4 of the Twenty-Fifth Amendment, ratified in 1967, provides a mechanism for presidential removal when the president is “unable to discharge the powers and duties of his office.”[^4] The Vice President and a majority of the cabinet can declare the president unable to serve, transferring power to the Vice President as Acting President.
The process is not simple. The president can contest the declaration, Congress must vote within 21 days, and a two-thirds vote in both houses is required to keep the president out of power.[^5] The threshold is high—higher than impeachment conviction.
But the mechanism exists. It has never been used. Why?
Political barriers: Invoking Section 4 requires the Vice President and cabinet to act against the president who appointed them. The incentive structure discourages use.
Vagueness: “Unable to discharge” is undefined. Physical incapacity is clear; mental incapacity or erratic behavior is contested.
Precedent fear: No one wants to be the first to invoke a constitutional mechanism that might be seen as a political coup.
The Twenty-Fifth Amendment provides a path for non-criminal removal—but a path that requires political will that has never materialized. It is a mechanism that exists on paper but not in practice.
The Parliamentary Alternative: Votes of No Confidence
Votes of no confidence are the standard removal mechanism in parliamentary democracies. They operate on a simple principle: the executive serves at the pleasure of the legislature. If the legislature withdraws that pleasure, the executive falls.[^6]
How It Works
The specifics vary by country. In the United Kingdom, a vote of no confidence is a parliamentary resolution stating that the House “has no confidence in Her Majesty’s Government.” If it passes, the prime minister must either resign or call a new election.[^7]
In Germany, the system is constructive: a vote of no confidence must simultaneously name a replacement. The outgoing chancellor is removed only when the incoming chancellor is identified and ready to serve.[^8]
In Japan, the process is informal: if the prime minister loses support of their party, they resign before a formal vote can be held.
The Successes
The record of votes of no confidence shows the mechanism functioning as intended:
Callaghan, 1979 (UK): James Callaghan’s Labour government fell on a single vote. The Scottish National Party, which had supported the government in confidence votes, withdrew its support over devolution legislation. Callaghan lost 311–310. A general election followed, which the Conservative Party under Margaret Thatcher won decisively.[^9]
Miyazawa, 1993 (Japan): Kiichi Miyazawa resigned as prime minister after the Liberal Democratic Party (LDP) lost seats in the lower house elections and Miyazawa himself admitted to accepting bribes during a scandal over the Sagawa Kyubin company.[^10] His resignation avoided a no-confidence vote that would have removed him.
Kohl, 1998 (Germany): Helmut Kohl lost a constructive no-confidence vote after his coalition collapsed. He was replaced by Gerhard Schröder.[^11]
May, 2019 (UK): Theresa May lost a vote of no confidence within her own Conservative Party, 211–148. She announced her resignation and remained as caretaker until a leadership election produced a replacement.[^12]
The Failures
But votes of no confidence do not always work smoothly:
May survives, January 2019: Jeremy Corbyn’s Labour Party moved a no-confidence vote against Theresa May’s government following the defeat of her Brexit withdrawal agreement. May’s government survived 325–306, with Democratic Unionist Party MPs supporting the government despite opposing the Brexit deal.[^13] The mechanism did not remove an unpopular prime minister because party dynamics favored survival.
Johnson survives, September 2019: Opposition parties moved a no-confidence motion against Boris Johnson’s government. Johnson responded by suspending parliament—a move later ruled unlawful by the Supreme Court—but no no-confidence vote was held before parliament was prorogued.[^14] The mechanism failed to produce accountability because the government controlled the timing.
Italy’s instability: Italy has had 68 governments since 1946.[^15] Frequent no-confidence votes produce constant turnover—prime ministers average little more than a year in office. This instability can prevent coherent policy-making and create opportunities for populist outsiders.
The pattern: Removal depends on party dynamics, not public will. Callaghan was removed even though no corruption allegations or criminal acts had been alleged—his crime was governing without a stable majority. Miyazawa remained in power as long as his party supported him—corruption allegations alone were not enough. May survived when party dynamics favored survival; she fell when party dynamics turned against her.
The mechanism doesn’t make leaders more responsive to public opinion. It makes them responsive to their party.
The Trade-offs
Accountability: Votes of no confidence provide a pathway to remove incompetent leaders without a criminal standard. The 1979 UK example shows a government removed not for corruption or criminal misconduct but for instability.
Democratic legitimacy: The mechanism is parliamentary, not public. The public does not vote directly on whether the leader should be removed. The leader serves at the pleasure of elected legislators.
Instability: In polarized systems, votes of no confidence could produce frequent government turnover. Italy’s history of frequent government changes—68 governments since 1946—is often cited as the cautionary example.[^15]
Party control: In parliamentary systems, party leaders select the prime minister and can remove them at will. In the UK, the Conservative Party removed Theresa May through an internal party vote before she ever faced a parliamentary confidence vote. The mechanism is democratic only insofar as the party itself is.
The Fundamental Problem
The United States cannot adopt a vote of no confidence without constitutional overhaul. The United States is not a parliamentary system.
The president is elected directly, not through the legislature. The president serves a fixed four-year term. The only constitutional mechanism for removal is impeachment.
Impeachment is designed for criminal misconduct. The House can impeach for “high crimes and misdemeanors,” but the Senate can convict only by a two-thirds vote. In a polarized system, no president will ever lose the support of 34 senators from their own party.
But impeachment has worked in other contexts. The Senate has convicted and removed eight federal judges.[^16] These cases demonstrate that the mechanism functions when evidence is clear and partisanship is lower. The problem is not impeachment itself—it is the application to presidents, where partisan loyalty overrides evidentiary standards.
The Nixon Precedent
Richard Nixon resigned in 1974 before impeachment proceedings could conclude. The House Judiciary Committee had approved articles of impeachment; the full House was expected to vote to impeach; and Republican senators had made clear they would vote to convict.[^17]
This is the closest the United States has come to removing a president deemed unfit. The mechanism worked—not through conviction, but through the credible threat of conviction producing resignation.
The Nixon precedent suggests that impeachment can function as a removal mechanism when: 1. The evidence is overwhelming 2. Public opinion turns decisively 3. Members of the president’s party conclude that defense is untenable
The question is whether current polarization makes these conditions impossible to achieve.
Feasibility for parliamentary-style removal: Near zero. The United States would have to become a parliamentary system—which would require rewriting the Constitution entirely.
Feasibility for improved impeachment function: Low but not zero. The mechanism exists; the political conditions for its use do not currently obtain. Whether those conditions can be recreated is an open question.
The Recall Elections Model
If the United States cannot adopt parliamentary removal, can it adopt direct democratic removal? Nineteen states provide for recall of state officials, including governors and state legislators.[^18]
How Recalls Work
Recall elections allow voters to petition for a vote on whether to remove an elected official before their term expires. If enough signatures are gathered—typically a percentage of votes cast in the previous election for that office—a recall election is held.[^19]
The ballot asks two questions: (1) Should the official be removed? (2) If so, who should replace them?
The Evidence
Davis recall, 2003 (California): Governor Gray Davis was recalled after signatures were gathered representing approximately 17% of votes cast in the previous election.[^20] Davis was removed by a 55-45 vote. Arnold Schwarzenegger, who won the replacement election with 48.6% of the vote, became governor.[^21]
The Davis recall illustrates a structural quirk of California’s system: the replacement candidate can win with far less support than the recalled official received. Schwarzenegger’s 48.6% was less than Davis’s 44.6% “remove” percentage—but the replacement election was a plurality contest, not a majority.
Wisconsin recalls, 2011-2012: Following the controversial Act 10 law limiting public worker collective bargaining, both Democrats and Republicans engaged in recall attempts. Thirteen state senators faced recall attempts—nine Democrats and four Republicans. Six Republicans were recalled (later, some were re-elected or replaced); two Democrats were recalled (later re-elected or replaced).[^22]
The Wisconsin experience illustrates the weaponization potential: recalls were used for policy disputes, not misconduct or incompetence.
Colorado, 2021: State Senator Julie Gonzales narrowly avoided recall, with 50.4% voting to retain her seat.[^23] The recall was driven by policy disagreement over police reform and pandemic restrictions.
The evidence shows recalls can remove officials who face a vote of no confidence from constituents. But recalls have downsides:
Weaponization: Recalls are available to anyone who can gather signatures. Well-funded opponents can use recalls as a political weapon, forcing incumbents to campaign perpetually.
Governance disruption: Recall campaigns can begin as soon as officials take office. The cost is perpetual campaigning, not governing.
Undermines electoral mandate: If officials can be removed mid-term, the election result—the mandate voters gave—becomes uncertain.
The Federal Problem
The Constitution contains no recall provision for federal officials. No president, senator, or representative has ever been recalled. The courts have consistently held that states cannot recall federal officials—the Constitution specifies the duration of their terms.[^24]
Implementing recall for the president would require a constitutional amendment. State-level recall for senators and representatives would require amendment of the Seventeenth Amendment or new constitutional text.
Feasibility: Low for federal recall (requires amendment). High for state recall (already in place in 19 states).
The Core Insight: Party, Not Public
These examinations—votes of no confidence, impeachment, the Twenty-Fifth Amendment, recall elections—reveal a fundamental truth.
Parliamentary systems remove leaders when they lose party support, not when they lose public support.
The 1979 removal of James Callaghan occurred because the Scottish National Party withdrew support—party dynamics, not public petition. The removal of Theresa May happened because the Conservative Party withdrew support. The failed removal of Theresa May in January 2019 occurred because party dynamics favored survival despite public opposition to her Brexit policy.
In every system, removal ultimately comes down to party dynamics. Public opinion matters only insofar as it shapes party calculations.
The question for American reform is not how to create a mechanism for public removal of incompetent leaders, but whether the United States can create a system where party loyalty does not always supersede the public interest.
The current system: party loyalty wins. The impeachment mechanism is technically available but functionally inert for presidents. The Twenty-Fifth Amendment provides a path for non-criminal removal but requires political will that has never materialized.
The reforms that would change this are either impossible (constitutional amendment to make impeachment work) or unacceptable (weakening the system’s protections against partisan removal).
Criteria for Evaluating Removal Mechanisms
The comparison between parliamentary and presidential systems requires defining what “better” means. Removal mechanisms can be evaluated across multiple dimensions:
| Criterion | Parliamentary (Vote of No Confidence) | Presidential (Impeachment) |
|———–|—————————————-|—————————|
| Responsiveness to misconduct | Moderate (party-dependent) | Low (party-dependent, high threshold) |
| Responsiveness to incompetence | Moderate (party-dependent) | Very Low (criminal threshold) |
| Stability | Variable (frequent turnover possible) | High (fixed terms) |
| Democratic legitimacy | Parliamentary (indirect) | Electoral (mandate respected) |
| Risk of partisan abuse | Moderate | High (impeachment weaponized) |
The question is not which system is objectively “better.” The question is what trade-offs a political system should accept—and whether those trade-offs can be adjusted within existing constitutional constraints.
Next: Anti-Capture Mechanisms
Removal reform addresses what happens when leaders prove unfit. The next part addresses what happens when leaders are captured before they are selected—or capture themselves once in power.
Campaign finance reform, lobbying restrictions, independent anti-corruption commissions—these are the mechanisms the world uses to prevent wealthy interests from buying influence.
Some work. Some don’t. The evidence matters more than the rhetoric.
Constitutional Reform: The Tree of Liberty’s Roots: Repairing What Is Rotten
Part 1: The Diagnosis: What’s Broken — Why the system can’t fix itself—and what the founders actually built
Part 2: What Works: Electoral Reform — RCV, open primaries, House expansion—what could pass tomorrow
Part 3: What Works: Removing Unfit Leaders — You are here
Part 4: What Works: Anti-Capture Mechanisms — Campaign finance, lobbying reform, democracy vouchers—what the evidence shows
Part 5: Novel Mechanisms: What Has Been Tried — Citizen assemblies, sortition, direct democracy—where evidence exists
Part 6: The Constitutional Question — What requires amendment, what can be done statutorily, and whether change is possible
Part 7: A Proposed Framework — How to prioritize reforms—and whether the system can fix itself
← Part 2: What Works: Electoral Reform | Part 4: What Works: Anti-Capture Mechanisms →
Notes
[^1]: U.S. Constitution, Article II, Section 4.
[^2]: U.S. Senate, “Impeachment,” https://www.senate.gov/artandhistory/history/common/briefing/Impeachment.htm. The four impeachments are: Andrew Johnson (1868), Bill Clinton (1998), Donald Trump (2019), and Donald Trump (2021).
[^3]: Ibid.
[^4]: U.S. Constitution, Amendment XXV, Section 4.
[^5]: Ibid.
[^6]: Kaare Strøm, “Parliamentary Democracy,” in The Oxford Handbook of Comparative Constitutions, eds. Rainer Bauböck and Thomas E. Bürklin (Oxford University Press, 2018), 872-894.
[^7]: House of Commons Library, “Votes of No Confidence,” https://commonslibrary.parliament.uk/research-briefings/cbp-8749/
[^8]: Basic Law for the Federal Republic of Germany, Article 68.
[^9]: David Butler and Gwen Ifill, Down to the Wire: The 1979 British Election (Croom Helm, 1980), 1-34.
[^10]: Richard Samuels, “Machiavelli’s Children: Leaders and Their Legacies in Italy and Japan” (Cornell University Press, 2003), 212-230.
[^11]: Helmut Kohl resignation coverage, Der Spiegel, October 27, 1998.
[^12]: BBC News, “Theresa May Resigns as Conservative Party Leader,” June 7, 2019.
[^13]: House of Commons Hansard, “Confidence Motion,” January 16, 2019.
[^14]: UK Supreme Court, R (Miller) v. The Prime Minister, [2019] UKSC 41.
[^15]: C.J. de Vreese, “Political Communication: Forcing the News Media into the Political Arena,” in The Sage Handbook of Political Communication, eds. Keren Tenenboim-Weinblatt and Sebastian Valenzuela (Sage, 2022), 451-468.
[^16]: U.S. Senate, “Impeachment of Federal Judges,” https://www.senate.gov/artandhistory/history/common/briefing/Senate_Impeachment_Trials.htm
[^17]: Stanley I. Kutler, The Wars of Watergate: The Last Crisis of Richard Nixon (Norton, 1990), 580-593.
[^18]: National Conference of State Legislatures, “State Recall Procedures,” https://www.ncsl.org/research/elections-and-campaigns/recall-of-state-officials.aspx
[^19]: Ibid.
[^20]: California Secretary of State, “2003 Recall Election Results.” The recall threshold was 12% of votes cast in the previous election for governor (897,158 signatures). Critics note that this represents approximately 17% of votes cast, not 12% of registered voters as sometimes claimed.
[^21]: Ibid.
[^22]: Wisconsin Elections Commission, “Recall Elections, 2011-2012,” https://elections.wi.gov/
[^23]: Colorado Secretary of State, “2021 General Election Results—State Senate District 34,” https://www.sos.state.co.us/
[^24]: U.S. Term Limits, Inc. v. Thornton, 514 U.S. 779 (1995).