50 Ways the Constitution Is Under-Maintained

The Constitution does not establish a national popular election for president. Article II and the Twelfth Amendment describe a system in which each state appoints electors “in such Manner as the Legislature thereof may direct,” and those electors cast the actual votes. The term “Electoral College” does not appear in the constitutional text. The modern expectation that Americans choose their president by popular vote is carried entirely by state law and party practice; a state legislature could reclaim direct appointment of electors without violating the federal Constitution. The Supreme Court acknowledged this in Bush v. Gore (2000). Five times in American history, including twice in the 21st century (2000 and 2016), the Electoral College has produced a president who lost the national popular vote.

Winner-take-all allocation, used by 48 of 50 states, compounds the problem. A vote in Wyoming (population 577,000, 3 electoral votes) carries roughly 3.6 times the Electoral College weight of a vote in California (population 39 million, 54 electoral votes), but that ratio captures only the allocation disparity. When analysts calculate effective voter power, accounting for the probability that a single vote changes the outcome, the ratio between swing-state and safe-state voters reaches roughly 10 million to one. Presidential candidates can safely ignore the tens of millions of Americans in non-competitive states; FairVote found that 96% of 2020 general-election campaign events took place in just 12 states. The system does not just miscount votes; it makes most of them irrelevant to governance. In 2016, Trump won the presidency while losing the popular vote by 2.9 million; the margin came from 77,000 combined votes across Wisconsin, Michigan, and Pennsylvania.

The elector system also creates a dangerous certification process. The Electoral Count Act of 1887 was exposed as dangerously ambiguous on January 6, 2021, when John Eastman’s memo argued that Vice President Pence could unilaterally reject electoral votes. 147 members of Congress voted to sustain objections. The Capitol was breached. Pence rejected the theory based on personal interpretation, not enforceable law. The Electoral Count Reform Act of 2022 improved procedures but remains an ordinary statute, repealable by simple majority.

The National Popular Vote Interstate Compact, under which participating states would allocate electors to the national popular-vote winner, has been enacted by states with 209 electoral votes. But its constitutional footing is dubious: the Compact Clause likely requires congressional consent under Virginia v. Tennessee (1893), and binding a state’s electors to other states’ votes may exceed what Article II’s “manner of appointing” means. The NPVIC is a statutory patch over a constitutional gap, and the gap is more likely to swallow the patch than the other way around.

The most important unwritten norm in a democracy is loser’s consent: the expectation that the losing side accepts the result and relinquishes power. In the United States, this norm held without exception from 1797 through 2017. Al Gore conceded after Bush v. Gore despite winning the popular vote. The norm was so strong Americans barely noticed it. It collapsed in 2020-2021. President Trump refused to concede, pressured state officials (the recorded call to Georgia Secretary of State Raffensperger asking him to “find 11,780 votes”), organized alternate elector slates in seven states, pressured Vice President Pence to refuse certification, and on January 6 addressed a rally that preceded the breach of the Capitol. For the first time in American history, the outgoing president did not attend the inauguration.

The operational transition machinery broke alongside the norm. The Presidential Transition Act of 1963 depends on a single political appointee, the GSA administrator, making an “ascertainment” that a winner exists. GSA Administrator Emily Murphy refused to ascertain Biden for 19 days, denying his team access to federal agencies, intelligence briefings, and pandemic response data while COVID-19 was killing over 1,500 Americans daily. Outgoing agencies cooperate with incoming teams because tradition demands it, not because law compels it. The 9/11 Commission found that the shortened 2000-2001 transition left the Bush administration weeks behind on national-security staffing. Other democracies treat transitions as mandatory legal processes: Australia codifies caretaker conventions, Germany requires the outgoing chancellor to brief the successor, Mexico mandates structured information sharing by constitutional reform. The United States gives the outgoing president ten weeks of unsupervised discretion.

Levitsky and Ziblatt’s How Democracies Die (2018) identified mutual toleration and institutional forbearance as the twin pillars of democratic stability, and argued that norm erosion, not formal constitutional breakdown, is how democracies typically collapse. The V-Dem Liberal Democracy Index dropped the United States from the 88th global percentile in 2015 to the 71st in 2021. A January 2024 CNN/SSRS poll found 35% of Americans still believe Biden did not legitimately win, four years after 60+ court challenges were dismissed. You cannot legislate good faith, but you can make bad faith less dangerous by replacing norms with enforceable law.

The Constitution does not guarantee the right to vote. The 15th, 19th, 24th, and 26th Amendments each prohibit denial of the vote on a specific basis (race, sex, poll taxes, age), but none affirmatively establishes a right to the franchise. Bush v. Gore (2000) acknowledged this: “the individual citizen has no federal constitutional right to vote for electors for the President of the United States.” The United States is the only peer democracy where the right to vote is defined by a patchwork of prohibitions rather than an affirmative guarantee. Any restriction that does not fall into a prohibited category, voter-ID requirements, registration purges, polling place closures, can proceed unless a court intervenes.

The country also has no national election system. Roughly 10,000 local jurisdictions run elections under 50 different state frameworks. The Help America Vote Act of 2002 created the Election Assistance Commission with voluntary guidelines and no enforcement power. In 2018, Georgia’s secretary of state administered the gubernatorial election in which he was a candidate, purging over 340,000 voter registrations and closing 214 polling places before narrowly winning. No federal authority had jurisdiction to intervene.

The Voting Rights Act of 1965 once provided a powerful federal check through preclearance (Section 5), requiring covered jurisdictions to obtain federal approval before changing election rules. The DOJ blocked over 1,000 discriminatory changes between 1965 and 2013. But Congress reauthorized the VRA four times without updating the coverage formula, which still used 1960s-era data. Shelby County v. Holder (2013) struck down the formula as unconstitutionally outdated. Within hours, Texas announced a voter-ID law that preclearance had blocked. States have since closed over 1,700 polling places, disproportionately in minority communities.

The VRA was always constitutionally precarious: it rested on Fifteenth Amendment enforcement power subject to the Court’s “congruence and proportionality” test (City of Boerne v. Flores, 1997). Efforts to restore preclearance, the For the People Act and the John Lewis Voting Rights Advancement Act, have failed in the Senate. The United States remains the only major democracy where rules for voting in a national election depend on which side of a county line you live on.

Political parties are the central organizing institution of American government. They determine who appears on ballots, structure legislative business in Congress, control committee assignments, coordinate fundraising, and shape voter identity for roughly 170 million registered Americans. The Constitution does not mention them. The Founders did not merely omit parties; they actively opposed them. Washington’s 1796 Farewell Address warned against “the baneful effects of the spirit of party.” Madison’s Federalist No. 10 treated faction as a disease to be controlled. Despite this hostility, a two-party system emerged almost immediately, driven by single-member, winner-take-all districts. Duverger’s Law, the political science finding that plurality voting in single-member districts tends to produce two dominant parties, explains why the United States has never sustained a viable third party. The system is a duopoly by mathematical inevitability, given the rules.

The constitutional silence on parties creates a regulatory vacuum. The Federal Election Commission, created in 1974, is structured as a six-member body evenly split between the two major parties, guaranteeing deadlock on any enforcement action either party objects to. Between 2010 and 2020, the FEC failed to reach agreement on enforcement in hundreds of cases. The agency tasked with policing the parties is controlled by the parties. Meanwhile, ballot-access laws in many states require third parties to gather tens of thousands of signatures while the two major parties are listed automatically. Gallup has found that the share of Americans identifying as independent has exceeded identification with either party every year since 2009, reaching 43% in 2023. Yet these voters have functionally no representation outside the two-party framework.

The process by which these constitutionally invisible parties select their nominees is equally unregulated. The primary system is not established by the Constitution, not standardized by federal law, and not consistent from state to state. It is a patchwork that varies in format (open, closed, semi-closed, blanket), timing (January through June), and method (secret ballot, show of hands, ranked preference). The modern system dates to the McGovern-Fraser Commission reforms following the chaotic 1968 Democratic National Convention, where Hubert Humphrey won the nomination without competing in a single primary. The reforms democratized nominations but created new problems: presidential primary turnout averaged roughly 27% in competitive contests between 2000 and 2020, according to the Bipartisan Policy Center. Low turnout means that a small, ideologically committed minority determines who appears on the general-election ballot. Primary electorates are consistently older, whiter, and more ideologically extreme than general-election voters in both parties, driving legislative polarization because nominees must satisfy their primary base before appealing to median voters.

The calendar compounds the problem. Early states exert disproportionate influence through the “momentum” dynamic: a win in Iowa or New Hampshire can make or break candidacies before most Americans have voted. In 2020, Pete Buttigieg and Amy Klobuchar dropped out and endorsed Joe Biden the day before Super Tuesday, consolidating the moderate lane in a coordinated maneuver that millions of voters in later states had no opportunity to weigh in on. No other advanced democracy selects its leaders through such a prolonged, fragmented, and low-participation process. British parties choose leaders through structured membership votes conducted over weeks. French parties hold compact national primaries. The American system extends over six months, costs billions of dollars, and systematically empowers the voters least representative of the broader electorate.

The legal regime governing money in American elections has been shaped more by the Supreme Court than by Congress, and the result is a system that changes with each shift in the Court’s composition. The starting point is Buckley v. Valeo (1976), in which the Court upheld contribution limits (donations to candidates) but struck down expenditure limits (spending by individuals and groups), reasoning that spending money is a form of speech protected by the First Amendment. This distinction, you can cap what someone gives a politician but not what someone spends to influence an election, has defined the landscape for five decades.

Citizens United v. FEC (2010) extended the logic of Buckley to its conclusion, ruling 5-4 that corporations and unions have First Amendment rights to make unlimited independent expenditures. Justice Kennedy’s majority opinion held that “independent expenditures, including those made by corporations, do not give rise to corruption or the appearance of corruption.” Within months, the D.C. Circuit’s SpeechNow.org v. FEC decision created super PACs, political action committees that can raise and spend unlimited sums as long as they do not “coordinate” with candidates. The coordination prohibition has proven almost unenforceable; in 2016 and 2020, super PACs were routinely staffed by former campaign officials and guided by publicly available strategic signals.

The spending numbers tell the story of what these decisions unleashed. The 2000 presidential election cost approximately $1.4 billion in total spending. The 2020 election cost $14.4 billion, a tenfold increase in two decades, according to OpenSecrets. The 2024 cycle exceeded $15.9 billion. Outside spending, money from super PACs, dark-money nonprofits, and other independent groups, rose from $338 million in 2008 to $4.5 billion in 2020. Much of this money is untraceable: 501(c)(4) “social welfare” organizations can spend on elections without disclosing their donors, a loophole neither Congress nor the Court has closed.

The instability of the regime is itself a problem. The campaign-finance rules in effect during any given election depend on the current composition of the Supreme Court and the willingness of the FEC to enforce existing law. Congress has not passed major campaign-finance legislation since the Bipartisan Campaign Reform Act of 2002 (McCain-Feingold), and the Court has struck down key provisions of that law. The DISCLOSE Act, which would require disclosure of large donors to politically active nonprofits, has been introduced in every Congress since 2010 and has never passed the Senate. A constitutional amendment to overturn Citizens United was supported by 54 senators in 2014, a majority, but short of the two-thirds required.

The practical effect is that the United States has a campaign-finance system that satisfies almost nobody: reformers see legalized corruption, free-speech advocates see remaining limits as unjustified, and ordinary voters see a system that amplifies the influence of the very wealthy while making accountability nearly impossible to trace.

Article I, Section 2 of the Constitution requires an “actual Enumeration” of the population every ten years. This is one of the few administrative obligations the Founders wrote directly into the constitutional text, reflecting their understanding that representative government depends on accurate population counts. The census determines the apportionment of House seats among the states, the drawing of congressional and state legislative districts, and the allocation of Electoral College votes. It also governs the distribution of roughly $2.8 trillion in annual federal funding: Medicaid, highway funds, Title I education grants, and hundreds of other programs that allocate money based on population formulas.

The Census Bureau sits within the Department of Commerce, an executive-branch agency headed by a political appointee. The Bureau’s director serves at the pleasure of the Secretary of Commerce, who serves at the pleasure of the president. There is no statutory guarantee of the Bureau’s independence comparable to what exists for the Bureau of Labor Statistics or the Federal Reserve. This structure makes the census vulnerable to political manipulation at precisely the moments when the stakes are highest.

The most prominent recent example is the citizenship question controversy of 2017-2019. Commerce Secretary Wilbur Ross directed the Census Bureau to add a question about citizenship status to the 2020 decennial census, a question that had not appeared on the short-form census since 1950. The Bureau’s own chief scientist estimated that adding the question would depress response rates among noncitizen households by 5.8%, disproportionately undercounting Hispanic communities and immigrant-heavy states like California, Texas, and New York. In Department of Commerce v. New York (2019), Chief Justice Roberts wrote for a 5-4 majority that Ross’s stated rationale, enforcing the Voting Rights Act, was “contrived” and “pretextual,” blocking the question. But the decision rested on administrative-law grounds, not constitutional ones; a better-crafted justification might have survived review.

Apportionment itself uses the Huntington-Hill method of equal proportions, adopted by Congress in 1941, which slightly favors smaller states in the allocation of marginal seats. After the 2020 census, New York lost a House seat by a margin of 89 people, a difference smaller than a single apartment building and well within the census’s margin of error. Minnesota kept its seventh seat over New York by that same razor-thin margin. These are not abstract statistical questions: each House seat carries roughly 760,000 constituents and one Electoral College vote, meaning that census accuracy directly determines the political power of millions of Americans.

The 2020 census also suffered significant operational disruptions from the COVID-19 pandemic, with field operations delayed and then compressed into a shortened timeline after the administration attempted to impose an early cutoff. The Census Bureau’s own post-enumeration survey found a 3.3% undercount of Hispanic residents, a 5.6% undercount of American Indians and Alaska Natives on reservations, and a 1.8% overcount of non-Hispanic white residents, the largest differential undercount in thirty years.

A separate apportionment problem is the fixed size of the House itself. The Permanent Apportionment Act of 1929 capped the House at 435 members, a number that has not changed since, even as the U.S. population has grown from 122 million to 335 million. Each House member now represents roughly 760,000 constituents, the highest ratio in the developed democratic world. The United Kingdom has one MP per 100,000 people; Germany has one Bundestag member per 115,000; France has one National Assembly deputy per 116,000. The cube root law, the empirical finding by political scientist Rein Taagepera that national legislatures across democracies tend to scale as roughly the cube root of population, would put the U.S. House at approximately 690 members. Larger chambers do not produce chaos; the European Parliament functions with 720 members, and the UK House of Commons with 650. A frozen House means that representation degrades with every census, districts grow too large for meaningful constituent service, and the Electoral College (which adds House seats to Senate seats) becomes increasingly distorted by the small-state Senate bonus as population concentrates in fewer states.

After the census determines how many House seats each state receives, someone must draw the district boundaries. The Constitution says nothing about how. In most states, the legislature draws the maps, which means the party in power draws the districts. The result is partisan gerrymandering: lines drawn to maximize one party’s seat share by packing the opposing party’s voters into a few districts and spreading the remainder thinly across the rest. Both parties do it when they have the opportunity. After the 2010 census, Republican-controlled legislatures in Pennsylvania, Ohio, North Carolina, Wisconsin, and other states drew maps that locked in durable majorities even in years when Democrats won more total votes statewide. Democrats did the same in Maryland and Illinois.

In Rucho v. Common Cause (2019), the Supreme Court held 5-4 that partisan gerrymandering presents a political question beyond the reach of federal courts. Chief Justice Roberts’s majority opinion acknowledged that “excessive partisanship in districting leads to results that reasonably seem unjust” but concluded that the Constitution provides no “limited and precise” standard for courts to apply. The decision closed the federal courthouse door. State courts in some states (Pennsylvania, North Carolina, Ohio) have struck down maps under state constitutional provisions, but the results are inconsistent, the legal standards vary from state to state, and the Supreme Court in Moore v. Harper (2023) came close to adopting the “independent state legislature” theory that would have stripped state courts of authority over federal redistricting entirely.

Only about a dozen states use independent redistricting commissions, and even these vary widely in structure and independence. California’s Citizens Redistricting Commission, created by ballot initiative in 2008 and 2010, is generally regarded as effective: its members are selected through a screening process designed to exclude partisan operatives, and the maps it produces have been more competitive than the legislature-drawn maps they replaced. Arizona’s commission has survived legal challenge (Arizona State Legislature v. Arizona Independent Redistricting Commission, 2015). But most states have no commission, no nonpartisan standard, and no mechanism to prevent the party in power from choosing its own voters.

The consequences are measurable. A 2020 analysis by the Brennan Center found that gerrymandering after the 2010 census gave Republicans a structural advantage of 16-17 House seats over the decade. The Associated Press calculated that in 2018, Democrats needed to win the national popular vote by roughly 7-8 percentage points just to win a bare majority of House seats. Gerrymandering also reduces electoral competitiveness: in the 2024 elections, fewer than 40 House races out of 435 were considered genuinely competitive. When general elections are predetermined, the only election that matters is the primary, which pulls representatives toward their base and away from the median voter, reinforcing polarization. Districts should be drawn once per decade, after the census, through a transparent process that the party in power does not control.

Article I, Section 9 of the Constitution gives Congress sole authority over the federal purse: no money shall be drawn from the Treasury except in consequence of appropriations made by law. The Congressional Budget Act of 1974 built a statutory framework around that power, establishing the annual budget resolution, the reconciliation process, and the October 1 start of the fiscal year. The system was designed to give Congress a comprehensive, coordinated method for setting national spending priorities rather than passing twelve separate appropriations bills in a vacuum.

The practice has diverged sharply from the design. The Government Accountability Office found that in all but three of the past thirty fiscal years, Congress failed to complete its appropriations work on time and funded the government through continuing resolutions instead. Brookings Institution scholars have called the appropriations process “in a state of collapse.” The consequences are concrete. CRs freeze spending at prior-year levels, preventing agencies from launching programs that Congress has already authorized. The Department of Defense has testified repeatedly that multi-month CRs cost billions in inefficiency because contractors cannot plan, multi-year procurements get disrupted, and last-minute omnibus bills are drafted in days and passed without meaningful review.

Government shutdowns illustrate the failure at its most acute. The 1995-96 shutdown lasted 21 days and furloughed roughly 800,000 federal workers. The 2013 shutdown cost the economy an estimated $24 billion according to Standard and Poor’s. The 2018-19 shutdown, lasting 35 days, was the longest in American history, delayed tax refunds, shuttered national parks, grounded food safety inspections, and left hundreds of thousands of workers without pay. None of these episodes resolved any underlying fiscal question; they were pure governance failures.

The root problem is structural. Congress has progressively surrendered its capacity to manage the detail of twelve distinct appropriations bills. The authorizing committees and the appropriations committees have been in institutional conflict for decades. Staff capacity has declined since the 1970s. The Congressional Research Service estimated in 2022 that total personal and committee staff is lower today in inflation-adjusted terms than it was forty years ago. Deadlines have no enforcement mechanism. A chamber that misses October 1 faces no consequence beyond the inconvenience of negotiating a CR, which is itself a form of majority-coalition management, not a penalty.

The debt ceiling was created by the Second Liberty Bond Act of 1917 as a wartime administrative convenience. Before that, Congress approved each bond issuance individually. Aggregating borrowing authority into a single statutory cap was meant to give Treasury operational flexibility while preserving congressional oversight of total indebtedness. The ceiling has been raised or suspended 78 times since 1960, under every president of both parties, because not raising it would mean defaulting on obligations Congress had already voted to incur through its spending and tax laws.

This creates a logical incoherence at the center of American fiscal governance. Congress votes to appropriate more than it taxes, then must vote a second time to permit borrowing the difference. The second vote adds no new fiscal information; the deficit was determined by the first votes. What it does create is a recurring opportunity for one faction to extract policy concessions by threatening to trigger a default on the full faith and credit of the United States. That threat was theoretical until 2011, when brinkmanship over the debt ceiling led Standard and Poor’s to downgrade U.S. sovereign debt for the first time in history, from AAA to AA+. The downgrade itself cost American taxpayers: Treasury yields rose, and the Bipartisan Policy Center estimated the episode added roughly $1.3 billion in interest costs in 2011 alone.

The 2023 episode came closer to a genuine default. Treasury Secretary Janet Yellen began deploying “extraordinary measures” in January, including suspending investments in the Civil Service Retirement and Disability Fund and the Postal Service Retiree Health Benefits Fund, to stay under the ceiling while Congress negotiated. By late May, Yellen’s public statements indicated the X-date (the day cash runs out) was as close as June 1. The Fiscal Responsibility Act passed with four days to spare. During that period, short-term Treasury bill yields spiked sharply because holders of bills maturing near the X-date demanded a default premium. The disruption in the world’s benchmark risk-free asset rippled through every credit market that prices off Treasuries.

The debt ceiling does not discipline spending. All spending that drives debt growth was authorized and appropriated before the ceiling became binding. It functions as a hostage mechanism rather than a fiscal constraint. Countries with Westminster systems simply appropriate and borrow; the authorization is implicit in the spending vote. Germany has a constitutional debt brake that operates at the point of appropriation, not after the fact. The American mechanism achieves neither genuine fiscal discipline nor a sensible legislative check: it menaces the credit of the country to adjudicate disputes that should have been resolved during appropriations.

Article II makes the president responsible for faithful execution of the laws, and the Appropriations Clause of Article I requires that money be spent only as Congress directs. For most of American history, presidents exercised modest discretion in spending down appropriations, deferring some expenditures for efficiency reasons. Richard Nixon converted that limited practice into a wholesale policy instrument, impounding billions in funds for water treatment, rural development, and housing that Congress had appropriated over his objections. Nixon argued impoundment was an inherent executive power; critics argued he was effectively exercising a line-item veto that the Constitution did not grant him.

Congress responded with the Impoundment Control Act of 1974, which created two categories. A “deferral” postpones spending and takes effect unless either chamber passes a resolution of disapproval within 45 days. A “rescission” cancels appropriated funds entirely but requires affirmative congressional approval within 45 days; if Congress does not act, the president must release the money. The Supreme Court’s decision in Train v. City of New York (1975) reinforced the obligation to spend: the administration could not impound funds Congress had mandated be spent on water pollution control.

The ICA’s limitations became vivid during the first Trump impeachment. In summer 2019, the Office of Management and Budget directed the Department of Defense to hold $391 million in security assistance to Ukraine that Congress had appropriated. GAO concluded in January 2020 that OMB had violated the ICA, because the hold exceeded 45 days and no rescission request had been submitted to Congress. The hold was eventually released under congressional pressure. The delay in aid delivery was one of the factual bases for the impeachment inquiry, and the GAO finding that the hold was unlawful was entered into the public record.

The 2025 controversy over executive branch spending freezes revisited this terrain on a larger scale. Early in the year, OMB issued guidance that appeared to pause all federal grants and loans pending review, a directive that a federal district court blocked within 48 hours as likely violating the ICA and the Appropriations Clause. The episode demonstrated that the ICA’s enforcement depends almost entirely on litigation by affected parties or GAO findings that the executive branch can simply contest. No automatic mechanism releases funds when a hold is found unlawful; someone must sue, and that takes time.

The Antideficiency Act, first passed in 1870 and substantially revised in 1950 and 1982, requires that appropriated funds be apportioned before agencies can obligate them. The purpose was protective: Congress was alarmed by agencies that spent their annual appropriations in the first months of the fiscal year and then returned for emergency supplements. Apportionment, administered by the Office of Management and Budget, breaks annual appropriations into time periods (quarterly) or project categories, ensuring that agencies do not exhaust funds prematurely and then plead insolvency.

The protective logic is sound. The power’s use as a policy instrument is a different matter. OMB can apportion funds in ways that slow, restrict, or condition spending on executive branch priorities without formally impounding the funds under the ICA. A quarterly apportionment can be set conservatively so that only a fraction of the full-year appropriation is released in the first quarter, creating a structural shortfall that forces agencies to defer activities Congress intended to fund. OMB apportionment footnotes have been used to condition the release of funds on compliance with executive orders, reorganization plans, and policy directives that were never debated in the appropriations process.

The distinction between a “deferral” under the ICA and an “apportionment condition” under the Antideficiency Act is legally contested and practically consequential. GAO has issued multiple opinions finding that certain apportionment conditions cross into ICA territory, but OMB is not bound by GAO legal opinions, and the Department of Justice Office of Legal Counsel has historically construed executive apportionment authority broadly. The result is that the same funds that Congress voted to appropriate pass through an administrative chokepoint where a small group of OMB career and political staff can reshape the timing and effective scope of spending without a single vote.

The problem is compounded by opacity. Apportionment documents are not routinely published in real time; agencies receive them internally, and public disclosure has been inconsistent. During the Ukraine aid hold in 2019, the existence of the apportionment order restricting DoD spending was not publicly disclosed until congressional investigators obtained it. A 2023 statute required OMB to publish apportionment data on USASpending.gov within two business days of issuance, a transparency improvement, but publication is not the same as enforceability.

The underlying structural issue is that Congress appropriates in gross while the executive controls the cash flow in detail. A legislature that can only appropriate but cannot enforce the rate and timing of spending has ceded more power than the constitutional text suggests it intended. The Appropriations Clause says no money shall be drawn from the Treasury except in consequence of appropriations made by law; it does not say the executive may withhold money that law has made available.

The Constitution specifies supermajority requirements in five places: treaty ratification, constitutional amendments, conviction after impeachment, expulsion of members, and overriding a veto. Ordinary legislation requires a simple majority of a quorum in each chamber. The filibuster is not among the constitutional requirements. It exists because in 1806, on Aaron Burr’s recommendation, the Senate dropped the previous question motion from its rulebook as redundant, inadvertently eliminating its only procedural mechanism for cutting off debate. Extended debate as a blocking tool became possible; it became a recognized legislative tactic by the 1830s.

The Senate adopted Rule XXII, the cloture rule, in 1917 after a group of eleven senators talked a bill authorizing the arming of merchant ships to death over President Wilson’s objections. Cloture initially required two-thirds of those present and voting. In 1975, the Senate changed the threshold to three-fifths of all senators duly chosen and sworn, currently 60 votes. The cloture rule has been further eroded by the “nuclear option”: in 2013, a Democratic majority eliminated the 60-vote threshold for executive nominations and non-Supreme Court judicial nominees; in 2017, a Republican majority extended that elimination to Supreme Court nominees.

The consequences of the filibuster on civil rights legislation alone are historically devastating. Strom Thurmond spoke for 24 hours and 18 minutes against the Civil Rights Act of 1957, the longest individual filibuster on record. The Civil Rights Act of 1964 was held up by a 60-day Southern filibuster, the longest in Senate history, requiring 67 votes for cloture under the then-applicable rule. Anti-lynching legislation passed the House three times before World War II and never received a floor vote in the Senate because of the filibuster threat. The Voting Rights Advancement Act, passed by the House in 2021, received no cloture vote because the outcome was known in advance.

The current “silent filibuster” requires no floor speech at all; the minority simply signals that it will not consent to cloture, and the majority leader pulls the bill rather than spend floor time on a vote it will lose. A single senator can effectively block legislation with no public accountability, no expenditure of political capital, and no record that can be used in a campaign. The procedural transparency of Thurmond’s 24-hour speech, whatever else one may say about its purpose, at least made the obstruction visible. The contemporary version does not.

The filibuster governs the affairs of the world’s most powerful legislature by virtue of a rule that the Senate can change by simple majority on any given day. The norm against changing it is itself not constitutionally grounded. It is a convention maintained by the members who benefit from it.

Article I, Section 7 gives the president two ways to kill a bill without a veto override fight. Under the regular veto, the president returns the bill unsigned with objections, and Congress may override by two-thirds in each chamber. Under the pocket veto, the president simply does not act within the 10-day window (Sundays excluded), and if Congress has adjourned so that it cannot receive the bill back, it dies. The pocket veto is absolute; there is no override mechanism because there is no bill to override. It is one of the few presidential powers that is genuinely unreviewable once exercised, which makes its trigger condition, what counts as an “adjournment” sufficient to prevent return, constitutionally significant.

The Supreme Court addressed the question in The Pocket Veto Case (1929), holding that a four-month recess between sessions counted as an adjournment preventing return, and thus that Coolidge’s non-signature had pocket-vetoed an Indian citizenship bill. The Court’s reasoning turned on whether Congress had provided an agent capable of receiving messages during the recess; it had not. Forty-five years later, in Kennedy v. Sampson (1974), the D.C. Circuit held that a brief intra-session recess did not trigger the pocket veto because the Senate had authorized its secretary to receive presidential messages and had therefore not adjourned in the constitutional sense. The two cases point in the same direction: the dispositive question is whether Congress has made itself functionally available, not whether a formal adjournment has occurred.

Presidents have not accepted this reading consistently. Nixon attempted pocket vetoes during congressional recesses of as few as 9 days. Ford claimed a pocket veto during a Christmas recess in 1974 on a bill concerning medical training. Courts mostly dismissed challenges to these on standing or mootness grounds rather than resolving the constitutional question, leaving the boundary undefined. Reagan and George H.W. Bush attempted pocket vetoes during intersession recesses when Congress had designated agents to receive messages, a practice that circuit courts found invalid but that the Supreme Court never definitively resolved.

The practical stakes are real. A pocket veto cannot be overridden. A regular veto of a bill with broad support can at least be publicly debated and potentially overridden. When a president pocket-vetoes legislation during a recess by claiming the recess qualifies as an adjournment, Congress is deprived not only of the law but of the public record of the president’s reasoning. The pocket veto was designed for genuine intersession gaps when the government effectively went dark; it was not designed to give the executive an absolute veto during routine legislative recesses.

Congress has tried to manage this by keeping itself technically in session through pro forma sessions every three days, the same mechanism that was litigated in NLRB v. Noel Canning (2014) in the context of recess appointments. Pro forma sessions are a workaround, not a resolution.

The Congressional Review Act, passed in 1996 as part of the Republican “Contract with America” agenda, created a fast-track procedure for Congress to disapprove agency rules. Under the CRA, agencies must submit all final rules to Congress before they take effect. Congress then has 60 “legislative days” (days the chamber is in session, not calendar days, which can stretch a 60-legislative-day window to eight or nine calendar months) to introduce a joint resolution of disapproval. If both chambers pass the resolution and the president signs it, or if Congress overrides a veto, the rule is nullified and the agency is prohibited from issuing any rule “substantially similar” to the disapproved one without new statutory authorization.

For its first twenty years, the CRA was used exactly once, to disapprove an ergonomics rule issued by OSHA in the final days of the Clinton administration. The Republican 115th Congress, beginning in January 2017, used the CRA 16 times in its first months to disapprove rules issued in the final year of the Obama administration, exploiting the extended lookback window that allowed rules finalized in the summer and fall of 2016 to remain within the 60-legislative-day window. The rules disapproved included the Social Security Administration’s reporting to the NICS background check database, a methane emissions rule for oil and gas operations on federal land, a rule requiring disclosure of payments to foreign governments by extractive industry companies, and a rule on broadband privacy. Each disapproval was permanent; the agencies could not simply reissue the regulations after the administration changed.

The “substantially similar” prohibition has become the CRA’s most consequential and least understood feature. It means that a CRA disapproval does not merely send the agency back to start the rulemaking process over; it removes the agency’s authority to regulate in that space absent new legislation. Courts have interpreted “substantially similar” inconsistently, with no circuit court having issued a comprehensive definition. The FCC, after its broadband privacy rule was disapproved in 2017, concluded internally that it lacked authority to reimpose equivalent protections through a different regulatory mechanism, leaving broadband customers without privacy rules for years.

The CRA also interacts poorly with the separation of powers. A joint resolution of disapproval looks like legislation and is subject to a presidential veto, which means the same president who oversees the agency can protect agency rules from CRA disapproval. Conversely, a president who dislikes a rule his own agency issued can effectively collude with a congressional majority to disapprove it, bypassing the notice-and-comment rulemaking requirements that the Administrative Procedure Act requires for rule revocation. The CRA is therefore neither a pure legislative check on the executive nor a neutral oversight tool; it can be used as a coordination mechanism between branches to circumvent the APA.

Congress’s power to investigate as an adjunct to its legislative function was recognized by the Supreme Court in McGrain v. Daugherty (1927) and has never been seriously questioned since. The constitutional basis is implied from the legislative power: Congress cannot legislate intelligently on matters it cannot investigate. The practical question is not whether Congress can investigate but whether it can compel production of testimony and documents when executive branch officials resist. Three mechanisms are available: inherent contempt, criminal contempt, and civil enforcement.

Inherent contempt, by which Congress could direct its own sergeant-at-arms to detain and fine a recalcitrant witness, was last exercised in 1935 against William MacCracken, an attorney charged with destroying documents subpoenaed by the Senate. The Supreme Court upheld the Senate’s power in Barry v. United States ex rel. Cunningham (1929). No chamber has used inherent contempt since, because the political costs of physically detaining a senior executive branch official are prohibitive and the legal proceedings would be complex.

Criminal contempt requires a referral to the Department of Justice, which has discretion to prosecute or decline. That discretion creates a circular dependency at the center of oversight enforcement: the branch being investigated controls the prosecution of those who defy the investigative subpoenas. The pattern is consistent across administrations. In 2012, the House voted to hold Attorney General Eric Holder in contempt over the “Fast and Furious” gun-walking investigation; the Obama DOJ declined to prosecute. In 2021, the House voted to hold Steve Bannon in contempt for defying a subpoena from the January 6th Select Committee; the Biden DOJ did prosecute, and Bannon was convicted and sentenced to four months. The difference was not legal; the facts were legally equivalent. The difference was the administration.

Civil enforcement through the courts takes years. The House Judiciary Committee subpoenaed former White House Counsel Don McGahn in April 2019 to testify about events described in the Mueller report. The resulting litigation, Committee on the Judiciary v. McGahn, wound through the D.C. Circuit for more than two years, ultimately producing a ruling that McGahn must appear but that executive privilege claims could still be litigated, rendering the subpoena practically moot by the time the question was resolved. By the time a congressional subpoena is adjudicated, the investigation that produced it is typically over, the relevant Congress has ended, and the facts the subpoena sought to develop are either stale or already established through other means.

No democratic legislature should depend on the goodwill of the branch it oversees for the enforcement of its own investigative power. The current system systematically advantages the executive in any confrontation, because the executive controls time and time is the variable that most affects whether oversight is meaningful.

The Constitution provides for member vacancies through death or resignation, for expulsion by two-thirds vote of the relevant chamber, and for the states to fill House and Senate vacancies through special elections or gubernatorial appointment. It provides nothing for incapacity short of expulsion. A member who is alive, has not resigned, has not been expelled, and has not been convicted of a disqualifying offense holds their seat regardless of their cognitive or physical condition. The 25th Amendment addresses presidential incapacity through a mechanism requiring the vice president and a majority of the cabinet to act; Congress has no parallel structure for its own members.

Senator Strom Thurmond of South Carolina served until age 100, completing his term in January 2003. Accounts from staff and colleagues in his final years describe a man who required substantial assistance to conduct his duties and who was often unaware of the substance of legislation he was voting on. His office functioned through staff management in ways that are not publicly documented but are widely reported among those who served in the Senate during the period. His vote counted the same as any other senator’s. Senator Robert Byrd of West Virginia served until his death at age 92 in 2010, with cognitive decline documented in the final two years of his service.

Senator Dianne Feinstein’s final months in 2023 provided the most publicly documented recent case. After a serious case of shingles complicated by encephalitis that kept her away from the Senate for three months, Feinstein returned in visibly diminished condition. Multiple news organizations reported that she did not remember having been absent, asked staff to explain where she was, and on at least one occasion was wheeled to the floor to cast a vote she appeared to receive direction on from a staff member seconds before casting it. The Senate Judiciary Committee, on which she served, could not advance several judicial nominees to the full Senate for months because she was unavailable to participate in the committee quorum. She declined repeated requests from Democratic colleagues to resign. She died in September 2023 having never transferred the seat voluntarily.

The harm extends beyond symbolism. Feinstein’s absence and incapacity affected the pace of judicial confirmations in a closely divided Senate where every vote on committee quorums mattered. A single incapacitated member can block committee action, prevent floor votes, and consume scheduling bandwidth that a functional majority needs for other business. The two-thirds expulsion threshold was designed for willful misconduct, not incapacity, and no chamber has ever expelled a member solely on grounds of disability. Voluntary resignation is the only realistic remedy, which means the institution is held hostage to the individual’s own judgment about their condition.

Article I, Section 5 gives each chamber authority to “punish its Members for disorderly Behaviour, and, with the Concurrence of two thirds, expel a Member.” Beyond this constitutional floor, Congress regulates its own members through internal rules, the House Ethics Committee, the Senate Select Committee on Ethics, and statutes passed by the members themselves. The investigators, the prosecutors, and the judges are all colleagues of the person under review, with reelection relationships, institutional loyalties, and shared interests in the continued flow of political donations that bind them to one another in ways that do not characterize independent regulatory bodies.

The STOCK Act of 2012 was passed after media reports revealed that members of Congress were trading stocks in industries they regulated and on which they received confidential briefings. A 2011 paper by economists Ziobrowski, Boyd, Cheng, and Ziobrowski found that U.S. senators’ personal stock portfolios outperformed the market by approximately 12 percent annually from 1993 to 1998, a result consistent with trading on non-public information. The STOCK Act required public disclosure of trades within 45 days. Enforcement has been limited to small fines for late disclosure. A 2022 analysis by Business Insider identified more than 50 members of Congress who had failed to disclose trades on time, with fines of $200 per violation, a sum that has no deterrent effect on a member with a multi-million-dollar investment portfolio.

The ethics committees illustrate the self-policing problem. The House Committee on Ethics has jurisdiction over violations of House rules, federal law as applied to members’ official conduct, and Standards of Official Conduct. Between 2009 and 2022, the committee completed investigations at a rate of roughly four to six per Congress, out of several dozen referrals and complaints. The Senate Ethics Committee has not issued a public letter of reprimand since 2008. The committees are known among members as places where investigations go to die; the institutional culture discourages vigorous enforcement because members are reluctant to create precedents that could be used against their own party’s members in the future.

When serious enforcement has occurred, it has come from outside. Senator Bob Menendez of New Jersey was indicted by the Department of Justice in 2023 on bribery charges involving gold bars and cash found in his home; he was convicted by a federal jury in 2024. The Senate Ethics Committee had been aware of concerns about Menendez’s conduct for years and had previously investigated and closed matters involving him. Representative George Santos of New York was indicted in 2023 for fraud and campaign finance violations; the House voted to expel him in December 2023, only after the DOJ indictment made his presence politically untenable. Representative William Jefferson of Louisiana had $90,000 in bribe money in his freezer in 2005; the Ethics Committee did not act; he served until 2009 when he lost reelection and was later convicted by DOJ.

Article II of the Constitution opens with a deceptively simple sentence: “The executive Power shall be vested in a President of the United States of America.” That sentence has generated more than two centuries of legal and political conflict because the Constitution never defines what “the executive Power” means. The Framers left the phrase open, and the gap has never been closed.

Two broad interpretive camps have formed. Unitary executive theorists, whose intellectual lineage runs through Alexander Hamilton and received its fullest modern articulation from scholars like Steven Calabresi, argue that the Vesting Clause grants the president a residual pool of inherent power over all executive functions, including the authority to direct or remove any officer in the executive branch and to act without explicit congressional authorization wherever foreign affairs or national security are implicated. The opposing view holds that the Vesting Clause is a structural assignment, not a substantive grant, and that presidential power must be grounded in specific constitutional text or statute.

The Supreme Court’s most durable framework for resolving these disputes came from Justice Robert Jackson’s concurrence in Youngstown Sheet & Tube Co. v. Sawyer (1952). President Truman had seized the nation’s steel mills during the Korean War to prevent a strike from disrupting munitions production, acting without any statutory authorization. The Court struck down the seizure 6-3. Jackson divided presidential action into three zones: where the president acts with congressional authorization (maximum authority), where Congress is silent (a “zone of twilight” where power is uncertain), and where the president acts against Congress’s expressed will (the lowest ebb of presidential power, limited to his own constitutional authorities). Jackson’s framework is elegant, but it tells you where to look, not what you will find. Courts applying it still have to decide what the Constitution actually authorizes.

In practice, the executive branch has filled the definitional vacuum with its own legal opinions. The Office of Legal Counsel, a Justice Department component that issues authoritative interpretations of law for the executive branch, has over decades produced opinions expanding presidential authority in ways that are never reviewed by courts. OLC’s post-September 11 “torture memos,” authorizing interrogation techniques that violated statute and treaty, illustrated how internal executive branch lawyering can substitute for public law. Signing statements, used aggressively since Ronald Reagan and systematized under George W. Bush, allow presidents to record their intent not to enforce statutory provisions they consider unconstitutional. Executive orders have multiplied across administrations regardless of party: Franklin Roosevelt issued 3,728; Barack Obama issued 276; Donald Trump’s first and second terms combined for over 200.

The cumulative effect is a presidency that operates far beyond what the constitutional text specifies. In parliamentary systems such as the United Kingdom’s, executive power is defined primarily by statute, and the government’s legal authority derives from Parliament. The American system has moved toward expansive presidential power almost entirely through executive self-assertion and judicial deference rather than democratic deliberation.

Article II, Section 2 gives the president power “to grant Reprieves and Pardons for Offences against the United States, except in Cases of Impeachment.” The text imposes no conditions, no procedures, no waiting periods, and no requirement of explanation. Hamilton in Federalist 74 argued that the pardon power needed to rest in a single hand, capable of acting with “benignity” during moments of crisis, and that subjecting it to Senate confirmation would create obstruction. What Hamilton could not have anticipated was a political culture in which the pardon power would become a tool for shielding political allies from accountability.

The history of presidential pardons has always included politically charged cases. George Washington pardoned participants in the 1794 Whiskey Rebellion. Abraham Lincoln extended amnesty to large categories of Confederate soldiers. Gerald Ford’s pardon of Richard Nixon in September 1974, issued before any indictment, cost Ford the 1976 election according to most political analysts and set a precedent that sitting and former presidents might be insulated from criminal prosecution. Bill Clinton’s pardon of financier Marc Rich in January 2001, on his last day in office, came after Rich’s ex-wife donated $450,000 to the Clinton Presidential Library. A federal investigation followed but produced no charges.

The pardons issued by Donald Trump in his first term were more systematic. He pardoned Joe Arpaio, the former Maricopa County sheriff, before Arpaio had exhausted his appeals on a criminal contempt conviction, short-circuiting the judicial process entirely. He pardoned Paul Manafort, Roger Stone, Michael Flynn, and Steve Bannon, each of whom had been prosecuted by or in connection with investigations touching on Trump himself. Trump’s second term included preemptive pardons for members of his family and for participants in the January 6, 2021 Capitol riot, some issued before any charges were filed.

The self-pardon question has never reached a court. The Office of Legal Counsel issued a memo in August 1974, as Nixon’s resignation became imminent, concluding that a president cannot pardon himself because “no man can be a judge in his own case.” The memo is not binding on any court, was never litigated, and represents only one administration’s legal opinion. No constitutional provision explicitly addresses the question.

Other democracies have imposed procedural guardrails on clemency. The United Kingdom’s royal prerogative of mercy operates through the Criminal Cases Review Commission, which investigates potential miscarriages of justice before recommending clemency. Germany’s Basic Law vests clemency in state governments with formal review procedures. France requires consultation with relevant judicial bodies. The American pardon power has none of these features. The current Department of Justice pardon regulations require a five-year waiting period after conviction and a petition process, but these regulations bind only DOJ’s recommendation process. They do not bind the president.

The pre-conviction pardon is a particularly dangerous feature. Ford’s pardon of Nixon and Lincoln’s amnesty proclamations are the canonical justifications: spare the country a divisive trial, or offer reconciliation after a civil war. But both cases involved a president pardoning someone else for reasons of national interest. The abuse case is different: a president who can pardon allies before they are even charged renders investigations meaningless, because subjects know they are immunized. Pre-conviction pardons eliminate the investigative leverage, cooperation agreements, plea deals, testimony incentives, that makes federal criminal enforcement work. Trump’s pardons of January 6 participants included people who had not yet been tried; the message to future participants in political violence is that presidential allies face no legal risk. If the concern motivating a pardon is a wrongful prosecution, the remedy is the courts: dismissal, acquittal, appeal. A pardon is supposed to be mercy after the legal process has produced a result, not a shield that prevents the legal process from occurring. Requiring conviction before a pardon can issue also solves the specificity problem: pardons would apply only to the specific crimes of conviction, not to blanket categories of conduct.

The Constitution divides military authority with apparent clarity. Article I gives Congress the power “to declare War, grant Letters of Marque and Reprisal, and make Rules concerning Captures on Land and Water,” as well as the power to raise and support armies and provide and maintain a navy. Article II makes the president “Commander in Chief of the Army and Navy.” The Framers understood these provisions as a check: Congress authorizes war, the president conducts it. That understanding has been almost completely reversed in practice.

The last formal declaration of war was against Romania, Bulgaria, and Hungary in June 1942. Every armed conflict since has been conducted under some combination of congressional authorizations short of a declaration, United Nations resolutions, inherent executive authority claims, or outright presidential unilateralism. The Korean War began in 1950 as a “police action” authorized by President Truman on the basis of a UN Security Council resolution, without congressional approval. Congress never declared war. Approximately 36,000 Americans died. Vietnam escalated through the Gulf of Tonkin Resolution of 1964, a joint resolution that authorized force but was not a declaration of war; Congress repealed it in 1971 while the war continued. Kosovo in 1999: President Clinton ordered 78 days of NATO airstrikes without congressional authorization; the House voted against authorizing the operation but also voted against requiring withdrawal. Libya in 2011: President Obama initiated airstrikes, and when the 60-day clock under the War Powers Resolution expired, his administration argued that the air campaign did not constitute “hostilities” and therefore the Resolution did not apply.

Congress passed the War Powers Resolution in 1973, over Nixon’s veto, as a structural constraint. The Resolution requires the president to notify Congress within 48 hours of introducing forces into hostilities and to withdraw them within 60 days (plus 30 days for withdrawal) absent congressional authorization. Every president from Nixon forward has declared the Resolution unconstitutional as an infringement on commander-in-chief authority. No president has ever acknowledged that the 60-day clock has run out. Courts have consistently declined to adjudicate the dispute on political question grounds.

The 2001 Authorization for Use of Military Force, passed three days after the September 11 attacks, authorized the president to use force against those who “planned, authorized, committed, or aided” the attacks. It contained no geographic limitation and no sunset clause. As of 2024, the executive branch has invoked it to justify operations in at least 22 countries, including Somalia, Libya, and Syria, against groups that did not exist in 2001. The 2002 AUMF authorizing force against Iraq has also never been formally repealed, despite the government it targeted having ceased to exist more than 20 years ago.

The National Emergencies Act of 1976 was Congress’s attempt, in the post-Watergate reform era, to impose order on an accumulation of emergency powers that had grown unchecked since the New Deal. Before 1976, emergency declarations from as far back as 1933 remained legally active. The Act established a formal declaration process, required annual renewal, and created a legislative veto mechanism allowing Congress to terminate emergencies by concurrent resolution. It was supposed to be a framework for constraint. Instead it became a framework for proliferation.

The legislative veto provision was effectively eliminated by INS v. Chadha (1983), in which the Supreme Court ruled that Congress cannot act with legal effect through a single chamber or without presenting legislation to the president. After Chadha, Congress can only terminate a national emergency through a joint resolution, which the president can veto, requiring a two-thirds override. No emergency declared by a sitting president has ever been terminated by Congress over that president’s objection. As of 2024, more than 40 national emergencies were active. Several dated to the 1970s. Jimmy Carter’s 1979 declaration regarding Iran, triggered by the hostage crisis, remains in effect.

A single emergency declaration can unlock dozens of statutory authorities. The International Emergency Economic Powers Act (IEEPA), enacted in 1977, allows the president to regulate international financial transactions, freeze assets, impose sanctions, and restrict trade during any declared national emergency involving an “unusual and extraordinary threat” from abroad. IEEPA has been used to impose sanctions on Iran, Russia, Venezuela, and dozens of other countries, and its scope has expanded with each administration. The Stafford Act governs disaster declarations, which activate federal disaster relief funds; as of the early 2020s, major disaster declarations were being issued at a rate approaching 60 per year, compared to roughly 25 per year in the 1980s.

The 2019 border wall episode illustrated how emergency powers can circumvent the appropriations process directly. Congress, including a Republican-controlled Senate, declined to appropriate the $5.7 billion Trump requested for a border wall in the fiscal year 2019 budget. Trump declared a national emergency under the National Emergencies Act, then invoked a separate statute, 10 U.S.C. § 2808, allowing the Secretary of Defense to redirect military construction funds to “construction projects necessary to support the use of the armed forces.” Approximately $3.6 billion in military construction funds were redirected, including projects previously funded for schools on military bases in Kentucky, Virginia, and Arizona. Courts issued conflicting rulings; the Supreme Court allowed the transfers to proceed while litigation continued.

The legal authority to classify national security information does not appear in any statute. It derives entirely from a series of presidential executive orders, the current governing order being Executive Order 13526, issued by President Obama in December 2009. The order establishes classification levels (Confidential, Secret, Top Secret), authorizes certain executive branch officials to make original classification decisions, sets rules for derivative classification, and creates a framework for declassification review. Because it is an executive order, the president can modify or revoke it unilaterally, without congressional input. The entire architecture of the nation’s classification system rests on a legal foundation that any president can dissolve with a signature.

The scale of the system is staggering. The Information Security Oversight Office (ISOO), which monitors classification activity across the executive branch, reported approximately 50 million classification decisions in fiscal year 2022 alone. The ISOO’s own annual reports consistently describe over-classification as a systemic problem: officials classify information that does not meet the legal standard for classification, often to avoid embarrassment or legal liability rather than to protect genuine national security equities. A 2012 Public Interest Declassification Board report found that “the current system for classifying and declassifying information does not serve the American people or the national security interests of the United States.”

The president’s declassification authority is equally unconstrained. Under EO 13526, the president is the ultimate classification authority and can declassify any information at any time. This has generated genuine legal confusion in cases where presidents claimed to have declassified information informally. During the Mar-a-Lago documents investigation, Trump’s lawyers initially argued that he had declassified documents through a “standing order” that automatically declassified any documents removed from the White House. No written record of such an order existed. Federal prosecutors charged Trump under the Espionage Act and obstruction statutes rather than any classification-specific criminal law, because no such law governs presidential conduct with classified materials.

Congress’s ability to compel declassification is limited. Intelligence authorization acts routinely include provisions requesting or directing declassification of specific materials, but courts have generally treated classification as a core executive function that Congress cannot mandate. The result is that members of Congress with oversight responsibilities routinely make decisions about intelligence programs and authorization levels without access to relevant classified information, and the public cannot verify whether declassification requests are being honored in substance or only in form.

The United Kingdom, Canada, and Australia all have statutory frameworks governing their classification systems, with independent oversight bodies that have binding authority to review and overturn classification decisions. The U.S. has the ISOO, which can report but cannot compel.

The federal conflict-of-interest statute, 18 U.S.C. § 208, prohibits executive branch officials from participating in government matters that affect their financial interests. The president and vice president are explicitly exempted. The exemption reflects a structural argument: the president’s duties are so comprehensive that any meaningful financial interest rule would paralyze governance. But it makes the president the only federal official who can legally make official decisions in which they have a direct personal financial stake. There is no legal requirement for the president to place assets in a blind trust, only a tradition followed by every president from Carter through Obama. The Office of Government Ethics can issue advisory opinions and refer matters to the Department of Justice, but it cannot compel divestiture, impose penalties, or enforce its own recommendations.

The Constitution does address one dimension of presidential financial conflicts, and even that has proven unenforceable. The Foreign Emoluments Clause (Article I, Section 9) prohibits any officeholder from accepting payments from foreign states without congressional consent. The Domestic Emoluments Clause (Article II, Section 1) prohibits the president from receiving compensation beyond the presidential salary from any federal or state source. But “emolument” has no settled judicial definition, Congress never established a consent mechanism for the Foreign Emoluments Clause, and no implementing statute defines covered transactions or designates an enforcement body.

Trump’s decision not to divest from the Trump Organization when he took office in January 2017 tested both the statutory and constitutional frameworks simultaneously. Instead of a blind trust, he placed his business interests in a revocable trust managed by his sons, with himself as sole beneficiary. OGE director Walter Shaub resigned in July 2017 citing frustration with the administration’s approach. Trump Organization properties including Trump International Hotel in Washington and Mar-a-Lago received payments from foreign governments, political campaigns, PACs, and domestic businesses seeking presidential favor. Three major emoluments lawsuits followed: CREW in the Southern District of New York, Blumenthal and 200+ members of Congress in D.C., and Maryland/D.C. over the hotel. The D.C. Circuit dismissed the congressional case for lack of standing. The Fourth Circuit dismissed Maryland/D.C. after Trump left office. The SDNY case evaporated. No court ruled on the merits.

The result is a two-layer failure. The statutory layer (conflict-of-interest law) explicitly exempts the president. The constitutional layer (Emoluments Clauses) lacks definitions, enforcement mechanisms, and a consent process. State governors are subject to stronger ethics rules than the president of the United States. The United Kingdom’s ministerial code requires disclosure and approval of outside income. Canada’s Conflict of Interest Act defines prohibited benefits and establishes an independent commissioner with investigation authority. The American system has nothing comparable for its most powerful office.

The phrase “high Crimes and Misdemeanors” in Article II, Section 4 derives from English parliamentary practice reaching back to 1386, when the House of Commons impeached William de la Pole, Duke of Suffolk, for “high crimes and misdemeanors” in his handling of English territories in France. The phrase carried a specific meaning in English constitutional history: grave abuses of public trust by officials in positions of power, not necessarily violations of the criminal law. James Madison at the Philadelphia Convention argued that impeachment was necessary to address situations where the executive “might pervert his administration into a scheme of peculation or oppression.” Hamilton wrote in Federalist 65 that impeachable offenses are those “which may with peculiar propriety be denominated POLITICAL, as they relate chiefly to injuries done immediately to the society itself.”

In practice, the standard has become whatever the House majority decides. Gerald Ford, then House Minority Leader, stated in 1970 during an unsuccessful attempt to impeach Supreme Court Justice William O. Douglas: “An impeachable offense is whatever a majority of the House of Representatives considers it to be at a given moment in history.” Ford’s formulation was offered as a political argument rather than a constitutional one, but subsequent history has largely vindicated his description of how impeachment actually works.

There have been four serious presidential impeachment proceedings. Andrew Johnson was impeached in 1868 for violating the Tenure of Office Act, a statute later held unconstitutional; he was acquitted in the Senate by one vote. Richard Nixon resigned in August 1974 before the full House voted on articles of impeachment that the Judiciary Committee had approved for obstruction of justice, abuse of power, and contempt of Congress; this is the closest any president has come to removal. Bill Clinton was impeached in December 1998 for perjury and obstruction of justice arising from testimony about his relationship with Monica Lewinsky; the Senate acquitted him, with no Democratic senator voting to convict. Trump was impeached twice: in December 2019 for abuse of power and obstruction of Congress in connection with pressure on Ukraine, and in January 2021 for incitement of insurrection following January 6. He was acquitted both times, with the 2021 trial producing 57 votes for conviction, a majority but short of the two-thirds required.

The partisan pattern across these cases is consistent: not a single president has been convicted by the Senate. Every acquittal has followed party lines with at most marginal defections. The two-thirds threshold in the Senate, combined with the political costs of crossing a president from one’s own party, means that impeachment functions as an accountability mechanism only for presidents who have already lost the support of their own political coalition.

Before 1978, presidential papers were treated as the personal property of presidents. Franklin Roosevelt donated his papers to the federal government voluntarily; Harry Truman took his with him to Independence, Missouri. Richard Nixon negotiated a deed of gift agreement with the General Services Administration that would have given him control over his White House records, and he attempted to have the tapes subpoenaed by the Watergate special prosecutor destroyed. Congress intervened in 1974, passing the Presidential Recordings and Materials Preservation Act specifically to preserve Nixon’s records, and then in 1978 enacted the Presidential Records Act as a permanent framework.

The PRA established that records created or received by the president and White House staff in the conduct of their official duties belong to the United States government from the moment of creation. When a president leaves office, the National Archives and Records Administration takes custody. The Act restricts public access for five years, after which the records are subject to the Freedom of Information Act, with additional restrictions for certain categories (national security, personal privacy) potentially lasting up to twelve years.

The Mar-a-Lago case exposed two significant weaknesses in the PRA framework. First, the Act has no criminal penalty specifically applicable to the president for mishandling presidential records. When federal investigators found hundreds of classified documents at Trump’s Palm Beach residence in August 2022, prosecutors did not charge violations of the PRA. They relied on the Espionage Act (18 U.S.C. § 793), which prohibits unauthorized retention of national defense information, and obstruction statutes. The Espionage Act was designed for espionage and sabotage cases; using it to address a former president’s failure to return government documents to the Archives creates legal and precedential complications the law was not designed to handle.

Second, the PRA’s definition of presidential records has been repeatedly contested at the margins. The deletion of emails, the use of personal phones and encrypted messaging applications, and the destruction of physical documents all test boundaries that the Act does not clearly resolve. The use of Signal and other disappearing-message applications by senior officials, which became a significant issue in early 2025 when a journalist was inadvertently added to a Signal group where senior Trump administration officials discussed military strikes on Yemen, highlighted the gap between the PRA’s archiving requirements and modern communications practices.

The Federal Records Act, which governs records for the rest of the executive branch, includes criminal penalties for unauthorized destruction of records. Its coverage of the presidency is separate and explicitly addressed by the PRA, creating a two-track system in which the president is governed by weaker rules than a mid-level agency employee.

The Twenty-Fifth Amendment, ratified in 1967 in the wake of John F. Kennedy’s assassination and concern about the absence of a constitutional mechanism for handling presidential disability, established the first clear procedures for presidential succession and temporary incapacity. Section 3 allows the president to voluntarily transfer power to the vice president by transmitting a written declaration to the Speaker of the House and the President pro tempore of the Senate. Section 4 allows the vice president and a majority of “the principal officers of the executive departments,” the cabinet, to declare the president unable to discharge his powers and duties, with the vice president assuming those duties as acting president. If the president contests the declaration, Congress decides the matter, with a two-thirds vote of both chambers required to sustain the incapacity finding.

Section 3 has been used three times: by George W. Bush before colonoscopies in June 2002 and July 2007, and by Joe Biden before a colonoscopy in November 2021. In each case, the president was unconscious for less than two hours and reclaimed his powers the same day. Section 4 has never been invoked for actual incapacity, and the gap between what the Amendment provides on paper and what it could accomplish in practice is substantial.

The historical episodes that prompted the Amendment illustrate why the gap matters. In July 1893, President Grover Cleveland secretly had a cancerous growth removed from his jaw aboard a yacht in the East River, concealing the operation and his recovery from the public and Congress for years. In October 1919, Woodrow Wilson suffered a massive stroke that left him partially paralyzed and cognitively impaired. Wilson’s wife, Edith Wilson, and his physician, Dr. Cary Grayson, managed access to the president for the remaining seventeen months of his term. Cabinet members were denied meetings. Legislation went unsigned. Secretary of State Robert Lansing convened informal cabinet meetings to manage government business and was later dismissed by Wilson, who accused him of usurping presidential authority.

The 1981 assassination attempt on Ronald Reagan tested the Amendment’s procedures under real pressure. Reagan underwent surgery at George Washington University Hospital for several hours. Vice President George H.W. Bush was traveling. Secretary of State Alexander Haig told reporters “I am in control here,” a statement that was legally incorrect, created public confusion, and illustrated the absence of any agreed protocol for public communication during a presidential incapacity. The Twenty-Fifth Amendment was not invoked.

Section 4 contains a provision allowing Congress to designate “such other body as Congress may by law provide” as an alternative to the cabinet for making an incapacity determination, recognizing that a president might be able to fire cabinet members who moved against them. Congress has never created this body. A president facing a Section 4 action could fire any cabinet member who voted to declare incapacity before the process was completed, potentially eliminating the majority needed to sustain the finding.

The Constitution says little about what happens when both the president and vice president are unable to serve. Article II, Section 1 authorizes Congress to “declare what Officer shall then act as President,” and Congress has exercised that authority three times, most recently in the Presidential Succession Act of 1947. Under that statute, the line runs: Speaker of the House, President pro tempore of the Senate, then cabinet officers in order of their departments’ creation, beginning with the Secretary of State. At first glance this looks like a sensible belt-and-suspenders arrangement. On examination, it has serious structural problems.

The constitutional defect starts with the word “Officer.” Article II uses that term to mean executive branch officers, and constitutional scholars including Akhil Reed Amar and Vikram Amar have argued at length that members of Congress are not “Officers” in the relevant sense, because they hold legislative rather than executive commissions. If correct, placing the Speaker and President pro tempore ahead of the cabinet was unconstitutional from the start. The 1947 Act reversed the 1886 Presidential Succession Act, which had placed only cabinet members in the line, precisely because President Truman wanted an elected official closer to the presidency. That political preference overrode the constitutional question, and the question has never been resolved by a court.

The deeper problem is simultaneous incapacitation. During the Cold War, continuity of government planning assumed nuclear decapitation scenarios. September 11 made the concern concrete and domestic: all four hijacked aircraft were in the air simultaneously, and had even one reached the Capitol during a joint session, the constitutional succession chain could have been broken or thrown into chaos. The 2003 Continuity of Government Commission, co-chaired by former White House Counsel Lloyd Cutler and former Attorney General Griffin Bell, concluded that the current line was constitutionally doubtful and operationally dangerous, and recommended replacing congressional officers with a deeper bench of pre-confirmed executive officers. Congress did not act.

There is also a partisan mischief problem the 1947 Act authors did not fully reckon with. In a divided government scenario, the Speaker of the House may be from the opposing party. If a president and vice president both died in office with a hostile Speaker next in line, the result would be a mid-term transfer of the executive branch to the opposing party, without any election. In January 2021, the Speaker was from a different party than the president, and the President pro tempore was 87 years old. The line of succession at that moment was constitutionally questionable, operationally fragile, and politically anomalous, all at once.

The Appointments Clause of Article II, Section 2 establishes two tracks for filling executive positions. Principal officers must be nominated by the president and confirmed by the Senate. Inferior officers may be appointed by the president alone, by heads of departments, or by courts of law, if Congress vests that power by statute. The framework looks clean on paper. In practice, the line between “principal” and “inferior” has been litigated for decades, the number of Senate-confirmed positions has grown to roughly 1,200, and the confirmation process has become a sustained institutional dysfunction that leaves the executive branch chronically understaffed at its upper levels.

The Supreme Court has tried to define the principal/inferior distinction without producing a stable test. In Edmond v. United States (1997), the Court held that Coast Guard Court of Criminal Appeals judges were inferior officers because their decisions were subject to review by a principal officer, and because they had limited jurisdiction and tenure. Justice Souter’s concurrence warned that the majority’s reasoning would be hard to apply consistently, and subsequent decisions have confirmed that prediction.

The sheer volume of Senate-confirmed positions has made the confirmation process a bottleneck that cripples new administrations. Under Ronald Reagan, the average time from nomination to confirmation was roughly two months. By the Obama administration it had stretched to over five months; under Biden it exceeded six months for many positions. A study by the Partnership for Public Service found that at the one-year mark of the Biden administration, roughly 40 percent of Senate-confirmed positions had not yet been filled. Agencies run on acting officials, career staff in temporary roles, or simply without leadership in key offices.

The consequences are concrete. The Federal Aviation Administration went without a Senate-confirmed administrator for nearly two years during a period that included the 737 MAX certification crisis. The Bureau of Alcohol, Tobacco, Firearms and Explosives went without a Senate-confirmed director for seven years between 2006 and 2013, operating under a series of acting directors during a period that included Fast and Furious. Agencies that lack confirmed leadership are more legally vulnerable, because adverse parties can challenge actions taken by officials whose authority is contested, and more practically constrained, because career staff are often reluctant to pursue major initiatives without confirmed principals to own the decision.

The Senate’s ability to block nominations through indefinite inaction is itself a structural defect. The Constitution says the president “shall nominate, and by and with the Advice and Consent of the Senate, shall appoint” officers. “Advice and Consent” implies a deliberative process that produces a decision, not an indefinite pocket veto. The Merrick Garland episode, in which the Senate refused to hold a hearing for eight months, demonstrated that nothing in the current system compels the Senate to act at all. The same dynamic plays out at lower levels: the ATF went seven years without a confirmed director not because the Senate rejected nominees but because it declined to vote on them.

When a Senate-confirmed position becomes vacant, the Federal Vacancies Reform Act of 1998 gives the president three options for filling it on a temporary basis: install a first assistant, designate a senior official already confirmed to a different position, or designate a senior employee who meets certain experience thresholds. Acting officials can serve for 210 days, with extensions available when a nomination is pending. Creative designation practices have repeatedly stretched the statute beyond its evident purpose. In November 2018, Matthew Whitaker was named Acting Attorney General despite having served only as chief of staff to Attorney General Sessions, a position that was not itself Senate-confirmed. The Supreme Court addressed some workarounds in NLRB v. SW General (2017), but administrations have continued to use “performing the functions of” designations and duty reassignments to install preferred officials without triggering the Act’s limits.

The Trump administration’s use of acting officials was extensive enough to constitute a governing strategy. The Department of Homeland Security cycled through acting secretaries for significant stretches of 2019 and 2020, during major immigration enforcement operations, the COVID-19 pandemic response, and the 2020 election security apparatus. The Intelligence Community had an acting Director of National Intelligence for an extended period. The accountability deficit is structural: acting officials can be replaced instantly if they displease the president, making them functionally more loyal than confirmed officers whose removal carries political costs.

The Recess Appointments Clause (Article II, Section 2) was designed as a safety valve for the same problem: ensuring the executive branch could function when the Senate was genuinely unavailable. Both parties escalated its use, with Clinton making 139 recess appointments and Bush 171. The Senate responded with pro forma sessions, brief meetings with no business conducted, specifically to prevent recesses long enough to trigger the power. In NLRB v. Noel Canning (2014), a unanimous Court held that pro forma sessions count as sessions and that a recess must last at least ten days. The practical consequence is that the Senate can block all recess appointments by holding pro forma sessions every three days, which requires only one senator to be physically present. A constitutional power the framers included to ensure executive branch functionality has been nullified by a procedure that takes thirty seconds of Senate floor time.

The two mechanisms, the Vacancies Reform Act and the Recess Appointments Clause, are responses to the same underlying dysfunction: the Senate confirmation process has become so slow and so weaponized that the executive branch cannot reliably staff itself through the constitutional process. Acting officials and recess appointments are both workarounds for a broken front door. The confirmation bottleneck drives the demand for acting officials; the Senate’s procedural blockade of recess appointments closes the safety valve that might have relieved the pressure. The result is an executive branch in which officials exercising cabinet-level authority may never have faced a Senate vote, while the constitutional mechanism designed to prevent exactly that situation has been rendered a dead letter.

The Constitution says nothing explicit about the president’s power to remove executive officers. Article II vests the executive power in the president and requires faithful execution of the laws, from which the Supreme Court has inferred a broad removal power, but the inference has been contested for a century.

The foundational case is Myers v. United States (1926), in which Chief Justice Taft held that the president has an inherent constitutional power to remove executive officers, and that a statute conditioning a postmaster’s removal on Senate consent was unconstitutional. Nine years later, in Humphrey’s Executor v. United States (1935), the Court upheld for-cause removal protections for Federal Trade Commission commissioners, holding that Congress could insulate “quasi-legislative” and “quasi-judicial” officers from at-will presidential removal. Humphrey’s Executor became the constitutional foundation for the modern independent agency, allowing Congress to create entities like the SEC, FTC, FCC, NLRB, and Federal Reserve Board with commissioners and governors who could not be fired simply because a new president wanted loyalists in those seats.

The Roberts Court has been chipping away at Humphrey’s Executor without overruling it directly. In Seila Law v. CFPB (2020), the Court held that a for-cause removal protection for the single director of the Consumer Financial Protection Bureau was unconstitutional, distinguishing Humphrey’s Executor on the grounds that it involved a multi-member commission rather than a single director. In Collins v. Yellen (2021), the same logic applied to the Federal Housing Finance Agency director. The Court’s reasoning creates a structural test that turns on whether an agency is led by one person or multiple commissioners, a distinction that is hard to defend as a matter of constitutional principle.

The practical stakes are considerable. The Federal Reserve Board of Governors has for-cause removal protections, and the Fed’s ability to conduct monetary policy independently of political pressure depends substantially on markets believing those protections are real. In 2025, public statements from the Trump administration questioning whether the president could remove Fed Chair Jerome Powell sent bond markets into brief turbulence and the dollar down, because investors understood that a politically controllable central bank is a different kind of institution than an independent one. The FTC, SEC, NLRB, and FCC face the same ambiguity. The constitutional status of the independent agency model, which has been the dominant structure for federal economic regulation for ninety years, is now genuinely in doubt.

The right to organize and strike is constitutionally protected in Germany, France, and most EU member states. In the United States, it rests entirely on the National Labor Relations Act of 1935, an ordinary statute whose enforcement depends on the NLRB, the same agency whose constitutional status is now uncertain after Seila Law. The US has not ratified several ILO core labor conventions that virtually every other developed nation has ratified, including those covering freedom of association and collective bargaining. Janus v. AFSCME (2018) further weakened organized labor by holding that public-sector unions cannot collect fees from non-members who benefit from collective bargaining agreements, reducing union financial capacity across states where public-sector unionism had been strongest. The result is a labor rights framework with no constitutional anchor, enforced by an agency of questionable independence, and progressively constrained by judicial doctrine.

The merit-based federal civil service dates to the Pendleton Civil Service Reform Act of 1883, passed after the assassination of President Garfield by a disappointed office-seeker, as a direct repudiation of the spoils system. The Hatch Act of 1939 added restrictions on federal employees’ partisan political activity, prohibiting most civilian employees from running for partisan office, soliciting political contributions from subordinates, and using their official authority to influence elections. Together, the Pendleton and Hatch frameworks created a professional career service of roughly 2.1 million civilian employees who are hired on merit, retain their positions across administrations, and are insulated from partisan pressure.

Schedule F, established by executive order in October 2020 and rescinded by President Biden in January 2021, would have reclassified federal employees in “policy-related” positions from the competitive service to a new “Schedule F” excepted service category, stripping them of civil service protections and making them at-will employees dismissible for any reason. The Office of Personnel Management estimated that between 50,000 and 100,000 positions could qualify. The Trump administration reinstated Schedule F by executive order in January 2025, and litigation over its scope and legality was proceeding in federal court as of early 2026.

The significance of Schedule F is about what kind of institution the federal government is. A civil service in which policy-related positions are at-will appointments becomes, over time, a civil service in which policy positions are staffed by political loyalists rather than professional experts. The IRS’s examination of tax returns, the FDA’s drug approval process, the EPA’s environmental impact assessments, the Social Security Administration’s disability determinations: all of these involve “policy-related” judgments by career employees. Reclassifying those employees as at-will does not immediately change any specific decision, but it changes the incentive structure within which decisions are made.

The historical record of pre-Pendleton spoils-system governance is instructive. Customs houses in New York City, the largest source of federal patronage in the 19th century, were reliably understaffed, corrupt, and inefficient because their employees owed their positions to political sponsors rather than competence. Chester Arthur, who became a civil service reformer after having been the collector of the New York Custom House under the spoils system, understood from personal experience that patronage-based agencies produce worse outcomes than merit-based ones. The post-Pendleton federal workforce built the Interstate Highway System, won two world wars, ran the Marshall Plan, and developed the internet.

The Hatch Act’s restrictions on political activity have also been under pressure. Multiple Trump administration officials, including Kellyanne Conway, received formal findings of Hatch Act violations from the Office of Special Counsel, an independent agency charged with enforcing the Act, and faced no meaningful consequences.

The Inspector General Act of 1978 was passed in the aftermath of Watergate, the Pentagon Papers, and a series of revelations about fraud and mismanagement across federal agencies. Congress concluded that internal agency management could not be trusted to police itself, and created a network of Inspectors General with authority to audit agency programs, investigate fraud and waste, and report findings both to agency heads and directly to Congress. There are now 74 federal Inspectors General, conducting roughly 15,000 audits and investigations annually and generating recoveries and savings that consistently exceed their combined budgets by a factor of several times.

The independence of IGs rests on two mechanisms. First, they are appointed by the president and confirmed by the Senate for establishment act IGs, which limits purely political appointments to those willing to withstand Senate scrutiny. Second, the IG Act Reform Act of 2008 required the president to give Congress 30 days’ notice before removing an IG, a provision added specifically in response to concerns about retaliatory firings. The 30-day notice requirement gives Congress time to investigate and publicize a firing but does not actually prevent it.

The weakness of that protection became visible in April 2020, when President Trump fired five Inspectors General within a period of a few weeks. The firings included Michael Atkinson, the Intelligence Community Inspector General who had transmitted to Congress the whistleblower complaint about Trump’s call with Ukrainian President Zelensky, the complaint that triggered the first impeachment. Atkinson’s firing was widely understood as retaliation. The other firings included Glenn Fine, who had been named chair of the Pandemic Response Accountability Committee charged with overseeing COVID-19 relief spending, and was removed before that oversight function could fully begin.

Congress responded with the Inspector General Independence and Empowerment Act of 2022, which tightened the notice requirement and required a written explanation of the reasons for removal. The explanation requirement is an improvement over the 2008 statute but does not impose a for-cause standard; a president can comply by providing a written explanation that amounts to a pretextual reason. The courts have not been asked to review whether a stated reason for IG removal is genuine.

The structural problem is that IGs derive their utility from their willingness to investigate and report on the agencies and officials they oversee, but their tenure depends on the goodwill of the president, who is the ultimate supervisor of all executive branch officials. An IG who knows that a politically sensitive investigation could lead to termination faces incentives that compromise independence even without any direct threat.

The National Security Council was created by the National Security Act of 1947, the same statute that established the CIA, unified the military under a Secretary of Defense, and restructured the national security architecture for the Cold War. The statute defines the NSC’s statutory membership: the president, the vice president, the Secretary of State, and the Secretary of Defense, with the Chairman of the Joint Chiefs of Staff and the Director of National Intelligence as statutory advisors. Every member of the statutory NSC either holds a Senate-confirmed position or is the elected president or vice president. That layer of accountability is the only layer the statute provides, because the statute says almost nothing about the NSC staff, the National Security Advisor, or the process by which NSC decisions are made and implemented.

The NSC staff has grown from a small secretariat in the Truman years to an operation of 200 to 400 people under recent administrations, drawn primarily from the military services, intelligence agencies, and State Department, supplemented by political appointees. The National Security Advisor, who chairs NSC meetings in the president’s absence, coordinates interagency policy processes, manages the staff, and in practice shapes the national security agenda, is appointed by the president without Senate confirmation and serves entirely at the president’s pleasure. There is no statutory job description, no Senate confirmation process, no legal obligation to testify before Congress, and no formal accountability structure beyond the president’s personal confidence.

The individuals who have held this position have included some of the most consequential foreign policy figures in postwar American history. Henry Kissinger ran American foreign policy toward China, Vietnam, and the Soviet Union from the NSC before moving to State. Zbigniew Brzezinski shaped the Carter administration’s response to the Soviet invasion of Afghanistan and managed the Iranian hostage crisis diplomacy from the NSC. John Bolton, who served for 17 months without Senate confirmation, engaged in direct diplomatic communications with foreign governments and shaped the administration’s Iran policy, before departing in acrimony and publishing a book alleging presidential misconduct. None of these officials was answerable to Congress in any formal sense.

The Iran-Contra affair illustrated what happens when NSC staff accountability is absent entirely. Oliver North, a Marine lieutenant colonel detailed to the NSC staff, operated a covert arms sale to Iran and used the proceeds to fund Nicaraguan Contra rebels, both in violation of congressional restrictions. North was not a Senate-confirmed officer; he operated within the NSC staff structure that the 1947 Act left unspecified; and his activities proceeded for years before the Tower Commission investigation uncovered them. The Tower Commission report, released in February 1987, explicitly criticized the NSC staff structure for enabling the operation and recommended reforms. Most of those reforms were not codified in statute.

The nondelegation doctrine holds that Congress, as the branch vested with “all legislative Powers” under Article I, cannot hand that power off to executive agencies. The rule has deep roots in constitutional theory and was affirmed in early Supreme Court decisions. In practice, however, Congress has spent the better part of a century doing exactly that, passing broad enabling statutes and leaving agencies to fill in almost every substantive detail through rulemaking. The Clean Air Act tells the EPA to set standards “requisite to protect the public health.” The Securities Exchange Act tells the SEC to adopt rules “necessary or appropriate in the public interest.” The actual content of American environmental law, financial regulation, and workplace safety law is written not by elected legislators but by career officials and political appointees in agencies that are constitutionally one step removed from democratic accountability.

The last time the Supreme Court struck down a federal statute on nondelegation grounds was 1935, in A.L.A. Schechter Poultry Corp. v. United States, which voided a provision of the National Industrial Recovery Act that let trade associations write binding “codes of fair competition” for entire industries. In the nine decades since, the Court has sustained delegation after delegation by invoking the “intelligible principle” test: if Congress supplies any discernible standard, the delegation is permissible. That test has been so permissive it is nearly a nullity.

Gundy v. United States (2019) came close to changing this. Four justices signaled they wanted to revive a stricter nondelegation doctrine; a fifth, Justice Alito, said he would reconsider if a majority were willing. Instead, in West Virginia v. EPA (2022), the Court developed the “major questions doctrine,” which holds that when agencies assert authority over decisions of vast economic and political significance, they must point to clear congressional authorization. The doctrine is a functional constraint on sweeping agency action, but it is applied case by case and provides no ex ante boundary.

The Administrative Procedure Act’s notice-and-comment rulemaking process is the primary procedural check on agency lawmaking. Agencies must publish proposed rules, accept public comment, and respond to significant objections. This generates a record and enables judicial review, but it does not require Congress to make the underlying policy choices. Agencies still determine whether to regulate, how stringently, and in favor of which interests. The result is a system in which the branch that most directly affects daily life derives its authority from statutes that often say little more than “regulate in the public interest.”

Article III of the Constitution vests “the judicial Power of the United States” in courts whose judges hold life tenure and salary protection, guarantees designed to ensure independence from political pressure. Yet the federal government adjudicates an enormous fraction of legally consequential disputes entirely outside Article III courts. The Social Security Administration processes roughly 700,000 disability hearings per year before Administrative Law Judges (ALJs). The Executive Office for Immigration Review (EOIR), sitting inside the Department of Justice, carries a backlog that exceeded 3.7 million pending cases as of late 2024. The Securities and Exchange Commission can bring enforcement actions in its own administrative tribunals rather than federal district court.

The legitimacy of this system depends on the independence of the judges who staff it. But administrative adjudication has a structural problem: the agency is simultaneously prosecutor, rulemaker, and judge in the same proceedings. Before Lucia v. SEC (2018), the SEC appointed its ALJs through staff-level hiring processes rather than by the agency head, a practice the Supreme Court held unconstitutional because ALJs exercise significant authority and are therefore “Officers of the United States” subject to the Appointments Clause. Lucia cleaned up one procedural defect, but it did not resolve the deeper due process question about structural bias when an agency adjudicates its own enforcement cases.

Immigration courts illustrate the problem at scale. Unlike Article III judges or even ALJs in most agencies, immigration judges are employees of the Department of Justice, a law enforcement agency that is simultaneously responsible for prosecuting removal cases. The Attorney General can certify immigration court decisions to himself and issue precedential rulings that bind every immigration judge in the country. Attorney General Jeff Sessions did this in Matter of A-B- (2018), restricting asylum claims based on gang and domestic violence, a decision that was later vacated in federal court but that reshaped hundreds of thousands of cases while it stood. The nation’s immigration law is, in material part, whatever the sitting Attorney General decides it is.

The volume argument is real: Article III courts could not absorb millions of Social Security and immigration adjudications. But volume does not justify eliminating structural independence. A system in which legal rights are determined by adjudicators who report to the enforcing agency, under doctrines shaped by the agency itself, is not a court system in any meaningful sense.

The Constitution creates three branches and assigns each a distinct function: Congress legislates, the President executes, courts adjudicate. The modern administrative state does not fit this scheme. A single agency routinely exercises all three functions: it promulgates binding rules (legislative), brings enforcement actions (executive), and adjudicates disputes (judicial). The Federal Trade Commission can define “unfair methods of competition” by rulemaking, investigate and prosecute companies under that definition, and adjudicate the resulting cases in its own proceedings. The founders’ separation of powers was designed precisely to prevent this concentration, but the consolidation happened gradually enough that no single moment triggered a decisive constitutional check.

The Administrative Procedure Act of 1946 was Congress’s attempt to impose procedural regularity on the agencies that had proliferated under the New Deal. For decades, the courts’ approach to judicial review was governed by Chevron U.S.A. v. Natural Resources Defense Council (1984), which instructed judges to defer to an agency’s reasonable interpretation of an ambiguous statute it administered. Chevron deference was justified on grounds of agency expertise and democratic accountability through the President, but it had the practical effect of concentrating interpretive authority in agencies and shielding their legal conclusions from independent judicial scrutiny.

In June 2024, the Supreme Court overturned Chevron in Loper Bright Enterprises v. Raimondo, holding that courts must exercise their own independent judgment in interpreting statutes, not defer to agency interpretations. Chief Justice Roberts’s majority opinion returned to the APA’s command that courts decide “all relevant questions of law.” The decision immediately unsettled decades of regulatory law: rules upholding fishing industry oversight, telecommunications policy, environmental standards, and immigration procedures all rested on statutory interpretations that had been sustained under Chevron deference. Lower courts are now working through what independent judgment means in practice when statutes are genuinely ambiguous and courts lack the technical expertise agencies possess.

The accountability gap is real regardless of how courts interpret statutes. Voters elect Congress and the President; they do not elect the Deputy Assistant Secretary for Energy Efficiency who writes the standard that determines which appliances are sold in American homes. The current post-Loper Bright moment, in which both the deference framework and the political legitimacy of administrative governance are contested simultaneously, is either an opportunity to build a more accountable system or a prelude to regulatory paralysis.

The Office of Legal Counsel is a division of the Department of Justice charged with providing authoritative legal advice to the President and executive branch agencies. Its opinions function as binding law within the executive branch: an agency that follows an OLC opinion has a complete defense to claims of legal error, and no mechanism exists for overriding an OLC opinion except a contrary opinion from the Attorney General or a new opinion from OLC itself. Courts rarely see OLC opinions because the conduct they authorize tends not to generate litigation, and because the executive branch historically has treated the opinions as privileged. OLC operates, in effect, as the executive branch’s supreme court, with jurisdiction over every legal question the branch faces, no adversarial process, no public record, and no external review.

The consequences of this arrangement became undeniable in 2002, when Deputy Assistant Attorney General John Yoo and Assistant Attorney General Jay Bybee authored a series of opinions concluding that enhanced interrogation techniques including waterboarding did not constitute torture under federal law, that the President’s Commander-in-Chief authority superseded any statutory prohibition, and that standard defenses applied to CIA interrogators. The “torture memos” were not published until their existence was revealed by press reporting in 2004. By then, the techniques had been used on dozens of detainees. The OLC opinions were subsequently withdrawn, but no criminal charges were brought against the officials who wrote or relied on them.

The pattern extends beyond interrogation policy. OLC opinions have authorized warrantless surveillance under the NSA’s Stellar Wind program (opinion kept secret for years), concluded that the President may launch military operations without congressional authorization in circumstances no statute addressed, interpreted recess appointment power in ways later rejected by the Supreme Court in NLRB v. Noel Canning (2014), and addressed the legality of drone strikes against American citizens overseas. The last category resulted in a 2010 memorandum authorizing the targeted killing of Anwar al-Awlaki, a U.S. citizen in Yemen, based on legal analysis that remained classified for four years and was released only through FOIA litigation.

OLC’s current publication policy is incomplete and inconsistent. The office publishes a selection of opinions on its website and in bound volumes, but significant opinions remain withheld, and there is no statutory requirement to publish or even to inventory what exists. An office that produces binding law governing the conduct of the most powerful government on earth, with no public record and no external check, is an accountability failure by definition.

Every administration relies on expert advice from outside the government: scientists, economists, engineers, and industry representatives who inform regulatory decisions, set research agendas, and evaluate grant applications. The Federal Advisory Committee Act of 1972 was Congress’s attempt to impose order on this process. FACA requires that advisory committees be chartered, that their membership be “fairly balanced in terms of the points of view represented,” that their meetings be open to the public, and that their records be made available. Today the General Services Administration reports approximately 1,000 active FACA committees with around 60,000 member slots.

The statute’s implementation has been uneven, and agencies have found multiple paths around it. The most direct is to rely on informal consultations that never produce a committee in the statutory sense: a cabinet secretary who holds a series of individual meetings with industry representatives, or a White House office that convenes stakeholders for “listening sessions” not subject to FACA’s open-meeting requirement. The Supreme Court addressed one such workaround in AAPS v. Clinton (1993), which involved Hillary Clinton’s health care task force.

Industry capture of formal advisory committees is well-documented. A 2006 Union of Concerned Scientists survey found that agency scientists frequently reported political interference in the composition of scientific advisory panels, including removal of members and substitution of industry-affiliated nominees. A 2006 analysis in the Journal of the American Medical Association found that more than half of voting members at ten FDA advisory committee meetings had financial ties to the company whose product was under review. FDA subsequently strengthened its conflict-of-interest waiver documentation, but a 2012 HHS Inspector General report found continued weaknesses in the process.

The National Academies of Sciences, Engineering, and Medicine occupy a separate and influential role. Their reports on topics from climate science to nuclear waste storage carry substantial scientific authority and are used by agencies and Congress to legitimize policy choices. The National Academies are not subject to FACA because they are chartered under a private congressional charter, not created by an agency. Given the weight their reports carry in regulatory proceedings, the gap between their practical authority and their formal accountability is meaningful.

The federal criminal code has expanded dramatically since the New Deal era. A 1982 Department of Justice study counted roughly 3,000 federal criminal offenses; by 2008, the Heritage Foundation and National Association of Criminal Defense Lawyers estimated the figure at over 4,450, with thousands more regulatory violations carrying criminal penalties scattered through the Code of Federal Regulations. No official, authoritative count exists, which is itself a problem. Congress adds crimes with each legislative session without repealing obsolete ones, creating a body of law so vast that no one can know its full extent.

Into this thicket, federal prosecutors carry nearly unlimited discretion. The decision to investigate, indict, offer a plea, or decline prosecution is made by United States Attorneys and their assistants with minimal external review. The Principles of Federal Prosecution, set out in the Justice Manual, instruct prosecutors to bring charges only when they believe the evidence is sufficient for conviction and that prosecution serves a “substantial federal interest,” but these are guidelines, not enforceable rules. The constitutional protection against selective prosecution requires a defendant to show both discriminatory effect and discriminatory intent, a burden so demanding that documented disparities in prosecution rates rarely satisfy it. In United States v. Armstrong (1996), the Supreme Court held that defendants seeking to prove selective prosecution cannot even obtain discovery to gather evidence of intent without first making a substantial threshold showing.

United States Attorneys are presidential appointees. The politicization risk is structural. The Trump administration’s firing of all 46 sitting U.S. Attorneys in March 2017 was legally permissible but unprecedented in its simultaneity. The subsequent investigations and prosecutions of the 2017-2021 period, the Durham investigation into the origins of the Russia probe (which resulted in one conviction and two acquittals after years and tens of millions in expenditure), the prosecution of Hunter Biden on gun and tax charges, and the federal prosecutions of Donald Trump himself all generated legitimate questions about whether charging decisions reflected the evidence or the politics. The problem is not that any particular decision was provably improper; the problem is that the structural conditions for politically motivated prosecution exist and no institutional mechanism catches or deters abuse.

A politically motivated federal prosecution, even one that ends in acquittal, can destroy a target’s finances, reputation, and professional life over years of proceedings. Federal law is both broader and harsher than most state law, and the investigative resources of the FBI, IRS, and other federal agencies amplify prosecutorial power in ways state offices cannot match.

The Freedom of Information Act was enacted in 1966 after years of pressure from journalists, congressional reformers, and advocates who argued that a democratic government owed its citizens access to the records of its operations. The statute establishes a presumption of disclosure and requires agencies to respond to requests within 20 business days, a timeline that is honored almost nowhere in the federal government. Nine statutory exemptions allow agencies to withhold records covering national security, internal personnel rules, trade secrets, inter-agency deliberations, personal privacy, law enforcement proceedings, financial regulation, oil well geology, and certain specifically exempted categories.

The gap between the statute’s promise and its reality is documented in the government’s own data. The 2023 Annual FOIA Report showed approximately 928,000 requests received government-wide, with over 200,000 pending at year-end. Average response times at major agencies ranged from several months to several years; the State Department’s average processing time for complex requests exceeded 800 days. The FBI backlog has at times stretched to five or more years. Agencies have financial and institutional incentives to delay: withheld records cannot be criticized, and requesters who wait long enough often give up. The FOIA Improvement Act of 2016 codified a “foreseeable harm” standard, requiring agencies to show that disclosure would actually cause harm rather than merely invoking an exemption categorically, but enforcement is weak because requesters must litigate to challenge wrongful withholding.

The practical burden falls heavily on the requesters the statute was designed to help. A Vaughn index, required when an agency withholds documents in litigation, is a detailed inventory explaining each exemption claimed for each withheld document or portion. Preparing and challenging a Vaughn index requires legal expertise and resources that individuals and small news organizations often lack. Major investigative stories built on FOIA requests typically required years of litigation, FOIA appeals, and in some cases parallel lawsuits under the Privacy Act and the Federal Records Act. The stories that never get reported because the requester lacks the resources to sustain a multi-year legal fight are invisible by definition.

Canada’s Access to Information Act, Sweden’s Freedom of the Press Act (dating to 1766), the United Kingdom’s Freedom of Information Act (2000), and New Zealand’s Official Information Act (1982) all offer comparison points. The U.S. system’s dysfunction is a choice, repeatedly confirmed by underfunding and narrow court interpretations, not a structural inevitability.

The Federal Reserve System was created by the Federal Reserve Act of 1913 as a response to the financial panics of the preceding decades, most recently the Panic of 1907, which had been contained only by the private intervention of J.P. Morgan. The Act created a central bank with a measure of independence from both the Treasury and elected officials: the Board of Governors serves staggered 14-year terms, members can be removed only “for cause,” and the Fed funds itself through its earnings on portfolio holdings rather than annual congressional appropriations.

The independence has always been incomplete and contested. The most consequential documented episode of political interference occurred under President Nixon, who in 1971-72 repeatedly pressured Fed Chairman Arthur Burns to keep monetary policy accommodative in the run-up to the 1972 election. White House tapes and Burns’s diary, published by historian Robert Ferrell in 2010, show Nixon explicitly telling Burns that the Fed should print more money and that Burns’s own future would depend on economic performance. Burns complied, keeping rates low through the 1972 election cycle; the resulting inflation contributed to the stagflation crisis of the mid-1970s that caused enormous economic damage and required Paul Volcker’s brutal 1980-81 tightening, which drove unemployment above 10 percent. The economic cost of that single episode of political interference ran into the trillions of dollars over the decade.

Modern threats to Fed independence have been primarily rhetorical but structurally significant. Between 2018 and 2019, President Trump publicly attacked Fed Chairman Jerome Powell over a dozen times, calling him an “enemy.” In 2025, Trump again raised the question of Powell’s tenure, publicly stating he wanted to remove him before his term expired in 2026. Unlike the Fed’s operational independence, which rests on statute and institutional norms, the legal durability of the “for cause” protection for the Chair specifically has not been definitively settled by the Supreme Court, especially after Seila Law v. CFPB (2020).

The Fed’s dual mandate, requiring it to pursue both maximum employment and stable prices, is statutory under the Humphrey-Hawkins Full Employment Act of 1978 and can be changed by a simple majority of Congress. The entire architecture of monetary policy independence, on which global currency markets, the bond market, and the dollar’s reserve currency status depend, rests on statutes that can be amended or repealed. No constitutional provision protects the Fed’s independence; the Burns episode shows how fragile that norm can be.

The power of federal courts to strike down legislation as unconstitutional is nowhere stated in the Constitution. Article III defines the judicial power and the courts’ jurisdiction; it does not say courts may nullify acts of Congress. Chief Justice John Marshall’s opinion in Marbury v. Madison (1803) established judicial review by reasoning that a written constitution is supreme law, that it is the province of courts to say what the law is, and that when a statute conflicts with the Constitution a court must apply the Constitution. Hamilton’s argument in Federalist No. 78 had anticipated it, but Jefferson and others disputed it from the start. Two centuries of institutional practice have made judicial review the settled foundation of American constitutionalism.

The Court’s 1958 opinion in Cooper v. Aaron, issued in response to Arkansas Governor Orval Faubus’s defiance of the desegregation order in Brown v. Board of Education, went further: all nine justices individually signed an opinion stating that the Supreme Court’s interpretation of the Constitution in Marbury is “the supreme law of the land” binding on all state officials. Cooper asserted not just judicial review but judicial supremacy, the claim that the Court’s reading of the Constitution is final and binding on the other branches.

The recent run of Supreme Court decisions has tested the system’s legitimacy in new ways. Dobbs v. Jackson Women’s Health Organization (2022) overturned Roe v. Wade after 49 years. New York State Rifle and Pistol Association v. Bruen (2022) established a new historical-analogues test for Second Amendment cases. Students for Fair Admissions v. Harvard (2023) eliminated race-conscious university admissions after 20 years. In each case, the majority resolved a contested political and moral question in a direction that tracked the ideological preferences of the justices appointed by Republican presidents, who hold a 6-3 supermajority. The pattern is that the most consequential policy questions in American life are resolved by a body selected through a process that has itself become the central arena of partisan conflict.

The appointment process reflects this. The Merrick Garland episode in 2016, when Senate Majority Leader Mitch McConnell refused to hold hearings on President Obama’s nominee for eight months, and the 2020 confirmation of Justice Amy Coney Barrett eight days before a presidential election, established that Senate majorities will use every procedural tool available to control the Court’s composition. Constitutional meaning is functionally determined by the partisan composition of the Senate in the years when vacancies happen to occur.

The size of the Court compounds the problem. The Constitution does not specify how many justices sit on the Supreme Court; that number is set by statute. Congress has changed it seven times, from six in 1789 to ten in 1863 to nine in 1869, where it has remained by convention. Nine is small enough that a single vacancy can shift the ideological balance of the entire institution, which is why each appointment has become a political crisis. FDR’s 1937 court-packing plan failed and established a norm against expanding the Court for partisan advantage, but that norm, like so many others in this catalog, has no legal backing. There are currently 13 federal circuit courts of appeals; tying the number of justices to the number of circuits, as was roughly the practice in the 19th century when each justice “rode circuit,” would produce a Court of 13, large enough to reduce the stakes of any single appointment and small enough to function as a deliberative body. Term limits would further depoliticize the process: if every president appoints a fixed number of justices on a predictable schedule, the winner-take-all dynamic of lifetime appointments during random vacancies disappears.

The constitutional architecture for judicial independence is clear: Article III judges hold their offices “during good Behaviour,” receive salary protection against congressional reduction, and can be removed only through impeachment. These protections were designed to insulate judges from political pressure. What the Constitution does not supply, and what Congress has been slow to provide, is a system of external accountability for the conduct judges engage in off the bench: financial relationships, public statements, attendance at partisan events, and failure to recuse from cases involving parties with whom they have personal or financial ties.

Lower federal court judges are subject to the Code of Conduct for United States Judges, administered by the Judicial Conference. Enforcement is handled through a complaint process under the Judicial Conduct and Disability Act of 1980. Roughly 1,000 complaints are filed annually against federal judges; the vast majority are dismissed, and public sanctions are rare.

The Supreme Court exempted itself from this system. No binding ethics code applied to Supreme Court justices until November 2023, when the Court adopted its own Code of Conduct, the first in its history. The code largely tracks the lower-court code’s substantive standards but contains no enforcement mechanism. The occasion for the code’s adoption was sustained reporting, primarily by ProPublica’s Justin Elliott and Joshua Kaplan, documenting that Justice Clarence Thomas had accepted more than two decades of undisclosed gifts, travel, and tuition payments from real estate developer and Republican donor Harlan Crow, with an estimated value exceeding $1 million. Justice Samuel Alito separately failed to disclose a luxury fishing trip with Republican donor Paul Singer, whose hedge fund had cases before the Court, and Alito declined to recuse from those cases.

The recusal statute, 28 U.S.C. § 455, requires any federal judge, including Supreme Court justices, to disqualify themselves in any proceeding in which their impartiality might reasonably be questioned, or in which they have a financial interest in any party. The statute is nominally mandatory, but for the Supreme Court it is entirely self-enforced. There is no mechanism by which a party can compel a justice’s recusal, no appellate review of a recusal decision, and no independent officer who reviews recusal decisions against the factual record.

The Posse Comitatus Act, enacted in 1878 in the aftermath of Reconstruction, prohibits using the Army (and by regulatory extension the Air Force, Navy, and Marine Corps) as a domestic police force. The statute was a direct response to the use of federal troops to supervise elections in Southern states during Reconstruction, and it reflects a deep strand of republican political thought: standing armies under executive command are incompatible with civilian governance and political liberty.

The exception swallows much of the rule. The Insurrection Act of 1807, a pre-existing statute that has never been repealed, expressly authorizes the President to deploy federal military forces domestically to suppress “any insurrection, domestic violence, unlawful combination, or conspiracy” that either a state government requests help suppressing or that “opposes or obstructs the execution of the laws of the United States.” The President acts unilaterally; there is no requirement of congressional approval, no defined sunset after which forces must be withdrawn, and no judicial review mechanism. The Act was invoked by President Eisenhower to deploy the 101st Airborne to enforce desegregation at Little Rock Central High School in 1957, a use almost universally regarded as legitimate. President George H.W. Bush invoked it to deploy Marines and Army troops to Los Angeles during the 1992 riots.

The more dangerous pattern is the use or threatened use of the Insurrection Act against political protest. In June 2020, as protests over the killing of George Floyd spread to every major American city, President Trump threatened publicly to invoke the Act and “deploy the United States military and quickly solve the problem.” Secretary of Defense Mark Esper and Chairman of the Joint Chiefs Mark Milley accompanied Trump during a photo opportunity in Lafayette Square immediately after federal law enforcement used chemical agents to clear peaceful protesters. Esper subsequently stated publicly that he opposed using active-duty forces under the Insurrection Act for domestic law enforcement; Milley later called the Lafayette Square incident “a mistake.” The episode illustrated how close the civilian-military boundary can come to breaking down when the President is willing to apply pressure and legal tools are available to justify it.

The National Guard adds another layer of complexity. Guard units are under state command in normal circumstances, but when federalized, they fall under the President. The 2021 Capitol attack created a situation in which the Pentagon delayed for hours before approving Guard deployment requests from Washington D.C. officials, raising questions about whether the delay was operational or political.

The Constitution constrains federal taxation through a set of categorical distinctions that made sense in an 18th-century agrarian economy and now function primarily as traps for modern fiscal policy. Article I, Sections 2 and 9 require that “direct Taxes” be apportioned among the states according to population, a rule designed to protect Southern slave wealth from Northern taxation under the Three-Fifths Compromise. The Sixteenth Amendment, ratified in 1913 after Pollock v. Farmers’ Loan & Trust Co. (1895) struck down the federal income tax, authorized Congress to tax “incomes, from whatever source derived, without apportionment.” That fixed one problem but left the underlying categorical architecture intact: direct vs. indirect, tax vs. penalty, income vs. wealth, apportioned vs. unapportioned. Each distinction limits Congress’s fiscal toolkit for reasons unrelated to sound governance.

The Affordable Care Act litigation made the costs of this architecture vivid. The individual mandate required Americans to obtain health insurance or pay a “shared responsibility payment.” Congress drafted it as a penalty under the Commerce Clause. In NFIB v. Sebelius (2012), five justices held that Congress could not compel individuals to purchase insurance under the Commerce Clause. Chief Justice Roberts saved the statute only by recharacterizing the payment as a tax under the taxing power, a category the statute’s own authors had avoided for political reasons. The constitutional validity of the ACA’s central mechanism turned not on whether the policy was sound, but on which 18th-century fiscal category it fell into. Roberts’s opinion acknowledged the absurdity: the payment “looks like a tax in many respects” and “produces at least some revenue for the Government,” but the constitutional question was whether calling it a penalty rather than a tax changed the legal analysis. It did.

The wealth tax problem is equally stark. A federal tax on accumulated assets rather than on income flow would almost certainly be challenged as an unapportioned direct tax. Apportionment would make such a tax absurd in practice: states with concentrated wealth like California would owe shares calculated by headcount, requiring wildly different effective rates across states. Moore v. United States (2024) upheld the mandatory repatriation tax on unrealized foreign earnings but explicitly declined to decide whether Congress can tax unrealized gains generally. Mark-to-market taxation of capital gains, a reform endorsed by economists across the political spectrum, faces the same constitutional uncertainty. A federal land value tax, which economists from Henry George to Milton Friedman have praised as among the most efficient possible taxes, would be a direct tax requiring apportionment. The constitutional categories foreclose fiscal options that have nothing to do with individual rights and everything to do with outdated economic distinctions.

Other democracies do not operate under these constraints. Germany, France, Switzerland, Norway, and Canada all have or have had wealth taxes, land value taxes, or broad-based asset levies implemented through ordinary legislation. None needed to navigate an apportionment requirement or a tax-vs.-penalty distinction. The only substantive constraint most modern constitutions impose on taxation is equal treatment: taxes cannot target specific individuals, and they must comply with anti-discrimination principles. The American system adds layers of formal categorization that protect no individual right but substantially limit Congress’s ability to design a fiscal system suited to a 21st-century economy in which wealth increasingly takes the form of unrealized capital gains, trust structures, and intangible assets that the constitutional categories were never designed to reach.

Federal revenue depends almost entirely on income and payroll taxes, which fall disproportionately on labor income. The Congressional Budget Office has documented that effective federal tax rates on the wealthiest Americans have declined substantially since the 1960s, in part because the tax base cannot reach unrealized appreciation and other forms of wealth accumulation that are not “income” in the constitutional sense. The constraint is not just about wealth taxes; it is about the basic question of whether Congress has the tools to fund the government it is expected to run.

The word “privacy” appears nowhere in the Constitution. The right was judicially inferred in Griswold v. Connecticut (1965), where Justice Douglas found it in “penumbras, formed by emanations” from the First, Third, Fourth, Fifth, and Ninth Amendments. That formulation was widely mocked even by sympathetic legal scholars, and the right it created has been built on and then partially demolished over the following six decades. Roe v. Wade (1973) grounded abortion rights in the privacy right from Griswold. Lawrence v. Texas (2003) struck down sodomy laws on the same basis. Obergefell v. Hodges (2015) established marriage equality partly through privacy and liberty interests under the Fourteenth Amendment. Each of these decisions rested on a constitutional foundation that had no textual anchor.

Dobbs v. Jackson Women’s Health Organization (2022) overturned Roe and Planned Parenthood v. Casey, holding that Roe was “egregiously wrong.” Justice Thomas’s concurrence argued that Griswold, Lawrence, and Obergefell should be reconsidered on the same grounds. Dobbs demonstrated what happens when a right that tens of millions relied on for fifty years rests on judicial inference: a change in Court composition eliminates it overnight.

Data privacy presents a second dimension. The “third-party doctrine” from Smith v. Maryland (1979) and United States v. Miller (1976) holds that information voluntarily disclosed to a third party carries no privacy expectation. Carpenter v. United States (2018) narrowed this for cell-site location data but did not overturn the doctrine. The vast majority of digital life, email metadata, browsing history, purchase records, app location data, exists in a constitutional gray zone. The EU has the GDPR and constitutional protections for personal data. The United States has a patchwork of sector-specific statutes (HIPAA, FERPA, COPPA) and state laws (California’s CCPA), but no constitutional floor and no comprehensive federal framework.

Section 230 of the Communications Decency Act (1996) provides that platforms are not liable for user-posted content. The statute was designed for an internet of bulletin boards. Modern platforms algorithmically select, rank, amplify, and recommend content to maximize engagement, making editorial choices functionally identical to a publisher’s, but Section 230 treats algorithmic amplification the same as passive hosting. The distinction between hosting and amplifying is the distinction between a telephone company and a newspaper, and the law treats both as the telephone company.

Foreign influence operations exploit this gap. The intelligence community documented systematic Russian use of American platforms in 2016 to distribute divisive content, create fictitious activist groups, and suppress voter turnout. Bluman v. FEC (2012) upheld the prohibition on foreign nationals spending in US elections, but there is no federal requirement to label foreign-origin political content, no obligation on platforms to flag coordinated foreign campaigns, and no liability when algorithmic amplification turns a foreign operation into a viral domestic event. The EU’s Digital Services Act (2022) requires platforms to assess and mitigate such risks; the United States has no equivalent.

The Fifth Amendment protects against compelled self-incrimination. Miranda v. Arizona (1966) required police to inform suspects of their rights before custodial interrogation. The Court has been eroding Miranda ever since: New York v. Quarles (1984) carved out a public safety exception; Berghuis v. Thompkins (2010) required suspects to affirmatively invoke their right to silence; and Vega v. Tekoh (2022) held 6-3 that Miranda violations do not support civil rights lawsuits under 42 U.S.C. § 1983. After Vega, officers who interrogate without warnings face no personal consequence. The deterrent Miranda was designed to create has been gutted without any legislative action.

No-knock warrants follow the same pattern. Wilson v. Arkansas (1995) held that knock-and-announce is part of the Fourth Amendment’s reasonableness inquiry. Hudson v. Michigan (2006) eliminated suppression as a remedy for violations. SWAT raids increased from roughly 3,000 per year in the 1980s to over 60,000 by the 2010s. In March 2020, Louisville police executing a no-knock warrant shot and killed Breonna Taylor in her apartment; the warrant targeted a different address and no drugs were found. There is no federal prohibition on no-knock warrants and, after Hudson, no constitutional remedy when the knock-and-announce rule is violated.

Civil asset forfeiture allows the government to seize property on the theory that it is connected to criminal activity, without charging the owner. The DOJ’s Asset Forfeiture Fund took in over $36 billion between 2000 and 2019. The DEA seized over $4 billion in cash from 2007 to 2016 without filing criminal charges or obtaining judicial warrants. Timbs v. Indiana (2019) applied the Excessive Fines Clause to the states but did not address the framework that permits forfeiture without conviction. Federal “equitable sharing” lets local police route seizures through federal agencies to bypass stricter state laws, returning up to 80% of proceeds to the seizing department.

Qualified immunity compounds all of these problems. Under Harlow v. Fitzgerald (1982), officers cannot be held liable under Section 1983 unless their conduct violated “clearly established” law, meaning a prior decision with nearly identical facts. The doctrine has no basis in the text of Section 1983. Pearson v. Callahan (2009) allowed courts to grant immunity without deciding whether a violation occurred, preventing the development of the precedents needed to overcome the “clearly established” requirement. A Reuters investigation found that appellate courts granted qualified immunity in over 57% of cases, shielding officers who shot a 10-year-old bystander, tased a surrendering driver, and sicced a police dog on a man sitting with his hands up. The George Floyd Justice in Policing Act, which would have abolished qualified immunity, passed the House in 2021 and failed in the Senate.

Germany’s Basic Law opens with Article 1: “Human dignity shall be inviolable.” The United States has no equivalent. The Eighth Amendment permits the death penalty, abolished by every EU member state. The UN Special Rapporteur on Torture has characterized US solitary confinement practices as torture; prolonged isolation lasting years is routine in federal and state prisons, imposed administratively without judicial review. The absence of a constitutional dignity right means these conditions are evaluated only under the Eighth Amendment’s deferential “deliberate indifference” standard.

The Constitution contains no right to a clean environment, no right to safe air or water, and no obligation on the government to protect natural resources for future generations. Environmental protection in the United States rests entirely on statutes: the Clean Air Act (1970), the Clean Water Act (1972), the National Environmental Policy Act (1970), the Endangered Species Act (1973). These laws were enacted during a brief bipartisan window when environmental degradation was visible enough to overcome legislative inertia. Richard Nixon created the EPA by executive order in 1970 and signed the foundational environmental statutes. That legislative window closed decades ago, and the statutory framework it produced is now under sustained judicial pressure.

West Virginia v. EPA (2022) held that the EPA could not use the Clean Air Act to require the electricity sector to shift generation from coal to gas and renewables, reasoning under the major questions doctrine that Congress must speak clearly before agencies can claim authority over decisions of such economic and political significance. The decision did not invalidate the Clean Air Act, but it established that the Court will read environmental statutes narrowly when agencies attempt to use them for transformative regulatory purposes. Sackett v. EPA (2023) narrowed the definition of “waters of the United States” under the Clean Water Act, stripping federal jurisdiction over wetlands that lack a continuous surface connection to navigable waters and reducing the geographic scope of federal water protection. Both decisions reduced the effective reach of environmental law without any legislative change, through judicial interpretation of statutes that Congress lacks the votes to update.

More than 100 national constitutions now include an explicit right to a clean or healthy environment. France added its Environmental Charter in 2004, granting constitutional status to the principle that “everyone has the right to live in a balanced environment that is respectful of health.” Germany amended Article 20a of its Basic Law in 1994 to make environmental protection a state objective binding on all branches of government. Ecuador’s 2008 constitution grants legal rights to nature itself. Even within the United States, Pennsylvania’s constitution (Article I, Section 27) declares that “the people have a right to clean air, pure water, and to the preservation of the natural, scenic, historic and esthetic values of the environment,” and the Pennsylvania Supreme Court has enforced this provision to block fracking in residential zones.

The absence of a federal constitutional environmental right has three consequences. First, environmental protection depends on agencies whose authority the Court is actively constraining through the major questions doctrine and the nondelegation revival (item 34). Second, Congress cannot update the statutes because of legislative gridlock, so the regulatory framework is frozen at its 1970s design while the problems it was built to address have changed dramatically. Third, there is no constitutional floor: a future Congress could repeal the Clean Air Act, the Clean Water Act, or the Endangered Species Act by simple majority, and no constitutional provision would prevent it. Climate change, the most consequential environmental challenge in human history, is being addressed by a government whose constitutional toolkit contains no environmental authority of any kind.

The United States Constitution contains no right to education. In San Antonio Independent School District v. Rodriguez (1973), the Supreme Court held 5-4 that education is not a fundamental right under the federal Constitution and that disparities in school funding between wealthy and poor districts do not violate the Equal Protection Clause. Justice Powell’s majority opinion reasoned that the Constitution does not explicitly or implicitly guarantee a right to education, and that the Court should defer to state and local control over school financing. Justice Marshall’s dissent argued that education is so closely linked to the exercise of other constitutional rights, free speech, voting, participation in civic life, that it should receive heightened judicial protection.

The practical consequence of Rodriguez is that school funding in the United States depends on local property taxes, producing enormous disparities between wealthy and poor districts within the same state. The highest-spending districts in the country spend more than three times per pupil what the lowest-spending districts spend. Children born in affluent suburbs receive a fundamentally different education from children born in rural poverty or urban disinvestment, and the federal Constitution provides no floor. State constitutions have partially filled the gap: 49 of 50 state constitutions contain education clauses, and state supreme courts in roughly 30 states have ordered funding reforms based on their own constitutional provisions. But the results are uneven, enforcement is inconsistent, and no federal baseline exists.

Most developed democracies treat education as a constitutional right. Germany’s Basic Law guarantees the right to freely choose an occupation and place of training (Article 12) and places the entire school system under state supervision (Article 7). The International Covenant on Economic, Social and Cultural Rights, ratified by 171 countries, recognizes “the right of everyone to education” and commits signatories to free and compulsory primary education; the United States signed the Covenant in 1977 but has never ratified it. Finland’s constitution guarantees the right to basic education free of charge. South Africa’s constitution (Section 29) guarantees the right to basic education and to further education that the state must make progressively available. Japan’s constitution (Article 26) declares that “all people shall have the right to receive an equal education correspondent to their ability.”

The absence of a federal right has consequences beyond funding. It means there is no constitutional basis for federal education standards, no floor below which states cannot fall, and no enforceable obligation to educate children in ways that prepare them for democratic participation. The connection between education and the functioning of a constitutional democracy is not abstract: voters who cannot evaluate candidates, jurors who cannot assess evidence, and citizens who cannot distinguish reliable information from propaganda are structural vulnerabilities for any self-governing system. Every other item in this catalog assumes a citizenry capable of understanding what its government is doing. Education is the precondition for that capability, and the Constitution does not guarantee it.

Every constitutional system depends on actors who operate within the spirit of its rules. Levitsky and Ziblatt’s How Democracies Die (2018) identifies two norms as essential: mutual toleration (accepting opponents as legitimate rivals) and institutional forbearance (not exercising every legal power to its destructive limit). The American record since 2010 documents both norms failing. The debt ceiling was used as a hostage in 2011, producing the first-ever U.S. credit downgrade. Four inspectors general were fired in a single week in 2020. Congressional subpoenas were systematically flouted across administrations. Pardons were issued to shield political allies. The V-Dem Liberal Democracy Index dropped the United States from the 88th global percentile in 2015 to the 71st in 2021. Freedom House reduced the U.S. score to 83 out of 100, its historical low.

Article V contains a structural timebomb. The Constitution allows a convention to be called by two-thirds of state legislatures (34 states), but says nothing about how such a convention would operate: no rules on delegates, scope, voting, or ratification timeline. The Convention of States Project claims 19 state applications; other balanced-budget applications bring the total to 28-30. No enforceable mechanism prevents a “runaway convention” from proposing anything; the 1787 convention was called to revise the Articles of Confederation and replaced them entirely. Ratification requires 38 states, but in a polarized environment with ideologically sorted state legislatures, that threshold may be lower than it appears.

The convention risk raises a question the Constitution does not answer: what happens to states that find its output unacceptable? Texas v. White (1869) holds that secession is unconstitutional, but that reasoning assumes the existing framework. A convention could propose amendments to Article V itself. Rhode Island and North Carolina refused to ratify the 1787 Constitution and remained outside the union for months; they had an exit option because the new charter replaced the old one. The current system provides no equivalent.

You cannot legislate good faith. What structural design can do is reduce the system’s dependence on it. Each of the preceding 49 items represents a specific instance in which structural reform can substitute for the forbearance that democratic backsliding erodes.