How a Veto Override Works (And Why It's So Rare)

A veto override is Congress forcing a bill into law over the President's objection, without a presidential signature — but only if two-thirds of the members present vote for it in both the House and the Senate. It's built into the Constitution as the tiebreaker for exactly this disagreement: the President says no, and the only way around that no is a supermajority of elected representatives in both chambers saying yes anyway.

The Two-Thirds Mechanic, Precisely

The process is set out in Article I, Section 7, and it runs in a fixed sequence. It's the same override path referenced in how a bill becomes a law as the fork after a bill reaches the President's desk.

1. The President returns the bill, unsigned, to the chamber where it originated, along with a written statement of objections — this is the veto itself.
2. That chamber holds a vote on repassing the bill despite the objections. If two-thirds of the members present and voting agree (assuming a quorum), it moves to the other chamber.
3. The second chamber votes the same way. If two-thirds there also agree, the bill becomes law — no presidential signature required, no further step needed.

Worth being precise about: this threshold is a constitutional requirement that applies to the whole House and the whole Senate. That's a different animal from the Senate's 60-vote cloture rule that ends a filibuster — a chamber rule, not a constitutional one, and one that only exists in the Senate. See what is a filibuster for how that mechanic works; the two get conflated constantly, but an override vote and a cloture vote aren't solving the same problem or answering to the same rulebook.

Why Overrides Are Rare

Since 1789, presidents have vetoed roughly 2,591 bills. Congress has overridden 112 of them — about four percent source. That low rate isn't an accident of arithmetic; it's the predictable result of how the veto gets used in the first place. A president who expects a bill to draw two-thirds opposition in both chambers usually doesn't need to veto it — it either doesn't pass in that form or gets negotiated down before it reaches the desk. Vetoes tend to land on bills where the president's own party controls, or nearly controls, at least one chamber, which is precisely the condition that makes assembling a cross-party two-thirds hard. You can see the pattern play out in real time by browsing vetoed and enacted bills side by side — overrides cluster around periods of genuinely split government and genuinely bipartisan bills, not routine partisan disputes.

The exception that proves the rule is Andrew Johnson, whose Reconstruction-era feuds with Congress produced 15 overrides out of 29 vetoes source — the most of any president in raw numbers, and by a wide margin. That's not a fluke of persuasion; it's what happens when a president's own party effectively disowns him and a hostile Congress has the votes, and the will, to override on a near-routine basis. Ordinary disagreement doesn't produce that alignment. It takes a genuinely broken relationship between the branches, which is exactly why the rate stays low outside episodes like it.

The Pocket Veto: A Dead End With No Override

There's a second way a bill dies at the President's desk, and it doesn't go through any of the steps above. If the President doesn't sign a bill and Congress adjourns within the 10-day window (Sundays excluded) that would otherwise let it become law without a signature, the bill dies unsigned — a pocket veto. No written objections get sent back. No chamber holds an override vote, because there's nothing to override: the bill was never formally returned. The only way forward is reintroducing it as a new bill in a later session and running it through the whole process again.

This isn't a loophole so much as a deliberate design choice: the override mechanism exists to answer objections, and a pocket veto raises none for Congress to answer. It's also why the pocket veto shows up on Bill100, and in bill status generally, as its own distinct dead-end rather than a variant of an ordinary veto — the two share a root cause (the President declining to sign) but nothing about what happens next.

A Real Override: JASTA, 2016

The clearest recent example is the Justice Against Sponsors of Terrorism Act (JASTA), which narrowed foreign sovereign immunity so 9/11 victims' families could sue state sponsors of terrorism. President Obama vetoed it on September 23, 2016. Five days later, on September 28, the Senate voted 97–1 to override — Senate Minority Leader Harry Reid cast the lone no — and the House followed the same day, 348–77. The bill became Public Law 114-222 over the President's objection. It was the only veto override of Obama's nearly eight years in office, which is itself a useful data point on just how high that bar sits even for a bill with overwhelming, bipartisan floor support.

Procedurally, nothing about JASTA's public law entry looks different from a bill the President happily signed. It got the same kind of public law number, took its place in the same Statutes at Large, and carries exactly the same legal force. The override shows up in the legislative history as the route it took to enactment — not as an asterisk on its authority once it's law.

Executive Order vs. Law: Not the Same Animal

People conflate these constantly, and the confusion is understandable — both produce headlines about a president "acting." But they're structurally nothing alike. An executive order is a directive from the President to the executive branch, exercising authority the President already has under the Constitution or an existing statute. It never goes to Congress. There's no vote, no committee, no bill status to track, because it was never a bill. It's not a statute — it's an instruction about how existing law and existing executive authority get applied.

That distinction is exactly why the two get reversed so differently. A statute — whether signed, allowed to become law without a signature, or enacted over a veto — stays on the books until Congress passes another law to repeal or amend it, or a court strikes it down. An executive order has no such durability: the same president can rescind it tomorrow, and the next president can revoke or replace it on day one with a signature, no override, no supermajority, no Congress involved at all. An override outcome is about as durable as law gets, precisely because it required both chambers to clear the highest voting bar in ordinary legislating. An executive order is, by design, only as durable as the president who signed it.

None of that makes executive orders weak, exactly — it makes them fast. An order signed on day one can redirect enforcement priorities across an entire agency before a comparable bill would even clear committee. What it trades away for that speed is precisely the staying power a statute gets from having survived Congress: durability and speed pull in opposite directions here, and an order that bypasses the legislature to move quickly is, for the same reason, unprotected from the next president doing the same thing in reverse.