Two-Thirds of the Senate, How the Treaty Clause Was Built to Be Hard

Every few years a diplomat returns from a summit with a signed agreement, a bouquet of adjectives ("historic", "landmark", "transformational") and a strong preference that nobody in the Senate read the thing too closely. The Constitution has a word for that preference. The word is no.

The Treaty Clause is one of the hardest hurdles in American government, and it was built that way on purpose. A treaty binds the whole nation, often for generations, and under the Supremacy Clause it becomes "the supreme Law of the Land." The Framers decided that a commitment that heavy should need more than a bare majority to make. Here is how the clause works, why it was designed so, what happened the few times it was really tested, and how modern presidents have learned to walk around it.

The text and the arithmetic

Article II, Section 2, Clause 2 begins:

"He shall have Power, by and with the Advice and Consent of the Senate, to make Treaties, provided two thirds of the Senators present concur"

Three features stand out.

1. The president makes treaties. He negotiates and signs them. The Senate does not negotiate. After the Senate consents, the president decides whether to ratify.
2. The Senate must consent, and only the Senate. The House, which shares nearly every other legislative power, has no formal role in making a treaty.
3. Two-thirds of the senators present must concur. Not a majority. Not two-thirds of the whole Senate, but two-thirds of those present and voting, assuming a quorum.

Do the arithmetic with a full Senate of 100. A treaty needs 67 votes if all are present. That means 34 senators can defeat it. With every state electing two senators, 34 senators can represent as few as 17 states. A treaty cannot pass over the firm opposition of a third of the states' representatives. That was the design.

Compare the Supremacy Clause in Article VI: "all Treaties made, or which shall be made, under the Authority of the United States, shall be the supreme Law of the Land." A treaty can carry the force of federal law. An ordinary federal statute needs a majority of both houses and the president's signature, or two-thirds of both houses to override his veto. A treaty skips the House but must clear a supermajority in the Senate. The Framers traded one hurdle for another.

Why the Framers made it hard

The decision grew out of hard experience. Under the Articles of Confederation, Article IX required nine of the thirteen states to agree to treaties, a supermajority of about seventy percent. The most bitter fight of the 1780s was over the Jay-Gardoqui negotiations of 1786, in which John Jay, the Confederation's foreign secretary, was prepared to give up American navigation rights on the Mississippi River for twenty-five or thirty years in exchange for a commercial treaty with Spain. Northern states, which wanted the trade, were willing. Southern states, whose western settlers depended on the river, were outraged. The vote in Congress split seven to five along sectional lines, short of the nine needed. Southerners never forgot how close they came.

At Philadelphia the delegates from those states wanted protection against a bare majority bargaining away a region's vital interests. The supermajority was that protection. It meant no treaty could pass on the strength of one section's votes against another's.

John Jay himself defended the arrangement in Federalist 64. He argued that the Senate, chosen then by state legislatures for six-year terms, would bring together men of experience and continuity, able to judge treaties with knowledge and to keep the necessary secrecy of negotiation, while the president handled the negotiating itself. Hamilton added Federalist 75, arguing that the treaty power is neither purely legislative nor purely executive, which is why it was given to the president and Senate together. He also argued that the House, larger and more frequently changing, was ill suited to the steady, secret and expert judgment treaties require.

The Jay Treaty, 1794 to 1796

The first great test came quickly, and it nearly tore the young Republic apart.

By 1794 the United States and Great Britain were near war. British forts still stood on American soil in the Northwest, despite the peace treaty of 1783, and the Royal Navy was seizing American merchant ships trading with France. President George Washington sent John Jay, by then Chief Justice, to London to negotiate.

The treaty Jay signed on November 19, 1794, got the British out of the western forts and set up arbitration commissions for debts and seizures. It did not stop the impressment of American sailors, and it gave Britain trading terms many Americans found humiliating. When its terms became public, it was burned in effigy across the country, and so was Jay.

The Senate debated in secret session in June 1795 and consented by a vote of 20 to 10, exactly two-thirds, after striking an article that restricted American trade in the West Indies. One vote fewer and the treaty would have failed.

Then came a constitutional fight that still matters. Carrying out parts of the treaty required money, and money requires appropriations, which need the House. In 1796 the House, led by Jeffersonian Republicans, demanded that Washington turn over the papers from Jay's negotiations, implying it could refuse funding if it disapproved the treaty. Washington refused in March 1796, writing that the House had no role in the treaty power, and that the assent of the House was not necessary to the validity of a treaty. After a long debate the House voted the money at the end of April 1796 by a margin of three votes.

The episode set two precedents. The president alone controls negotiations and need not share diplomatic papers with the House as a condition of a treaty. And the House, through the power of the purse, retains real leverage over treaties that require money. Both precedents survive.

The Treaty of Versailles

The most famous defeat came after the First World War. President Woodrow Wilson personally negotiated the Treaty of Versailles in Paris in 1919, including the Covenant of the League of Nations. Article 10 of the Covenant committed members to respect and preserve one another's territorial integrity against external aggression, which critics read as a commitment that could draw the United States into war without a vote of Congress.

Senator Henry Cabot Lodge of Massachusetts, chairman of the Foreign Relations Committee, proposed reservations protecting Congress's war power and American sovereignty. Wilson refused to accept them and urged his supporters to vote against the treaty with the reservations attached. The Senate voted on November 19, 1919, and the treaty failed both with and without reservations. A final vote came on March 19, 1920. With the Lodge reservations attached, the treaty received 49 votes in favor and 35 against: a majority, but seven votes short of two-thirds. The United States never joined the League.

People argue to this day about whether that was wise. The point for our purposes is narrower. A treaty with a majority of the Senate behind it still failed, because a bare majority was not enough. The clause worked exactly as designed, protecting the war power of Congress from a commitment made by a president abroad. There was a compromise available, and the president would not take it.

The Senate has rejected few treaties outright since. One notable example came in October 1999, when it rejected the Comprehensive Test Ban Treaty, 48 to 51.

Missouri v. Holland (1920): a treaty and the states

The treaty power also raised a question about federalism. Can a treaty give Congress power it would not otherwise have?

In 1916 the United States and Great Britain, acting for Canada, signed a treaty protecting migratory birds. Congress passed the Migratory Bird Treaty Act of 1918 to carry it out. An earlier federal statute regulating bird hunting, passed without a treaty, had been struck down by lower federal courts as beyond Congress's powers. Missouri sued, arguing that a treaty could not expand federal power over matters reserved to the states.

Justice Oliver Wendell Holmes, writing for the Court, upheld the statute. A valid treaty, he reasoned, can support legislation under the Necessary and Proper Clause, even on subjects Congress could not otherwise reach. Holmes did not say treaties had no limits, but he declined to find one here.

Conservatives have worried about Missouri v. Holland for a century, and with reason. Read broadly, it would let the president and two-thirds of the Senate expand federal power over the states through an international agreement, sidestepping the limits of Article I. The Court sidestepped that question in Bond v. United States (2014), deciding the case on statutory grounds, a story told in Bond v. United States, a Jilted Wife and a Chemical Weapons Treaty. The structural worry remains, and it is one more reason the two-thirds requirement matters. A high bar for treaties is a safeguard for federalism as well as for sovereignty.

Medellin v. Texas (2008): treaties are not self-executing by default

The second modern landmark limits the reach of treaties in a different way. Jose Medellin, a Mexican national, was convicted of murder in Texas and sentenced to death. He had not been told of his right under the Vienna Convention on Consular Relations to contact his consulate. In 2004, the International Court of Justice, in the Avena case, held that the United States had violated the convention as to Medellin and dozens of other Mexican nationals and must review their convictions. President George W. Bush issued a memorandum directing state courts to give effect to the ruling.

The Supreme Court held, six to three, in an opinion by Chief Justice John Roberts, that neither the ICJ judgment nor the president's memorandum was binding federal law that displaced Texas's procedural rules. The relevant treaty commitments were not self-executing: they did not become enforceable domestic law without implementing legislation from Congress. And the president could not unilaterally turn a non-self-executing treaty into domestic law, because that would be lawmaking, which belongs to Congress.

Medellin is a quiet victory for self-government. It means an international tribunal cannot rewrite American law, and a president cannot make it do so by memorandum. If the nation wants to be bound domestically, Congress has to pass a law.

The rise of executive agreements

Here is the problem the modern presidency found with the Treaty Clause: it is hard. So presidents have increasingly avoided it.

An executive agreement is an international agreement made by the president without the Senate's two-thirds consent. Some rest on a statute Congress passed in advance, sometimes called congressional-executive agreements, approved by a simple majority of both houses. Trade agreements have generally been handled this way. Others rest on the president's own constitutional authority, called sole executive agreements.

The Supreme Court gave sole executive agreements real force in United States v. Belmont (1937) and United States v. Pink (1942), both arising from the Litvinov Assignment of 1933, by which President Franklin Roosevelt recognized the Soviet Union and settled claims. The Court held that such agreements, within the president's power, could override conflicting state law.

The number of executive agreements now dwarfs the number of treaties. Some of the most consequential recent commitments were made without a two-thirds vote. The 2015 nuclear arrangement with Iran was never submitted to the Senate as a treaty, and the 2015 Paris climate agreement was joined by executive action. Both were later abandoned by executive action, which illustrates the problem neatly: an agreement made by one president's signature can be unmade by the next one's. In 1997, before the Kyoto Protocol was even signed, the Senate passed the Byrd-Hagel resolution 95 to 0, warning against such a commitment, a story covered in The Senate Voted 95 to 0 Against a Climate Treaty Before Kyoto Was Written.

The objection, and the answer

The strongest objection runs this way. The two-thirds rule gives a minority a veto over American foreign policy. In a world of fast-moving crises and complex trade, requiring 67 votes makes it nearly impossible to commit to anything. Executive agreements are a sensible adaptation, and congressional-executive agreements even get majorities of both houses, arguably a more democratic approval than the Senate alone.

Here is the answer.

1. The difficulty is the point. A treaty binds the nation as supreme law, often for decades. Commitments that cannot win two-thirds of the Senate are commitments the country does not broadly share, and they will not survive the next election anyway, as Iran and Paris showed.
2. Congressional-executive agreements, where Congress actually votes, are a defensible tool, at least on subjects within Congress's own powers like trade. Sole executive agreements on weighty matters are not. They give the president alone a power the Constitution gave him jointly with the Senate.
3. Durability is a virtue in foreign affairs. Allies and adversaries alike prefer commitments that survive a change of administration. A two-thirds treaty is one. A signature on a press release is not.
4. The remedy for a difficult hurdle is persuasion, not evasion. If a president cannot persuade a third of the Senate, he should keep trying, narrow the agreement or accept reservations, as Wilson would not.

The broader principle connects to everything in Article II. Power should be exercised where it can be checked and where the public knows who answers for it. A treaty debated and voted on by the Senate is a public act. An executive agreement filed in a drawer is not. The related question of the president's war powers is examined in Commander in Chief Means Less Than Presidents Claim and More Than Critics Say, and the Senate's role as the states' house in Calling the Senate Undemocratic Misses Why It Was Built to Represent States.

The case in brief

1. The Treaty Clause requires two-thirds of the senators present to concur, a bar set high because a treaty becomes "the supreme Law of the Land."
2. The Framers built it on the lesson of the Jay-Gardoqui fight, and Jay and Hamilton defended it in Federalist 64 and 75.
3. The Jay Treaty passed with exactly two-thirds in 1795, and the Treaty of Versailles failed in 1920 with a majority but not two-thirds; the clause worked both times.
4. Missouri v. Holland raised the risk of treaties expanding federal power, and Medellin v. Texas held that treaties need Congress to become domestic law unless they execute themselves.
5. Executive agreements have grown to avoid the clause, and commitments made by one signature are unmade by the next.

What you can do: read Article II, Section 2 and Article VI on our Constitution page, then ask your senators, through our representatives page, whether they will insist that major international commitments come to the Senate as treaties.