The NATO Treaty Does Not Give Congress a Bye on World War III

On March 13, Russian missiles hit a military base in Ukraine, about 15 miles from the Polish border. Poland is a NATO member. Suppose the airstrike had occurred 15 miles on the other side of the border, in Poland. Would the NATO Treaty automatically place the United States and all other NATO members at war with Russia?

The prevailing opinion seems to be yes. The NATO treaty is widely thought to be an ironclad guarantee that the allies will automatically come to the defense of any member state subject to an armed attack. In the United States, the treaty is popularly understood, further, as authorizing the president to use force under such circumstances without congressional approval. In other words, the belief is that neither the president nor Congress has a choice if a NATO ally is attacked—the United States, like it or not, is then at war.

That understanding is mistaken. The NATO treaty does not require the United States or any other party automatically to go to war if a party is attacked. The treaty provides that an attack on one is an attack on all—but it leaves each nation free, in accordance with its own “constitutional processes,” to determine whether an armed attack has occurred and to take such action “as it deems necessary” in response. Nor does the treaty expand the president’s war power. “The distribution of power within the United States government is precisely what it would have been in the absence of the treaty,” the Senate Foreign Relations Committee said in 1979. In the event of an armed attack, the United States “reserves the right to determine for itself what military action, if any, is appropriate.”

The NATO Treaty’s limited obligation reflects a bedrock principle for the United States: The Constitution vests the power to declare war in both houses of Congress, acting jointly. Reflecting on this principle in 1978, the committee said that the president and the Senate cannot make a treaty that would “unconstitutionally divest the House of Representatives of its share of the warmaking power.” It was in part this concern that led the Senate in 1919 to reject the Covenant of the League of Nations—which, opponents argued, would have required the United States to go to war if the league ordered it to do so, cutting the House of Representatives out of the process.

The United States has thus never been party to a treaty that automatically placed the nation at war. Each of the seven mutual security treaties that the United States entered into after World War II provided, like the NATO Treaty, that it would be carried out by the United States in accordance with its constitutional processes or contained other language to make clear that military action is not required. This intent was underscored at every stage when the NATO Treaty was considered.

The day the text of the proposed NATO Treaty was made public in 1949, Secretary of State Dean Acheson addressed the nation. The treaty, he said, “does not mean that the United States would be automatically at war if one of the nations covered by the pact is subjected to armed attack. Under our Constitution, the Congress alone has the power to declare war.” Acheson reiterated that in his letter transmitting the treaty to President Truman, adding that parties’ obligations “might or might not include the use of armed force depending upon the circumstances and gravity of the attack. … Each party retains for itself the right of determination as to whether an armed attack has in fact occurred and what action it deems necessary to take.”

Shortly afterward, Acheson emphasized those points to the Senate Foreign Relations Committee, which summarized his testimony in its report on the treaty. The “question was repeatedly asked,” the report recounted, whether the United States was “obligated to react to an attack on Paris or Copenhagen in the same way it would react to an attack on New York City? In such an event does the treaty give the President the power to take any action, without specific congressional authorization, which he could not take in the absence of the treaty?  The answer to both these questions,” it said, as had Acheson, “is ‘No.’” “Nothing in the treaty,” the committee said, “increases or decreases the constitutional powers of either the President or the Congress or changes the relationship between them.”

These limitations on the U.S. commitment were reaffirmed repeatedly on the Senate floor by the committee chairman, Sen. Tom Connally, in response to charges that the NATO Treaty would automatically bind the United States to go to war in the event of an armed attack. “I challenge anyone to find such a commitment,” he said. “The full authority of the Congress to declare war, with all the discretion that power implies, remains unimpaired.”

Despite the intensity with which such concerns were voiced in the Senate during consideration of the League of Nations covenant and the NATO Treaty, in two notable instances presidents afterward claimed authority to use force to enforce the terms of a treaty. President Truman in 1951 argued that the U.N. Charter authorized him to prosecute the war in Korea without congressional authorization, and the Johnson administration claimed in 1966 that the SEATO treaty committed the United States to defend South Vietnam and authorized the president to undertake independent military action toward that end.

Congress responded to these claims in the 1973 War Powers Resolution. Section 8(a)(2) of the resolution precludes inferring authority to use force from any treaty that is not implemented by legislation specifically authorizing the use of force. Subsequent administrations have seemingly ruled out the arguments made by Presidents Truman and Johnson. During the Ford administration, for example, Sen. Dick Clark asked in 1975 whether “any treaty authorize[s] the introduction of U.S. armed forces into hostilities.” The administration replied:

[T]he answer is “no.” Treaties of the United States which express defense commitments to other nations commit the United States to act only in accordance with its constitutional processes. Such treaties do not confer authority which would not otherwise be available through the constitutional processes of the United States.

The Carter administration, in response to a 1977 letter from Sen. George McGovern to Secretary of State Cyrus Vance, said the following:

[A]lthough our mutual security agreements entail a legal obligation to respond to an armed attack on another party, the nature and scope of that response is left to the discretion of the responding party …. Accordingly, such treaties do not confer  “Authority to introduce United States Armed Forces into hostilities” within the meaning of Section 8(a)(1) of the War Powers Resolution.

What, then, does the NATO Treaty require in obliging each party to take such action “as it deems necessary” in accordance with its “constitutional processes”? At the most obvious level, the notion of “process” presupposes the possibility of different outcomes. If only one result were preordained, there would be no point in working through a process to reach that result; the process would be only a sham. Process implies choice—a deliberative procedure that can produce alternative outcomes. Thus the Senate Foreign Relations Committee said in 1978 (referring to a requirement in the Panama Canal Treaty to enter into a prisoner exchange agreement) that “authority is meaningless if it is required to be given; the authority to disapprove is implied if our ‘constitutional processes’ are to be upheld.”

The right of each party to decide what action to take under its own constitutional process hardly makes its NATO commitment meaningless. Three elements are significant. First, the treaty obliges each party to assess in good faith whether an armed attack on a NATO ally has occurred, as NATO members did after the 9/11 attacks. Members may disagree. They may interpret “armed attack” differently, or they may view the facts differently.  Resolving those disagreements may take time. When NATO ambassadors discussed the possibility of invoking Article 5 in Brussels on Sept. 11 and 12, 2001, for example, it was not yet clear whether the attacks on the United States had been externally directed or were internal acts of terrorism, like the Oklahoma City bombing. NATO’s North Atlantic Council issued an initial statement on Sept. 11 condemning the “barbaric acts” committed against the United States, with no reference to attacks, while a statement issued the next day indicated that Article 5 would cover the acts only “if it is determined that this attack was directed from abroad against the United States.” On Oct. 2, according to NATO, after the council had been briefed on the results of investigations into the 9/11 attacks, it determined that the attacks were regarded as an action covered by Article 5. Subsequently, on Oct. 4, NATO took specific steps to “operationalise Article 5” of the treaty.  This gradually unfolding process has led to disagreement as to the date on which Article 5 was actually “invoked.”

Second, if a party concludes that an armed attack has occurred, that finding requires it to view the attack as gravely as it would an attack on itself and all other NATO members. But the treaty recognizes that, in doing so, each party will honor its own constitutional processes. Following the 9/11 attacks, German Chancellor Gerhard Schröder—“in keeping with the constitution,” he said—sought and received approval from the Bundestag, on Sept. 19, before Germany came to the aid of the United States. Like other NATO members, Germany had every right to determine for itself what had occurred, to assess who directed the attack, and to weigh the costs and benefits of all available options.

Third, taking action “it deems necessary” might mean use of force, but any party might also conclude that such measures are not necessary and that some other response, such as economic, diplomatic, or trade sanctions, might be more conducive to meeting the treaty’s end—“to restore and maintain the security of the North Atlantic area.” Chancellor Schröder, for example, announced that Germany would “not participate in air strikes or deploy combat troops on the ground,” which under the treaty was a matter within Germany’s sole discretion.

As this history suggests, the NATO Treaty creates not only legal but also political imperatives. These are not always co-extensive. It is entirely possible that a member might escape the treaty’s narrowly construed legal obligations while defaulting on its political undertakings. Doing nothing in response to an unequivocal, unprovoked armed attack on a NATO ally would fall within that realm. Reliance and reliability matter politically in international affairs. 

Still, it would be idle to deny that the NATO Treaty seeks to mesh competing values that do not neatly reconcile. On the one hand, its aim is to deter aggression against its parties. Deterrence is effective, in international relations as in domestic criminal justice, largely to the extent that the penalties it threatens are swift and certain. On the other hand, the treaty also seeks to respect the deliberative, inclusive decision-making processes of the democracies it is designed to protect. The history of Western constitutionalism is in no small part a history of the transfer of the war power from monarchies and executives to parliaments and legislatures.  To have discarded that value in favor of air-tight, option-free deterrence would almost certainly have guaranteed the treaty’s rejection by the Senate, and probably other countries as well. The treaty therefore rejects swift and sure deterrence in favor of the right to decide, to choose through a meaningful process—to examine the facts of a supposed attack, to determine whether it was provoked, to weigh its gravity, to identify its true perpetrators and sponsors, to assess its likely consequences, and to evaluate the risks and benefits of all options, both military and nonmilitary. It hardly bears noting that, had questions concerning such matters been asked assertively at the outset of several wars the United States has fought—from the Spanish-American War through Vietnam and Iraq—history might have unfolded very differently.

To return to the original question: In the event of a Russian attack on a NATO member, each member of the alliance would face the urgent question of how to respond. The choice to confront Russia militarily would, as President Biden has said, lead to World War III. In the U.S. system, the Constitution does not assign that choice to one person, and the NATO Treaty surely does not eliminate that choice altogether. In fact, the treaty leaves the choice for war or peace exactly where the Constitution placed it—in Congress. A treaty aimed at protecting 30 of the world’s leading democracies could hardly do otherwise.