Armed Conflict Executive Branch Foreign Relations & International Law

Some Thoughts on the Weak U.S.-Ukraine Security Agreement

Jack Goldsmith
Friday, June 14, 2024, 6:11 PM
It does not, as the national security adviser says, signal U.S. resolve. Quite the opposite.
President Joseph Biden hosted Ukraine’s President Volodymyr Zelenskyy at the White House on December 12–13, 2023 (usembassykyiv, public domain, via Wikimedia Commons)

Published by The Lawfare Institute
in Cooperation With

National Security Adviser Jake Sullivan said that the Bilateral Security Agreement Between the United States of America and Ukraine signed yesterday by President Biden is a “real marker of our commitment, not just for this month, this year, but for many years to support Ukraine, both in defending against Russian aggression and in deterring future aggression so that Ukraine can be a sovereign, viable, thriving democracy,” and added that the agreement was “a signal of our resolve.” David Sanger in the New York Times reports that the agreement was “designed to demonstrate to President Vladimir V. Putin of Russia that the United States and its allies had no intention of packing up and leaving.”

I believe the agreement sends something approaching the opposite signal of the one identified by Sullivan and Sanger. It demonstrates the Biden administration’s and the United States’s surprisingly weak commitment and resolve in defending Ukraine against Russian aggression.


The agreement appears to be an executive agreement made solely on the president’s authority, though the precise authority is not specified. It is unclear whether it is a legally binding or nonbinding agreement. It is probably best seen as a binding agreement with mostly nonbinding terms. Some provisions use the language of binding agreements (such as the parties “shall … meet” within 24 hours, if possible, if one is attacked; and the entry into force clause) and others seem nonbinding (such as “the policy of the Parties to work together to help deter … aggression”).

But I do not think the agreement’s status as binding or nonbinding matters at all since the substantive commitments are so weak.


The only real obligation upon attack is to “immediately meet, where possible within 24 hours, at the most senior levels to determine appropriate next steps and additional defense needs.” At that meeting the parties “may” (emphasis added) decide to “implement additional appropriate defense and deterrent responses.” Or, by implication, they may not. Compare, for example, the North Atlantic Treaty, which states that the treaty parties, if attacked, “will assist” (emphasis added) the attacked party “by taking forthwith ... such action as it deems necessary, including the use of armed force.” President Biden could not have made a binding agreement of this sort on his own authority (more on this point below), but the contrast is edifying.

The key commitments to “help deter and confront any future aggression against the territorial integrity of either Party” and to “assist Ukraine in maintaining a credible defense and deterrence capability” are stated as “polic[ies]” and are not backed up by any material commitments. Future aggression against Ukraine is described in the agreement as a matter of “grave concern[,]” but no obligation to act is triggered.

The parties also agree to “further implement[] … existing security agreements,” to “meet on a regular basis,” and to “further bolster their defense and security cooperation” in ways that “may,” but need not, include training, provision of defense articles, joint planning, and cybersecurity cooperation, among other means of coordination.

The agreement later states that the parties “agree to advance the appropriate sharing of intelligence and to promote enhanced cooperation between their intelligence services,” but the nature and scope of this sharing and cooperation is entirely unspecified and thus imposes no real commitment. (The U.S. and Ukraine engage in significant intelligence sharing, the details of which are largely classified.)

Perhaps most importantly, “providing sustainable levels of security assistance for Ukraine” is stated as a mere policy, not obligation, and rather than commit funds, the United States “intends” (but is not even obliged) “to seek from … Congress appropriation of funds to help sustain a Ukrainian credible defense and deterrent capability, in war and peace.”

The “cooperation on economic recovery and reform” provision is stated in terms of an “inten[t] to cooperate,” not an obligation. But in perhaps the most consequential commitment, the parties “shall cooperate” on institutional reforms to advance Euro-Atlantic cooperation, but there’s no commitment to Ukraine’s entrance into NATO, although the agreement “reaffirms that Ukraine’s future is in NATO.”

There are some very thin and vague obligations assumed later in the document. For example, the agreement says that the parties “shall cooperate to advance a just and lasting peace that has broad global support.” Any disagreements about the meaning or application of the agreement “shall be resolved only through consultation between the Parties” and not in any court: If the parties disagree on the agreement, the remedy is to talk about it.

And then there is the “entry into force” provision: “This Agreement shall enter into force upon signature by both Parties. This Agreement shall remain in force 10 years from entry into force and may be extended by mutual written agreement of the Parties.” The parties also intend to register the agreement with the UN. These provisions aim to give the document a patina of bindingness and seriousness. But the agreement that “enters into force” imposes no commitments of substance.

The “termination” provision warrants a similar conclusion and is in many ways the most puzzling provision in the entire document. The 10-year agreement may be extended by mutual agreement, but importantly, it also permits any party to terminate the agreement at any point six months after notification. I am very surprised that the Biden team insisted on a termination clause with such a short window. (Ukraine surely would have preferred a longer one.) This is especially surprising when one contrasts this agreement with last year’s U.S.-Bahrain “Comprehensive Security Integration and Prosperity Agreement,” which also provides (with emphasis added) for a six-month withdrawal period, but in contrast to the Ukraine agreement, the possibility of withdrawal in the Bahrain agreement does not arise until “[a]fter this Agreement has been in force for five years.” The contrast between the five-year commitment to Bahrain followed by a six-month withdrawal period, and the naked six-month termination provision in the Ukraine deal that is invocable at any point, is notable. (Each of the 16 bilateral security agreements that U.S. partners have reached with Ukraine since the Joint Declaration of Support for Ukraine last July have included six-month termination provisions, but the president could have made the notice period longer in the U.S.-Ukraine agreement if he had liked. He also could have included an “in force” period preceding a termination possibility, as in the Bahrain agreement.)

But in truth the length of the termination window doesn’t matter much because the termination clause is basically a ruse; the agreement imposes no obligations of substance to actually help Ukraine, so during the six-month termination period the U.S. need not do anything of substance.


President Biden said that the U.S. goal in the agreement was “to strengthen Ukraine’s credible defense and deterrence capabilities for the long term.” The agreement does not achieve this goal or commit the United States to any real steps toward achieving this goal. One might have thought that the Biden team would make the most precise possible set of commitments to Ukraine’s security to make it more difficult for future, less committed presidents to back away. But it did not.

It is fair to ask what more the Biden administration could have done. A large part of the problem, of course, is that the administration appeared not to think that Congress or two-thirds of the Senate would support a binding international agreement, so Biden was left to rely on unilateral tools. Yet even with this serious constraint, there are at least three ways in which the administration might have made the agreement stronger.

First, as noted, it could have (as in the Bahrain agreement) made the agreement binding for a period before allowing the possibility of termination to kick in, and/or it could have made the termination or withdrawal period longer.

Second, the president could have made stronger executive branch commitments that do not depend on Congress. The president almost certainly cannot pledge in a binding executive agreement to commit the United States to the security defense of another nation. But he can in a legally nonbinding commitment personally pledge that during his presidency he will use his Article II powers to defend (and provide other security commitments to) other nations. (See earlier analysis of the relevant precedents for the president’s Article II power of “collective self-defense” or “collective self-defense of Coalition partnered forces.” And see here analysis of the president’s very broad constitutional authority to make legally nonbinding commitments.) The same Article II power that has permitted President Biden to defend Israel without congressional authorization can be invoked to defend Ukraine and can be so pledged in a nonbinding agreement. This would have been a much stronger pledge for the remainder of Biden’s term, including during his next term, if he wins the 2024 presidential election. It would have been politically controversial, to be sure, and arguably imprudent, but it would have been legally available nonetheless. The president chose not to go there.

Third, there are many very broad open-ended statutory authorizations for military and security assistance and cooperation and related forms of assistance that can be the basis for binding congressional-executive agreements. Moreover, the State Department has sometimes appeared to take the view that the president can make a binding international agreement on any topic as long as it is consistent with domestic statutes. It seems for these reasons that the Biden administration might have pursued a binding congressional-executive agreement here that went further than this agreement. Perhaps the State Department has tapped out all of these sources; I doubt it, but I do not know. But the department knows how to be super aggressive in asserting its international agreement authorities when it wants to.

Whatever President Biden’s legally available options were, the agreement he signed makes no material commitments to Ukraine, and it will be very easy for a future president to ignore, or to interpret away, or to terminate, or all of the above, if that president wishes to not support, or to provide less support to, Ukraine. In that sense, the security agreement definitely does not send a strong deterrent or resolve signal to Putin.

Separate from this agreement, however, “the G7 signed a plan to finalize and unlock $50 billion from the proceeds of those frozen assets, to put that money to work for Ukraine,” as President Biden put it. This is a real commitment and a consequential one that is separate from the security pact.

(An earlier and less nuanced version of this article appeared on X.)

Jack Goldsmith is the Learned Hand Professor at Harvard Law School, co-founder of Lawfare, and a Non-Resident Senior Fellow at the American Enterprise Institute. Before coming to Harvard, Professor Goldsmith served as Assistant Attorney General, Office of Legal Counsel from 2003-2004, and Special Counsel to the Department of Defense from 2002-2003.

Subscribe to Lawfare