Debate: America’s Military Role in the World

Last updated 3 weeks ago. Our resources are updated regularly but please keep in mind that links, programs, policies, and contact information do change.

On October 9, 2001, radar planes began patrolling the skies over American cities, and most of the crews inside them were not American. Seven NATO aircraft flew more than 360 sorties between then and mid-May 2002, crewed by 830 crew members from 13 NATO countries. The operation was called Eagle Assist, and it ran because the North Atlantic Treaty Organization (NATO) had invoked Article 5, its mutual defense clause, for the first and only time in its history. The ally under attack was the United States.

Most people have the direction of that promise backwards, and they have its contents backwards too. If a war starts tomorrow on the other side of the world, you are not already committed to fight it: every mutual defense treaty the United States has ratified leaves the choice of response to the United States, and routes it through what the texts call constitutional processes. Someone still has to decide, and in those texts the someone is the American government, acting under American law.

So the live argument is not really about the treaties. It is about the posture built on top of them: the bases, the people stationed at them, what that costs, and how much a president can already do without asking anyone.

The sentence everyone quotes, and the words they skip

Article 5 obliges each member to help the attacked ally by taking such action as it deems necessary, including the use of armed force. Read that clause slowly. The action is whatever the member judges necessary, and armed force is one item on a list rather than the requirement. Article 6 then draws the map: the clause covers an armed attack on the territory of any of the Parties in Europe or North America. Geography is part of the bargain, not an afterthought.

The Pacific treaties narrow the promise twice over. Under the 1953 treaty with South Korea, the trigger is an armed attack in the Pacific area on territory under the parties’ administrative control. Each side then declares only that it would act to meet the common danger in accordance with its constitutional processes. In the American system, the constitutional process for going to war runs through Congress, so the treaty’s own wording hands the decision back to the legislature.

Japan’s treaty uses the same phrase and the same fence, covering an armed attack against either party in the territories under the administration of Japan. It also states plainly what the United States gets back. Under Article VI, the United States is granted the use by its land, air and naval forces of facilities and areas in Japan. That is the shape of nearly every alliance the United States has: a conditional promise in one direction, real estate in the other.

Beyond NATO, the Congressional Research Service (CRS) names the Pacific instruments one at a time. In 1960, the United States signed a mutual defense treaty with Japan. In 1953, it signed a mutual defense treaty with the ROK, the Republic of Korea.

The promises nobody ratified

For Taiwan there is no treaty at all. There is a statute, and it is careful. The Taiwan Relations Act directs the United States to provide Taiwan with arms of a defensive character, and to maintain the capacity of the United States to resist any resort to force or coercion against the island.

Maintaining a capacity is not a promise to use it. The strongest words in the law are that a non-peaceful attempt to settle Taiwan’s future would be of grave concern to the United States, which is a statement of feeling, not of obligation.

Ukraine’s case is the one people most often misremember. On December 5, 1994, at a summit in Budapest, four leaders signed a memorandum: Leonid D. Kuchma for Ukraine, Boris N. Yeltsin for the Russian Federation, John Major for the United Kingdom and William J. Clinton for the United States. Ukraine’s side of it was a commitment to eliminate all nuclear weapons from its territory within a specified period of time.

What the United States gave back was thinner than the trade suggests. It promised to refrain from economic coercion designed to subordinate Ukraine’s exercise of its sovereign rights. And if nuclear weapons were ever used or threatened against Ukraine, the signatories would seek immediate United Nations Security Council action to provide assistance. Asking the Security Council for a vote is the ceiling of that document. The word defend never appears.

The 2024 arrangement that gets called a security guarantee is a different animal again. President Biden and Ukrainian President Volodymyr Zelensky signed a 10-year bilateral security agreement during the G7 summit on June 13, 2024. It is an executive agreement, and it does not require congressional approval.

Congress touches it only from outside, through oversight and the use of appropriations. An instrument a later president can leave and a later Congress can defund is a policy, not a commitment in the treaty sense, and the distinction is the whole reason the Senate exists in this area.

Where the troops actually are

As of March 2024, roughly 67,200 active-duty servicemembers were permanently assigned to overseas bases in Europe. The largest contingents were in Germany (35,068), Italy (12,375) and the United Kingdom (10,058). That tally leaves people out on purpose: thousands more are in Europe on rotational deployments or other temporary assignments, and the permanent count does not see them. Which is how two honest figures for the same continent can differ by tens of thousands, and why the White House letter’s 80,000 is larger than the Pentagon’s file.

Counting bases is harder than counting people, because the government’s own inventory and outside tallies use different rules for what a base is. The anthropologist David Vine, working with Patterson Deppen and Leah Bolger, built the best-known independent count in a 2021 report called Drawdown, documenting an estimated 750 US bases in 80 foreign countries and colonies. That report is advocacy as well as arithmetic; it argues for closures. Read it as a claim with a method attached, which is more than most numbers in this argument come with.

What the posture costs, and what nobody counts

Start with the whole, so the parts have a scale. Total national defense, budget function 050, ran to $926,763 million for fiscal 2025 in outlays, measured in fiscal 2025 constant dollars, against $860,626 million in fiscal 2023. That figure covers everything from a submarine to a private’s pay. The overseas share is a slice of it, and a smaller slice than the argument’s volume implies.

When the Government Accountability Office (GAO) went looking for that slice, it found a hole in the books. The Defense Department (DOD) does not formally track or estimate the value of Japan’s and South Korea’s indirect financial support, and that support is not reflected in the total amount of direct financial support the department reports. So the official contribution totals understate what those two allies provide, by an amount the Pentagon has never measured.

GAO also priced one concrete move, which is rarer than it should be. Before the 2010 Quadrennial Defense Review, the Army planned to bring two of its four brigade combat teams in Europe home, which would have saved millions annually in overseas stationing costs by allowing two installations in Germany to close.

Army analysis put the long-term incremental cost of keeping four brigades in Europe rather than two at between $1 billion and $2 billion across fiscal years 2012 through 2021. Those are real sums and, set beside total national defense outlays, small ones. The posture abroad is not where most of the money goes. It is where most of the commitments live.

Whether the allies pay their share

The famous 2 percent is written as an aspiration, in the grammar of a communique. At the 2014 Wales summit, allies below the line said they would aim to move towards the 2 percent guideline within a decade. A second aim covered hardware: 20 percent or more of total defence expenditures on major new equipment, including research and development. No ally owes NATO money for missing either. That does not make the complaint unfair, but it does mean nobody is in arrears.

The bigger trouble is that three institutions measure the same thing three ways, and each yardstick flatters a different side.

The third row is the one that rarely makes the argument, and Korea supplies its clearest example. On June 29, 2018, Army Gen. Vincent K. Brooks, commander of US Forces Korea and United Nations Command, opened a new headquarters building at Camp Humphreys and dedicated it to Gen. John William Vessey Jr., the first commander of Combined Forces Command.

In posture testimony, the command had told senators the relocation would triple the size of Camp Humphreys. It also told them who was paying: the Republic of Korea is bearing the majority of the relocation’s cost, committing $10 billion to what the command called the department’s largest peacetime relocation project. A reader who has only heard that allies freeload has not heard that.

The case for staying, in the words of the people who make it

The doctrine’s own term is integrated deterrence, and its practical half is a verb. The 2022 National Defense Strategy says the department will campaign day-to-day to gain and sustain military advantages, counter acute forms of competitors’ coercion, and complicate their military preparations. Day-to-day is the load-bearing phrase. A brigade at Fort Bliss cannot do anything daily in the Taiwan Strait. The command responsible for that ocean states the objective in its own grammar: to deter China by prioritizing combat credible military forces postured in the Western Pacific.

The sharpest version of the argument comes with a clock. Across RAND’s Baltic wargames, the longest it took Russian forces to reach the outskirts of Tallinn and Riga was 60 hours. The remedy the study identified was a force of about seven brigades, three of them heavy armored, ready to fight at the onset of hostilities. Sixty hours is shorter than any reinforcement plan on earth. That is the entire deterrence-by-denial case in one number: forces already standing there, or a promise that arrives after the country is gone.

The American Enterprise Institute (AEI) turns the cost question inside out. The expensive outcome, on its reading, is not the alliance but its collapse: the true danger lies in the breakdown of Western deterrence. Preventing that means making clear the United States will in fact fight to defend treaty allies such as the Baltic states, and holding the military power to make the threat credible. Deterrence, in this view, is a bill you pay in peacetime or a bill you pay in war.

To the reply that Europe should simply take over, the Atlantic Council answers with plumbing rather than willpower. NATO relies on US enablers for theater integrated C2, ISR, and mission networks, meaning command and control, and intelligence, surveillance and reconnaissance.

Without the United States, it argues, NATO remains superior in numbers and technology on paper but lacks the operational integration, logistics and joint force capabilities to match Russian mass and tempo near its borders quickly. Gen. Christopher G. Cavoli brought the same point to his posture hearing, telling senators that the European Deterrence Initiative has been vital to this posture by supporting rotational deployments and infrastructure at key locations.

What that looks like on a given night is the USS Carney. On October 19, 2023, over about ten hours, the destroyer shot down 15 drones and four cruise missiles in the Red Sea, in a Houthi attack the Pentagon later said was headed for Israel. Her commanding officer for that cruise, Cmdr. Jeremy Robertson, told Navy Times the ship had not expected a near-daily battle to keep the claustrophobic Red Sea lanes open for commerce.

“None of us really could have known what we were going to get into once Oct. 7 happened,” he said. Both readings of that night are available on the same facts. A ship on station stopped an attack that would otherwise have landed. A ship on station was also in a fight nobody in Washington had voted on.

Presence is also a place where people live, which is why leaving is never only a budget line. When US Army Garrison Schweinfurt closed on September 19, 2014, the lord mayor, Sebastian Remelé, spoke at the ceremony. After almost 70 years, he said, the last American Soldiers and their Families were leaving the location.

At its peak the American community there numbered 12,000 people, a major part of the population of the city and district alike. Their withdrawal left 300 hectares of built-up land behind and marked, in his words, an end of an era in the post-war history of Schweinfurt.

The case for coming home, in the words of the people who make it

Offshore balancing is a rule about when, not a rule against ever. John Mearsheimer and Stephen Walt state it flatly: US troops would be stationed on foreign soil only when a country was in a vital region and threatened by a would-be hegemon.

In that case, they write, the potential victim would view the United States as a savior rather than an occupier. Its advocates get impatient when this is called isolationism, and they have a point. America still comes if one power threatens to dominate Europe, Northeast Asia or the Gulf. It just stops garrisoning the world while waiting to find out.

Their explanation for allied under-spending is not ingratitude but incentives. Defense Priorities calls it free-riding and traces it to the guarantee itself, which is made possible by a state’s commitment to another’s defense.

Once one state says another’s security is vital enough that it will risk national suicide to defend its ally, the ally can pass much of the cost and risk of its own defense to the patron. That is the burden-sharing complaint at its strongest, and it says under-spending is exactly what a rational ally does when someone else has promised to bleed first.

Congress has heard the argument before and voted it down. On May 11, 1971, Senate Majority Leader Michael Mansfield of Montana introduced an amendment calling for a one-half reduction in the United States military presence in Europe. The Senate defeated it 61 to 36 on May 19. He came back in the autumn, and on November 23 the Senate voted 39 to 54 against a ceiling of 250,000 troops in Europe. Half a century later the pattern is unchanged: the drawdown argument reaches the floor, and loses there.

Restrainers also point at what these places cost when they end, and the accounting is not pretty. On August 5, 2024, the Defense Department announced in a joint statement with Niger’s defense ministry that the withdrawal of US forces and assets from Air Base 201 in Agadez is complete.

The pullout came five years after the formal opening of the $110 million base, touted as the largest US Air Force-led construction project in history. At one point roughly 1,000 US troops were in Niger, at Agadez and at a drone installation in the capital, Niamey.

Leaving can also look like Bagram. The Americans left Afghanistan’s Bagram Airfield after nearly 20 years by shutting off the electricity and slipping away in the night, without notifying the base’s new Afghan commander, Gen. Mir Asadullah Kohistani. He discovered the departure more than two hours after they left, Afghan military officials said. Each side reads that night as its own proof, one that withdrawal is a decision that can actually be executed, the other that the consequences of it arrive later and land on someone else.

And a base has neighbors with votes. In February 2019, Okinawan voters rejected the Henoko landfill plan by 72.1 percent. The 434,273 voters who opposed the relocation outnumbered those who had elected Governor Denny Tamaki the previous September.

The result was legally nonbinding on the central government, though it obliged the governor to respect it and relay it to Tokyo and Washington, which he did. Construction went ahead. Whatever you conclude about the alliance, hosting it is a thing done to particular people in particular places, and they do not always get the last word.

Pick a theater: the prioritizers

A third camp wants neither the current posture nor a general retreat. The Heritage Foundation argues for a new distribution of labor among NATO allies. Because the United States must focus tightly on homeland defense and on strengthening deterrence against China, it says, allies must take primary responsibility for Europe’s conventional defense, providing most if not all of the conventional forces needed to defeat Russian aggression, while America keeps the nuclear guarantee over them.

The hard fact underneath that argument is inventory. A Center for Strategic and International Studies (CSIS) wargame of a Chinese invasion of Taiwan found that in three to four weeks of fighting, US forces usually expended about 5,000 long-range precision missiles. In every scenario, the United States expended its global LRASM inventory, its worldwide stock of long-range anti-ship missiles, within the first few days. You cannot fire the same missile in two oceans, and no amount of resolve manufactures a second one.

Secretary of Defense Pete Hegseth said the quiet part to the Ukraine Defense Contact Group in Brussels on February 12, 2025: stark strategic realities prevent the United States of America from being primarily focused on the security of Europe. He named the homeland, the American border, and a peer competitor in the Communist Chinese with the capability and intent to threaten it. Europe, he added, must provide the overwhelming share of future lethal and nonlethal aid to Ukraine.

Within days, European leaders were in emergency talks in Paris, trying to form a united front. A once-solid alliance had been thrown into turmoil, and the Europeans were left questioning the reliability of their key transatlantic partner.

Shortly before that meeting, French President Emmanuel Macron spoke with US President Donald Trump for 20 minutes, and his office would not say what was discussed. That sequence is the prioritizers’ theory and its risk in the same week: allies asked to do more, and allies wondering whether they are being left.

Deter or entangle: what the evidence can and cannot settle

The fear that alliances drag the United States into other people’s wars is old, and one political scientist went looking for the bodies. Michael Beckley’s argument is that a superpower with many allies escapes entanglement by inserting loopholes into alliance agreements, sidestepping costly commitments, and playing allied demands against one another. He finds only five ostensible episodes of US entanglement, and calls even those far from clear-cut.

There is a second effect the guarantee is meant to buy, and it is easier to document than deterrence. CRS records that during negotiations on the Treaty on the Non-Proliferation of Nuclear Weapons in the 1960s, several US allies sought additional American security assurances while pledging that they would not develop nuclear weapons.

American policy has since extended the nuclear umbrella over allies in Europe and Asia partly to reduce the incentive for them to acquire their own weapons. That inverts the usual framing: the alternative to a guarantee is not only fewer American obligations, it may also be more nuclear-armed states.

Neither finding closes the argument, and you should be suspicious of anyone who says it does. Beckley is counting one kind of event, and a war that deterrence prevented leaves nothing behind to count. The honest position is that the entanglement risk is smaller than the rhetoric and the deterrent effect is less measurable than the doctrine.

Who can actually order a war

On paper the split is clean. Article I, section 8 gives Congress the power To declare War, grant Letters of Marque and Reprisal, to raise and support armies, and to provide a navy. Article II makes the President Commander in Chief of the Army and Navy of the United States, which is a command role rather than a power to begin. In practice that division has not held since 1945, and the reason is a legal test the executive branch wrote for itself.

The Justice Department’s Office of Legal Counsel applies two questions. First, whether the operation is supported by sufficiently important national interests to fall within the President’s constitutional power. Second, whether it is extensive enough in nature, scope and duration to be a war in the constitutional sense requiring prior congressional approval. On the Libya deployment, it answered yes to the first and no to the second. Read plainly, that test permits anything short of a war, and lets the executive branch decide where the line sits.

Congress’s counter is a clock. Within sixty calendar days after a report is submitted or required under section 1543(a)(1), the President shall terminate any use of United States Armed Forces covered by that report, unless Congress declares war, passes a specific authorization, or extends the deadline by law. In plain terms: two months, then bring them home or get a vote.

The clock has a defect, and it is not subtle. Presidents have submitted 136 reports to Congress as a result of the War Powers Resolution. Exactly one, the Mayaguez situation, cited Section 4(a)(1) or stated that forces had been introduced into hostilities. Only a 4(a)(1) report starts the countdown. So the countdown has, in practice, never begun.

One officer tried to make a court say otherwise. Army Captain Nathan Michael Smith, an intelligence officer serving in the campaign against the Islamic State, asked the federal district court in Washington to rule that the war against the Islamic State broke the sixty-day limit and had never been authorized by the 2001 or 2002 statutes.

Judge Colleen Kollar-Kotelly dismissed the case on November 21, 2016, finding that the individual service member plaintiff has not alleged personal injury from the alleged failure. She held in addition that dismissal was appropriate under the political question doctrine. Standing and the political question doctrine are the two doors most war powers suits close behind. The sixty-day rule exists; it has no bailiff.

The authorizations still on the books

The 2001 authorization, Public Law 107-40, is a single sentence doing an enormous amount of work. It authorizes all necessary and appropriate force against those nations, organizations, or persons he determines planned, authorized, committed or aided the attacks of September 11, 2001, or harbored those responsible. No country is named. No group is named. The President determines who is covered, and the determination is the authorization.

One member of Congress voted no. Representative Barbara Lee of California told the House she rose with a very heavy heart, one filled with sorrow for the families killed and injured that week. The attack, she said, had forced her to rely on my moral compass, my conscience, and my God for direction.

She was convinced, she told the chamber, that military action would not prevent further acts of international terrorism against the United States. She was the only vote against, and the sentence she voted against is still law.

Where that sentence now reaches is published annually. The section 1264 report on the legal and policy frameworks for the use of military force lists, theater by theater, the groups the executive branch treats as covered, with further detail in a classified annex.

For the 2023 report year, that included strikes and other operations against ISIS and al-Qa’ida in Syria and Iraq. A statute written to answer one morning in 2001 now covers organizations that did not exist that morning, under an associated forces theory the executive branch defines and then reports on to the body that wrote the statute.

The Senate has tried to tidy part of this. On March 29, 2023, it passed S. 316, a bill to repeal the authorizations for use of military force against Iraq, by 66 yeas to 30 nays with four not voting. The House did not take it up, so neither Iraq authorization was repealed.

In an April 2014 report, CRS listed the major statutory authorizations Congress has passed across two centuries, from France 1798 and Tripoli 1802 through Formosa 1955, the Middle East 1957, Southeast Asia 1964, Lebanon 1983, Iraq 1991 and the two modern ones. Of the pre-2001 entries, the one still on the books is the 1991 Iraq authorization.

How to check any of this yourself

Every claim above sits in a document you can open. For the treaties, the State Department publishes Treaties in Force, arranged as bilateral agreements by country and multilateral agreements by subject, and it may be updated periodically throughout the year on the Treaty Affairs webpage. If someone tells you a country is a treaty ally, that list either has it or does not.

For troop numbers, the official file is the Defense Manpower Data Center’s quarterly workbook, Military and Civilian Personnel by Service/Agency by State/Country, downloadable from the department’s workforce reports page. Read it knowing what it counts: permanently assigned people and nobody else, so rotational deployments have to be added from other reporting before the figure is compared with a headline.

For allied spending, go to NATO’s own tables rather than a summary of them, and read the notes first. NATO warns that its figures may considerably diverge from those referenced by media, published by national authorities or given in national budgets. The same notes record that personnel expenditure includes pensions paid to retirees, and that the two most recent years in any edition are estimates rather than actuals.

And for the use of force, the president’s own reports are public. They appear in the Compilation of Presidential Documents on govinfo.gov under communications to Congress, each with a document number. One from June 2025 begins by stating that on the night of June 21, 2025, at the President’s direction, United States forces conducted a precision strike against three nuclear facilities in Iran.

Search that collection for the phrase consistent with the War Powers Resolution and the latest letter surfaces. That phrase is the tell worth knowing: it is how a president reports a use of force while declining to concede that the sixty-day clock has started.

None of these documents is hard to read. A mutual defense treaty is about a page, the 2001 authorization is a paragraph, and the spending tables are published by the institutions being argued over. The argument about America’s role in the world stays loud partly because so few of the people having it have read the sentences that leave the decision still to be made.