The United States and Britain launched strikes against Houthi targets in Yemen on Friday, hitting sites associated with missiles, drones, radar and other military capabilities after weeks of attacks on commercial shipping in the Red Sea. Washington and London describe the operation as a defensive response intended to restore freedom of navigation through one of the world’s most important maritime trade routes. The Houthis have promised that the strikes will not end their campaign, which they say is intended to pressure Israel over Gaza. A confrontation that began with drones and missiles threatening commercial vessels has therefore moved into direct military exchanges between Western powers and an armed movement that has controlled much of northern Yemen for years.
Last week, I argued that the principal danger across the Middle East was escalation produced by actors who insist they do not want a larger war. This week’s strikes demonstrate exactly how that process operates. The United States did not begin bombing Yemen because it wanted another Middle Eastern conflict. It did so because repeated warnings and naval interceptions had not stopped Houthi attacks, creating pressure to demonstrate that threatening international shipping carries costs. The military action is therefore intended not to expand the conflict but to restore deterrence. Whether it succeeds depends on something more complicated than how many Houthi missiles are destroyed.
Deterrence is ultimately an attempt to change another actor’s calculation. The basic logic appears simple: if behaviour produces sufficiently large costs, rational actors should stop the behaviour. Governments therefore threaten sanctions, military retaliation or other punishment to increase the expected price of actions they want to prevent. But that logic works only if policymakers understand what the adversary actually values.
The Houthis are not a shipping company attempting to maximise profits. They are an armed political movement whose power emerged from Yemen’s civil war and whose legitimacy depends partly on resistance narratives, domestic political control and opposition to foreign intervention. Being attacked by the United States therefore creates military costs while potentially producing political benefits.
That makes the deterrence calculation unusual.
Destroying a missile launcher clearly reduces Houthi military capability.
American bombs falling on Yemen can simultaneously strengthen the Houthis’ claim that they are confronting the United States and Israel on behalf of Palestinians.
Military punishment can therefore reduce one component of the adversary’s utility while increasing another.
The net effect determines behaviour.
This is why the assumption that “more punishment produces more deterrence” can fail. If an armed organisation values ideological credibility, recruitment and political legitimacy highly enough, limited military retaliation may actually reinforce the incentives sustaining its campaign.
The same mechanism has appeared repeatedly in asymmetric warfare. Militarily weaker organisations cannot defeat great powers conventionally, so they redefine success. Surviving becomes success. Provoking an overreaction can become success. Generating publicity becomes success. Demonstrating that a superpower needs to devote expensive military resources to them becomes success.
The weaker actor therefore operates under a different victory condition.
The United States wins conventionally by destroying military capabilities.
The Houthis can claim victory simply by continuing to fire after those capabilities are attacked.
This creates an asymmetric burden of proof. Washington needs attacks on shipping to decline substantially before it can demonstrate that deterrence worked. The Houthis need one missile launch after the strikes to demonstrate publicly that they were not intimidated.
The political contest is therefore much easier for the weaker actor to narrate.
This is one reason wars against insurgent and non-state organisations become frustrating for conventional militaries. A state can destroy substantially more equipment, kill more fighters and control more territory while struggling to translate those advantages into political compliance.
Military superiority determines the ability to impose costs.
It does not determine how the opponent interprets those costs.
The Houthis’ history makes this particularly important. They have already survived years of war against a Saudi-led coalition possessing vastly greater air power and supported by sophisticated Western weapons. Yemen endured devastating humanitarian consequences, yet the Houthis ultimately strengthened their territorial position across much of northern Yemen rather than disappearing.
That experience presumably affects their expectations today.
An organisation that survived years of Saudi air strikes may not consider several nights of American and British bombing sufficient reason to abandon a politically valuable campaign.
Past survival changes present risk tolerance.
This is why credibility operates on both sides. Washington wants the Houthis to believe attacks will produce escalating costs. The Houthis want Washington to believe those costs will not change their behaviour.
Each therefore needs to demonstrate resolve.
That creates the escalation mechanism from last week.
The United States strikes because warnings failed.
The Houthis respond because failing to respond could make the strikes appear successful.
Washington then faces another choice if attacks continue: accept that the first strikes failed or impose additional costs.
The first option risks weakening deterrence.
The second moves further up the escalation ladder.
Nobody needs to want war.
Each side only needs to believe backing down first is more costly than another round.
The Red Sea adds another complication because the United States is defending a public good rather than simply defending American territory. Commercial freedom of navigation benefits Europe, Asia, Gulf states, Egypt and global shipping companies. If one armed organisation can substantially close a major maritime corridor through relatively cheap missiles and drones, similar strategies become more attractive elsewhere.
Washington therefore worries about precedent.
Allowing the attacks to continue could signal that strategically positioned non-state groups can impose enormous global economic costs without facing meaningful consequences.
That concern is legitimate.
The problem is that defending a principle can create reputational escalation. Once governments describe an issue as a test of credibility, backing down becomes harder because the consequences are no longer evaluated only through the immediate dispute. Leaders begin worrying about what China, Iran, Russia or other adversaries will infer.
A missile launch in Yemen becomes connected conceptually to deterrence in completely different regions.
That can lead governments to overinvest in small confrontations because they fear every compromise will be interpreted globally.
Credibility is rarely that transferable.
China’s judgement about American willingness to defend Taiwan depends much more on American interests and capabilities around Taiwan than on whether Washington destroys every Houthi missile launcher in Yemen.
Yet political leaders frequently fear reputational consequences because they cannot know exactly how adversaries interpret restraint.
This creates another incentive towards escalation.
The economic stakes remain significant. The Red Sea normally carries roughly 12% of global trade, while the Suez route dramatically reduces travel time between Asia and Europe. Shipping companies diverting around the Cape of Good Hope face longer voyages, greater fuel consumption and reduced effective fleet capacity. The disruption therefore imposes costs on businesses and consumers with no involvement in the Gaza war.
This gives the international community a strong interest in restoring navigation.
But it also creates a difficult question about who should pay for global security.
The United States and Britain are currently providing much of the military capability required to intercept attacks and strike launch infrastructure. European and Asian economies receive substantial benefits from secure shipping.
This resembles the burden-sharing problem inside NATO.
Security is a public good.
Every country benefits from somebody protecting the route.
Each country would prefer somebody else bear more of the military and financial cost.
The multinational nature of the naval coalition partly addresses this, but participation remains uneven. Governments differ over whether military strikes will actually improve security or risk expanding the Gaza conflict.
That disagreement is not necessarily free-riding.
It can reflect different estimates of the same strategic probabilities.
One government believes strikes reduce future attacks.
Another believes strikes increase Houthi mobilisation and therefore increase future attacks.
The policy disagreement depends on assumptions about behaviour rather than simply willingness to contribute.
This is why evaluating the operation requires identifying the actual objective. If the objective is destroying Houthi military infrastructure, the United States and Britain possess overwhelming capacity to do so repeatedly. If the objective is stopping attacks on commercial shipping, military destruction matters only insofar as it changes Houthi behaviour.
Outputs and outcomes are different.
Bombs dropped are outputs.
Launch sites destroyed are outputs.
Ships passing safely through the Red Sea are the outcome.
Governments often measure the first two because they are easier to count.
Strategy should be judged through the third.
This distinction becomes even more important because the Houthis can adapt. Mobile missile systems can move. Drones can be launched from dispersed locations. Equipment can be concealed among civilian infrastructure. Iran can potentially provide additional technology or components.
The marginal cost of eliminating each additional capability may therefore rise.
The first wave hits known fixed sites.
Subsequent waves search for increasingly dispersed targets.
Military campaigns frequently experience diminishing returns.
Meanwhile, civilian risks increase if targeting expands.
Yemen has already experienced one of the world’s worst humanitarian crises after years of civil war. Renewed large-scale conflict could further damage infrastructure and undermine the fragile reduction in fighting achieved between the Houthis and Saudi Arabia.
This creates another strategic externality.
The United States wants to secure international shipping.
Escalation could destabilise Yemen.
Instability could create more armed activity around the same shipping route.
A policy intended to protect the Red Sea could therefore undermine itself if military expansion recreates broader war inside Yemen.
Again, tactical success can coexist with strategic failure.
This is why Saudi Arabia appears particularly cautious. Riyadh spent years fighting the Houthis and has increasingly sought to disengage from Yemen through negotiations. A renewed American-Houthi confrontation risks drawing Saudi territory back into a conflict it has been trying to exit.
Saudi incentives therefore differ from American incentives despite both opposing Houthi attacks.
This is a useful reminder that alliances do not erase geography.
The United States can project military power into the region and eventually leave.
Saudi Arabia remains next door.
Countries closer to conflicts often possess lower tolerance for escalation because they bear more of the consequences if deterrence fails.
The same principle explains European caution around Russia and South Korean caution around North Korea.
Geographic exposure changes strategic preferences.
Iran’s role again complicates deterrence. Tehran supports the Houthis but does not necessarily command every Houthi decision. Striking Iran directly could potentially reduce material support while massively increasing escalation risk. Striking only the Houthis may leave the wider supply network intact.
This is the proxy deterrence problem.
Do you deter the agent?
The sponsor?
Both?
Punishing the sponsor can create a larger war.
Punishing only the agent may not change the sponsor’s incentives.
The ambiguity itself benefits Iran because Tehran can obtain strategic pressure while maintaining some distance from individual attacks.
This is why networks of armed partners have become such an effective component of Iranian regional strategy.
They distribute coercive capacity.
They also distribute responsibility.
The United States then needs to determine which level of the network should bear consequences.
Mistakes can escalate rapidly.
There is also a connection to Gaza that military action in Yemen cannot eliminate. The Houthis explicitly frame their campaign around Palestinian suffering and the Israeli operation in Gaza. Whether that justification accurately describes every strategic motive is less important than the fact that it provides political legitimacy for their actions across parts of the Arab and Muslim world.
A reduction in the Gaza war could therefore weaken one source of political support for Red Sea attacks.
That does not mean Hamas or the Houthis would disappear.
It means military fronts are politically connected even when operational command is separate.
This is why regional de-escalation ultimately requires addressing the central political conflict feeding mobilisation, not merely intercepting its secondary manifestations.
Military power can suppress symptoms.
Political settlements alter incentives.
The difficulty is that political settlement in Gaza currently appears extraordinarily distant.
The United States therefore confronts a familiar problem: it needs to manage the consequences of a conflict it cannot resolve quickly.
Naval defence becomes necessary because diplomacy has not removed the threat.
Air strikes become attractive because naval defence alone has not stopped attacks.
Every layer compensates for the failure of the previous one.
That is how limited interventions expand.
Mission creep does not always begin with governments secretly wanting a larger mission.
It can emerge through incremental problem-solving.
Step one does not work completely.
Step two becomes necessary.
Step two creates another problem.
Step three addresses that.
Months later, the intervention looks radically different from what policymakers initially intended.
The history of Western involvement in the Middle East contains enough examples that Washington should be especially sensitive to this dynamic.
That does not mean military action is always wrong.
It means every escalation needs a clearly defined condition for stopping.
What would count as success?
A week without attacks?
A month?
A formal Houthi declaration?
Safe passage restored sufficiently that insurers reduce premiums?
Without a measurable objective, military operations can continue because the adversary still exists rather than because the original strategic purpose remains unmet.
The distinction is essential.
The United States does not need to defeat the Houthis as a political movement to restore Red Sea shipping.
Trying to accomplish the larger objective could turn a maritime-security operation into another Yemen war.
Strategic discipline therefore requires matching the scale of force to the scale of the objective.
The Houthis have incentives to blur that distinction. The larger the confrontation becomes, the more important they become internationally.
Before the Red Sea campaign, Yemen’s conflict had received declining global attention.
Now the Houthis are confronting the United States and Britain, affecting global shipping and appearing in international headlines daily.
Escalation increases their political relevance.
That itself can be a reward.
This creates another reason punishment may fail to deter.
The Houthis can lose physical assets while gaining status.
For an armed political organisation, status can generate recruitment, funding and domestic legitimacy.
Again, the cost-benefit calculation is multidimensional.
This does not mean Western governments should allow attacks simply because retaliation might strengthen the Houthis politically. That would give armed groups a convenient immunity: become sufficiently capable of benefiting from retaliation and nobody can punish you.
The more sophisticated strategy is reducing the political payoff while increasing the operational difficulty of attacks.
Defensive interception can do part of this because a failed missile receives less political value than a successful strike.
Diplomacy with regional governments can reduce material support.
Economic measures can target financing.
Military action can degrade specific capabilities where the expected benefit exceeds escalation risk.
No single instrument needs to solve the entire problem.
This is layered deterrence.
It resembles infrastructure resilience.
Several imperfect defences can collectively produce a robust system.
The mistake is expecting one dramatic strike to change an adversary’s preferences permanently.
Military power is often most effective when integrated into a broader incentive structure rather than used as a substitute for one.
The next several weeks will therefore reveal more than whether American and British weapons can destroy Houthi facilities. That outcome is not seriously in doubt.
The important information will come from Houthi behaviour and commercial behaviour.
Do attacks decline?
Do shipping companies return?
Do insurance costs fall?
Does Iran restrain support?
Does the Yemen peace process survive?
Does regional escalation remain contained?
Those are strategic indicators.
If the strikes destroy dozens of targets while shipping remains diverted and attacks continue, tactical success will not have produced the desired outcome.
That distinction is important because powerful states can become trapped by their own military competence.
When you possess the ability to strike almost any target, striking becomes an available answer to many problems.
The harder question is whether the target’s destruction changes the political mechanism producing the problem.
The United States and Britain have demonstrated that Houthi attacks will not remain costless.
Now the Houthis will decide whether those costs exceed the benefits they believe the campaign provides.
That is the actual deterrence test.
Last week, the problem was how a relatively weak movement could exploit geographical leverage to impose global economic costs.
This week reveals the next layer: economic leverage can be easier to destroy physically than politically.
The West can bomb missile sites.
It cannot bomb the incentive that makes firing the next missile worthwhile.
Changing that incentive is what determines whether deterrence succeeds.
