How International Sanctions Work, and What the Evidence Says About Them

Sanctions occupy a peculiar position in statecraft. They are stronger than a diplomatic protest and weaker than military action, which makes them the natural choice whenever a government wants to do something visible without doing something irreversible. That convenience has made them ubiquitous. It has not made them well understood.

The main categories

Comprehensive sanctions target an entire economy, restricting most trade and financial dealings with a country. They were the dominant model into the 1990s and have largely fallen out of favour, because the humanitarian consequences proved severe and the political consequences proved limited.

Targeted or smart sanctions replaced them. These aim at specific individuals and entities through asset freezes, travel bans and prohibitions on dealing with named parties. Sectoral sanctions sit between the two, restricting particular industries — defence, energy technology, sovereign debt — rather than the whole economy. Secondary sanctions extend the reach further, penalising third-country firms that continue to deal with the primary target.

Why financial measures do the heavy lifting

The most consequential sanctions are rarely trade restrictions. They are restrictions on access to the financial system, and specifically on the ability to clear transactions in major reserve currencies. A bank that loses correspondent relationships cannot settle international payments regardless of what the underlying trade rules permit.

This creates a mechanism economists call overcompliance. Faced with severe penalties for getting it wrong and thin margins for getting it right, banks and insurers withdraw from anything adjacent to a sanctioned party rather than assessing each transaction. The practical effect of a sanctions regime is therefore usually broader than its legal text, which matters a great deal for humanitarian exemptions that exist on paper but cannot be used because no institution will process the payment.

What the evidence shows

The most cited body of research finds that sanctions achieve at least partial success in something in the range of a third of cases. That figure has been contested in both directions — some scholars argue it overstates success by counting ambiguous outcomes, others that it understates it by ignoring cases where the threat alone produced concessions and no sanctions were imposed.

The findings that hold up most consistently across studies are about conditions rather than headline rates. Sanctions work better when the objective is modest and specific rather than regime change. They work better when imposed multilaterally, because unilateral measures are more easily circumvented through third countries. They work better when applied quickly and at full strength rather than escalated gradually, since gradual escalation gives the target time to adapt supply chains and financial routing. And they work better against countries with close prior economic ties to the sanctioning state, simply because there is more to withdraw.

The adaptation problem

Targets are not passive. Over time, sanctioned economies reroute trade through intermediary jurisdictions, shift to alternative payment channels, substitute domestic production for restricted imports at higher cost, and build reserves in assets that are harder to freeze. None of this is costless — substitution is generally inefficient and the discount demanded by intermediaries is real — but it converts a sharp shock into a chronic drag.

Adaptation also has a political dimension. Sanctions can strengthen a targeted government’s position domestically by supplying an external explanation for economic hardship and by concentrating control over scarce foreign exchange in the hands of the state. This rally effect is well documented, though it tends to weaken as hardship persists.

The measurement difficulty

Assessing whether sanctions worked is harder than it appears, because the counterfactual is unobservable. If a targeted state does not do the thing it was warned against, that may be deterrence or it may be that it never intended to. If it does, sanctions may have failed or may have delayed and raised the cost.

There is also the question of what the objective actually was. Stated aims and operative aims often diverge. A sanctions package may be publicly justified as changing a government’s behaviour while actually being designed to constrain its capabilities, signal resolve to allies, or satisfy domestic political demand for action. Judged against the stated aim it fails; judged against the operative one it may not.

Costs to the sanctioning side

Sanctions are not free for those imposing them. Exporters lose markets, and the loss is concentrated in specific firms and regions while the benefit is diffuse — a distribution that generates persistent domestic lobbying for exemptions. Financial sanctions carry a longer-term cost as well: each high-profile use gives other states an incentive to develop alternative payment infrastructure and to diversify reserves, gradually reducing the leverage the measures depend on.

How to read a sanctions announcement

Three questions usually reveal how much a new package will actually do. Does it restrict financial clearing, or only trade in goods? Is it multilateral, or will the target be able to route around it through a major economy that has not joined? And does it include secondary sanctions, which determine whether third-country firms face a genuine choice between markets? A package that is trade-only, unilateral and free of secondary provisions is a signal. One that hits clearing, is jointly imposed and reaches third parties is a policy.