The gap between the announcement and the outcome
You are standing in the briefing room. A finance minister steps to the podium, reads the list: frozen assets, blocked transactions, named individuals. The cameras are rolling. The message is legible and firm. This behaviour will have consequences. And then, in the country being targeted, the government carries on.
So why don't sanctions work? The short answer is that they are designed around an economic logic that assumes political leaders behave like rational consumers: apply enough financial pain and they will change course to relieve it. The longer answer is that this assumption is wrong in almost every meaningful case. The history of sanctions is largely a history of that assumption colliding with reality, losing, and being quietly shelved before the next press conference.
The mechanism that looks iron-clad on paper
Here is the theory, stated plainly. Country A imposes costs on Country B's economy: frozen central bank reserves, export controls on critical components, exclusion from international payment systems. The population of Country B suffers. The government of Country B, facing popular discontent and elite defection, calculates that compliance is cheaper than resistance and changes course.
Every link in that chain can break.
In practice, most of them do. Start with the pain-transmission problem. Sanctions hurt an economy broadly, but authoritarian governments have a well-practised tool for redirecting that hurt: concentrate it on populations who were never going to threaten them anyway, and insulate the elites whose loyalty they actually need. Cuba under decades of American sanctions is the textbook case. The shortages were real. The regime was not dislodged, not reformed, not meaningfully pressured into political liberalisation. The government controlled the distribution of scarcity just as deliberately as it had once controlled the distribution of abundance. Sanctions, in that context, functioned less like a vice and more like a weather system: damaging, widespread, and navigated around.
Then there is the substitution problem. Sanctioned economies don't freeze; they reroute. A country cut off from Western semiconductor suppliers starts sourcing older-generation chips through intermediary states, paying a premium but maintaining function. A central bank excluded from dollar clearing shifts reserves into bilateral swap arrangements with sympathetic neighbours. The rerouting is inefficient and it costs money. But it rarely stops anything. The economist Gary Clyde Hufbauer, who spent decades cataloguing sanctions episodes with colleagues at the Peterson Institute for International Economics, found that sanctions achieved their stated policy goals in only about a third of cases, and that figure was generous. It included many instances where the causal role of sanctions was disputed and where the goal had been quietly redefined downward after the fact.
The part most guides skip: what counts as success
Here is the wrinkle the press conference never addresses. Governments announcing sanctions almost never define in advance what specific behavioural change, by what mechanism, would constitute success. This matters enormously, because it means sanctions can be maintained indefinitely without ever being evaluated honestly.
Consider a plausible scenario. Two neighbouring states: call them Arvenia and Beldor. Arvenia imposes sanctions on Beldor following a disputed election. The stated goal is democratic reform. Three years later, Beldor has held another election under the same disputed conditions but has also signed a modest trade agreement with a third country that Arvenia values. Arvenia quietly eases some restrictions, describes the situation as progress, and both governments move on. Was that success? By the original standard, no. By the operational standard that actually governed the decision, possibly yes. The ambiguity is a feature, not a bug. It lets governments claim credit without having to demonstrate change.
This is the part that deserves more honest scrutiny, and it is the part that almost nobody in the room is willing to say plainly. Sanctions serve domestic audiences at least as much as they serve foreign policy objectives. A government that does nothing in response to an atrocity faces political costs at home. Sanctions are the thing you do that looks like a response and carries limited military risk. That is a perfectly legitimate political calculation. It just shouldn't be confused with an effective instrument of behaviour change. Those are two different things, and conflating them has cost the concept whatever analytical credibility it once had.
When they do bite, and why that's rarer than it sounds
Sanctions are not useless in every configuration. The research, including Hufbauer's work and subsequent scholarship by Daniel Drezner at Tufts, identifies the conditions under which they have a reasonable track record: when the target state is small and trade-dependent, when the sanctioning coalition is broad and unified, when the goal is modest and specific rather than sweeping regime change, and when a face-saving off-ramp is built into the design from the start.
South Africa in the 1980s is the most-cited partial success. A broad multilateral coalition, a target economy genuinely exposed to international trade and finance, a domestic business community that ultimately concluded the costs of apartheid were exceeding its benefits. Even there, historians debate how much the sanctions themselves drove the transition versus the internal political dynamics that would have shifted regardless.
The conditions for that kind of success are rarely present in the cases where sanctions get imposed most loudly. Large, commodity-rich states with authoritarian governments and existing relationships with non-Western trading partners fail almost every criterion. Which, frankly, describes the majority of targets in recent decades.
The honest reckoning
What people get wrong about sanctions is the direction of causation. The assumption is that governments impose them when they calculate they will work. The reality is closer to the reverse: sanctions get imposed when governments want to signal resolve but lack the will or ability to do anything costlier, and the question of whether they will actually change behaviour is secondary.
That doesn't make them irrational as a policy instrument. Signalling matters. Coalition-building matters. There is a reasonable argument that sustained economic pressure, even when it fails to produce its stated goal, shapes the long-run cost calculations of regimes and their neighbours in ways that are hard to measure but real. That argument deserves respect.
Still, the gap between the rhetoric at the podium and the outcomes in the targeted country is not a gap that closes with better implementation or stricter enforcement. It is structural, built into the theory of change itself. Ask yourself: when did you last hear a government official explain, before imposing sanctions, exactly which actor inside the target state would feel which pressure, through which channel, and respond in which specific way? The answer is almost never, because spelling that out would expose how thin the causal chain actually is.
Sanctions assume that leaders behave like consumers. Most of the leaders who attract sanctions have spent their careers proving they don't. The press conference will keep happening, the frozen assets will keep being announced, and the honest question nobody asks in the room is this: the mechanism requires someone inside the target government to feel pain and choose compliance. Who, precisely, is that person? And what, in their history, suggests they will?