Can financial sanctions work?
In 2022 the West froze a great power’s reserves and called it a financial nuclear strike. The war didn’t end. So which was it — the decisive weapon, or expensive theater?
Voir comme graphe de débatThe weapon, and the shrug
“We are causing the collapse of the Russian economy.”
— Bruno Le Maire, French Finance Minister, on the freezing of Russia’s reserves, March 1, 2022
“The ruble is back to pre-war levels. Russian GDP fell about 2% in 2022 and grew in 2023. The oil is still selling. By every measure that was supposed to break the regime, the sanctions failed.”
— the rebuttal that hardened within months, drawn from IMF and Bank of Russia data, 2022–2023
In February and March 2022, the United States and its allies froze roughly $300 billion of the Russian central bank’s reserves and cut major Russian banks off the SWIFT messaging system. No one had ever immobilized a G20 central bank’s war chest before. It was hailed as the weaponization of finance — and within months the standing rebuttal was that the patient hadn’t died. The era of policy this sits in — the post-2008 order where central banks and finance ministries learned to improvise — is History Ch.19 §19.7 (Institutional improvisation as the new normal). Both of the claims above are true. That is the problem.
What does it even mean to “weaponize finance”? Concretely, a controlling state can deny a target four things: access to dollar-clearing (the pipes through which almost every cross-border payment ultimately settles), correspondent-banking relationships (the accounts foreign banks must hold to move dollars at all), the SWIFT messaging rails that tell those banks what to do, and — the unprecedented 2022 move — the target’s own reserves, frozen where they sit on the books of Western central banks. That is the whole apparatus, named. Whether naming it settles anything is the rest of this walkthrough.
“The dollar is the most powerful weapon we have. It is more powerful than any aircraft carrier.”
— the maximum-pressure thesis, as voiced across a decade of Treasury and national-security officials
“Finance is the decisive weapon”
One country sits at the center of the world’s payment system and can switch any other country off. On this view, 2022 wasn’t an experiment that failed — it was the opening move of a kind of power no army can match.
The blow and the recovery
“The freezing of the assets of the Russian central bank is the most significant economic-statecraft action of the twenty-first century. We have shown that the reserves a country builds for safety can be turned, in a weekend, into a liability.”
— the Western coalition’s framing of the 2022 reserve freeze, February–March 2022
The weaponization camp has a real case, and it is worth stating in its own voice. The 2022 package was a genuine turning point in statecraft: it didn’t target a firm or a sector, it targeted the monetary defense of an entire nation-state. Freezing the reserves meant Russia couldn’t spend its own savings to prop up the ruble through the usual channels. And it was delivered by a coalition — the United States, the European Union, the United Kingdom, Japan, Canada — acting in days, which is itself a demonstration of what coordinated financial power can do. The move was unprecedented, and it bit.
“And yet: the ruble is stronger than before the war, oil revenues funded the budget, and Moscow rebuilt its import supply through third countries within a year. If this was a knockout punch, the opponent is still standing.”
— the adaptation rebuttal, drawn from 2022–2024 trade and currency data
The other voice is, for now, deliberately incomplete — the full case comes in Stage 3. But it has to be planted here, because it is the fact the whole walkthrough has to metabolize. Russia adapted. The reserves were frozen, but the central bank deployed capital controls and a punishing interest-rate hike, the ruble recovered within weeks, and the oil kept flowing to buyers in Asia through a re-flagged shadow fleet. Cost was imposed; collapse was not. The question Stages 2 and 3 will adjudicate with apparatus and evidence is the one this voice raises and can’t yet answer: when does the blow translate into the goal?
Where this leaves us
Both framings are true at once. The 2022 package really was unprecedented, and it really did impose cost — and Russia really didn’t collapse, and the war didn’t end. Which means the headline question, “did it work?”, is the wrong question until two prior ones are answered: can financial sanctions bite at all, and when does biting translate into the goal? The rest of this walkthrough refuses to score the Russia case until those are settled. To settle them, we build the strongest case that sanctions work — then the strongest case that they don’t — and only then return to score the freeze.
Start with the optimists at their best — because the financial weapon really is more powerful than the skeptics admit, and there is one canonical case where it brought a determined adversary to the negotiating table.
The case that they work
“Iran’s oil exports fell by more than half — from about 2.5 million barrels a day to under 1 million — and its banks were cut out of the global system. That pressure is what brought Tehran to the table.”
— the documented 2012–2015 collapse of Iranian oil exports that preceded the nuclear deal
Before scoring failures, give the optimists their strongest exhibit. The leverage is real and the mechanism is specific — and the Iran case is the canonical example of sanctions coercing a determined adversary into a genuine concession. To see why financial sanctions can bite, you have to see the plumbing.
The financial-chokepoint mechanism. The international payments system is not a neutral network. Nearly every cross-border dollar payment ultimately clears through a small set of correspondent accounts that touch the United States, and a comparable concentration holds for euro-clearing through Europe. A controlling state can deny a target access to those rails — no clearing, no correspondent accounts, no messaging. The reserve-freeze is the newer and sharper escalation: a target central bank holds much of its safety buffer abroad, in the form of claims on Western central banks and governments, and freezing those claims disables the target’s own monetary defense. It cannot deploy reserves it cannot reach. The accounting of who holds claims on whom across borders is the home of Ch 17 §17.1 (Balance of Payments Accounting); the demand for reserve balances that makes a central bank’s buffer matter is Ch 16 §16.1 (Why Hold Money?).
Secondary sanctions as the force-multiplier. The chokepoint extends extraterritorially. A bank in a third country that keeps doing business with the target can itself be cut off from dollar access. That single rule turns a unilateral American measure into a near-global one without needing anyone else to agree: faced with a choice between a sanctioned customer and the dollar, almost every bank picks the dollar. This is why a financially-integrated target is so much more squeezable than an isolated one — the more wired into the system it is, the more surfaces there are to squeeze.
Targeted versus comprehensive. The optimists’ case rests specifically on targeted financial measures — precise strikes on named banks, elites, and transactions, coordinated across allies — not the blunt comprehensive trade embargoes of an earlier era. The distinction matters enormously, because Stage 3’s sobering base rate codes the whole sanctions universe, embargoes included. Whether targeted financial sanctions outperform that embargo-heavy average is the live question; the works-case is the claim that they do, precisely because they hit the one layer — dollar-clearing — that is hardest to route around.
The third-party bank’s choice is a simple comparison. Let $V_d$ be the value of retained dollar access and $V_t$ the value of the target-related business. The bank complies with the sanction whenever
$$V_d - V_t > 0$$For all but a handful of niche institutions, $V_d \gg V_t$, which is why secondary sanctions cascade through the system without an enforcement army. The target’s capitulation, by contrast, turns on whether the imposed cost exceeds the value it places on the contested behavior — a far higher and more variable threshold, which is exactly why the chokepoint is reliable and the outcome is not.
Picture every road to the global economy passing through a single tollbooth, and the United States owns the tollbooth. That is the chokepoint. Secondary sanctions are the rule that anyone caught giving the banned traveler a ride loses their own toll pass — so the whole world enforces the ban to protect its access. The catch: controlling the tollbooth lets you make a target poorer. It does not, by itself, let you decide what the target does next.
One more piece belongs here, and it is older than the dollar system. Albert Hirschman showed in 1945 that asymmetric trade dependence is itself coercive leverage: the state that can do without the relationship holds power over the state that can’t. Financial sanctions are that insight applied to the payment system rather than the goods trade. The same logic — that openness creates dependence, and dependence creates a lever — runs through the trade-policy debate; the efficiency-and-distribution side of it is the home of the free-trade walkthrough, and the specific mechanics of freezing a central bank’s reserves connect to the central-banks walkthrough, which owns how those reserves are held and managed in the first place.
“It was sanctions that brought Iran to the table. The pressure was real, it was biting, and it produced a negotiation that would not have happened otherwise.”
— the case for the 2015 nuclear deal as a sanctions success, as argued by its negotiators
“The Iran deal is the proof”
Sanctions cut Iran’s oil exports in half and locked its banks out of the system. Then Iran signed a deal capping its nuclear program. If that isn’t sanctions working, what would be?
What the success-rate metric misses
“Counting only the cases where a target fully capitulates ignores everything sanctions do in between: they raise the cost of a behavior, they signal resolve at a price the bluffer won’t pay, and they bind a coalition. The metric understates the tool.”
— the economic-statecraft case that efficacy is wider than stated-goal success
The works-camp, at full strength, makes two arguments. First, the financial-centrality leverage is unmatched and the Iran case proves it can coerce. Second — and this is the deeper point — the usual scorecard asks only whether the target fully gave in, which throws away most of what sanctions accomplish. A costly signal is valuable precisely because it is costly: a state that imposes real economic pain on itself and its allies to punish a behavior communicates resolve that a speech cannot. The 2022 package, even granting it didn’t end the war, did something the metric misses entirely — it bound a sprawling coalition into common action and raised the standing cost of aggression for the next aggressor watching. Coercion at the margin, signaling, coalition-building: these are real functions, and a tally of capitulations counts none of them.
“Cost imposed is not the same as the goal achieved. Iran absorbed the pain, signed a deal, and then the deal fell apart — and Iran is closer to a weapon now than it was in 2015. Where exactly is the win?”
— the skeptic’s opening question, argued in full in Stage 3
This voice is, again, deliberately partial here — its full case is Stage 3. But it lands one blow that the works-camp has to absorb: the gap between imposing cost and changing behavior. Yes, Iran felt the squeeze; yes, it negotiated; and yes, within three years the arrangement had collapsed and the underlying problem was worse. If the measure of success is the strategic outcome rather than the economic bruise, the Iran case looks less like a clean win and more like a temporary, reversible pause. That objection is the doorway into the case that sanctions don’t work — which is where we go next.
Where this leaves us
Financial sanctions can bite, and harder than a generation ago, because the world’s payments plumbing runs through a chokepoint the United States controls — and secondary sanctions extend that chokepoint across the globe. And they have genuinely worked: the Iran case shows a determined adversary coerced into a real concession, and the 2022 coalition shows the signaling-and-binding function the success-rate metric never counts. But look closely at what worked — a narrow goal, a strong coalition, a financially-integrated target — and the shape of where they don’t is already visible. The conditions that made Iran a win are exactly the conditions most campaigns lack.
Because for every Iran-2015 there is a Cuba blockaded since 1960, an Iran that survived four decades of pressure, a North Korea that built a bomb under the heaviest sanctions in history. The skeptics have a number, and it is brutal.
The case that they don’t
“Across the largest catalog of cases ever assembled, economic sanctions achieved their stated goals at best about a third of the time — and that third is dominated by the modest goals. For ambitious goals, the success rate collapses.”
— the Hufbauer–Schott–Elliott coding in Economic Sanctions Reconsidered, the canonical empirical anchor
The strongest case against the optimists is not an argument. It is a record. Name a single regime that sanctions toppled — the list is short to the point of empty, while the list of regimes that endured decades of pressure is long: Cuba, North Korea, Iran, Venezuela. Now give the failure-case its own full-strength stage, the way Stage 2 gave the works-case its own.
The base rate, and the selection problem. The Hufbauer–Schott–Elliott project coded the modern sanctions universe and found partial success in roughly a third of cases — with success heavily skewed toward modest goals. Robert Pape’s reanalysis pushed the rate far lower by insisting on stricter causal attribution; Daniel Drezner and others defended and extended the coding while conceding the ambitious-goal failure. Underneath the dispute sits a selection problem that cuts both ways: sanctions are imposed on the hard cases — the targets that wouldn’t back down to a threat — which biases the observed success rate downward, and they are often threatened but not imposed precisely when they would have worked, which hides their successes from the count entirely. The coded rate, in short, is not the causal rate.
A scope seam that matters. That base rate codes the whole sanctions universe, dominated by comprehensive trade embargoes of the old kind. The question this walkthrough is built around is narrower — targeted financial sanctions — and whether the precise, dollar-clearing-layer tools of the 2010s and 2020s beat the embargo-heavy average is itself contested. It is not honest to swing the embargo-inclusive number at financial sanctions as if they were the same instrument; nor is it honest to assume the new tools escape the old record. The truthful position is that financial sanctions are a subset whose efficacy may differ, and the magnitude of any difference is exactly what the profession argues about.
Targets adapt. A financially-integrated target does not sit still under pressure — it re-routes. Russia after 2022 is the live demonstration: parallel imports through neighboring states, a shadow fleet of tankers moving oil to Asian buyers outside the Western insurance and shipping system, settlement in non-dollar currencies, and a central bank that defended the ruble with capital controls and rate hikes. Iran built four decades of workarounds. Each evasion costs efficiency — the target pays more, sells at a discount, runs slower — but it blunts the chokepoint, and over time the blunting compounds. The open-economy mechanics of this re-routing — trade diversion, capital flight, the search for new counterparties — are the subject of Ch 17 §17.7 (Global Imbalances and Capital Flows).
The harm lands on civilians, and can strengthen the regime. There is a paradox at the heart of sanctioning an authoritarian state. The pain falls first on ordinary people — medicine gets scarce, prices spike, savings evaporate — while the regime, which controls the now-precious hard currency, can ration it to loyalists and tighten its grip. Scarcity that the state monopolizes is a tool of control, not a constraint on it. And the population, watching a foreign power impose that scarcity, may rally to the flag rather than against the ruler. The more autarkic the regime — the more it can wall its people off — the more sanctions hit civilians and the less they bind the leadership. Venezuela is one entry in this record; the full causal story of why its collapse did not produce reform, and how much weight the post-2017 financial sanctions on its oil company carry in that story, belongs to the Venezuela walkthrough, which treats sanctions as one disputed cause of a single collapse rather than as a general instrument.
And over-use erodes the weapon itself. Every time the financial weapon is fired, every target and every nervous bystander gets another reason to build an exit — to hold reserves in gold rather than dollars, to settle in other currencies, to construct payment rails outside Western reach. Central banks have been buying gold at record pace; non-dollar settlement is growing at the edges. This is the self-attenuation channel, and it is where this walkthrough hands off to the dollar-reserve-status walkthrough: that one owns the question of whether weaponizing the dollar erodes its reserve status; this one owns whether the weapon achieves its goals in the first place. Same 2022 event, two consequences.
“Name one regime that sanctions toppled. Cuba, North Korea, Iran, Venezuela — all sanctioned for decades, all still standing. The modal outcome is a hurt population and an entrenched ruler.”
— the strong-skeptic challenge, in the Pape tradition
“The record is the rebuttal”
The optimists have anecdotes. The skeptics have a base rate — success roughly a third of the time, and almost never for the goals that matter most. The survivors are the argument.
“Broad sanctions function as collective punishment of an entire civilian population for the conduct of a government they do not control.”
— the humanitarian-harm critique, voiced by UN Special Rapporteurs and the mortality-econometrics literature (CEPR / Weisbrot)
“The bill goes to the wrong people”
Sanctions are aimed at rulers and land on the ruled. The moral charge is serious — and it is a separate axis from whether the tool achieves its goal.
Theater, or a different animal?
“Sanctions rarely succeed in coercing target states. They persist because they let leaders be seen to do something forceful short of war — the appeal is domestic and symbolic, not strategic.”
— the strong-skeptic position, in the tradition of Robert Pape
The strong skeptic, at full strength, makes a deflationary case: the success rate is dismal once you demand real causal attribution, cost imposed is routinely mistaken for goal achieved, and sanctions endure not because they work but because they are politically useful at home — a way to register outrage and project resolve without the costs and risks of force. On this reading, the 2022 package and the Iran campaign alike are better understood as expressive politics than as instruments that reliably bend adversaries. The graveyard of decades-long campaigns against unbowed regimes is the evidence.
“The base rate understates the tool. It is biased by selection toward the hard cases, it misses the wins extracted by threat alone, and it lumps surgical financial measures together with the blunt embargoes of the past. Targeted financial sanctions are a different animal.”
— the calibrated defense, in the tradition of Daniel Drezner
The calibrated defender does not deny the record — the ambitious-goal failure is conceded. The argument is that the headline number is the wrong measure of the modern tool. Selection bias drags it down; the threatened-but-not-imposed wins never enter it; and crucially, the coding mixes comprehensive trade embargoes with the precise, dollar-clearing-layer financial measures that are the actual subject here. Whether targeted financial sanctions beat the embargo-heavy average is the genuinely open empirical question — and the honest answer is that we do not yet know the magnitude, only that treating the old number as the verdict on the new instrument is a category error.
Where this leaves us
The skeptics’ number is real and it is brutal: ambitious goals — regime change, war reversal, broad capitulation — almost never come, and the modal outcome of a maximum-pressure campaign against an autarkic regime is a hurt population and an entrenched ruler. But “rarely achieves stated goals” is not “never works.” The third that works is real; it skews toward modest, well-defined, multilaterally-coordinated goals against integrated targets; and the success-rate metric systematically misses the coercion-at-the-margin, signaling, and coalition-binding the optimists were half-right about. Stage 2 and Stage 3 don’t cancel — they locate the answer. It is conditional, and the conditions are now visible.
Which lets us finally answer the question we started with — and score the Russia-2022 freeze honestly, including a cost the optimists never put on the ledger.
A real but limited tool
“Every time we use the financial weapon, we give the rest of the world one more reason to build something we can’t touch. We are sawing off the branch we sit on.”
— the de-dollarization-blowback warning, voiced by analysts including Zoltan Pozsar and former Treasury officials
The honest answer to “can financial sanctions work?” is not a yes or a no. It is a conditional — plus a long-run cost the headlines never count. A scalpel, not a hammer; and a scalpel that dulls itself a little with every cut.
No new machinery is needed here — only a way to read the two preceding stages together. Take the Stage-2 chokepoint leverage and the Stage-3 base-rate-and-adaptation record, and run both through four conditions: how narrow the goal is, how strong the multilateral coordination is, how financially integrated the target is, and whether the aim is marginal coercion or maximalist capitulation. Iran-2015 scores high on all four and succeeded; a maximum-pressure campaign against an autarkic regime bent on survival scores low and fails. The conditions are not a hedge. They are the answer.
There is one more layer, and it is the verdict’s distinctive payoff: the tool is self-attenuating. The chokepoint works because the world’s finance runs through dollar-clearing — that centrality is the source of the power. But every use teaches targets and bystanders to build around it, which over time erodes the very centrality the weapon depends on. The breakdown of a shared, integrated financial order into rival blocs — the fragmentation logic — is treated in Ch 17 §17.5 (International Policy Coordination), where coordination’s failure is the mirror image of the integration that made the chokepoint possible. The reserve-status side of this erosion is owned by the dollar-reserve-status walkthrough; here the point is narrower — the weapon’s power supplies the seed of its own decline.
“Money is being weaponized, and the response will be to hold less of the weaponizer’s money. The cost of using the tool is paid in the tool’s own future power.”
— the self-attenuating-tool thesis, in the Pozsar tradition
“The weapon dulls itself”
The most-discussed cost of sanctions is the harm to the target. The least-discussed cost is the harm to the sanctioner’s own future leverage — and it may be the one that matters most.
The genuine open edge
“The failures are failures of over-ambition, not of the tool. Used for narrow, multilateral goals against integrated targets, financial sanctions work — Iran and the 2022 coalition prove it. The answer is to use the instrument better, not to abandon it.”
— the conditional-defender position
The conditional defender accepts everything Stage 3 established and draws a constructive conclusion: the record is an indictment of how sanctions are used, not of what they are. Aim them at narrow, well-defined goals; build and hold a real coalition; pick integrated targets; and accept that the win will be bounded and may be reversible. Do that, and the tool earns its place in the statecraft kit. The Iran deal and the coalition-binding of 2022 are the model; the maximum-pressure fantasies are the deviation.
“The tool over-promises, hurts civilians, entrenches the regimes it targets, and erodes the dollar centrality that gives it power. Weigh the diffuse long-run cost against the marginal short-run gain and the ledger looks worse than its advocates admit.”
— the skeptic-plus-blowback position
The skeptic-plus-blowback voice presses on the part the defender waves past: how do you weigh a concrete coercive gain you can see against a diffuse erosion you can’t yet measure? If over-use steadily builds the rails that would neutralize the weapon, then even “successful” narrow uses carry a hidden charge against future leverage — and the harm to civilians is paid in full regardless of whether the goal is met. This is the genuine open edge of the debate, and it is not resolved here, because it turns on a frame-level disagreement — what counts as “working,” and how to price a slow cost against a fast benefit — that the evidence underdetermines. Reporting it honestly as unresolved is the position; pretending the ledger nets out cleanly would be the evasion.
The verdict
Financial sanctions are a real but limited tool — effective for coercion-at-the-margin, signaling, and coalition-coordination; ineffective for regime change and broad capitulation. The answer is conditional, and the layers are worth naming separately. That sanctions rarely achieve ambitious goals is near-consensus across the statecraft literature — Hufbauer–Schott–Elliott, Pape, and Drezner all agree the ambitious-goal failure rate is high, and the public maximum-pressure framing sits well outside that professional consensus. That they do achieve narrow, marginal, signaling, and coalition goals is mainstream, but the magnitudes are genuinely disputed: how often, how much, and whether targeted financial sanctions beat the embargo-dominated base rate are all open.
Which lets us finally score the Russia-2022 freeze — and the honest scorecard is that it is genuinely ambiguous, reported here as a frame-level, definitional split rather than forced into a column. It worked at imposing real cost and binding a sprawling coalition into common action; it failed at the strategic capitulation it was implicitly sold as — the war did not end and the regime did not fall; and it accelerated de-dollarization-hedging, a cost the optimists never ledgered. Whether it “worked” depends on the goal you assign it, and the profession genuinely disagrees because the underlying question is partly “what was it for?” That is not a refusal to commit — the conditional structure is the commitment.
And the long-run cost is real and shared with the reserve-status debate: over-use erodes the dollar centrality the tool depends on. The channel is agreed; the speed is what reasonable analysts disagree about. The strategic implication follows directly — use the weapon sparingly, multilaterally, and for narrow goals. One line: a real but limited tool — effective for coercion-at-the-margin, signaling, and coalition-coordination; ineffective for regime-change; with rising long-run costs to dollar centrality. The reserve-status consequence of all this belongs to the dollar-reserve-status walkthrough; the trade-as-coercion lineage to the free-trade walkthrough.
Where this leaves us
- The freeze was ambiguous from day one. Immobilizing $300 billion of a great power’s reserves was unprecedented and it bit — and Russia adapted, the ruble recovered, the war went on. Both framings were true, which made “did it work?” the wrong first question.
- The chokepoint is real, and Iran is the win. The world’s payments run through a tollbooth the United States owns, secondary sanctions make the whole world enforce the ban, and freezing a target’s reserves disables its monetary defense. That leverage coerced Iran to the table in 2015 — a genuine success, narrow and reversible.
- The base rate is brutal, and targets adapt. Sanctions achieve stated goals at best about a third of the time, almost never the ambitious ones; Cuba, North Korea, Iran, and Venezuela all endured; the harm lands on civilians and can entrench the regime. “Rarely” is not “never,” but the record is sobering.
- The answer is conditional, and there is a hidden cost. Sanctions work for narrow, multilateral, marginal goals against integrated targets, and fail for maximalist goals against adaptable or autarkic ones — and every use erodes the dollar centrality that makes them possible.
So: can financial sanctions work? Yes — for the right job. They are a scalpel, effective at coercion-at-the-margin, at signaling resolve at a price, and at binding coalitions, against targets wired into the system the United States controls. They are not a hammer: against a determined, adaptable, or walled-off regime pursuing survival, the ambitious goal almost never arrives, and the population pays a bill the regime largely escapes. The Russia-2022 freeze is the whole argument in one event — the most ambitious financial-statecraft action ever taken, and a demonstration of the tool’s limits, scored honestly as ambiguous because it genuinely is.
And the deepest point is the one the headlines miss: the weapon dulls itself. Its power comes from the dollar’s centrality, and every use gives the world a reason to build around it. That is the real case for restraint — not that finance is a weak instrument, but that it is a finite one, and spending it on goals it cannot reach wastes both the campaign and the leverage. Use it sparingly, multilaterally, and for narrow goals — or watch it lose the power that made it worth using at all.