Sanctions and economic statecraft

Name the channel a sanction runs through, and the workaround that erodes it.

A sanction is a price rather than a wall. It does not stop a trade from happening. It makes the trade more expensive, and everything that follows is about who ends up paying the difference and how long they keep paying it.

The mechanism

Cutting a target off from its usual buyers forces it to sell through somebody else at a discount. That discount is the bite, and in the first year it is usually large.

A discount is also a profit opportunity. Somebody who is willing to handle the trade can buy cheap and sell at the world price, so intermediaries appear, capacity is built, and the discount narrows. Nothing about the sanction changed. The market routed around it.

Bite in year one

30%

Bite twelve years on

Halved by year

Target loses per unit the coalition gives up

2.1

A fungible good finds another buyer. The bite halves by year 3 and settles at 7%, and 65% of what the target lost went to whoever handled the rerouting rather than disappearing.

Share of buyers enforcing60%Makes the bite bigger. Watch how little it does to the shape.How replaceable the good isoil, sayThis is the one that decides whether the bite lasts.

Five ways to cut somebody off, and how each gets around

Two sliders, and they do different jobs. One makes the bite bigger. The other decides whether it lasts. Finding out which is which is the useful part of this lesson.

Worked example

Breadth buys depth, specificity buys durability

I expected a broad coalition to hold up better than a narrow one. Building the model showed that it does not, and the reason is worth following.

Widening the coalition makes the bite bigger. Thirty percent of buyers enforcing gives a 15% discount. Everybody enforcing gives 50%.
And it barely changes how fast the bite erodes. Both halve by year three. The erosion lands within four percentage points across the whole range.
Because a bigger discount is a bigger margin. A harsher sanction is a more profitable one to work around, so it pulls in exactly as much capacity as the wider coalition was suppressing. The two effects cancel.
Only replaceability changes the shape. Take the good from fungible to single-source and the bite is still near its original size after twelve years, because there is nowhere to route it.

There is an uncomfortable symmetry underneath. The goods where a sanction holds are the goods the target cannot easily sell elsewhere, which are usually the goods the coalition cannot easily buy elsewhere either. The sanctions that work are the expensive ones.

The empirical literature on this is unusually contested. Hufbauer, Schott, and Elliott's Economic Sanctions Reconsideredputs the success rate around a third, and Robert Pape's rereading of the same cases puts it near five percent, largely by disagreeing about what counts as success. The model on this page is stylized and unit tested; it is not an empirical claim about any particular sanctions regime.

ConventionSecondary sanctionsare the response to the leakage. Rather than only punishing the target, they punish anybody who trades with the target, which is an attempt to widen the enforcing coalition without persuading anybody to join it. They work because access to the dollar payment system is worth more to most banks than the business they would lose, which makes them a use of the reserve currency's position rather than of diplomacy.

Five channels, five different clocks

"Sanctions" covers instruments that behave completely differently, and naming which one is in play tells you most of what to expect. The buttons under the chart set out each one. In short:

Trade bans erode in months to a couple of years, faster the more fungible the good.
Cutting off the payment system takes years to work around and builds permanent alternative plumbing on the way, which is a cost to the sanctioner that outlasts the episode.
Individual designations erode immediately unless enforcement follows the beneficial owner rather than the name.
Export controls on specific technology are the slowest, because the thing being denied is not fungible. This is where sanctions actually bite.
Freezing central bank reserves never erodes for the money already frozen, and the deterrent erodes the moment it is used.

What breaks

Quoting the first-year effect. It is the largest number and the least informative one. What matters is the level three years on, and the difference between those two figures is the entire subject.

Assuming the loss is destroyed. Most of what the target gives up is captured by whoever handles the rerouting. That creates a constituency with a direct financial interest in the sanction staying in place, which is not the constituency anybody intended to create.

Confusing pressure with change. Sanctions reliably impose costs and unreliably change behavior, and the gap between those is where the arguing happens. A regime that treats the sanction as an attack often becomes more consolidated rather than less, and the burden lands on people with no ability to alter the policy being punished.

What to do with this

When a sanctions package is announced, ask three things. Which channel is it using? How replaceable is the thing being denied? And who outside the coalition has both the appetite and the capacity to intermediate?

Those three answers give you the shape of the curve before anybody reports a number. And when the first-year figures arrive, treat them as the start of a series rather than a result, because the series is the finding and the first point is a headline.

Test yourself

01A sanctioned country's exports fall 30% in the first year and are back near normal volumes by year three, at a lower price. Did the sanctions work?

It depends entirely on what they were for. If the goal was to stop the exports, no: the volume came back. If the goal was to reduce the revenue, partly: the discount survives even when the volume does, because rerouting is never free.

The question worth asking is where the difference went. Most of what the target gave up was not destroyed. It was captured by whoever was willing to handle the trade, which means the sanction created an industry with a direct financial interest in the sanction continuing.

02Why do export controls on advanced chipmaking equipment hold up better than an oil embargo?

Because oil is oil. A barrel refused by one buyer is identical to a barrel any other buyer wants, so the workaround is a shipping route and a broker, and both exist within a couple of years.

Advanced lithography has essentially one manufacturer. There is no third country to route through because no third country makes the thing. The workaround is not a trade route, it is building a domestic industry, and that takes a decade or more if it works at all.

03Freezing a central bank's reserves is the most severe financial measure available. What does it cost the country doing the freezing?

The reserves themselves are gone for the target and that part does not erode. What erodes instantly is the deterrent, because every other country in the world watches it happen and starts asking where its own reserves are held. The measure works perfectly once and teaches everybody to reduce their exposure to the currency it was denominated in, which is a cost paid by the issuer over decades.

A tutor that knows this lesson. It asks before it explains, and it will not tell you what to do with your own money.

Educational material, not investment or policy advice. Figures are cited where they come from a filing or a statistical series, and labelled as illustrative where they do not.