Finding a Substitute Is Not the Same as Switching to One
 |
Whenever one supplier, one route or one component becomes uncomfortably important, the same reassurance appears: this is temporary. Prices this high will bring alternatives. The dependency will resolve itself.
It sometimes does. It also frequently does not, and the reason has almost nothing to do with whether an alternative exists.
A Chokepoint Is a Price, Not a Fact
Nothing is truly irreplaceable. There is another supplier, another route, another material, another vendor. The question is never availability. It is what the alternative costs compared with continuing to pay the toll.
That framing is useful because it turns a fixed-sounding problem into a moving one. A dependency that looks permanent is a dependency where the detour is currently more expensive. Raise the toll far enough and the arithmetic flips.
Which is why holders of a chokepoint face a constraint that gets underappreciated: charging too much is the one thing that reliably ends the arrangement. The toll works only while paying it stays cheaper than leaving.
So far this is the standard story, and it is where most analysis stops. It is also where it goes wrong, because it treats the switch as a single decision about price.
A Detour Has to Clear Three Gates, Not One
In practice a substitute has to satisfy three separate conditions, and price is only the first.
It has to be cheaper than the toll. The obvious one, and the only one most commentary models.
It has to actually work. Equivalent specification, equivalent quality, equivalent reliability in the process it feeds. A cheaper input that fails one batch in twenty is not cheaper.
It has to still be there next year. A supplier who appeared during a shortage and disappears afterwards has not solved anything. Committing to them means rebuilding the relationship, the qualification process and the logistics a second time, at your own cost.
The third gate is the one that quietly kills most switches. Substitution is not a purchase. It is a qualification programme — testing, approvals, contracts, tooling, sometimes regulatory sign-off — with costs paid upfront against savings that only arrive if the new supplier survives.
And here is the trap. The moment the toll is highest is exactly the moment alternatives look most attractive and are least likely to be durable, because half of them exist only because the toll is high.
What Happens When the Emergency Ends
Follow the sequence and the failure mode becomes visible.
A disruption pushes the cost of the incumbent supply up sharply. New suppliers appear, because at that price marginal producers become viable. Buyers, urgently, place orders. Officials announce diversification. Everyone reports that the dependency is being reduced.
Then the disruption eases. The incumbent's price falls back. The marginal suppliers — who were only ever economic at the peak — become uneconomic again and exit. Buyers who were mid-qualification abandon the process, because finishing it now costs more than simply going back.
Six months later the concentration figure is where it started, or higher, and the episode is remembered as a scare rather than as a structural change. The timing is what does the damage: the decision to commit has to be made while the crisis price is still visible, and the payoff only lands long after it is gone — the same lag that governs how long a raw-material move takes to reach anything a household actually pays.
Nothing failed here. Every party behaved rationally at every step. The outcome was determined by the fact that the alternative's viability and the crisis had the same lifespan.
What This Looks Like From Korea
Korea ran this experiment in full view, and the results are unusually well documented — including the part that did not work.
In 2021 Korea imported roughly 97.6% of its industrial urea from a single country. Urea is not a glamorous input; it is the base for the diesel exhaust fluid that heavy vehicles legally require to run. When export restrictions appeared at the source, the effect was not financial. It was logistical — trucks that could not legally operate, on a supply measured in weeks.
The response was exactly the textbook one. Diversify. Sourcing was arranged from several other producing countries. Concentration fell: from roughly 71% in 2021 to about 66% in 2022.
Then it went back up. By the first half of 2023 the single-country share had risen to roughly 89% — higher than the trough and higher than the target. The temporary suppliers from Southeast Asia, which had briefly held single-digit shares, fell back to almost nothing.
The explanations given by people in the industry were not about politics. They were about specification and economics — the substitute product was assessed as lower quality, and once the emergency premium was gone, it could not compete on cost.
⚠️ That is the finding, and it is uncomfortable: a country identified the dependency, treated it as a national priority, publicly committed to fixing it, and the number went the wrong way. Not through negligence. Through the second and third gates.
It is worth sitting with, because the same episode gets retold as a success — the immediate shortage was resolved, so the story ends there. The structural position, which is the part that matters for the next time, ended up worse than before.
The Same Shape Outside Commodities
This is not a commodity phenomenon. It shows up anywhere switching has a fixed cost.
Software and cloud. Migrating a platform is a project with a budget and a completion risk, while staying is a line item. The vendor's pricing power is bounded by the migration cost, not by the market price of the service.
Proprietary consumables. Printers, cartridges, coffee systems and medical devices are frequently sold cheap because the recurring supply is captive. A third-party refill exists. Whether it voids the warranty is the real question.
Specialised components. Where a part is qualified into a design, replacing it means requalifying the design. The supplier's leverage comes from that process, not from the part — and the price actually paid sits under a contract, which is why a published benchmark and a company's real input cost are different objects.
Distribution. Any platform taking a percentage is exposed to the same limit — the take rate is safe until it exceeds what building a direct channel would cost, and that threshold moves as the alternative gets easier to build.
In each case the pattern holds: the incumbent's power is bounded by the total cost of leaving, and that total is dominated by the one-off costs, not by the recurring price. Which is why comparing the recurring prices — the thing everyone does — answers the wrong question. It is the same asymmetry that decides whether a purchase with a lower running cost actually gets made: a large single payment and a small repeating one are not interchangeable, however the arithmetic nets out.
What This Page Used to Say
This URL previously held a post arguing that an armed conflict was economically benefiting one country, and nearly all of it has been removed.
It named the states involved, characterised an administration's strategy and intentions, described one government's leverage over a shipping route as a "card" it was playing, and closed by suggesting that country would eventually be rendered irrelevant. It carried employment and trade figures from one particular month and a set of things to watch when markets opened on a specific Monday.
None of that belongs here. Which states benefit from a conflict, what any government intends by a policy, and how long a country retains strategic leverage are not questions this blog can answer responsibly — the sourcing does not fix it, and neither does hedging the language.
What survives is one observation the original made in passing, and it was the best sentence in the piece:
Leverage over a chokepoint is not infinite. It is a function of how badly the rest of the world needs a shortcut, and how long they're willing to pay for one.
The original then assumed the answer — that given enough time, the alternatives win. Korea's urea episode suggests that assumption deserves more scepticism than it usually gets.
Where This Stops Being Useful
Substitution does succeed. Plenty of dependencies have been broken permanently, usually where the alternative was genuinely equivalent and the switching cost was paid once by someone with a long horizon.
This does not predict which way any particular case goes. The three gates tell you what to examine. They do not tell you the outcome.
And it is not an argument against diversification. Paying more for a second source is often correct precisely because the value is insurance, not price. The error is counting a temporary second source as though the problem were solved — treating a figure as measured when it was constructed is the same mistake in a different setting.
Where You Are Standing
Ask what the switch costs, not what the alternative costs. For any dependency — a supplier, a platform, a service you pay for personally — the number that matters is the total cost of leaving, including the work. That figure, not the monthly price, is what the incumbent's pricing power actually rests on.
Check whether the alternative outlives the emergency. A second source that only exists at crisis prices is a hedge with an expiry date on it, which is the general problem with any price whose conditions are not printed alongside it.
And treat "diversification announced" as an input, not an outcome. The concentration figure a year later is the outcome. It is published, it is checkable, and it disagrees with the announcement often enough to be worth the look.
A Closing Note
The comfortable version of a dependency story ends with the alternative arriving. High prices summon supply, supply breaks the grip, the problem solves itself.
Sometimes it does. But an alternative existing, an alternative working, and an alternative still being there when the panic subsides are three different facts, and only the first one shows up in the coverage.
A shortage gets solved in weeks. A dependency gets solved in years, or not at all — and the difference is rarely visible while it is happening.
Data as of 2026-08-18. Sources: Korean trade and industry reporting on 2021 industrial urea import concentration and subsequent import-source diversification, including reported single-country shares for 2021, 2022 and the first half of 2023. This post describes supply-chain substitution mechanics and is not a view on any country, company or commodity. This post is observation, not investment advice.
Related Posts:
Supply Doesn't Vanish. It Becomes Unreachable.
Oil Doesn't Reach You All at Once
Every Price Has an Expiry Date. None of Them Print It.
Visuals on this post are AI-generated. The author works with AI as a research and drafting assistant; topics, judgments, and final edits are the author's own. This post is observation, not investment advice. See full Disclaimer for details.
