ethiopia

The Queue

One of the most important markets in Ethiopia is not a market at all.

It is a queue.

If you wanted dollars to import machinery, medicine, spare parts, or inputs, you often did not mainly need price. You needed patience, paperwork, and the right relationships. Time became a currency of its own.

That’s an ineffective way to manage an economy.

When foreign exchange is scarce and the state goes ahead and sets the price below reality, demand will not disappear. It spills somewhere else. Into waiting lists. Into discretion. Into favors. Into the parallel market.

and the system starts looking like this:

flowchart TD
A["Importer needs dollars"] --> B["Official bank request"]
B --> C["Long queue"]
C --> D["Delay, uncertainty, discretion"]
D --> E["Production slows or stops"]
D --> F["Parallel market premium rises"]
F --> G["Higher prices and distortion"]

This is not solely just a finance problem. It is a progress problem.

A factory cannot plan when it has no idea if it will get inputs in two weeks or six months. A pharmacy cannot serve patients with with a queue-based currency regime. A serious exporter will not invest wholeheartedly if the whole trade systems looks and feels like a waiting room.

In his work titled “Industrial Policy for the Twenty-First Century” Dani Rodrik lists some arguments. One of his big arguments is that governments should identify the real bottlenecks holding productive firms back and remove them. You cannot ask firms to become globally competitive while forcing them to beg for foreign exchange through an obscure administrative funnel.

The archaic system did not allocate dollars to the most productive uses. It more often than not allocated them to the most persistent, the most connected, or the least time-sensitive.

A queue is a hidden tax on speed.

And speed matters, it matters more than people think. Fast firms learn faster. They adjust faster. They compound faster. Slow systems do the opposite. They pressure businesses to spend energy on waiting, hedging, and surviving instead of expanding.

Which is why Ethiopia’s July 29, 2024 foreign exchange reform mattered. The National Bank of Ethiopia (NBE) moved toward a market based-exchange regime. Since then, the IMF and World Bank have both reported a stronger export performance, better reserve accumulation, and narrower distortions.

This is the right direction.

The solution is simple to state, even if it is politically hard: do not rebuild the queue.

Keep foreign exchange allocation as rule-based and price-based as possible. Let banks compete. Let prices carry information. Publish spreads. Publish volumes. Reduce the room for invisible discretion.

In other words:

flowchart TD
A["Importer needs dollars"] --> B["Competitive bank market"]
B --> C["Transparent price and spread"]
C --> D["Faster allocation"]
D --> E["More predictable imports"]
E --> F["Better planning, output, and trade"]

This is not costless. A more market-based system can mean a weaker birr, higher import prices, and real pain in the short run. That pain is not imaginary.

And by no means is this not costless. A more market-based system can mean a weaker birr, higher import prices and real pain in the short run. And that pain is neither imaginary nor hypothetical.

But the second-order effects of the old system were worse. Queues reward lobbying. They punish newcomers. They protect low-productivity incumbents. They make corruption rational. They slow every part of the economy that depends on imported inputs, which is to say, much of the economy.

A country doesn’t industrialize by making time cheaper for bureaucrats and costlier for producers.

It industrializes by doing the reverse.

If Ethiopia is to progress faster, one practical rule should be this: whenever scarcity appears, resist the temptation to turn it into a queue.

What looks like order in a queue is often simply a shortage disguised as such.


written for the Overdue Progress essay competition

Sources


Previous I created a podcast site, so when I feel like yapping Next What if LLMs thought in Lisp