Ethiopia has dismantled the last of its emergency controls and handed the economy to a market-based framework held together by interlocking parts. Every piece now depends on every other piece and the National Bank of Ethiopia must keep all of them balanced at once, in high wind.
In August 1974, Philippe Petit stepped onto a wire strung between the Twin Towers with nothing beneath him but 400 metres of air. What made the walk possible was not courage alone. It was the rigging, a lattice of cavalletti guy-lines anchoring the wire against sway, each one tensioned against the others. Cut any single line and the wire does not merely wobble. It whips.
Ethiopia’s monetary policy has just stepped onto its own wire. With the seventh Monetary Policy Committee meeting, the National Bank of Ethiopia removed the credit growth ceiling that had disciplined bank lending for nearly three years and raised the National Bank Rate from 15 to 16 percent, the first hike since the interest-rate framework was born in July 2024. The message was deliberate: administrative controls are finished; the price of money now does the work.
It is the most consequential monetary decision Ethiopia has made this decade, and it is also the most fragile moment in the reform program so far. Because the system the NBE now operates is not a single instrument. It is a machine of interdependent parts, the policy rate, the interbank market, the Treasury bill market, the foreign exchange market, fiscal discipline, external buffers, and an IMF program that finances and certifies all of it.
The machine works beautifully when every part holds. But it has a property the old command system never had: failure propagates. A shock that enters through one channel travels through all of them.
This is the deep architecture of Ethiopia’s new quantitative economy and the map of how it could come apart.
The Machine: What Replaced the Cap
Under the credit ceiling, monetary control was blunt but simple. The NBE told banks how fast their loan books could grow, 14 percent at the cap’s introduction in August 2023, loosened to 18 percent in December 2024 and 24 percent in September 2025 and lending obeyed by decree. Inflation, which had been running near 30 percent, was starved of credit fuel directly.
The new framework replaces that single lever with a transmission chain. The NBE sets the policy rate at 16 percent. That rate is supposed to anchor the interbank market, where banks now lend to each other, a market that did not exist before October 2024 and has since cleared a cumulative Birr 3 trillion. The interbank rate, in turn, is supposed to shape deposit and lending rates across the banking system, which shape credit demand, which shapes aggregate demand, which shapes inflation.
Alongside it sits the Treasury bill market, where 91-day yields have fallen to 11 percent on excess demand of roughly Birr 667.8 billion, the government’s financing lifeline now that it no longer borrows from the central bank. And beneath everything sits the liberalized foreign exchange market, freshly loosened again with the FX commission cut from 2.5 to 1.5 percent and the exporter surrender requirement reduced from 50 to 30 percent.
Each of these is an achievement. None of them stands alone.
Transmission only works if the interbank and T-bill markets are deep and liquid and they are only deep because fiscal policy is disciplined enough to fund itself through them rather than through the printing press. The exchange rate only holds because reserves have been rebuilt and exports are growing. Inflation expectations only anchor because the policy rate stays credibly above inflation. And the entire structure is financed, monitored, and certified by the IMF’s $3.4 billion Extended Credit Facility, whose seal of approval is what keeps official creditors, Eurobond holders, and development partners at the table.
That is the rigging. Now consider the wind.
The Shock Problem: Why This Machine Is Exposed
The timing of the cap’s removal is not comfortable. It comes precisely as inflation has broken its two-year downtrend. After touching 9.7 percent in December 2025, single digits for the first time in nearly a decade, headline inflation climbed to 11.7 percent in April and 13.4 percent in May, with food inflation at 15 percent. The proximate cause is external: the US–Iran conflict and the wider Middle East war disrupted trade routes and spiked the price of imported fuel and fertilizer, feeding directly into transport and production costs in an economy that imports both.
This is exactly the kind of shock the new framework handles worst. A credit cap does not care where inflation comes from; it simply throttles lending. An interest rate framework, by contrast, must make a judgment call: is this inflation demand-driven, in which case rates should bite, or supply-driven, in which case over-tightening strangles growth for nothing?
The NBE has judged the current episode to be mostly imported and it is probably right, but supply shocks have a habit of becoming demand problems when they de-anchor expectations. In a country where the memory of 30-percent inflation is only three years old, expectations are not deeply anchored. They are on parole.
The internal shock channel is just as live. The IMF’s own projections show credit to the private sector and state enterprises expanding by as much as 55.9 percent in FY2025/26, a violent reversal from the 9.7 percent contraction a year earlier, and a number the Fund itself flagged as requiring “close monitoring.” Ethiopia’s banking system is sitting on enormous liquidity: T-bill auctions oversubscribed by hundreds of billions of birr, an interbank rate that drifted down from 18 percent in March to 14.6 percent in May.
Banks that were legally forbidden from lending freely for three years now face no ceiling at all, at the exact moment the fiscal calendar approaches an election cycle with its usual appetite for spending, and insecurity in several regions continues to disrupt agricultural trade flows. If pent-up lending erupts into an economy already absorbing a fuel shock, the NBE will be fighting inflation on two fronts with a transmission mechanism that is barely eighteen months old.
And then there is the structural fragility no policy rate can fix quickly: reserves of about 2.1 months of import cover. That is a dramatic recovery from the 0.7 months of two years ago, but still below the three-month threshold considered the minimum safe buffer for an import-dependent economy.
Public debt jumped from 35.5 to 50.5 percent of GDP in a single year, largely the cost of recapitalizing the Commercial Bank of Ethiopia and the projected decline back toward 45 percent depends entirely on debt restructuring negotiations that are advanced but not concluded. The wire, in other words, is anchored to guy-lines that are themselves still being tensioned.
The IMF: External Anchor, and Single Point of Failure
It is impossible to understand how this system stays upright without understanding the role of the Fund. On July 1, the IMF Executive Board completed the fifth review of Ethiopia’s ECF arrangement, releasing about $464 million and unusually, rephasing roughly $200 million of future program money forward specifically to cushion the Middle East war shock. Total disbursements now stand near $2.65 billion of the $3.4 billion envelope.
The money matters, but the money is the smaller part. The program’s real function is certification. Every quantitative performance criterion Ethiopia meets, the zero limit on FX intervention outside auctions, the ban on central bank financing of the deficit, the reserve accumulation targets, is a public, verifiable commitment that substitutes for the credibility a two-year-old monetary framework has not yet earned on its own.
The Official Creditor Committee memorandum, the agreement-in-principle with Eurobond holders, the flow of development partner financing: all of it is conditioned, formally or informally, on the program staying on track. The IMF’s financing assurances remain explicitly provisional until the debt restructuring is finalized.
This is the “one hand lets go” problem in its purest form. Consider the chain. Suppose fiscal discipline slips, election spending, a security emergency, a subsidy reversal under social pressure. The deficit widens beyond what the T-bill market will absorb at tolerable yields. The temptation returns to lean on the central bank, or on the state-owned banking system, for financing.
The moment that happens, three things break at once: the monetary anchor (base money growth, already at 43 percent, re-accelerates), the program (a breached performance criterion stalls the next review and the next disbursement), and the debt deal (creditors who signed on the assumption of an IMF-certified adjustment path reassess).
A stalled program means a financing gap; a financing gap means pressure on reserves; pressure on 2.1 months of reserves means pressure on the exchange rate; a sliding birr means imported inflation on top of imported inflation; and the parallel-market premium, the ghost Ethiopia spent two years exorcising, walks back through the door. Each failure finances the next. The machine that transmits policy in good times transmits collapse in bad ones.
That is not a prediction. It is a description of the system’s topology. Quantitative, market-based economies are efficient precisely because everything is connected to everything and fragile for exactly the same reason.
What the NBE Must Actually Do
The good news is that the NBE’s recent decisions suggest it understands the topology. The playbook for the next eighteen months writes itself from the structure of the risks.
First, defend the real rate at all costs. At 16 percent against 13.4 percent inflation, the policy rate is positive in real terms by roughly 2.6 percentage points, thin, but positive. That margin is the single number on which the framework’s credibility rests. If May’s inflation print becomes a trend and the real rate goes negative, the NBE must hike again, quickly and without apology, even into a supply shock. A young framework cannot afford to be tested and found bluffing. The IMF has already signaled the same: a tight stance remains appropriate to anchor expectations, and further increases should be on the table if disinflation stalls.
Second, use the targeted tools early, not late. The NBE’s new safeguard, the power to impose additional reserve requirements on individual banks whose loan-to-deposit ratios run hot, is well designed precisely because it is surgical. But macroprudential tools only work preemptively. If credit growth genuinely tracks toward the 55.9 percent projection, the first targeted reserve requirement should land while the expansion is months old, not after a year of deteriorating credit quality. The thematic review of bank lending practices the NBE has undertaken should feed directly into this trigger.
Third, keep building the plumbing. The transmission mechanism is only as strong as the markets that carry it. That means pushing secondary trading of government securities, operationalizing the repo framework under the Ethiopian Master Repurchase Agreement, easing the structural concentration of liquidity in the Commercial Bank of Ethiopia, and continuing to publish auction results and rate decisions with boring, metronomic transparency. Every basis point of the interbank curve that starts responding predictably to the policy rate is a guy-line added to the wire.
Fourth, hoard foreign exchange like the shock is not over, because it may not be. The surrender-requirement cut to 30 percent trades short-term reserve accumulation for long-term export competitiveness. That is a defensible trade only if the NBE continues to manage the foreign exchange market effectively, uses auctions to provide liquidity when needed, and takes advantage of the current gold and remittance windfall to strengthen reserves while it lasts. The distance between 2.1 and 3.0 months of import cover is the distance between a shock that is absorbed and a shock that becomes a crisis.
Fifth, protect the fiscal-monetary boundary as if the entire reform depends on it, because it does. The government’s record here is genuinely strong: no central bank borrowing since the reforms began, a deficit of 0.9 percent of GDP, Birr 206.5 billion raised through the T-bill market. That discipline, more than any interest rate decision, is why inflation fell. The NBE’s job is partly institutional: to make the cost of breaching that boundary politically visible before anyone is tempted, especially as the electoral calendar heats up.
The Wire Holds Only While Every Line Holds
Ethiopia has done something few frontier economies attempt and fewer complete: it has replaced a command monetary system with a market one in the middle of a debt restructuring, a currency float, and a global commodity shock, and so far it has kept its balance. Growth of 9.2 percent, a current account deficit compressed from $6.2 billion to $1.8 billion, reserves tripled, an inflation rate that, even after the recent rebound, is less than half its 2023 peak. The rigging was built well.
But the defining feature of the new system is that it has no single margin of safety. Its resilience is the product of every anchor holding simultaneously: real rates positive, fiscal borrowing market-based, reserves rising, the IMF program on track, the debt deal closing, and banks lending freely but not recklessly. The credit cap was a crude instrument, but it was a net. The net is gone now.
Petit crossed the wire eight times that morning, and the walk is remembered as artistry. The part nobody photographs is the rigging check, the obsessive, unglamorous verification of every line before each step. That is the NBE’s job now: not the drama of the walk, but the discipline of the anchors. Ethiopia’s monetary policy has never been more modern. It has also never been more connected, and the price of connection is that the central bank can no longer afford to let go with either hand.
Sources: National Bank of Ethiopia MPC (7th meeting), IMF fifth ECF review (July 2026), Ethiopian Statistical Service.
