This piece builds on two arguments we have made in previous pieces: first, that Ethiopian Airlines illustrates how bureaucracy can work when insulated from planner logic—a mindset that prioritizes rigid blueprints, input targets, and centralized control over real-world problem solving; and second, that most development failures stem not from technical gaps, but from the persistence of such planning mindsets in dynamic environments. By contrast, searchers adapt through iteration, feedback, and user responsiveness. What, then, do we make of the exceptions—those public institutions in Ethiopia that, while not matching the scale or success of EAL, nonetheless deliver results in complex policy domains?

They are few, but not accidental. From Information Network Security Agency (INSA)’s agile cybersecurity operations to Ethio Telecom’s rapid digital expansion through Telebirr, these entities have managed to produce real outcomes in a system otherwise marred by rigidity, opacity, and institutional inertia. Their success does not disprove the planner vs. searcher distinction. On the contrary, it sharpens it. These institutions have survived not because the system works—but because, in critical ways, they have escaped it.

What follows is an anatomy of that escape: a set of vignettes illustrating how certain public entities managed to carve space for adaptation, responsiveness, and performance. Then we turn to the shared conditions that made this possible—not to romanticize these outliers, but to understand what their stories reveal about the broader reform challenge.

Exceptions That Prove the Rule: Adaptive Public Institutions in Ethiopia

Despite the dominance of planner logic across Ethiopia’s public sector, a small number of public institutions have managed to deliver greater than expected results. Their success is not because they rewrote the rules of the system but because, in critical moments, they found ways to sidestep them. These institutions do not represent systemic transformation—they are exceptions that prove the rule. But in doing so, they help clarify what the rule actually is, and under what conditions it can be breached. By examining how these outliers escaped the gravitational pull of rigid plans and meaningless metrics, we can begin to sketch what adaptive governance might look like.

Vignette 1: A Mobile Money Platform That Listened

The most illustrative example is Ethio Telecom’s mobile money platform, Telebirr. Launched in 2021 to expand digital financial inclusion, it reached over 40 million users within two years—a rare feat given Ethiopia’s limited banking infrastructure and uneven digital literacy. Unlike typical e-government platforms that focus on devices or app installations, Telebirr tracked real usage: transactions, failures, and user complaints. When rural users struggled with the interface, Amharic-language voice commands were added within weeks—an unusually fast turnaround for a public platform.

Telebirr’s success stemmed from two key design features that let it bypass traditional bureaucracy. First, it had high-level political backing. Tied to the Prime Minister, it gained fast-tracking and waivers usually reserved for prestige projects—giving it rare agility among public Information and Communication Technology (ICT) rollouts. Second, it operated autonomously. Rather than fitting into Ethio Telecom’s civil service structure, it ran from a separate site under private-sector HR rules. Engineers were merit-based hires, well-compensated, and empowered to iterate using real-time data. For a brief window, the institution was built to search, not just to comply.

But the story also shows the fragility of exception-based governance. After a leadership reshuffle in 2023, transaction failures rose and update cycles slowed. Without structural safeguards—rules that preserve feedback loops and insulate technical teams from political volatility—adaptation collapses. Telebirr’s success remains real but precarious, dependent on political will and informal arrangements rather than lasting institutional design.

Vignette 2: A Dam Built on Destiny

The Grand Ethiopian Renaissance Dam (GERD) is a different kind of exception—defined by scale, persistence, and elite backing. GERD is Africa’s largest hydropower project, which was framed not as an infrastructure project but as national destiny. This framing allowed it to sidestep many of the institutional constraints that typically derail large-scale investments. Environmental reviews were fast-tracked under national security exemptions. Budget reallocations prioritized GERD even during fiscal crises. When foreign financing faltered, a domestic bond campaign filled the gap—not because of financial innovation, but because political symbolism made compliance near-mandatory.

GERD avoided planner logic in part because it wasn’t managed by a conventional bureaucracy. Its oversight body reported directly to the Prime Minister’s Office and was staffed by engineers with decades of dam construction experience. This tight organizational focus—combined with a clear output metric (megawatts delivered)—meant that success could be judged in real terms, not through process indicators. When unexpected geological conditions required design modifications, decisions were made quickly, without months of inter-ministerial coordination.

Yet GERD’s success came at a cost. Procurement irregularities and cost overruns escaped scrutiny, transparency was sacrificed for momentum, and diplomatic tensions grew due to poor policy coordination. Domestically, the singular focus on the dam diverted resources from rural electrification and grid reliability—critical needs that remain underfunded. In bypassing bureaucracy, GERD delivered—but also revealed what gets left behind.

GERD proves that Ethiopia’s state can deliver world-class infrastructure under the right conditions. But it also shows how narrow those conditions are: political centralization, elite cohesion, and symbolic stakes that override normal bureaucratic incentives. In that sense, GERD is not a model but an outlier—powerful, instructive, and ultimately fragile.

Vignette 3: Digitizing Procurement to Disrupt Patronage

If GERD succeeded through political will, the Electronic Government Procurement (eGP) platform achieved impact through something far rarer in Ethiopia’s public sector: insulation. Originally piloted with donor support, eGP was designed to digitize and centralize public tenders—a notoriously opaque domain long dominated by personal networks and discretionary decisions. By 2023, it had processed thousands of tenders across health, construction, and agriculture, reducing procurement time from 120 days to 45 and lowering corruption indicators by nearly a third, according to World Bank reviews.

eGP’s escape from planner logic came not from technology, but from institutional design. It operated under an independent directorate with its own hiring protocols, outside the Ministry of Finance’s civil service rules. Officers were trained in specialized digital workflows, while oversight bodies used real-time dashboards. Crucially, benchmarks tied to donor disbursements and third-party audits enforced accountability—leaving little room to hide behind process indicators like “manuals distributed” or “trainings held.”

The platform’s success also stemmed from its tightly bounded mandate. Unlike broader reform efforts that try to transform entire ministries, eGP did one thing: digitize tenders. It did not try to reform the procurement law, overhaul budgeting systems, or retrain the entire civil service. That narrow focus made its gains real, even if modest.

Yet even eGP is not immune to the gravity of the system around it. Small and medium enterprises still struggle to access the platform due to onerous compliance demands. Attempts to scale it to subnational governments have faced resistance from officials who benefit from discretionary procurement. And without a constituency to defend it, eGP risks becoming a technical shell—digitally efficient, but easily sidelined by political actors with other priorities.

In this sense, eGP shows what is possible when planner logic is neutralized: when success is judged by transactions completed, not committees formed; and when donor leverage is used not to impose blueprints, but to create feedback loops. Its future, however, remains uncertain.

Vignette 4: Surveillance Agency, Adaptive by Design

Few Ethiopian institutions are as opaque—or as surprisingly adaptive—as INSA. Formed in 2006 with military ties, it has faced criticism for secrecy and domestic surveillance. Yet it has also developed unusual technical strength, especially in defending critical infrastructure. In 2023, it claimed to block 82% of foreign cyber intrusions targeting government and banking systems.

INSA’s effectiveness lies in its structural divergence from mainstream bureaucracies. Unlike ministries that must route procurement through Ministry of Finance (MoF) or follow standardized hiring protocols, INSA operates with operational autonomy. It recruits engineers, cryptographers, and analysts through its own channels—often from military academies or technical universities—paying above civil service grades and offering specialized training. Its AI-driven border surveillance programs, used to interdict trafficking and unauthorized crossings, evolved not through five-year plans but through iterative field testing with field agents and software teams working in tandem.

Like other high-performing entities, INSA enjoys political priority. Designated a “sovereignty institution,” it receives top budget status even during contractions. Cabinet-level interest shields it from the delays that cripple ministries. During cyberattacks, INSA doesn’t wait for ministerial clearance—it acts.

INSA’s autonomy comes at a price. Lacking full parliamentary oversight and transparent budgets, it escapes scrutiny. Civil society has raised concerns about digital rights violations under state-of-emergency powers. Its performance rests on technocratic authority—but without democratic checks.

What INSA reveals is that institutional adaptation doesn’t always emerge from reform. Sometimes, it is born out of security logics—narrow mandates, elite protection, and insulation from democratic scrutiny. In such cases, planner logic is disrupted not by participatory search, but by technocratic autonomy under political patronage. The result is performance—but at a price.

Vignette 5: Fixing Freight, Not Just Building Roads

The Integrated Freight Transport Management System (IFTMS) is a quietly effective logistics initiative. While industrial strategies often fixate on infrastructure—rails, trucks, ports—IFTMS targets coordination: digitally managing freight along the Djibouti–Addis corridor. Launched under the World Bank–funded TRANSIP project, it cut port clearance times from fourteen to three days and is estimated to save $180 million annually by reducing demurrage, spoilage, and idle time.

IFTMS escaped planner logic in three ways. First, it had a narrow, technical mandate—focused purely on freight tracking. It didn’t manage roads or regulate licenses, which let engineers optimize without bureaucratic drift. Second, its design embedded feedback loops: truck movements, delays, and bottlenecks were continuously analyzed and used to improve routing. When data showed surges in Kombolcha, the system adjusted within weeks—something few agencies could manage. Third, its hybrid model tied World Bank financing to specific outcomes: reduced clearance times, interoperability, and anti-corruption. Procurement exemptions enabled rapid import of needed equipment like radar scanners and GPS modules.

Still, IFTMS is not immune to risk. Its scope remains urban-focused, with no mandate over rural hubs, and its systems don’t yet integrate with border customs in Moyale or Metema.. More critically, its success hasn’t shifted broader infrastructure culture—railways and industrial parks are still managed through rigid, construction-driven blueprints.

IFTMS shows that planner logic can be interrupted not by institutional reform alone, but by narrow mandates, embedded data systems, and strategic donor alignment. It demonstrates what’s possible when Ethiopia’s logistics challenges are approached as problems to be solved—not symbols to be built.

The Anatomy of Escape

What distinguishes Ethiopia’s most adaptive public institutions is not charisma or luck—it is structural deviation from the prevailing logic that governs much of the public sector. While each vignette charts a unique trajectory for a particular institution, the anatomy of escape reveals a shared set of enabling conditions.

Most decisive is political prioritization. When projects are deemed too important to fail—whether because they symbolize national strength, bolster elite legitimacy, or address urgent state capacities—they are insulated from the bureaucratic drag that immobilizes less visible programs. GERD, INSA, and Telebirr did not succeed because Ethiopia’s bureaucracy was reformed, but because, in these instances, the bureaucracy was selectively bypassed.

Operational autonomy follows. These institutions were structurally uncoupled from civil service routines. Fast-tracked hiring, flexible procurement, and delegated decisions let teams iterate in real time rather than wait for inter-ministerial clearance. Without such autonomy, feedback loops are cut off before they can trigger adaptation.

Third is a narrow mandate. The institutions that broke through did not try to solve everything. IFTMS focused solely on freight movement. eGP digitized procurement. GERD built a dam. With a singular focus, it became possible to define success in terms of outcomes rather than processes, to attribute accountability clearly, and to learn without drowning in complexity.

Hybrid finance—when structured with conditions—also played a catalytic role. Donor funding did not merely fill resource gaps. In the case of eGP and IFTMS, it imposed performance-linked discipline that domestic systems alone could not enforce. That discipline was most effective when aligned with local priorities; when misaligned, it produced confusion or token compliance.

Finally, embedded feedback mechanisms helped sustain adaptation. Institutions that measured what mattered—actual transactions, delivery times, system failures—could correct course. But feedback only mattered when coupled with the authority to respond. In organizations that lacked political cover or autonomy, data often accumulated without consequence. These five elements—prioritization, autonomy, focus, financing, and feedback—do not guarantee success. But their absence almost always guarantees failure. What makes these institutions notable is not that they rewrote the rules of government, but that the rules bent around them. Until Ethiopia’s public sector normalizes these deviations—until adaptation is systemically rewarded rather than selectively tolerated—such cases will remain exceptions, not evidence of change.

These enabling conditions do not operate in isolation. Their power lies in how they interact—how one creates the space for the next.

Most adaptive institutions begin as niches: narrowly scoped projects with clear functional goals and minimal exposure to the full weight of bureaucratic machinery. This boundedness makes it possible to establish operational autonomy, since the risk of disruption to other systems is low. Autonomy, in turn, requires and reinforces political shielding—the formal or informal protection that allows a unit to bend rules without triggering institutional antibodies.

With autonomy and shielding in place, feedback becomes actionable. Iteration is no longer blocked by rigid protocols or veto players. Teams can adjust based on real-world data, not theoretical targets. And when that feedback produces results—faster procurement, higher adoption, visible savings—it earns political credit, which can then be reinvested into scale.

Without this sequence, the parts break down: feedback without shielding provokes backlash, autonomy without focus leads to drift, shielding without results breeds impunity. Sequence isn’t optional—it’s the system.

Conclusion: The Limits of Escape

The anatomy of escape explains how Ethiopia’s exceptional institutions temporarily sidestepped planner logic. Yet their fragility exposes a harder truth: these conditions remain accidental, not systemic. Where they are absent, as in Ethiopia’s clean cooking crisis, planner logic persists with lethal consequences.

Consider the dissonance: Telebirr’s transaction-driven iteration cut user adoption barriers, while clean cooking interventions trap themselves in measuring inputs—stoves distributed rather than smoke eliminated. GERD’s political shielding fast-tracked geological adaptations, yet clean cooking engineers wait for approvals to modify stove designs to suit local diets. IFTMS’s real-time freight data flows into weekly system optimizations, but clean cooking programs ignore field reports of 83% stove abandonment in Oromia. The difference isn’t technical capacity; it’s institutional architecture.

Clean cooking thus becomes the critical test case for the following piece. Its failures crystallize the stakes: more than 60,000 annual deaths from household air pollution persist precisely because the enabling conditions observed in Telebirr, GERD, or INSA—elite prioritization, operational autonomy, feedback discipline—are wholly absent here. This raises the pivotal question our next piece must confront: Can the anatomy of escape be deliberately engineered for domains starved of political oxygen? Or is adaptive capacity forever hostage to contingency?

We will examine clean cooking not as a problem to be planned, but as a probe into system change. Can a niche pilot with shielded autonomy and user-driven feedback crack this graveyard of intentions? Or does replicating escape require dismantling the planner logic itself? The viability of reform—not just for stoves, but for Ethiopian governance—hangs on the answer.

Tsegaye Nega is Professor Emeritus at Carleton College (USA) and Founder & CEO of Anega Energies Manufacturing, an Ethiopian clean cookstove enterprise.

Contributed by Tsegaye Nega