On paper, Ethiopia is having an extraordinary run. The International Monetary Fund’s April 2026 Regional Economic Outlook put Ethiopia’s projected 2026 real GDP growth at 9.2 percent, the highest of any large economy in Sub-Saharan Africa, more than double the regional average of 4.3 percent, and ahead of every other economy the Fund tracks on the continent except tiny, war-recovering South Sudan. Three years after a civil war that displaced millions and after a sovereign default that took nearly six years to unwind, that is a striking reversal.

But sit with two other numbers. Ethiopia’s headline inflation rate accelerated to 15.3 percent in July 2026, its fourth consecutive monthly increase and the highest reading since January 2025. And on the World Bank’s own reckoning, the share of Ethiopians living below the $3-a-day poverty line is estimated to have risen to roughly 43 percent in 2025, up from 33 percent in 2016.

Put differently: an economy that the IMF says is expanding faster than almost anywhere else in the world is simultaneously seeing inflation reaccelerate and poverty rise. That combination is the real subject of this article. It is not a story about collapse; the evidence does not support that word and using it would be an exaggeration. It is a story about whether headline growth is a reliable signal of Ethiopia becoming a stronger, more resilient economy, or whether it is concealing fragility that a growth number alone cannot show.

What the growth number actually says — and doesn’t say

The 9.2 percent figure comes from the IMF’s April 2026 Regional Economic Outlook for Sub-Saharan Africa, and it has since been repeated widely in financial and general media. It’s worth noting immediately that forecasts for Ethiopia are not fully consistent across sources: the IMF’s own country page and Article IV-linked material, alongside the UNDP’s Ethiopia economic profile, cite a somewhat lower, GDP growth figure closer to 7.1–7.2 percent for 2026, and consensus trackers such as Focus Economics have cited a range of roughly 7.1–7.3 percent. Ethiopia’s own Ministry of Finance has at various points pointed to an even more optimistic target near 8.9 percent. These are not necessarily contradictions, different vintages of IMF forecasts, different program documents, and government projections routinely diverge by a percentage point or more, but the spread itself (7.1 to 9.2 percent) is a reminder that even the “headline” growth figure is really a band of estimates, not a single hard fact, and it should be read that way rather than quoted as settled truth.

There is a further, more structural reason for caution, and it is one that independent economists, not just critics, raise: Ethiopia’s national accounts are still built on a 2015/16 base year, a decade old, which the Ministry of Planning and Development has itself acknowledged is overdue for updating. Ethiopia’s authorities have announced plans to rebase GDP to a 2024/25 base year by around September 2026, partly under a “statistical sovereignty” initiative. Rebasing exercises (Nigeria’s 2014 and 2025 rebasing are a well-known regional precedent) frequently produce large, mechanical revisions to GDP levels and can also alter measured growth rates, simply by better capturing the size of services, the informal economy, and sectors that have grown since the base year was set. The World Bank has separately noted, in a country document, that “limitations are particularly noticeable in the estimation of expenditure components” in Ethiopia’s GDP data. None of this means the 9-percent-plus figure is fabricated. But it does mean the number carries wider uncertainty bands than a comparable figure from a country with modern, high-frequency national accounts, a point a well-informed economist reviewing Ethiopia’s data would insist be made explicit rather than glossed over.

What is more solid is the composition of growth. The IMF and independent trackers agree that growth in 2025–26 has been broad-based rather than reliant on a single commodity: mining (notably gold), construction, manufacturing and agriculture have all contributed, which several analysts argue makes the expansion structurally more durable than in economies leaning on one export. That is a genuine point in Ethiopia’s favour and distinguishes its story from, say, a pure gold- or oil-price windfall.

The other half of the picture: prices are accelerating again

Ethiopia’s disinflation story through late 2025 was real and worth taking seriously. Annual CPI inflation, which had been above 30 percent in 2022–2023, fell to single digits by December 2025 (9.7 percent), a genuine policy achievement that Prime Minister Abiy Ahmed’s government has credited to subsidy targeting, income adjustments and supply-chain reforms, alongside the National Bank of Ethiopia’s tight 15 percent policy rate.

That disinflation has now reversed. According to Ethiopia’s Statistical Service, annual inflation rose for four straight months through mid-2026, 13.4 percent in May, 13.9 percent in June, and 15.3 percent in July, the highest since January 2025. Non-food inflation jumped particularly sharply, to 14.8 percent in July from 12.2 percent in June, while food inflation reached 15.7 percent, driven by outsized increases in sugar (nearly 40 percent year-on-year), coffee and other staples. Ethiopia recorded the largest inflation increase of any of the ten major African economies tracked in the first half of 2026.

The proximate trigger, according to the IMF and multiple financial outlets, is external: the war in the Middle East disrupted fuel and fertiliser shipments and pushed global energy prices higher, and Addis Ababa responded to the resulting supply shock by raising domestic petrol prices by roughly 35 percent over a matter of weeks in spring 2026, a decision that fed directly into transport fares and food prices. The IMF’s July 2026 review explicitly frames the war as “a significant external shock” that disrupted trade and caused “temporary fuel shortages and sharp increases in the price of imported fuel and fertilizer,” while judging that the impact on output and inflation had, at that point, been “modest.”

It is worth flagging a genuine and unresolved dispute here rather than picking a side. The government’s official inflation figures, the ones cited above, put 2025/26 inflation in a 9–15 percent range depending on the month. A March 2026 commentary circulated by independent Ethiopian economists and monetary analysts argued that Ethiopia’s “underlying” inflation rate may be closer to 30 percent, roughly three times the official estimate, a claim tied to broader scepticism about the reliability of official statistics generally. This is an assertion from independent commentators, not a verified alternative index, and no comparable independent CPI series (of the kind a group like Johns Hopkins-Cato’s “Troubled Currencies Project” or similar produces for countries such as Venezuela or Zimbabwe) was found to corroborate the 30 percent figure specifically for Ethiopia in 2026. Readers should treat it as a contested claim rather than an established fact, but its existence, from analysts who are not simply reflexively hostile to the government, is itself informative about how much confidence outside observers place in the official CPI.

The currency: liberalized, but still not settled

Ethiopia’s July 2024 exchange-rate reform, a roughly 30 percent devaluation and a shift toward a market-determined birr, undertaken as an IMF program condition — was the single largest structural change underpinning the current reform narrative. Two years on, the currency is still adjusting, and not smoothly.

As of late August 2026, the official/commercial-bank rate for the birr was trading in the neighbourhood of 155–163 to the US dollar, while the parallel (informal) market rate was running well above that, reported at roughly 174–180 birr per dollar in various accounts through August, implying a persistent gap of somewhere between 10 and 20 percent depending on the source and week. The IMF’s own July 2026 review put the average spread at around 11 percent as of May, an improvement from wider gaps earlier, attributing the narrowing partly to a more active central-bank FX auction system. Independent market reporting (Bloomberg, cited via regional outlets) suggested the gap was wider again by mid-August, with some banks bidding as low as 163 birr per dollar at a 12 August auction against a parallel rate near 180 — and Citigroup’s Africa economist was quoted forecasting the birr could weaken to 185–195 per dollar by year-end, though he expected the authorities to try to prevent it crossing 200.

The persistence of a double-digit parallel-market premium two years after a “market-based” reform is itself a data point worth taking seriously: it suggests dollar demand still substantially outstrips the supply the formal banking system can offer, a classic symptom of an economy where foreign-exchange scarcity, rather than having been resolved, has merely been repriced and partially formalized. The National Bank of Ethiopia has reportedly absorbed foreign-exchange losses estimated at about $2.6 billion since the reform, according to audited financial statements reported by Addis Standard and The EastAfrican, a cost of defending the currency that is separate from, and additional to, ordinary reserve depletion.

On reserves, the picture is one of gradual, IMF-monitored improvement from a very low base rather than either crisis or comfort. The IMF’s fifth ECF review (completed July 2026) projected gross international reserves reaching about $5.9 billion by the end of Ethiopia’s 2025/26 fiscal year, providing import cover of a little over two months, up from roughly 0.7 months two years earlier, but still thin by conventional benchmarks, which generally regard three months of import cover as a minimum comfort threshold. The Fund projects that cover only reaches roughly 3.5–3.8 months by 2030/31, several years away. In other words: real progress, from a genuinely dangerous starting point, but not yet a robust buffer.

Debt: a restructuring finally nearing completion, not a debt crisis resolved

This is the area where there has been the most unambiguous, verifiable progress in 2026 — and it deserves to be reported as such.

Ethiopia defaulted on its sole, $1 billion Eurobond in December 2023 after missing a coupon payment, having already requested treatment under the G20’s Common Framework in January 2021. What followed was one of the framework’s slowest and most contested cases: an agreement in principle with the Official Creditor Committee (OCC, co-chaired by France and China) was reached in March 2025 and formalized in a Memorandum of Understanding that July; a preliminary bondholder deal announced in January 2026 was then rejected by the OCC for failing the framework’s “Comparability of Treatment” test; renegotiation followed, with talks collapsing again in May before a new agreement in principle was reached with bondholders on 29 June 2026. On 21 August 2026, the OCC’s co-chairs formally confirmed that this deal was consistent with Comparability of Treatment, clearing the way for implementation. Under the terms reported by Reuters and regional outlets, Ethiopia will exchange the defaulted bond for a new three-year, $880 million instrument carrying a 6.15 percent coupon, maturing July 2029, alongside payment in full of roughly $99.4 million in missed coupons and a modest consent fee — reporting describes this as roughly a 12–15 percent face-value haircut for bondholders, alongside a warrant tied to Ethiopia’s future market access. Ethiopia’s government has set an October 2026 target to finalize the remaining commercial-creditor agreements, which make up roughly a tenth of total external debt.

This is genuinely significant: Ethiopia is the last country still working through the Common Framework since its 2021 launch, and its case has been widely described (including by Foreign Policy) as a test of whether the mechanism can function at all. Completion would restore a path back to international capital markets that has been closed for years.

It would be a mistake, however, to read debt-restructuring completion as debt-sustainability achieved. The IMF’s own Debt Sustainability Analysis, published alongside the fourth ECF review in early 2026, shows the present-value of public debt-to-GDP ratio rising to about 44 percent in 2026/27 under a combined contingent-liabilities stress scenario, remaining above safer thresholds until roughly 2029/30. External public and publicly guaranteed debt jumped from 15.7 percent of GDP to 31.7 percent of GDP in a single year (2023/24 to 2024/25) — a jump the IMF attributes mainly to currency depreciation following the FX reform rather than new borrowing, which is a useful and important distinction, but one that also illustrates how mechanically sensitive Ethiopia’s debt ratios are to further birr weakness. Separately, independent commentary (an economic-analysis piece published on the pan-African outlet Africa Is a Country in July 2026) put debt service at “roughly 24 percent of the national budget” — a claim from an opinion/analysis source rather than an official IMF or Ministry of Finance figure, and one this article could not independently verify against a primary document, but plausible enough, given the scale of restructuring underway, to be worth flagging as a serious claimed fiscal constraint rather than dismissing it.

The IMF’s own framing, in its July 2026 press release, is measured rather than triumphalist: it describes “strong macroeconomic performance to date” that “has created resilience,” while explicitly warning that “debt vulnerabilities remained significant” and that the Middle East war represents “a substantial external shock.” That is closer to the honest state of play than either “debt crisis resolved” or “debt crisis unresolved” would be on its own.

Growth, poverty, and the transmission problem

This is arguably the sharpest tension in the whole picture, and it is documented by the World Bank rather than by critics with an axe to grind. The World Bank’s poverty-and-equity assessment for Ethiopia projected the poverty rate (measured at the $3-a-day, 2021 PPP line) rising to 43 percent in 2025, up from 33 percent in 2016 and 39 percent in 2021 — a reversal of roughly two decades of poverty reduction. The Bank attributes this to a combination of factors: the COVID-19 pandemic, the Tigray conflict, drought, a slowdown in growth in the early 2020s, and high inflation, which it notes hit urban households (who buy most of their food) harder than rural households (many of whom are net food sellers, though largely disconnected from off-farm income opportunities). The Bank itself projects poverty may only begin to decline gradually from 2026, and only “assuming peace and stability return” — a conditional forecast, not a promise.

Separately, Ethiopian development economists quoted in domestic reporting (The Reporter, via its magazine coverage) have pointed to structural features compounding the growth-poverty gap: what one economist termed “premature deindustrialization,” continuing insecurity in parts of the country limiting the free movement of goods and labour, and climate shocks. Youth unemployment is estimated near 27 percent in early 2026 according to Trading Economics data cited in regional coverage, with commentators describing a pattern of “jobless growth” in which capital-intensive sectors such as construction and mining expand output without proportionately absorbing the roughly two million young Ethiopians entering the labour force each year.

This is the empirical core of the “growth versus resilience” question this article set out to investigate, and it is where the evidence most clearly supports genuine concern rather than either alarmism or complacency. GDP growth measures the expansion of aggregate output; it does not, by construction, measure how that output is distributed, whether real wages are keeping pace with inflation, or whether job creation matches the size of the labour force. Ethiopia’s own numbers, a rising poverty rate concurrent with strong reported GDP growth, are consistent with a “transmission problem”: growth concentrated in capital-intensive, often geographically or sectorally narrow activity (large infrastructure projects, mining, gold, an urban construction boom) that has not yet broadly lifted household incomes, particularly for the roughly three-quarters of Ethiopians who live in rural areas and, per the World Bank’s 2021 human-capital data cited in its own report, face very low rates of primary-education completion and persistently high rates of child stunting.

It’s worth being precise about what this does not establish. It does not establish that the GDP growth figures are fictitious, nor that Ethiopia is undergoing “economic collapse”, inflation, while elevated, is roughly half its 2022 peak; reserves and exports are improving from a genuinely dangerous 2023 starting point; the debt default is close to being resolved through an internationally recognized process; and independent institutions (the IMF, World Bank, AfDB) continue to describe Ethiopia’s trajectory as one of gradual, hard-won stabilization rather than deterioration. What the evidence does establish is a real and currently unresolved gap between top-line output growth and household-level living standards — precisely the distinction the World Bank itself draws when it writes that “growth alone cannot end poverty unless it is broad-based and stable.”

So: growing, or getting stronger?

Weighing all of this, a fair summary looks something like this:

What the evidence supports.

Ethiopia’s economy is genuinely expanding, probably somewhere in a 7–9 percent range depending on which vintage of estimate is used, and that growth is reasonably diversified across mining, manufacturing, construction and agriculture rather than dependent on one commodity. A years-long, complicated sovereign-debt restructuring, the hardest test case yet of the G20 Common Framework, appears close to completion on genuinely improved terms relative to Ethiopia’s 2023 default. Foreign reserves, exports and government revenue have all improved from very low, dangerous 2023 levels, according to the IMF’s own repeated program reviews. Disinflation from over 30 percent in 2022 to single digits by December 2025 was real, even if partial and now reversing.

What the evidence also supports, and what growth alone does not capture.

Inflation has reaccelerated sharply through mid-2026, driven substantially by an external shock (the Middle East war) but amplified by domestic fuel-price pass-through. The birr, two years after being “liberalized,” still trades with a persistent double-digit gap against its informal-market rate, evidence that foreign-exchange scarcity has not been fully resolved. Foreign reserves, while improving, remain thin by international benchmarks. Public debt ratios remain vulnerable to further currency depreciation and contingent liabilities, even as the restructuring nears completion. And most importantly, by the World Bank’s own reckoning, poverty has risen, not fallen, over the past decade, even as GDP has grown — a pattern that should trouble anyone treating the growth rate as a proxy for how ordinary Ethiopians are actually faring.

Growth and resilience are not the same thing, and neither is stable. An economy can grow while its currency stays fragile, its inflation stays volatile, and its poorest households fall further behind — and that, on the balance of evidence assembled from the IMF, World Bank, G20 creditor process documents, and independent Ethiopian and international analysts, is closer to where Ethiopia stands in August 2026 than either a triumphant “Africa’s fastest-growing economy” narrative or a “collapse” narrative would suggest. The more useful question, going forward, is not whether the growth number for 2026 lands nearer 7 percent or 9 percent. It’s whether the reforms now largely complete on paper — currency liberalization, debt restructuring, fiscal consolidation — translate, in the next two or three years, into a currency that trades near one rate rather than two, reserves that comfortably clear import-cover benchmarks, and a poverty rate that finally turns back downward. Ethiopia’s own data suggests that transition has not yet happened. Whether it does is the real test of whether this is a growth story or a resilience story.

Why Ethiopian microfinance must build digital readiness before making digital commitments

What concerns me is not that Ethiopian microfinance institutions are moving too slowly into digital finance. My concern is that some of us may be trying to become digital before we are digitally ready.

I began the African Microfinance Weekly Media Digest nearly six months ago for a simple reason. I was already following developments in microfinance, financial inclusion and technology for my own learning, and I felt those developments were worth sharing with others in the profession. As the weeks passed, however, the reading began to reveal something more than news. One concern kept returning. Technology is moving quickly, while many of the institutions expected to make decisions about it still lack the knowledge, systems and governance needed to lead those decisions.

I am not an information technology expert, and I do not write as one. I write as a business development practitioner who has spent the past five years working within the microfinance sector, with opportunities to engage closely with institutional leaders, regulators, technology providers and development partners. I have also sat inside discussions where major technology decisions were being considered and watched what happens when technical knowledge, management confidence and governance capacity are badly unbalanced. I have seen how easily a product can be discussed before the problem is clearly defined, how a provider can control the conversation because it controls the language, and how easily an institution can become too dependent on one technical specialist when others either lack the knowledge or are not given enough time, space and authority to question or independently verify the advice being given.

These experiences have left me with a difficult question. Are Ethiopian MFIs building digital capability, or are we acquiring the appearance of being digital?

The appearance of transformation

Digital transformation is increasingly confused with acquiring visible signs of technology. A mobile wallet, an agent banking channel, a mobile application or a fintech partnership can easily be presented as evidence that an institution is becoming modern. These are visible developments. They are easy to mention in a board report, a social media post or a speech, and they can create a strong sense that the institution is moving forward.

But visibility is not transformation. A new channel is not a digital strategy, a signed agreement is not an institutional capability, and a software licence is not a transformed business.

Real digital readiness is more basic than launching a visible customer facing product. An MFI may have a core banking system and still conduct much of its daily work manually. Approvals may depend on paper, staff records may sit outside integrated systems, routine workflows may remain unautomated, and employees may rely on outdated computers that slow even ordinary tasks. In that environment, adding a wallet, an agent banking channel or a mobile application does not transform the institution. It simply places a digital layer over operations that are still largely manual.

I have observed discussions in which a wallet, agent banking, a mobile application and other digital channels were already being treated as priorities before the institution had clearly defined the problems they were meant to solve. Different products with different purposes, operational requirements and regulatory implications could be discussed almost as one digital package. Who needed each service, what problem it would solve, whether the institution needed to build something new at all, and whether it was ready to manage it had not been properly tested. Yet attention had already moved to providers, approvals and launch dates.

That is not digital transformation. It is digital ambition without institutional preparation.

The first act of digital leadership should therefore be simple. The institution should begin by deciding what it actually wants to improve. It may be faster loan collection, easier deposits and withdrawals, lower cost outreach, better information for credit decisions or less cash in the hands of field officers. Different problems may require different solutions. Starting with the technology instead of the problem risks spending scarce resources on something the institution never truly needed.

When the seller also defines the problem

The second concern is the knowledge imbalance between many MFIs and the companies approaching them.

A fintech or technology vendor normally arrives prepared. It understands its product, the architecture, the commercial model and the language of the negotiation. It has a presentation, a demonstration, financial projections and answers to expected questions. It may have negotiated similar arrangements before. The MFI, on the other hand, may be entering this kind of discussion for the first time. Its board and management may not have enough independent knowledge to test what they are hearing.

This creates an unhealthy situation. The company selling the solution can end up defining the problem, designing the solution, estimating the opportunity, valuing what each party contributes and explaining why the agreement is attractive. The MFI appears to be making a decision, but most of the decision has already been framed by the other side.

The issue is not that technology companies are automatically dishonest. Many fintechs have valuable knowledge and can help MFIs serve clients better. But every provider also has its own commercial interest and will naturally present its solution in the strongest possible light. The difficulty begins when the MFI does not have enough independent knowledge to test that case properly. In that situation, the provider’s understanding of the problem, the solution and even the value of the deal can quietly become the basis on which the MFI makes its decision.

Consider a profit sharing proposal. A fifty fifty split may look balanced, and a sixty forty split may appear even more attractive to the MFI. But neither percentage tells the institution whether the deal is fair. The real question is how that share was arrived at. The MFI may contribute the lending capital, cost of funds, customer base, field presence, collections capacity, regulatory responsibility and years of trust in the market. The fintech may contribute the platform, product design, technical operation and other specialised capabilities. Each side may also carry different costs and risks. Until those contributions, costs and risks are properly understood and valued, a percentage is only a number.

The institution also needs to understand which costs are deducted before profit is calculated, and who controls pricing, credit assessment, loan approval, collections and customer communication. It must also understand how financial, operational, regulatory, customer and reputational risks are divided between the parties. An attractive commercial term should never substitute for understanding the full allocation of risk and responsibility.

If these questions are not answered in language that the board and management can understand without the vendor in the room, the MFI is not ready to sign.

When one person becomes the strategy

The stories I have followed through the Media Digest show how quickly digital finance is expanding around African microfinance. New platforms, partnerships and technology driven models are becoming difficult for MFIs to ignore. But the pressure to move quickly creates another risk inside the institution. When a specialist is brought in to lead a major digital change, and management and the board are not equipped to properly question the direction being proposed, that person can gradually become more powerful than the governance system around the change, even when the specialist is capable and acting in good faith.

Hiring an experienced IT leader may be necessary, especially where specialised skills are scarce. But hiring one strong person is not the same as building a strong IT function. Neither is changing the organisational structure. If the wider team remains poorly equipped, insufficiently trained and unable to question or carry major work independently, knowledge simply becomes concentrated at the top. The institution has not solved its capability gap. It has concentrated it.

That dependence becomes more serious once major decisions and investments begin to accumulate. Systems may be changed, contracts signed, data migrated and new services placed on the roadmap. The same person may then become the main source of explanation for what has been done, what remains unfinished and what should happen next. Losing that person can begin to look like a risk to the entire programme. Management may then become reluctant to challenge decisions it cannot independently verify because the cost of losing the specialist appears greater than the cost of continuing the dependency. At that point, this is no longer an ordinary staffing issue. It is a governance risk.

The problem becomes worse when legitimate questions are treated as resistance to change. Someone who asks whether a product is really needed, whether the institution is ready, whether the cost is justified or whether another approach should be considered can easily be seen as protecting old ways of working or standing in the way of progress. That is a dangerous confusion. Questioning how transformation is being done is not the same as opposing transformation. A change programme becomes difficult to govern when challenging the approach is treated as rejecting the goal itself.

The answer is not weaker IT leadership. Ethiopian MFIs need stronger technology professionals, stronger IT teams and much greater investment in specialised skills. But knowledge must be shared, staff must be developed, major decisions must be documented, and management must have access to more than one informed view. Business, operations, finance, risk, compliance and internal audit must also have enough understanding and authority to challenge decisions from their own areas of responsibility. Strong technical leadership should make the institution more capable, not more dependent on the person leading it.

The test is simple. If an MFI cannot confidently explain, question or continue its digital direction without one person in the room, it has not yet built digital capacity. It has built digital dependence.

A shared system is not a sign of weakness

The question of core banking deserves particular care. Some MFIs may believe that leaving a shared system and acquiring a system directly is proof of greater independence or digital maturity. This is not necessarily true.

A shared core banking arrangement can be a sensible choice for smaller and medium sized MFIs when the service is reliable, secure, responsive and well governed. It can spread costs across institutions, provide access to specialised technical capacity, support common standards and strengthen bargaining power with technology providers. But shared systems can also create problems when responsibilities are unclear, service quality is weak or participating institutions have little influence. The arrangement should therefore be judged by how well it works and how well it fits the institution, not by whether it appears more or less independent.

Shared core banking models should not be dismissed as temporary arrangements for institutions that lack ambition or digital maturity. They can be deliberately designed to automate core processes, respond to different institutional needs and allow future integration with fintechs and other digital service providers. When properly structured, they can also bring together planning, system selection, testing, configuration, migration and technical support that individual MFIs may struggle to manage alone. The value of the model should therefore be judged by its capability, governance and service quality, not by the fact that the system is shared.

Whether a shared arrangement is actually serving each institution well is a separate question and should be judged honestly. Poor service, weak support or limited flexibility are valid reasons for an MFI to consider leaving the shared arrangement. But acquiring the same or a similar system independently does not automatically create greater digital maturity. It may simply transfer more cost, implementation work, security exposure and technical responsibility to one institution. If that institution is not ready to manage those responsibilities, greater ownership of the system can actually create greater dependence.

Ownership of a copy of software is not ownership of digital capability. Before leaving a shared arrangement, an MFI should be able to show what will genuinely improve for the institution and its clients. It should understand the full cost, the demands of migration and integration, the capacity required to manage the system, and how operations will continue if implementation fails or is delayed. Independence is valuable only when the institution has the capacity to carry the responsibilities that come with it.

What the MFI may slowly give away

An MFI brings assets to a digital partnership that do not always appear clearly in a financial model. It brings its licence, capital, customers, community trust, local knowledge, field presence, collection experience and access to a market that may have taken decades to build. A fintech brings different strengths, including technology, speed, specialised knowledge, product capability and a better digital customer experience. A good partnership can combine these strengths. But before discussing how the benefits will be shared, the MFI must first understand exactly what kind of partnership it is entering.

Not every fintech partnership is the same. In one arrangement, the MFI may simply purchase technology while retaining control of the product and customer relationship. In another, both parties may jointly distribute and manage a service. In another, the fintech may control the customer interface and much of the product experience while the MFI provides the regulated lending capacity behind it. These arrangements create very different positions for the MFI. The danger begins when an institution enters one model while believing it has entered another.

The balance can also shift gradually after the partnership begins. If the fintech controls the interface, the data, the customer journey, the credit model and communication with the client, the MFI may slowly lose the ability to understand and manage its own business. It may continue to provide capital and carry regulatory responsibility while another company learns from its customers, strengthens the relationship and builds the more valuable part of the business. The MFI can eventually become little more than the regulated balance sheet behind somebody else’s financial service.

This is why questions about the customer relationship, data access, product decisions, pricing, credit assessment, complaints, regulatory accountability, knowledge transfer and exit rights are not legal details to be settled after the commercial agreement. They are the business model itself. The MFI needs to know what it will continue to control, what it is willing to share and what capability will remain inside the institution if the partnership ends.

Collaboration is necessary. Most MFIs cannot and should not build every technology themselves. But a partnership should leave the institution better able to understand its customers, make decisions, manage risk and serve its mission. It should not make the MFI more dependent on the partner with every year that passes.

Regulation does not allow us to outsource responsibility

The most useful Ethiopian reference for this discussion is the National Bank of Ethiopia Directive MFI/33/2022 on information technology management for microfinance institutions. The directive does not tell an MFI to begin with a wallet or an application. It requires the institution to define the role of technology within its business strategy and to develop an IT strategy that supports that direction. It also places clear responsibilities on the board, senior management, the IT function, risk management and internal audit.

The requirements go much deeper than having a system in place. The directive calls for proper planning, adequate financial and human resources, project and vendor management, automation of core business processes, reliable management information, technology risk assessment, disaster recovery, training and IT audit. It also requires the board and senior management to review technology matters regularly. Taken together, these are not simply compliance requirements. They describe what institutional readiness should look like.

The wider national direction reinforces the same message. The draft National Digital Payments Strategy for 2026 to 2030 places human capacity, trust, resilience, supervision, consumer protection and shared infrastructure alongside innovation. It does not treat digital progress as a race to acquire more technology. It treats it as a system in which infrastructure, people, governance and responsible use must develop together.

The same principle applies when an MFI works with an outside technology provider. International guidance is clear that using a third party does not remove the institution’s responsibility. The MFI still has to understand the risk, assess the provider, define responsibilities clearly, monitor performance, prepare for failure and retain a practical way to exit. It must also keep enough knowledge inside the institution to manage the relationship itself. A provider can supply technology and expertise. It cannot replace the judgement that properly belongs to the institution.

What real digital readiness requires

The more I have followed digital developments through the Media Digest, the more one thing has become clear to me. An MFI is not digitally ready because it can buy a system, sign with a fintech or obtain approval for a new channel. It is digitally ready when it understands the problem before choosing the technology, understands the technology before committing to it, and understands the consequences before putting its clients and institution behind it. Being able to launch something is not the same as being ready for it.

Readiness begins inside the institution. If approvals still move on paper, records are difficult to retrieve, basic processes remain manual, reports cannot always be trusted and staff struggle with inadequate tools, a new digital product does not make those weaknesses disappear. It carries them into the new system. Technology can make a strong process faster, but it can also make a weak process fail faster and at a larger scale. Digital transformation cannot begin at the customer interface while the institution behind that interface remains largely unchanged.

It also begins with knowing what you want. Before asking which wallet, platform, application or provider to choose, the institution should be able to explain the problem in plain language. Whose problem are we solving? What is difficult for the client today? What will become easier tomorrow? Why does this particular technology solve it better than the alternatives? If those questions are difficult to answer without the vendor in the room, the institution is already moving too fast.

Real readiness is also visible in how confidently an institution can question a proposal. Management should understand enough to challenge the economics. The board should understand enough to challenge the direction. The IT team should be strong enough that knowledge does not sit with one person. Business, operations, finance, risk and audit should understand what the change means for their own responsibilities. An MFI does not need to know everything a technology company knows. But it must know enough to decide for itself. Otherwise, the institution may own the licence and provide the capital while somebody else effectively owns the thinking.

Readiness also means having the confidence to say no, or not yet. Digital transformation should not become a race to collect channels and products. An MFI can start with the need that matters most, pilot one solution, let staff and clients live with it, learn where it works and where it strains the institution, then improve and expand. The institution should feel in control of one step before rushing into the next. That is not moving slowly. It is building the capacity to move further without losing control.

For me, this is the real test of digital readiness. Can the MFI explain what it wants, choose deliberately, challenge what it is being told, support what it builds, learn from what happens and change direction when necessary? Can it do all of this without becoming dependent on one vendor, one system or one individual? If the answer is no, the institution may be moving toward digital finance, but it is not yet ready to lead its own digital transformation.

The greater danger is not being late

Ethiopian MFIs do need to change. Clients expect easier access, faster service and greater control over their money. Digital finance can help us reach further, operate better and serve people in ways the traditional model could not. Standing still is not an option.

But neither is moving simply because everyone else is moving. The greater danger may not be arriving late. It may be committing to systems, products and partnerships that the institution does not yet understand well enough to govern. An MFI is not weak because it needs a fintech partner, uses shared infrastructure or takes time to prepare. It becomes weak when it can no longer explain its own choices, challenge what it is being told or operate without those it has become dependent on.

Perhaps every major digital decision should face one simple test. If the system or partnership ended three years from now, would the MFI be left with stronger people, better knowledge, better data and greater ability to serve its clients? Or would it discover that much of the capability it thought it was building never truly became its own?

For me, this is what digital maturity ultimately means. Not having the newest technology, but becoming an institution that knows what it wants, understands what it is committing to and remains capable of making its own decisions. The MFI should understand the deal at least as well as the vendor. Until it does, it should not sign.