Keep Colonial State intact: do Donor Work through Companies

Set up in 2014 as a Section 42 not-for-profit, it was built around the UK Foreign, Commonwealth and Development Office’s £187 million Enterprise and Asset Growth Program. By March 2021, according to FCDO’s own mid-term assessment, £111 million — PKR 23.42 billion at the program’s conversion — had already been deployed through it. Its audited accounts show it took Rs5.6 billion in fresh grants in 2022, none at all in 2023, and did not shrink. It kept running on accumulated funds, investment income, recyclable capital and program assets. A body that receives no new money in a year and keeps its footprint intact is no longer a program. It is an endowment with a board.

And the recent money is no longer buying microloans. In December 2025 the Gates Foundation committed $13.7 million, over four years, to strengthen the governance of Raast — Pakistan’s sovereign instant-payment rail — and to wire it into cross-border payments across six Gulf states. In November 2025 it funded Karandaaz to help the State Bank of Pakistan make “data-driven decisions” about its agent networks. Further 2026 commitments extend toward digital-government systems. Together these recent Gates commitments run to roughly $18 million. Karandaaz already moves Benazir Income Support Program payments to the poorest households through Raast. A foreign foundation is now financing the governance of the national payment infrastructure, the central bank’s network strategy, and the plumbing of digital government — through a private vehicle no citizen elected and no legislature audits.

But the clearest illustration of what this entity has become is the tax authority. In 2024, on the Prime Minister’s instructions, the Federal Board of Revenue signed a memorandum of understanding with Karandaaz to modernise Pakistan’s tax system. Using Gates Foundation money, Karandaaz then hired the global consultancy McKinsey & Company to build the FBR’s digital and IT transformation strategy. Karandaaz ran the procurement: its own digital-services director briefed the government steering committee on the tender, and the Gates Foundation’s country lead sat at the table as a “key stakeholder” in the affairs of the country’s revenue authority.

Stop and look at what that sentence describes. A company whose mandate is financial inclusion, and whose expertise in tax administration is open to question, was placed inside the FBR to redesign how the state collects its revenue — and to procure a foreign consultancy to do it, with a foreign foundation’s grant, with that foundation seated as a stakeholder in the process. Senators noticed. On the floor of the house, one asked who had authorised hiring McKinsey while bypassing the Public Procurement Regulatory Authority’s rules, and how a foundation that came to Pakistan for micro-financing had become a stakeholder in tax matters. Others worried, reasonably, about a foreign firm’s access to the most sensitive data the state holds.

The defenders will say Karandaaz simply moved faster than a government tender ever could. That is precisely the problem, and it is worth naming plainly. Routing the contract through a Section 42 company is what makes the procurement law not apply. PPRA binds government agencies; it does not bind a private company. So the state did not merely hire a helper — it used a private vehicle to place a sovereign function outside its own rules. The conduit is not incidental to the story. The conduit is the mechanism.

And here the affair meets its perfect description. In The Big Con, Mariana Mazzucato and Rosie Collington anatomise how governments become dependent on consultancies such as McKinsey, which operate as advisors, legitimators and outsourcers while cultivating the illusion that they are objective sources of expertise and capacity. Their central claim is a warning we should have heeded: the more governments outsource, the less they know how to do. The state stops learning by doing, because someone else is always doing the doing. The FBR case is that thesis in miniature — with an added twist Mazzucato does not quite anticipate. In her account, a government hires the consultant directly. In ours, the revenue authority does not hire McKinsey at all. A donor-funded company hires McKinsey, on a foundation’s grant, with the foundation in the room. It is outsourcing squared: the state outsources to a vehicle built to be an outsourcer, which then outsources to a consultancy. Each layer puts more distance between the sovereign function of taxation and any citizen who might hold it to account. A tax authority that collects trillions of rupees a year could not, apparently, find the money or the legal room to modernise itself without a foreign foundation’s grant and a private company’s procurement. That is not a capacity gap. It is a sovereignty question.

None of this is a Pakistani eccentricity. It has a name in the scholarship. Daniela Gabor and Sally Brooks call it the fintech–philanthropy–development complex: development organised through networks of state agencies, donors, philanthropic capital and consultancies, with “financial inclusion” and “digital transformation” as the unifying frame (Gabor & Brooks, 2017). Karandaaz is a textbook node in exactly that network. Recognising the type matters, because it tells us the questions to ask are not about intentions. They are about power, ownership and accountability.

The first issue is accountability — or its absence. In her study of the Gates Foundation, Linsey McGoey makes the point plainly: large foundations are unelected, opaque, and answerable to no electorate for their failures, yet they increasingly shape policy in the countries they fund (McGoey, 2015). Now apply that lens at home. Karandaaz sits off-budget. It is outside the Auditor General’s remit and outside the Public Accounts Committee’s scrutiny. Its board is self-selecting. Its accountability runs upward, to FCDO and Gates, not downward, to the citizens whose payment rail and tax system it now helps run. When Parliament tried to ask basic questions about the FBR contract, it discovered how little purchase it had. That difficulty is not a glitch. It is the design.

The second issue is sovereignty. Tax collection and the national payment rail are sovereign functions, in the same category as the currency. Who governs them is a question of constitutional weight, not procurement. Deborah Bräutigam and Stephen Knack showed two decades ago that aid can quietly corrode governance precisely by loosening the state’s dependence on its own taxpayers — the fiscal link that forces governments to answer to citizens (Bräutigam & Knack, 2004). The World Bank admitted the mechanism in its own Assessing Aid review: donors have repeatedly done “end runs” around local institutions because bypassing them is the easiest route to a successful project (World Bank, 1998). The FBR episode is that end run made literal — a foreign foundation and a foreign consultancy placed inside the sovereign revenue function, with the citizen as a spectator to the operation of his own state.

The third issue is the one I want to dwell on, because it is the most revealing: autonomy. Karandaaz enjoys a degree of operational independence that almost no attached department in Pakistan can claim. An attached department must live inside the government’s budget cycle, its procurement rules, its audit objections, its establishment controls and pay scales, its transfer-posting politics, its ministerial direction and its parliamentary scrutiny. Karandaaz, by contrast, operates as a Section 42 company with donor funds, its own board, its own hiring and salary structures, its own procurement practices, its own investment vehicles, and its own program priorities negotiated directly with donors and partner agencies. This gives it speed and flexibility. It also gives it an autonomy that the state’s own implementation arms are simply not allowed to possess.

The irony is total, and it should be the headline of this whole debate. Donors claim to build state capacity. Yet they give their preferred delivery vehicle exactly the freedoms — to hire, to pay, to procure, to invest, to experiment — that no Pakistani public agency is permitted to exercise. This is what Matt Andrews, Lant Pritchett and Michael Woolcock call a capability trap, sustained by isomorphic mimicry: the state acquires the outward appearance of modern delivery — instant payments, a digital tax stack, agent networks — while the actual capacity to build and run these things migrates outside the state and comes to rest in a company the state does not control (Andrews, Pritchett & Woolcock, 2017). Every workaround SPV is one more reason never to fix the thing that made the workaround necessary. The scaffolding becomes the building.

And here the argument turns from complaint to prescription. If such autonomy is genuinely necessary for performance — and it plainly is — then the lesson is not to keep multiplying donor SPVs until the real state is a hollow shell surrounded by nimble substitutes. The lesson is to reform the state so that its own departments can hire, procure, invest, experiment and deliver — with accountability rather than suffocation. This is not a lonely conclusion. Mazzucato and Collington, arriving from the opposite end of the development spectrum, reach the same cure: stop treating consultants and outside vehicles as substitutes for public capacity, and rebuild the capability in-house. The freedoms we have quietly granted to a private company are an unanswerable indictment of the rules we impose on our own public agencies. We have proved, with Karandaaz, that a Pakistani institution can be fast, flexible and effective. We have simply decided that our own state may not be.

The honest counterargument deserves a hearing. Raast works. The FBR did need help the bureaucracy could not supply on its own timetable. Financial inclusion has widened. Better a functioning system delivered through a donor vehicle, the argument runs, than another decade of committees and nothing built. I concede the delivery. But it answers the wrong question. The question was never whether the work gets done. It is who owns the capacity when the grant ends, who answers when the system fails, who sets its priorities while it runs, and whether the sovereign functions of the state should be operated by entities accountable to foreign funders. Delivery is not institution-building. A rail that works but that the state cannot govern, audit or reproduce is not a national asset. It is a dependency with good uptime.

This is why Karandaaz stops being a story about one organisation and becomes a story about us. It fits, almost too neatly, the structural dysfunction I have argued for years — in the Framework for Economic Growth (2011) and the PIDE Reform Agenda (2022): a state that outsources its thinking to donors and its building to their vehicles; a permission economy in which capacity is rented rather than owned; capital deployed at scale with no domestic principal behind it. Karandaaz is not the disease. It is the most sophisticated symptom yet — the donor-funded parallel state, no longer confined to a microfinance window, now operating the country’s most modern and most sensitive infrastructure.

What should be done does not require shutting anything down. It requires ending the pretence.

First, transparency and public audit. An entity of this scale, now helping run a national payment rail and the tax system, should publish its audited accounts, board decisions and grant agreements, and should fall within the reach of the Auditor General and the Public Accounts Committee. Public functions demand public scrutiny.

Second, sovereign functions must be governed by the sovereign. The governance of Raast belongs with the State Bank; the design of tax administration belongs with the FBR and the Finance Division. Donor roles in these domains should carry an explicit sunset, not an open-ended sequence of renewals.

Third — and most important — reform the state’s own arms instead of multiplying substitutes. Give public agencies the autonomy Karandaaz already enjoys: to hire on merit, to pay competitively, to procure and experiment quickly — but paired with real accountability, not the current mixture of suffocating control and negligible results. If PPRA is too rigid for modern delivery, the answer is to fix PPRA, not to route sovereign contracts around it through private companies.

Fourth, a parliamentary debate — and a real one. Parliament has already begun to ask the right questions about the FBR contract. It should not stop at a single exchange on the floor. A £187 million, Gates-co-funded body governing national payments and modernising the revenue authority deserves a sustained hearing.

The issue is not the competence of Karandaaz. It is the ownership of the state. We have built a company to do the things we forbid our own government to do, funded it from abroad, and let it operate our most important public functions with freedoms no public servant is allowed. The task is not to build more such companies. It is to build a state that no longer needs them.

Sources and further reading

• Andrews, M., Pritchett, L., & Woolcock, M. (2017). Building State Capability: Evidence, Analysis, Action. Oxford University Press. (See also Pritchett, Woolcock & Andrews, “Capability Traps? The Mechanisms of Persistent Implementation Failure,” CGD Working Paper 234, 2010.)

• Bräutigam, D. A., & Knack, S. (2004). “Foreign Aid, Institutions, and Governance in Sub-Saharan Africa.” Economic Development and Cultural Change, 52(2), 255–285. (See also Bräutigam, Aid Dependence and Governance, 2000.)

• Gabor, D., & Brooks, S. (2017). “The Digital Revolution in Financial Inclusion: International Development in the Fintech Era.” New Political Economy, 22(4), 423–436.

• Mazzucato, M., & Collington, R. (2023). The Big Con: How the Consulting Industry Weakens Our Businesses, Infantilizes Our Governments, and Warps Our Economies. Allen Lane / Penguin.

• McGoey, L. (2015). No Such Thing as a Free Gift: The Gates Foundation and the Price of Philanthropy. Verso.

• World Bank (1998). Assessing Aid: What Works, What Doesn’t, and Why. Oxford University Press.

• FCDO / A&M (2021). Mid-Term Assessment of the EAGR Program (£111m / PKR 23.42bn figure).

• Bill & Melinda Gates Foundation, Committed 2025Grants Database (December 2025, November 2025 and 2026 Karandaaz grants).

• On the FBR–Karandaaz–McKinsey engagement and the Senate objections: Express Tribune (”Pakistan partners with Gates’ Foundation to overhaul FBR,” March 2024); The Nation / Business Recorder / Mettis Global (steering committee approval of McKinsey, April 2024); Dawn (”Senators question US firm’s role in FBR digitisation,” April 2024; “Quest for economic sovereignty,” September).