
The owners put in 57 percent of the money. Across the 28 agribusiness projects the START facility in northern Uganda had approved by 2021, they contributed UGX 23.7 billion of UGX 41.5 billion, against a design assumption of 25 percent. Each project then went to Uganda Development Bank for credit analysis “as per the bank’s standards”, where many owners struggled with appraisal fees, collateral and valuation. That is the answer to who allocates in an African ecosystem: a relay of allocators, each deciding on evidence produced for the allocator above it, with the founder’s own money doing more of the work than any of them records.
A development finance institution in Europe lends to a Kampala bank on the strength of the bank. The bank lends to a founder on the strength of a land title. A donor facility counts approved projects against a target. And the owners, who know the business best, carry most of the cost.
Key Takeaways
- Capital reaches East African founders through a chain: a DFI or donor funds an intermediary, and the intermediary funds the founder, each on different evidence.
- In the 28 projects the START facility had approved by 2021, owners contributed UGX 23.7 billion of UGX 41.5 billion, about 57 percent, against a design assumption of 25 percent.
- The latest Uganda figure in the World Bank’s indicator database, from the 2013 survey, found collateral required on 86.7 percent of loans, at a value of 161.8 percent of the loan.
- A 2025 survey of 104 African local capital providers found that 90 percent of DFI funding to small-business growth funds had gone to fewer than 10 funds.
- French research in the book shows allocators deciding on personal standing and off-the-record doubts; the Ugandan relay adds a collateral gate on top of that relational one.
- Judging panels carry real information, but in Ghana they were better at screening out the weakest applicants than at ranking the strongest.
One facility, three sources of money
UNCDF reported a total project cost of UGX 41,535,883,827 across the 28 approved projects. The facility itself put in UGX 8.7 billion. Uganda Development Bank and other lenders put in UGX 9.1 billion. The owners put in UGX 23.7 billion, and the report notes that the design had assumed a 25 percent owner share.
Three parties allocated, each reading something different. The START board read a business plan. The six projects in its 2021 round had initially failed to make the pipeline and received pre-investment business development support before they reached the board. The bank read its standard credit file: appraisal, collateral, a valuation. The owners read their own business, which they knew better than anyone in the chain, and backed it with contributions that appear in the report only as a total.
The facility reported progress as 28 projects, 70 percent of a target to finance 40 small businesses between 2018 and 2022. What such counts leave out is the subject of Resource Flow: Why Counting Programs Misses What Reaches Founders; here the question is who decided, on what.
Stephen Kakonge, Senior Investments Manager for agriculture at Uganda Development Bank, said at the time that the bank was revising its requirements “so as to allow more SMEs [to] benefit from the bank’s financing options.” The bank saw the problem in its own file and moved, which deserves credit and also shows where the binding gate sat.
The money above the bank
One step further up sits the lender to the lender. FMO, the Dutch development bank, describes on its project page a USD 20 million senior unsecured loan to I&M Bank Uganda, half committed and half uncommitted, to grow SME lending. The page tells you what FMO read: a bank ranked 14th of 24 in Uganda with a 2.06 percent market share, part of a regional group, and a strategy to “deepen relationships with strong regional banking groups.” FMO lends to the bank unsecured; the bank’s standing is the evidence. The bank, in turn, will ask the founder for security.
The Collaborative for Frontier Finance surveyed 104 local capital providers across Africa for its October 2025 State of Play report and found the same relay from the fund manager’s seat. The largest share of DFI portfolios sits with local regulated banks. Ninety percent of DFI funding to small-business growth funds has gone to fewer than 10 of them. The managers cite access to DFI and institutional money as their single biggest barrier, having raised about $790 million against an estimated need of $2.25 billion.
What do these upstream allocators read? The report lists team capability and experience, vehicle structure and governance among the criteria. A first-time manager with a sharp local thesis lacks a fund track record by definition, so 81 percent of surveyed funds bring external members onto their investment committees, which the authors say adds confidence for limited partners. The local manager borrows credibility to satisfy the evidence test of the allocator above.
Lower down, local growth funds source deals through referral networks and their own accelerators. Commercial banks rely materially on walk-in clients who are strangers to them. And when the European Union reviewed 16 financing vehicles it had built in Uganda, one key barrier to uptake was the requirement for land titles as collateral.
What a French region’s allocators read
Two chapters of Entrepreneurial Ecosystems: Resource Acquisition, Flow and Management, edited by Parisa Baig and four colleagues (Springer, 2025), answer the same question for France.
Laura Bidaut, Karim Messeghem and Sophie Casanova interviewed 21 actors in the Burgundy ecosystem between January and July 2024, including two banks, two investors and an accountant. Most money for founders there comes from the regional council and departmental funds, and one investor summed up the base: “Today the entire innovation ecosystem is financed by public money.”
The evidence individual allocators read is personal. An investor said the ecosystem works on an intuitu personae basis, in person and on personal standing, so some people simply decline to work with each other. A lab director described growing wary of a start-up on the slightest doubt about its ability to finish, even a doubt voiced off the record by another operator. And a bank interviewee asked for something every founder would also want: “We need to clarify who does what and what each person’s tasks and powers are.” The bank was asking the ecosystem for a list of its allocators.
Pauline Brunner and Véronique Schaeffer, surveying student entrepreneurs in French university entrepreneurship centers between 2019 and 2021, counted who put up the money. Of 281 financial contributions from ecosystem actors outside the founder’s family and university, public funders made 73, banks 41, customers 37 and investors 30. Above €50,000 the order changed: investors 10, banks 8, public funders 7, customers 6. Customers allocated money too, and at the top end almost as often as banks.
Honor loans in Dijon, land titles in northern Uganda
What transfers is the personal gate. Burgundy’s investor working intuitu personae and the lab director wary of an off-the-record doubt read the same kind of evidence as a Ugandan growth fund sourcing deals through referrals and its own accelerator. In both places a founder with a trusted introducer is legible and a stranger is close to invisible, and the French bank’s request for a map of who does what would be just as welcome in Kampala. The customer as allocator transfers too, as a question: the French survey found customers making large contributions, and START’s owners show that money from inside the business, its owners and its buyers, carries much of the load.
What fails is the public base. Burgundy’s support system is financed mostly from regional and departmental budgets, and one of its mainstays is the honor loan for early-stage founders, backed by regional funds, the Caisse des Dépôts, Bpifrance and regional banks. A Burgundy founder’s barrier is mainly relational: knowing someone, and keeping anyone from voicing a doubt.
A Ugandan founder meets the relational gate and a collateral gate behind it. The latest Uganda figure in the World Bank’s Enterprise Surveys database, from the 2013 round, found collateral required on 86.7 percent of loans, at an average value of 161.8 percent of the loan. Kenya’s 2018 figures were 81.4 percent and 252.7 percent; Rwanda’s 2019 figures were 99.4 percent and 231.3 percent. A 2025 Uganda round has released its microdata; until its headline figure is published, treat 2013 as old and the pattern as regional. And the top of our relay often sits outside the country, in The Hague or Brussels, so the evidence the founder must produce is shaped by allocators who will never meet her.
A gold ring at the assembly door
James describes an allocation scene. A man with a gold ring and fine clothes walks into the assembly, and a poor man in filthy clothes walks in behind him. Someone in the room offers the first “a good seat” and tells the second to stand, doing what the bank does with a land title: reading the visible proxy because it is quick and usually correlates with something. James calls it becoming “judges with evil thoughts.” The Burgundy chapter and the Enterprise Surveys show allocators reading the ring for want of a better instrument, and they leave open why anyone should pay for a better one. James answers in his first line, where he names whose assembly it is, “our glorious Lord Jesus Christ”. Whoever hands out the seats is handing out someone else’s, and answers to Him for looking past the clothes to the person. That is a reason to build the better instrument even when the proxy is cheaper.
Judges in Accra, Tema and Kumasi
Marcel Fafchamps and Christopher Woodruff launched a business plan competition in Accra and Tema, Ghana, in January 2010, extended it to Kumasi, and surveyed the firms again one and two years after the competition. Panels of three or four judges, successful business owners and consultants, scored written plans and 30-minute presentations. The panel scores predicted later growth, and they added predictive power even after the researchers controlled for detailed survey measures of the owners’ ability. The judges rarely got feedback on such calls, yet their judgment carried information.
That answers anyone who wants allocators to discard judgment, referrals and personal standing. Experienced people sitting across from a founder do see something, and the Burgundy investor working intuitu personae is reading real signals.
The limit is specific. The survey measures explained more of the variance in growth than the panels did. When the researchers split the sample, the panels proved better at cleaving off the bottom of the distribution than at distinguishing within the top. Judgment screens out weak applications well. The money, though, goes to the top, and there the panels’ edge thinned.
The Evidence Relay
The Evidence Relay is a way for anyone who holds a funding window to see the chain they sit in. It has four handoffs, and each is a question you answer in writing for one live window.
Who funds you, and what do they count? Name the allocator above you and the number they ask for each quarter. If it is firms financed, you will be pulled toward approvals; if it is portfolio at risk, toward security.
What do you read about the founder? List every document and signal in your requirements: title deed, valuation, audited accounts, referral, panel score.
Is each item a direct read or a stand-in? A land title stands in for the ability to repay. A referral stands in for character. Twelve months of sales receipts, signed customer orders or an owner contribution already paid in are closer to the thing itself.
Which stand-in can you swap this cycle? Pick one and replace it with direct evidence the founder can produce, then tell your upstream funder you changed it and why.
AVODA holds a place in this relay. Across 18 programs, 1,247+ people have applied to us and 600+ have been trained, so most of what we allocate is seats, and every applicant is entitled to ask what evidence filled them.
Frequently Asked Questions
Who put up most of the money in Uganda’s START facility projects?
The owners. In the 28 agribusiness projects the START facility in northern Uganda had approved by 2021, owners contributed UGX 23.7 billion of UGX 41.5 billion, about 57 percent. The facility, and Uganda Development Bank with other lenders, each contributed roughly a fifth. The UNCDF report records the owners’ share as a single total.
Why do development finance institutions lend to banks instead of founders?
Institutional lenders want scale, governance and proven risk management they can assess from abroad, and a regulated bank supplies all three. FMO’s USD 20 million loan to I&M Bank Uganda is unsecured; the bank’s standing is the security. The October 2025 CFF survey found the largest share of DFI portfolios sits with local regulated banks.
How much collateral do East African banks ask for?
On the World Bank Enterprise Surveys, collateral was required on 86.7 percent of Ugandan loans in 2013, valued at 161.8 percent of the loan. Kenya in 2018 showed 81.4 percent and 252.7 percent, and Rwanda in 2019 showed 99.4 percent and 231.3 percent. A newer Uganda round is pending publication.
Do judging panels pick the right founders?
Partly. In Ghana, Fafchamps and Woodruff found that panels of business owners and consultants predicted later firm growth, even after survey controls for ability. Survey measures explained more of the variance, though, and panels were better at screening out the weakest applicants than at ranking the strongest ones, who receive the awards.
What is the Evidence Relay?
A four-question exercise for anyone running a funding window. Name who funds you and what they count, list what you read about founders, mark each item as a direct read or a stand-in, and swap one stand-in for direct evidence this cycle. It makes the chain of allocators visible to the people inside it.
Next Step
Write the four handoffs in the margin of the application form you use now. If the stand-in you want to swap is one your own funder insists on, use a strategy call with Joshua to draft the case you send upstream, with the funder’s reporting template open beside the form so the facility stays in place.
Related Reading
- Lending in Shillings: The DFI Local-Currency Pivot
- The ESO Funding Crisis in Africa: After the Aid Collapse
- The Parish Development Model as a Capital Allocator
- 77 Angel Networks: Africa’s Quietest Capital Revolution
Sources and Evidence
- Baig, Montoya-Martinez, Belitski, Theodoraki and Godley (eds.), Entrepreneurial Ecosystems: Resource Acquisition, Flow and Management, Springer, 2025: the book tested here.
- Bidaut, Messeghem and Casanova, “Local Policies and Institutional Frameworks: Shaping the Landscape of Resource Accessibility in the Burgundy Entrepreneurial Ecosystem”, chapter 5: 21 interviews, 2024; public funding base, intuitu personae, off-the-record doubts, “who does what”, honor loans.
- Brunner and Schaeffer, “Which Actors Provide which Resources to Student Entrepreneurs in Entrepreneurial Ecosystems?”, chapter 4: French student entrepreneurs; financial contributions by public funders, banks, customers and investors, including those above €50,000.
- UNCDF, “START Management Board Approves 6 Additional Projects for Financing, Spurs Policy Change”, September 2021: 28 projects, funding split, Uganda Development Bank requirements, 40-SME target.
- FMO, project detail: I&M Bank Uganda Limited: USD 20 million senior unsecured loan for SME lending; bank rank and market share; regional banking group strategy.
- Collaborative for Frontier Finance, Small Business Finance in Africa: State of Play, Issue No. 1, October 2025: 104 local capital providers; DFI concentration; LP criteria; external investment committees; EU review of 16 Ugandan vehicles.
- World Bank Enterprise Surveys, proportion of loans requiring collateral, Uganda, Kenya, Tanzania, Rwanda: share of loans requiring collateral by country and survey year.
- World Bank Enterprise Surveys, value of collateral needed for a loan, Uganda, Kenya, Tanzania, Rwanda: collateral as a percent of the loan amount by country and survey year.
- Fafchamps and Woodruff, “Identifying Gazelles: Expert Panels vs. Surveys as a Means to Identify Firms with Rapid Growth Potential”, World Bank Policy Research Working Paper 7647: Ghana business plan competition; panel and survey prediction of growth.
- James 2:1-4, NIV: the usher, the gold ring and the seat.
