Why we should underwrite what lenders can see, not what borrowers own
Ask a traditional lender what it needs before extending credit and the answer is always the same: a title deed, a guarantor, a credit history, a bank statement. Every item is a proxy for one question. If this person can’t repay, what can we take?
It’s a fair question. But in African agriculture, it quietly shuts out the very people the sector needs most. Most of the tractor owners we finance are first-time business owners. Many have no credit history and no bank account. The land they farm is usually held through family or customary arrangements that no bank will accept as security. By conventional measures, they are unfinanceable. And yet a smallholder will never be mechanized unless someone owns the machine and runs it as a business.
So if we want to close the mechanization gap, we have to change the question. Ours is simple: not what does this person own, but what can we see?
We’ve been measuring the wrong thing
For decades, access to finance has depended on what people can already prove they have. I understand why. Lenders have to manage risk, and they need some assurance they’ll be repaid.
But we should ask what happens when our way of measuring risk excludes the people with the greatest potential to create value. Across Africa, there are entrepreneurs with the ambition, the skills and the customers to build real agricultural businesses. What they lack is the paperwork to convince a bank. The result is a system that rewards what someone already owns over what they are capable of building.
That should bother all of us.
Visibility gives lenders what collateral can’t
Land became the default security for two reasons: it’s easy to find, and it can’t run away. Strip out the legal machinery and that is all collateral does. It tells the lender the asset, and therefore the loan, won’t disappear.
Land is just one way to deliver that confidence, and a poor fit where land rights are informal. A productive asset can do the same job if you can see it. An untracked tractor in a rural district is a machine nobody can verify. It might be working, parked, sold or stripped for parts, and the lender finds out which only when payments stop. A tracked tractor is a different proposition. When you know where it is, whether it’s working and how much, the asset stops being a risk you hope will behave and becomes one you can follow.
What matters most is timing. In our portfolio, a tractor that goes quiet mid-season isn’t a default. It’s a reason to pick up the phone that week and ask about a breakdown, a spare part, a dry spell or a gap in bookings. Collateral can only be seized after the loss. Visibility works while there’s still something to save. For investors, that’s the real shift: from enforcing against losses to managing them before they happen.
The record is the more valuable asset
When repayment is tied to work completed, every job and every installment adds a line to a history that didn’t exist before. Someone who arrived with no credit file, no bank account and no property now has a season-by-season account of what they earned and what they repaid.
The industry treats the loan record as internal paperwork. We’ve come to see it as the most valuable thing a customer takes away. It’s a credential that lets an owner who started with nothing be taken seriously by the next financier, the next supplier and the next opportunity. Financial institutions that want to reach this segment aren’t held back by a lack of demand. They’re held back by a lack of verifiable information. Transparent repayment histories close that gap, and each one lowers the cost of lending to the next owner in line.
Our Nigerian portfolio shows what this looks like. These are five-year contracts, built on the assumption that owners would need the full term to reach ownership. Some have cleared theirs in about three years. They treated the tractor as a business: they showed up for the work, took bookings seriously and paid consistently, season after season. That matters for two reasons. First, it challenges the core assumption of collateral-based lending, that borrowers without assets are inherently riskier. An owner with no collateral, no credit file and no bank account repaid in full, and ahead of schedule, once the structure gave them a fair chance to earn. Second, it changes what happens next. An owner who finishes early walks away with a documented history of a business that performed, which is exactly what a bank or supplier looks for and exactly what they couldn’t show at the start.
It also reveals something structural. When repayment follows work completed, the borrower’s success and the lender’s success are the same event. The faster an owner earns, the faster they repay.
Visibility only works if it runs both ways
There’s a version of this approach that is just surveillance: a lender that watches everything and shows nothing. That isn’t trust, and it won’t last. What we’ve found is that the model works only when owners understand what’s expected of them before they take possession, and can see afterward exactly what they owe and what they’ve earned. Data on where demand sits and where their peers are finding work should flow back to them. If you ask a borrower to be open, you have to be open in return.
The same logic is why the human layer can’t be automated away. A tracker can tell you a tractor has been idle for two weeks. Only a conversation tells you whether it’s the machine, the weather or something going on in the owner’s life. Lenders adopting technology-led models should resist treating data as a substitute for relationships. Data is what makes the relationship timely.
What this approach can’t do
It would be dishonest to pretend visibility cures risk. This year, abnormal rainfall across East and West Africa and rising fuel costs squeezed owners from both sides at once, and no dataset could have prevented that. What visibility did was show us early where the pressure was building, so we could work out how to support owners through a hard stretch instead of pushing them to keep working through it.
It’s also why we measure the model differently. Repossession is not a success. When a lender takes back the machine, the owner loses a livelihood, the farmers they served lose a service, and the lender loses a customer and ends up with a depreciating asset. The case for this approach rests on making that outcome rarer by spotting trouble while there’s still room to act.
What this means for lenders and investors
If this experience generalizes, and we believe it does, the implications start with how risk is assessed.
Underwrite behavior as seriously as balance sheets. Where borrowers have no formal credit history, observed performance is better evidence than the proxies lenders usually rely on. A season of completed jobs and consistent payments says more about reliability than a bank statement that doesn’t exist.
Tie repayment to earning. When repayment follows productive work, the gap between what a borrower can pay and what they owe narrows, and early signs of stress become far easier to spot. That’s not only fair to the borrower. It builds a portfolio that tells you where it’s struggling long before a missed payment would.
Fund the visibility, not just the loan. Tracking and account management are usually treated as costs to minimize. In our experience they are risk-management infrastructure, the mechanism that turns a warning signal into a conversation while there’s still something to save. Investors who fund the loan but not the visibility around it are leaving their best protection unfunded.
Make repayment records portable. Every verifiable repayment history makes it cheaper for the next lender to serve the next owner, and the whole market benefits when those records travel with the borrower. Treat them as assets, not internal paperwork, and a handful of successful loans becomes a more financeable segment.
Trust is built, not pledged
A title deed is a document you present once. Trust is a record built over many seasons: every job completed, every payment made, every hard stretch navigated together. You can’t copy it the way you can copy a pricing model or an app, because it exists only where someone has shown up long enough for it to accumulate.
That’s why I believe the next generation of agricultural finance won’t be defined by who accepts the most kinds of collateral. It will be defined by who is best at making reliable people visible. For an entrepreneur with no land to pledge, a verifiable record of showing up and getting paid is worth more than any deed.
The people who most need mechanization aren’t the ones with the most to pledge. They’re the ones who, given the chance and the support, will do the work. Our job is to make that visible, and then to keep showing up.