Most conversations about lending in Latin America start with the same conclusion: credit is expensive because borrowers are risky.
Low income volatility, large informal sectors, thin credit files, and macro instability are usually cited as the root causes. From that view, high interest rates and low approval rates are simply the rational price of higher default risk.
But that explanation is incomplete.
The real constraint in LATAM credit is not risk, it’s lack of data. Lenders are not consistently bad at pricing risk. They are often blind to it.
And when you can’t see clearly, you price defensively.
The Wrong Starting Point: “LATAM Borrowers Are Risky”
The traditional narrative goes like this:
Many consumers and SMEs operate informally
Income is unstable or undocumented
Credit bureau files are thin or nonexistent
Financial histories are fragmented
Therefore, borrowers are structurally high risk
This logic leads to predictable outcomes:
High rejection rates
Wide risk buckets
High average interest rates (Annual Percentage Rates)
Conservative limits
Slow credit expansion
But this mixes up two very different concepts: risk and uncertainty.
Risk can be priced. Uncertainty cannot.
Risk vs Uncertainty: A More Useful Framework
A simple distinction changes the lens.
Risk = you know the variables and can estimate probabilities.
Uncertainty = you lack enough variables to model probabilities reliably.
Example:
A borrower with 10 years of bureau data, tax records, and bank history has measurable risk
A borrower with no bureau file but steady platform income will be uncertain, not necessarily risky
Many LATAM borrowers have not high risk, they are highly uncertain.
Uncertainty forces lenders to assume worst-case scenarios. That assumption gets embedded into pricing and approvals.
In other words: spreads are often driven more by missing data than by actual behavior.
Why Traditional Credit Data Breaks in LATAM
Classic underwriting models were built around formal financial footprints:
Bureau scores
Payroll records
Tax filings
Bank loan history
Credit card usage
In much of LATAM, these signals are incomplete or delayed. Large segments of the population and SME base generate economic value, but outside traditional reporting rails.
What legacy data often misses:
Informal or hybrid income streams
Wallet transaction flows
Marketplace seller revenues
Gig platform earnings
Cross-border remittances
Prepaid and debit usage patterns
Utility and telecom payment history
When core signals are missing, lenders fall back to rough guesses. That creates large, heterogeneous buckets where good and bad borrowers are mixed together.
The result: good borrowers subsidize unknown borrowers, and both get expensive credit.
LATAM Is Not Data-Poor, It’s Data-Fragmented
A common misconception is that the region lacks usable data.
Today, many borrowers leave rich digital trails, such as the one I mentioned in the previous section.
The challenge is not data creation, it’s data integration and permissioned access.
This is why lenders embedded in ecosystems (wallets, marketplaces, payment processors, etc.) often outperform standalone lenders. They sit closer to the transaction layer and see behavior in real time, not through delayed reports.
When Data Improves, Credit Economics Improve
Better data doesn’t just reduce losses. It changes the full lending equation:
Approval rates increase
Loss rates become more predictable
Pricing can narrow by cohort
Limits can scale faster for good payers
Repeat usage rises
Customer lifetime value expands
CAC payback improves
This is why transaction-embedded lenders often show stronger unit economics than traditional institutions serving the same segments.
They are not lending to different people. They are seeing the same people more clearly.
The Investor Misread
Many investors still evaluate LATAM lenders with a simplified heuristic:
High APR (anual interest rates) = predatory or fragile.
But in many cases, high APR reflects an uncertainty premium, not purely a risk premium.
Two lenders can serve similar borrower segments with very different outcomes depending on their data advantage.
Better diligence questions include:
What percentage of underwriting uses first-party transaction data?
How fast does the model retrain with repayment events?
How granular are risk cohorts?
Is pricing dynamic or static?
Does the lender sit inside the user’s financial workflow?
How tight is the feedback loop between behavior and limits?
The core asset is not just the loan book, it’s the data loop.
The Second-Order Insight: Distribution + Data Beats Capital
The next generation of credit winners in LATAM will not be defined primarily by:
Lowest cost of capital
Biggest balance sheet
Most aggressive marketing
They will be defined by:
Proprietary behavioral data
Embedded distribution
Real-time visibility into cash flows
Continuous model learning
In that sense, the strongest credit companies increasingly look like data companies that lend, not lenders that collect data.
What This Means Going Forward
Several shifts are likely:
Bureau data becomes a baseline, not the core signal
Platform and transaction data become primary inputs
Embedded credit outperforms standalone credit
Risk spreads compress, selectively
Margin differences track data advantage
Credit access expands through workflow platforms
The key separator will not be who takes more risk, but who measures it better.
Portfolio Companies
News about our Portfolio Companies
Caja Los Andes Invests in Startup Soyio: A Historic Debut
Leracom AI raises US $1 million in a Pre-Seed round to transform contact centers with Artificial Intelligence
Portfolio companies currently raising investment rounds
Wisecricket: A platform that automates financial audits using AI: monitoring invoices, contracts, and fiscal compliance in real time.
Please don’t hesitate to contact us in case you are interested in connecting with any of these amazing companies.
Great by Choice: Uncertainty, Chaos, and Luck; Why Some Thrive Despite Them All - Jim Collins. Him reframes success not as a straight line of skill, but as a living tapestry woven out of uncertainty, luck, and the choices we make in the gaps between them.
Daniel Kahneman: Algorithms Make Better Decisions Than You | The Knowledge Problem