Latest / Fintech Conversations with Fexingo: Banking, Payments, and Financial Software Companies / How Fintech Is Using Climate Risk Data for Lending
Transcript
- Lucas: If you're building a lending product in 2026, you probably have a pretty sophisticated model for credit risk. You're pulling in bank transaction data, maybe payroll data, utility payments, even psychometric scoring. Luna: Right — the alternative data revolution has been the big story for a decade. But I'm hearing a new variable creeping into underwriting models lately, and it's not behavioral. It's physical. Lucas: Exactly. Climate risk. Lenders are starting to ask: is this asset going to be standing in ten years? And if it is, what will it cost to insure? That's a fundamentally new question for credit analysis. Luna: And before we go deep on this — Lucas, these kinds of conversations are exactly why we keep this show ad-free. We want to talk about what's actually reshaping finance, not interrupt you with a mattress ad. Lucas: Totally. And if you find that useful, all we ask is that you consider tossing a few dollars our way. It's buy me a coffee dot com slash fexingo. That's it. No tiers, no perks, just supporting the work. Luna: And it genuinely helps us keep the lights on and spend time on episodes like this. Now — back to climate. Lucas, what's actually changed that makes climate risk relevant to a lender today? Lucas: Three things. First, data quality. Companies like Jupiter Intelligence and Riskiq now offer property-level flood and wildfire risk scores at a resolution of a few meters. Second, insurance premiums for high-risk properties have jumped 20 to 50 percent in some regions, and that directly impacts a borrower's cash flow. Luna: So the cost of ownership is changing, even if the mortgage hasn't moved. And the third factor? Lucas: Regulation. The OCC in the US and the ECB in Europe have started asking banks to stress-test their loan portfolios against climate scenarios. JPMorgan, Bank of America — they're all piloting climate-adjusted loan pricing. Luna: Let's talk about that pricing adjustment. How do you even quantify climate risk for a loan? It's not like a FICO score where you have decades of historical defaults. Lucas: Right. So the approach is probabilistic. You use climate models to estimate the likelihood of, say, a one-in-100-year flood over the life of a 30-year mortgage. Then you estimate the expected loss — damage to collateral, business interruption, increased insurance costs. Luna: And that feeds into loss given default? I've seen estimates that climate-adjusted LGD could increase by 30 basis points by 2030 for portfolios with heavy exposure to coastal properties. Lucas: Yeah, and 30 basis points is significant when you're managing a billion-dollar loan book. But the real concern is concentration risk. A lender might have a diversified credit portfolio on paper, but if 30 percent of their commercial real estate loans are in high-risk flood zones — and we're seeing numbers like 12 percent nationally — that's a systemic blind spot. Luna: So it's not just about pricing individual loans, it's about aggregate risk. Which brings up a fascinating question: does this lead to climate redlining? If a bank decides not to lend in certain ZIP codes, that's a real social consequence. Lucas: It's already happening, though it's not called that. Some banks have tightened loan to value ratios for properties in wildfire-prone areas in California. And if you're a small business in a floodplain, your access to credit is shrinking even if your personal credit is stellar. Luna: What about the data vendors? Jupiter Intelligence — they're a private company, right? How transparent are their models? Lucas: That's a huge open question. These models are black boxes to most lenders. They buy a score, they build it into their decision engine, but they don't fully understand the assumptions. And different vendors give different scores for the same property. There's no standard yet. Luna: So we have a situation where lenders are making high-stakes decisions based on opaque models, with limited historical validation, and the potential for discriminatory outcomes. Sounds like a recipe for regulatory scrutiny. Lucas: It does. The OCC has already issued guidance on climate risk management, and I expect they'll start asking banks to validate their climate data sources. But there's also an opportunity here for fintechs to build transparent, auditable climate risk models. Luna: Are any startups doing that? I'm thinking of companies like Riskiq that specialize in climate analytics for insurance. Could they pivot into lending? Lucas: They're already doing it. Riskiq has a product that gives lenders a climate-adjusted expected loss for commercial properties. And there's a startup called Zesty.ai that uses aerial imagery and AI to assess wildfire risk at the individual structure level. They're working with a few regional banks. Luna: What about the data? Satellite imagery is one thing, but do they incorporate things like local fire department response times, or building materials? Lucas: They do. Zesty.ai's model includes over a hundred variables — roof type, vegetation proximity, historical fire perimeters, even wind patterns. It's incredibly granular. But again, the lack of standardization means a lender using Zesty might get a different risk score than one using Jupiter for the same property. Luna: So if I'm a community bank with a portfolio of loans in the Southeast, where do I start? Do I need to hire a climate scientist? Lucas: You don't, but you should probably start by mapping your existing loan portfolio against publicly available flood and fire risk maps from FEMA or NOAA. That's a low-cost first step. Then you can prioritize which loans to re-evaluate with a commercial model. Luna: And if you find that 20 percent of your portfolio is in a high-risk zone, what do you actually do? You can't just call up those borrowers and say 'your rate is going up because of climate change.' Lucas: No, and that's the next frontier. Some lenders are experimenting with dynamic pricing at origination, but for existing loans, the main tool is insurance requirements. You can require borrowers to carry flood insurance, and if premiums rise, that impacts their cash flow. But you're right — you can't unilaterally adjust rates. Luna: What about the mortgage secondary market? If Fannie Mae and Freddie Mac start incorporating climate risk into their purchase criteria, that would force the entire market to adapt. Lucas: That's the big domino. In 2025, Fannie Mae started requiring climate risk disclosures for multifamily properties. I expect that to expand to single-family. Once the GSEs are involved, climate-adjusted pricing becomes standard, not experimental. Luna: And that's huge for fintech lenders who originate loans to sell to the GSEs. They'll need to have climate data built into their origination systems from day one. Lucas: Exactly. And we're already seeing some fintechs, like Blend and Better, integrate climate risk data into their digital mortgage platforms. It's early, but it's happening. Luna: Let's go back to the international angle. The ECB has been pushing climate stress tests for European banks since 2022. Are European fintechs ahead on this? Lucas: In some ways, yes. European regulators have been more prescriptive. The ECB's 2024 climate stress test showed that banks' loan losses could be 15 percent higher under a severe climate scenario. So European banks are further along in integrating climate data into credit risk models. Luna: But American banks are playing catch-up. The OCC's 2025 guidance was a nudge, not a mandate. So the US is more of a patchwork — some large banks are doing sophisticated work, while community banks are largely ignoring it. Lucas: That's where fintechs can really add value. They can offer climate risk analytics as a service, so smaller lenders don't have to build it themselves. And the pricing models they develop could become the de facto standard. Luna: There's also an interesting competitive angle. If a fintech lender can offer lower rates for low-climate-risk properties — because their model shows those loans are safer — they could steal market share from incumbents who aren't differentiating on climate. Lucas: That's the optimistic view. The cynical view is that climate risk data becomes another tool for cherry-picking the best borrowers and leaving the rest to the government or credit unions. That's the redlining concern. Luna: So regulation is going to be critical. I think we'll see something like the Equal Credit Opportunity Act amended to address climate discrimination. Or maybe a new regulation requiring lenders to disclose their climate risk methodology. Lucas: I think you're right. And that's actually good for the fintechs that are building transparent models — they'll have a compliance advantage. The ones selling black-box scores might find themselves on the wrong side of a regulatory investigation. Luna: So where does this go in the next two years? Let's say it's mid-2028. What does the lending landscape look like? Lucas: I think climate risk data becomes a standard input in every commercial real estate loan underwriting. Residential mortgages will follow, but more slowly. And I think we'll see a few climate-focused fintech lenders emerge — companies that only lend to properties with low climate risk, and offer better rates because of it. Luna: Essentially a green mortgage product. That could be a powerful brand differentiator, especially with younger borrowers who care about climate. Lucas: Exactly. And it's not just about marketing — it's about genuinely lower default risk. If your model is right, you can offer better terms and still have a safer portfolio. That's a win-win. Luna: Alright, so to wrap up: climate risk is moving from an ESG consideration to a core credit risk factor. The data is getting good enough to use, the regulators are pushing, and the first movers are already integrating it. Lucas: And for lenders who ignore it, the risk isn't just reputational — it's financial. A few bad hurricane seasons could wipe out a decade of underwriting profits if you're concentrated in the wrong areas. Luna: A sobering thought. But for fintechs, it's an opportunity to build smarter, fairer, more resilient lending systems. And that's what this show is about. Lucas: Thanks for listening. If today's episode got you thinking, share it with a colleague. And if you want to support the work, you know where to find us.