Your Reporting Data Is a Gold Mine
Causal inference unlocks strategic insights from the data you're already collecting
Imagine you’re sitting in a board meeting, and the CFO presents the quarterly results. Claims are up 12% year-over-year. The room shifts uncomfortably. Someone asks the obvious question: “Why?”
The actuarial team has an answer ready. They pull up a slide showing that claims inflation is running at 8%, mix shift accounts for 2%, and the remaining 2% is unexplained volatility. The CFO nods. The board nods. Everyone seems satisfied.
But here’s what nobody says out loud: they have no idea if this explanation is actually true.
The actuarial team has correlated claims with inflation indices, looked at how the customer mix has changed, and attributed the remainder to noise. But correlation isn’t causation. And in insurance, the difference between “we think this is what’s happening” and “we know this is what’s happening” can cost millions.
This is the reporting paradox. Insurers invest enormous resources collecting detailed data—claims, underwriting decisions, reserving assumptions, exposure characteristics—all meticulously validated for regulatory compliance. Yet when it comes to the strategic questions that matter most, they stop at description and prediction. They ask “How many claims will we have?” and “Are our reserves adequate?” But they rarely ask “Why did this happen, and what should we do about it?”
That’s where causal inference comes in. And here’s the thing: you’re already collecting the data you need.
The Cost of Not Knowing Why
Let me make this concrete with a scenario that’s probably happening at your company right now.
A mid-sized regional insurer has been growing steadily, expanding into new geographies and adding new product lines. Over the past three years, their loss ratios have deteriorated across the board. The underwriting team is concerned. The finance team is concerned. The board is asking questions.
The standard response: “We’re seeing claims inflation in medical costs, and we’ve also had some bad luck with large losses.” They adjust their assumptions, tighten underwriting slightly, and move on.
But what if the real story is different? What if the problem isn’t external inflation or bad luck, but internal? What if their underwriting standards have gradually loosened—not deliberately, but through a combination of competitive pressure, staff turnover, and model drift? What if they’re attracting riskier customers than they used to, and that’s the real driver of the loss ratio deterioration?
The strategic response to these two scenarios is completely different. If it’s external inflation, you adjust your pricing assumptions and move forward. If it’s internal underwriting quality, you need to fix your underwriting process, retrain your team, and potentially reprice your entire book. The data you’re collecting—claims, underwriting characteristics, customer profiles—contains the answer. But you need to ask a causal question to find it.
This is where most insurers get stuck. They have the data. They have smart actuaries. But they don’t have a framework for asking “what causes what?” Instead, they ask “what correlates with what?” And those are not the same question.
Reporting Data Is Perfect for Causal Inference
Here’s the good news: if you’re an insurer, you’re already in a better position than most industries to do causal inference well.
Reporting data is uniquely suited to causal analysis. First, it’s detailed. You’re not working with aggregate summaries or sampled data; you have transaction-level information on underwriting decisions, claims outcomes, and reserving assumptions. You know what was underwritten, what was claimed, and what was reserved. That granularity is gold for causal inference.
Second, it’s historical. You have years—sometimes decades—of data. Causal inference works best when you can observe relationships over time and across different conditions. Insurance data gives you that.
Third, it’s validated. Every number in your reporting data has been audited, reconciled, and verified for regulatory compliance. You don’t have to worry about data quality in the same way you would with, say, web analytics or customer survey data. The data has already been vetted.
Fourth, and this is important, it’s already being collected. You’re not asking your business to invest in new data infrastructure or hire new teams to gather information. The data exists. It’s sitting in your data warehouse right now, waiting to be analyzed in a different way.
Most companies that want to do causal inference face a chicken-and-egg problem: they need data to learn causal relationships, but they don’t have the data. Insurers don’t have this problem. You have the data. You just need to ask different questions.
From Description to Decision
The shift from correlational thinking to causal thinking is fundamentally a shift in how you frame strategic questions.
Correlational thinking asks: “What variables are associated with claims?” You build a model, you find the strongest predictors, and you use them for pricing or reserving. This is valuable, but it has a ceiling. It tells you what to measure, but not what to change.
Causal thinking asks: “If I change X, what happens to Y?” This is the language of strategy. If we tighten underwriting standards, what happens to claims? If we shift our geographic mix, what happens to profitability? If we change our reserving methodology, what happens to our balance sheet?
These are the questions that matter in the boardroom. And they require causal reasoning, not just correlation.
The reserve adequacy question I mentioned earlier is a perfect example. Suppose your reserves are volatile. You could correlate reserve changes with inflation indices, loss development patterns, and underwriting variables. You’d find correlations. But you wouldn’t know which one is the real driver.
With causal reasoning, you can ask: “Here’s my hypothesis about how reserves are determined. Underwriting decisions lead to claims outcomes. Claims outcomes, combined with inflation assumptions, determine reserves. Is this causal chain correct? And if it is, which link is the weakest?” Now you can actually intervene. If the problem is in the underwriting link, you fix underwriting. If it’s in the inflation assumptions, you adjust those. Different diagnosis, different treatment.
This is what causal inference brings to the table. It’s not just a better way to analyze data; it’s a better way to think about strategy.
Why This Matters Now
Insurance is changing. Regulatory pressure is increasing. Competitive pressure is increasing. The cost of getting strategy wrong—either by misallocating capital, mispricing risk, or missing market shifts—is higher than ever.
At the same time, your reporting data is becoming richer. You’re collecting more variables, at higher granularity, with better validation. The infrastructure is there. The data is there. What’s missing is the framework for asking causal questions.
This is where you have an advantage. Your competitors are probably doing what you’re doing: treating reporting as a compliance function, extracting prediction models, and making strategic decisions based on intuition and historical patterns. But if you shift your mindset—if you start treating your reporting data as a strategic asset and ask causal questions—you can see things they can’t.
You can identify the real drivers of profitability, not just the correlates. You can validate your strategic assumptions against your own data. You can make decisions with conviction, not just correlation.
This isn’t about building fancy models or hiring data scientists. It’s about asking better questions of the data you already have.
On Friday, we’ll get into the technical side: how to actually build causal models from your reporting data using tools like Bayesian networks. But the real insight starts here, in the boardroom. It starts with recognizing that your reporting data isn’t just a compliance burden. It’s a strategic asset. And causal inference is how you unlock it.
The question is: are you ready to ask different questions?
Reads of the Week
Gordon Aitken’s deep dive into the UK life insurance sector reveals how 2025 became a pivotal “re-rating” year after a long stretch of investor scepticism. The article is a sharp lesson in how undervalued sectors can rebound when fundamentals, regulatory signals, and capital flows align—especially when private equity starts circling. As 2026 begins, Aitken’s analysis is a timely reminder to watch not just performance, but how competition, capital use, and narrative shifts shape markets.
This breakdown of permanent versus term life insurance by Quantor Capital Perspectives is a high quality read for anyone weighing their financial protection options. With clear-eyed analysis, they explain why the popular “buy term and invest the rest” strategy usually beats permanent life policies, which are often expensive, complex, and underperforming. For Wangari Digest readers navigating personal finance decisions, this piece offers a grounded, numbers-first look at a topic often clouded by sales pitches.
This week’s In Force brief from The Intelligence Council is a sharp snapshot of how risk in P&C insurance is evolving faster than products, pricing, and governance structures. It’s especially valuable in showing how cyber risk remains dominant while AI rapidly climbs the risk agenda—blurring the lines between technology failure, liability, and board-level accountability. The piece matters because it reframes familiar headlines (Florida stabilization, Lloyd’s discipline, cyber threats) as strategic signals about where control, not growth or appetite, will define winners in the next cycle.