Why Your NPL Ratio Lies: The Hidden Cost of Decision Inconsistency

Why Your NPL Ratio Lies: The Hidden Cost of Decision Inconsistency Here's the uncomfortable truth: your NPL ratio tells you almost nothing about future portfolio performance. It's a snapshot of the past. By the time your NPL ratio rises, the damage is already done. ## The NPL Lag Problem A borrower enters default after 90-180 days of missed payments. But the decision to lend to that borrower—the underwriting decision that led to the default—was made 18-36 months earlier. Your NPL ratio reflects decisions made when credit conditions, management philosophy, and lending standards were entirely different. It's not predictive. It's diagnostic. And diagnosis comes too late. Financial institutions report looking at NPL ratios as their primary credit health metric. Yet institutions with identical NPL ratios can have wildly different future performance: - Institution A: 3% NPL ratio, but rising 50 bps per quarter - Institution B: 3% NPL ratio, but stable for 24 months Same metric. Completely different stories. ## What NPL Ratios Actually Measure Your NPL ratio measures the **frequency of a past decision error**—but it doesn't measure the size of current risk. It can't answer: - Are new loan decisions getting worse? - Is our portfolio trending toward more defaults or fewer? - Is our underwriting quality improving or deteriorating? - What's the default risk for borrowers who haven't missed payments yet? These are the questions that matter. NPL ratios answer none of them. ## The Decision Consistency Alternative Instead of waiting for defaults to become visible (the NPL approach), decision consistency asks: Are we lending to the same quality of borrower, on the same terms, regardless of market conditions? This matters because: **Market sentiment drives most lending decisions.** When times are good, standards loosen. When times are tough, standards tighten. This creates clustering—multiple borrowers default together because they were all underwritten during the same optimistic (or pessimistic) period. **Consistency breaks clustering.** If you lend to the same credit quality always, defaults are distributed across time, not clustered. Your portfolio is more stable. **Early signals are possible.** If you define "consistent lending" operationally (same approval criteria, same pricing for similar profiles), then deviations from consistency are immediate early warnings. You don't wait 18 months to see that standards have drifted. ## Measuring Decision Consistency Here's how to measure it: 1. **Segment your portfolio** by borrower type (MSMEs, agriculture, trade, etc.) 2. **Define a "standard" borrower** for each segment (e.g., 2-year trading history, 40% debt-to-income ratio, $5K monthly turnover) 3. **Track the decision** given to standard borrowers over time - Did this borrower get approved in Month 1 and rejected in Month 12? - Did the approval rate for standard borrowers drop 40%? 4. **When consistency breaks, investigate immediately** - Has management changed risk appetite? - Has new staff been hired without proper training? - Is the market environment driving irrational decisions? Decision consistency metrics give you 12-18 month warning of NPL increases. NPL ratios give you the diagnosis after the problem is entrenched. ## The VALR Application We built our risk OS around this insight: **Behavioral assessment is more predictive than balance sheet assessment.** By tracking whether borrowers behave like other borrowers with identical profiles—not just whether they're making payments—we catch distress signals while they're still reversible. On KES 1.4B+ under live surveillance: - 25% average NPL reduction within 90 days of pilot - 30% improvement in Portfolio at Risk (PAR) through early warning - 12-18 month advance warning vs. traditional NPL metrics Your NPL ratio is a hospital report. Decision consistency is a preventive health check. One tells you what already happened. The other tells you what's about to.