By the time your actuary’s report shows adverse development, the development already happened. The claims that drove it were reported late, reserved slowly, lawyered up, or left open too long. The TPA data that would have warned you was sitting in the claims system twelve months ago, but nobody ran the right query.
A reserve study is a lagging indicator. It tells you what already happened to the dollars. The five metrics in this article are leading indicators. They describe operational signals at the claim level that move six to eighteen months before the aggregate triangle does. You can request all five from your TPA on Monday morning without involving your actuary, and any one of them trending in the wrong direction is a reason to pick up the phone before the next reserve review.
This article is the operational companion to the diagnostic framework that maps adverse development to its three root causes. If that article told you where to look, this one tells you what to measure.
What “leading indicator” means in a reserving context
The distinction matters. A lagging indicator describes what already happened: the development factors in your loss triangle, the actuarial point estimate, the prior-period adjustment your auditor wants explained. The chain ladder methodology that produces most reserve estimates is fundamentally lagging. It projects forward by assuming that future development will look like historical development (Friedland, p. 84). When the assumption holds, the method works. When it breaks, you find out at the next valuation.
A leading indicator describes a condition at the claim level that predicts the lagging result. Late-reported claims are more expensive. Slow case reserving leads to case adequacy problems. Rising attorney involvement drives longer duration and higher severity. Changing closure rates signal settlement pattern shifts. Large-loss frequency predicts tail development. Each of these is observable in TPA data months or quarters before the aggregate triangle reflects it.
The five indicators below are not actuarial calculations. They are operational metrics. Your TPA’s claims management system already tracks them. The gap is not data availability; it is that nobody routinely delivers these metrics to the person who signs the reserve on the balance sheet.
Indicator 1: Average days from incident to first report
What it measures. The elapsed time between the date of loss and the date the claim is first reported to your TPA. In workers compensation, it is the gap between the workplace injury and the employer’s first report of injury filing. In general liability, it is the gap between the occurrence and the claimant’s notice to the insured.
Why it predicts adverse development. Late-reported claims are systematically more expensive than promptly reported ones. The delay between incident and reporting is time during which injuries go untreated, witnesses become unavailable, evidence degrades, and claimants engage attorneys. By the time a late-reported claim enters the system, it arrives with a longer expected tail, higher expected severity, and less investigative leverage. NCCI data on workers compensation shows that claims reported after four weeks have attorney involvement rates exceeding 30 percent, compared to 13 percent on day zero. The cost implication compounds from there.
When average days-to-report rises across your book, two things happen in the triangle. First, pure IBNR increases because more claims are unreported at any given valuation date. Second, the claims that do arrive are more expensive on average, which inflates the development pattern. Both effects push the actuary’s estimate up at the next review.
The chain ladder’s core assumption is that claims recorded to date will continue to develop consistently (Friedland, p. 84). A rising reporting lag violates that assumption directly, because the claims now entering the triangle are not the same quality as the ones the historical pattern was built on.
How to query it. Ask your TPA for a monthly report of average days-to-report by accident month for the trailing twenty-four months. Segment by line of business and, if possible, by location. You are looking for a trend, not a snapshot. A single month’s spike is noise. Three or more months trending up is a signal that something changed in your reporting chain.
Concrete action. If the trend is rising, investigate whether the change is in employee reporting behavior, supervisor reporting protocols, or the TPA’s own intake process. Each has a different fix. The operational remedy is almost always faster and cheaper than the reserve impact you are trying to avoid.
Indicator 2: Average days from first report to first case reserve
What it measures. The elapsed time between the date a claim enters the TPA’s system and the date an adjuster sets an initial case reserve. This is the TPA’s response time to assess the claim and establish an initial dollar estimate of its ultimate cost.
Why it predicts case reserve adequacy slippage. When adjusters are overloaded, initial reserves get rushed. An adjuster handling 150 open files does not investigate each new claim with the same depth as one handling 80. The result is initial reserves that are systematically too low, which look fine in the incurred triangle for the first twelve months and then develop adversely as the adjuster catches up and strengthens the file. This is exactly the case reserve adequacy problem described in the diagnostic framework as Diagnostic 1.
A rising days-to-first-reserve metric is a proxy for adjuster workload stress. It tells you the claims team is falling behind before the reserve impact shows up in the triangle. It also correlates directly with larger ultimate claim costs, because delays in initial assessment mean delays in return-to-work coordination, medical management, and subrogation, all of which are time-sensitive.
How to query it. Ask your TPA for the average calendar days between first report date and first reserve posting date, by accident month, trailing twenty-four months. Benchmark the result against the TPA’s stated service standard. Most TPAs commit to initial contact within twenty-four to forty-eight hours and an initial reserve within five to ten business days. If the actual number is drifting past those benchmarks, the TPA’s staffing is not keeping up with claim volume.
Concrete action. If the metric is rising, ask the TPA directly: what has changed in adjuster caseloads over the last twelve months? Is the increase in caseload driven by new claim volume (a risk management problem) or by staffing turnover (a TPA management problem)? The answer determines whether you need to manage your exposure differently or renegotiate the TPA service agreement.
Indicator 3: Attorney representation rate by accident month
What it measures. The percentage of claims in which the claimant has retained legal counsel, measured by accident month. In workers compensation, this is the share of lost-time claims with attorney representation. In general liability, it is the share of third-party claims in suit or with counsel of record.
Why it predicts severity and duration shifts. Attorney-represented claims cost more and take longer to close. This relationship is one of the most extensively documented in workers compensation research. WCRI’s studies on attorney involvement have found that attorney representation is associated with significantly higher indemnity and medical payments and substantially longer claim durations. In a study of nearly one million claims across thirty-one states, attorney involvement was associated with payment increases of $7,700 to $12,400 per claim and a nearly threefold increase in lost-time days.
The trend matters more than the level. A stable attorney representation rate, even a relatively high one, is baked into your historical development pattern. The actuary’s chain ladder is calibrated to it. The problem arises when the rate changes. A representation rate climbing from 18 percent to 26 percent over six quarters is a freight train heading for your next reserve review, because every additional represented claim carries longer duration and higher severity that the historical pattern did not anticipate. Friedland’s failure-mode table flags changes in the mix of claims as a condition that breaks the development technique’s assumptions (Friedland, Section 4). A rising attorney rate is a mix change, and it maps directly to longer tails and fatter development factors.
How to query it. Ask your TPA for attorney representation rate by accident month for the trailing twenty-four months, segmented by line of business. In workers compensation, also segment by injury type if possible. Overlay the trend against your aggregate paid-loss development and look for whether the representation rate leads the severity trend by two to four quarters. If it does, you have your early warning.
Concrete action. If the rate is trending up, investigate why. Are injuries getting more severe? Are employees dissatisfied with the claims process? Is there a new plaintiff firm marketing in your geography? Each of these has a different operational response. The one response that never works is ignoring the trend and letting the actuary discover it twelve months later in the triangle.
Indicator 4: Closure rate by maturity
What it measures. The percentage of claims closed within defined maturity windows: 90 days, 180 days, and 365 days from the date of first report. This is the speed at which your claims inventory resolves, and it is one of the most direct operational proxies for the settlement pattern that drives the paid chain ladder.
Why it predicts settlement-pattern shifts. Friedland’s failure-mode table explicitly flags settlement speed-up and settlement slowdown as conditions that break the paid chain ladder (Friedland, Section 4). The CAS research literature reinforces this: a CAS paper on the value of claim closure information to loss reserving found that claim closure counts carry significant predictive information for ultimate loss estimation beyond what paid and incurred triangles alone provide.
Both directions are signals. When closure rates slow down, claims are usually getting more complex, more litigated, or less actively managed. Any of these predicts higher ultimate costs. When closure rates speed up, the interpretation is more nuanced. It can mean the TPA improved its settlement process, which is genuinely good. But it can also mean claims are being closed prematurely and will reopen, or that small claims are being pushed out the door while large claims sit open longer, which changes the mix of what the triangle sees at each maturity. You need to know which interpretation applies.
How to query it. Ask your TPA for the percentage of claims closed within 90, 180, and 365 days of first report, by accident quarter, for the last eight to twelve quarters. Look at the trend in each maturity bucket independently. A 90-day closure rate that drops five percentage points over six quarters while the 365-day rate is stable tells you something different from a uniform slowdown across all maturities. The first suggests early-stage claims are getting more complex; the second suggests a systemic capacity problem.
Concrete action. If closure rates are decelerating, ask the TPA for a breakdown of open claims by age bucket. How many claims are open past twelve months? Past twenty-four? What is the average reserve on those aged claims? The answers will tell you whether the slowdown is benign (complex claims properly managed) or a staffing problem that will generate adverse development on the next diagonal.
Indicator 5: Frequency of large-loss notices and reserve changes above threshold
What it measures. The count and aggregate dollar amount of claims that exceed a predefined severity threshold, typically your stop-loss or specific retention level, or a round-number internal benchmark such as $100,000 or $250,000 in incurred losses. Most TPAs maintain a large-loss notification workflow that flags these claims automatically when the reserve or total incurred crosses the threshold.
Why it matters. Large losses are leveraged in the chain ladder. Friedland illustrates this with a bodily injury accident year whose paid development factor at early maturities can be as high as 9.00, meaning a small number of paid dollars at twelve months projects to nine times that amount at ultimate (Friedland, p. 134). A single large claim entering the triangle at an immature accident year can move the projected ultimate by hundreds of thousands of dollars. The leverage is the reason large losses carry outsized tail risk and the reason their frequency is a leading indicator of aggregate development.
A rising frequency of large-loss notices tells you that either claim severity is genuinely increasing (a risk management problem) or your retention threshold is no longer appropriate for the current book (a program design problem). Both predict adverse development, but the response is different.
How to query it. Ask your TPA for a monthly report of large-loss notice count and aggregate reserve change above your threshold, by accident month, for the trailing twenty-four months. Also ask for the current reserve on each open large loss, with a flag for any reserve increase in the last quarter. Your actuary may be capping large claims in the triangle, but if the cap is based on last year’s threshold while this year’s losses are routinely reaching it, the cap is stale and the projection understates the tail.
Concrete action. Set a standing weekly or biweekly large-loss review with the TPA. This is not a reactive call when something blows up. It is a scheduled review of every open claim above threshold: current status, reserve adequacy, litigation posture, settlement strategy. The goal is to catch reserve changes early and feed them to the actuary before they surprise the triangle.
The dashboard you should actually build
Five metrics. Monthly. Trailing twenty-four months. Segmented by line of business.
You do not need a business intelligence platform to do this. A single spreadsheet with five tabs is enough. One tab per indicator, one row per month, one column per line of business, with a simple chart showing the trailing trend. The point is routine review, not sophistication.
Set a calendar reminder to update the dashboard quarterly, before your TPA stewardship meeting and before your actuarial review. The value is not in any single data point. It is in the pattern across quarters. Most adverse development is invisible until somebody actually looks at the operational data that predicts it.
What this does and does not do for IBNR
These five indicators warn you that adverse development is coming. They do not by themselves change the actuary’s estimate, and they do not lower IBNR.
What they do is give you six to eighteen months of runway to act before the development shows up in the triangle. That runway is the difference between managing a problem and discovering one. When a leading indicator signals trouble, you have time to address the operational driver: push back on case reserve philosophy if Indicator 2 is drifting, renegotiate the TPA service standard if caseloads are the root cause, commission an interim actuarial review if multiple indicators are trending adversely at once. Friedland notes that interim monitoring between full studies adds value precisely because conditions can change between annual valuations (Friedland, Section 8).
The IBNR reduction happens through the actions the indicators trigger, not through the indicators themselves. A falling days-to-report metric does not lower IBNR directly. But the return-to-work program or the supervisor training initiative or the TPA staffing adjustment that caused the metric to fall will, over time, produce a triangle with less adverse development. The indicator is the instrument panel. The risk management is the steering wheel.
Future articles in this series will cover specific risk management interventions and how to measure their impact on reserve development.
Three concrete actions for this week
1. Email your TPA and request the five metrics. Ask for average days-to-report, average days-to-first-reserve, attorney representation rate, closure rates at 90/180/365 days, and large-loss notice count and reserve changes above threshold. Request them as a recurring monthly report, by accident month, segmented by line of business, for the trailing twenty-four months. Most TPAs can produce this from their standard claims management system. If your TPA says it cannot produce one or more of these metrics, that itself is diagnostic information about how well instrumented your claims program is.
2. Review the trailing twenty-four months yourself before your next quarterly TPA review meeting. You are looking for any indicator that has changed direction. A metric that was stable for eighteen months and then moved in the last two or three quarters is a signal. Circle it, write down a question, and bring it to the meeting. You do not need to diagnose the cause yourself. You need to ask the question that forces the TPA to diagnose it.
3. If any indicator is trending in the wrong direction by 20 percent or more over the trailing period, raise it formally. Document the trend in writing, ask the TPA for a root-cause explanation, and document their response. If the explanation is “claim volume increased” or “we had adjuster turnover,” ask what the remediation plan is and when you should expect the metric to normalize. This is not adversarial. It is the governance conversation that should happen at every TPA stewardship meeting and almost never does.
For the diagnostic framework that maps these indicators back to the three root causes of adverse development, see the claim-level diagnostic guide. If you want a second set of eyes on your program’s reserve adequacy, the independent review note explains why an independent actuarial assessment adds clarity when leading indicators are flashing.