Is your mod projection on track? A mid-year check with unit stat data

A mod projection is a forecast, and forecasts age. The projection you ran at renewal — the one the client's budget and bill rates were built on — rested on open-claim reserves as of a valuation date that is now months old. The mid-year question is simple: is the number still true? Here is the four-part check we run, with the math that turns each answer into dollars.

Throughout, we'll use the worked account from our mod-calculation post: expected losses of $607,240 across three years, eleven claims totaling $594,000 incurred, four of them open, projecting to a 0.975 mod (filed 0.97) at current reserves.

1. Re-value the open claims

Open claims enter the mod at paid-plus-reserve, so reserve movement is mod movement. Pull a fresh loss run and compare each open claim to the valuation your projection used. On the worked account:

That's the whole case for the mid-year check in one table: between "reserves hold" and "+50% development" sits seven points of mod — crossing 1.00, which for some contracts and prequalification thresholds is a cliff, not a gradient. If the fresh loss run shows adverse development, you want the reserve conversations (post 3's checklist) started now, while there's calendar left before the valuation date.

2. Price the new claims — against the split point

Claims incurred since renewal will enter a future rating year, but their shape tells you where the mod is heading, and the split point tells you how to read them. Each claim counts toward primary losses only up to the split point ($18,500 in our 2026 NCCI-state modeling; it varies by state and year) — so a cluster of small claims moves a future mod far more than one large claim of the same total. Mid-year is when a frequency problem is still a trend you can interrupt rather than a year you have to explain. Count the year-to-date claims, compare to the same period last year, and if the count is up, the response (and its documentation) starts now.

3. Check payroll against the estimate

Expected losses — the denominator of the mod — are built from payroll. If the business is running 20% ahead of the payroll estimate the projection used, expected losses are understated and the projected mod is pessimistic; if payroll is falling short, the projection is flattering you. For staffing firms this compounds with mix: payroll shifting toward higher-rated classes changes both the premium and the expected-loss build. Mid-year is also when a payroll true-up conversation with the carrier avoids an ugly audit, which is its own post.

4. Confirm the calendar

The unit statistical report for each policy year is filed roughly six months after that year ends, and the reserves as of that valuation are what the bureau rates on. Mark two dates: the valuation date for the policy year about to report, and your next renewal. Everything you want reflected in the next mod — closures, reserve reductions, coding corrections, subrogation credits — must land before the first date. A mid-year check that finds problems with ninety days of runway is a plan; the same findings discovered at renewal are a eulogy.

What "on track" looks like

A projection passes the mid-year check when: open-claim reserves are at or below the projected values, year-to-date claim frequency is flat or down against the prior period, payroll is within ~10% of estimate with a stable class mix, and no fix you're counting on is scheduled after the valuation date. Fail any one, and you re-run the projection with current data and reset the client's expectations mid-year — which is a far easier conversation than the one where the filed mod arrives twenty points above the budget.

ModIQ re-runs the projection continuously from current loss-run data and flags the moment any open claim's development moves the projected mod — with the valuation calendar built in. Run your own numbers.