Earned Value With a One-Time Insurance Claim

Basic indices

• CPI = EV/AC = 5.10/5.70 = 0.895
• SPI = EV/PV = 5.10/6.00 = 0.850
• CV = EV − AC = −$0.60M
• SV = EV − PV = −$0.90M

The PM's EAC. BAC/CPI = 12.0/0.895 = $13.41M. The arithmetic is right; the input is contaminated.

The $0.40M of rework. It is in AC but not in the baseline and not the contractor's fault. Leaving it in CPI computes a performance index that includes a one-time event as if it were a recurring productivity trend. Strip it: adjusted AC = $5.30M, adjusted CPI = 5.10/5.30 = 0.962.

Then decide what happens to the $0.40M.

• If the insurer pays in full, it is a receivable, not a cost overrun: exclude from the forecast entirely and carry it on the balance sheet.
• If recovery is partial or uncertain, carry the unrecovered portion as a discrete forecast item on top of the performance-based estimate - do not let it distort CPI.

Forecast with the adjusted index: EAC = 5.70 + (12.00 − 5.10)/0.962 = $12.87M including the rework already spent, or about $12.47M excluding it, less any insurance recovery.

Choosing an EAC. The candidates:

• BAC/CPI = $13.41M - assumes the cost variance is permanent and repeats. Reasonable default, but it ignores the schedule variance and treats a one-off as recurring.
• Composite: AC + (BAC − EV)/(CPI × SPI) = 5.70 + 6.90/0.760 = $14.77M - assumes cost and schedule variance are both permanent. Appropriate when the schedule slip is driving the cost growth (rework, overtime, extended supervision).
• Adjusted-CPI EAC ≈ $12.5–12.9M - removes the one-time event.

Selection logic, not formula recall. Look at why the schedule and cost variance exist. Here the SPI of 0.85 is partly procurement front-loading and partly the panel event. If the delay is genuinely driving indirect cost growth, the composite EAC is right despite looking punitive. If the schedule variance is a measurement artifact, use the adjusted CPI. Justify it with evidence - the cost ledger by account, not the summary index.

Earned schedule (linear approximation). ES = 6 × (5.10/6.00) = 5.1 months; AT = 6; SPI(t) = 5.1/6 = 0.85; independent EAC(t) = PD/SPI(t) = 12.0/0.85 = 14.1 months, i.e. roughly 2 months late.

Assumption to state: linear PV, which is often wrong on a project with front-loaded procurement - the true ES is found by interpolating the PV curve. Report the interval as well as the point.

Leading indicators

• Budgeted cost of work remaining (BCWR) and its burn rate against the remaining plan, by discipline.
• Float consumption on the critical and near-critical paths - this is the earliest schedule warning.
• Open RFIs, change orders and submittals ageing, plus their committed schedule impact.
• Rework hours as a percentage of total hours - directly relevant here.
• Procurement: vendor schedule performance on the automation hardware, where EVM lags reality by weeks.
• Staffing versus plan, and overtime percentage as an early cost-growth signal.

Corrective actions

1. Resolve the insurance claim on a date certain; carry it as a receivable with a defined owner inside the team.
2. Recover the critical path by re-sequencing commissioning and testing and by adding a second shift at the panel integration stage.
3. Separate owner-caused delay from contractor-caused delay and start the notice trail now, before the milestone dates are missed.
4. Reset the EAC once, with a documented basis, and stop re-forecasting monthly - repeated small revisions destroy the forecast's credibility.
5. Re-baseline only if the client approves a scope or date change; otherwise keep measuring against the contract baseline.