A Class 4 Pipeline Estimate Built on a Simple Average
A midstream operator needs a feasibility-phase (AACE Class 4) estimate for 62 km of new 12-inch gathering line. The estimating team proposes the simple average of three internal analogues: two benign-terrain projects at $1.05M/km and $1.15M/km, and a Rocky Mountain project at $2.20M/km that was 70% rock. Geotechnical borings on the new route indicate rock over roughly 30% of its length. The analogues are in Q3 2022 dollars; the estimate base date is Q1 2025. Management has applied a flat 4% per year escalation for 2.5 years (treated as 10%, simple), added 10% contingency, and produced a single number of $110M that the project director wants to 'put in the authorization package' by Friday.
1. Evaluate the $110M figure. What is wrong with the analogy basis, and what unit rate and base cost would you defend? Show your arithmetic.
Read the worked answer
The average hides the driver. A simple mean of $1.05, $1.15 and $2.20 gives $1.467M/km, but the third point is not a different project so much as a different terrain mix. Averaging it with benign projects smears the rock effect across every kilometre.
Decompose.
- Benign rate = (1.05 + 1.15)/2 = $1.10M/km
- Rocky Mountain project at 70% rock: implied 100%-rock premium = (2.20 − 1.10)/0.70 = $1.571M/km
- New route at 30% rock: 1.10 + 0.30 × 1.571 = $1.571M/km
- Base cost = 62 km × $1.571M = $97.4M (Q3 2022 dollars)
State the assumptions the decomposition rests on. It assumes the rock premium is linear in rock fraction and that the Rocky Mountain project differs from the others only in rock. Normalize for pipe diameter, wall thickness, pressure class, right-of-way width, labour agreement, and any scope the analogue carried (compression, launchers, valves). If any differ materially, the derived premium is contaminated and you fall back to range estimating or a parametric model with terrain as an input.
2. Management's escalation and contingency treatment is wrong in more than one way. Set out each error and what should replace it.
Read the worked answer
Escalation errors
- Simple instead of compounded. 2.5 years at 4% compounded = 1.04^2.5 = 1.103, not 1.10. Small here, material on a 5-year, back-loaded project.
- Applied to the total as a lump. Escalation should be applied to the cost when it is actually spent - labour in the field, major equipment at PO placement, materials at delivery. A flat uplift on the whole estimate over-escalates early spend and under-escalates late spend.
- Escalated to the wrong date. The estimate should escalate to the midpoint of construction, not to authorization. Nothing in the calculation shows which date was used.
Contingency errors
- 10% is not a Class 4 contingency. Accuracy at this stage is typically on the order of −30%/+50%; a flat 10% add-on contradicts the very class the estimate claims.
- Percent-of-base is not risk-based. Contingency should be derived from an identified risk register or a range-estimating/Monte Carlo exercise, so the number carries a confidence level.
- Double counting risk. Escalation is not a contingency item; neither is scope the owner has not decided. Those belong in the base or in a separate owner reserve, not commingled in project contingency.
Bottom line: escalation is a deterministic cost of time and belongs in the base estimate; contingency is an uncertainty allowance and belongs on top, sized to a stated confidence level.
3. What do you actually hand the project director on Friday? Specify the number, the range, and the caveats.
Read the worked answer
The number
- Base, Q3 2022: $97.4M
- Escalated to construction midpoint (1.04^2.5 = 1.103): $107.5M
- Risk-based contingency of ~20% to reach roughly P50: $129M
- Class 4 range around that P50: approximately $90M to $195M
The $110M already in his package sits below the escalated base cost and is not a fundable number.
Caveats that must travel with it
- Basis of estimate: analogues used, the rock-premium derivation, assumed linearity, escalation index and target date.
- Exclusions: right-of-way acquisition, permitting, owner's own costs, landowner settlements, any compression scope.
- Accuracy statement: this is Class 4; the range is the message, not the point estimate.
What would buy him a tighter number Geotechnical coverage along the full alignment rather than spot borings, ~30% engineering, firm line pipe quotations, and a permit schedule from the authority. Each of those moves the estimate toward Class 3 and narrows the range far more than any spreadsheet refinement will.