2.1 AACE Recommended Practice 18R-97 Estimate Classification System

AACE Recommended Practice 18R-97 Estimate Classification System

In capital project development, cost estimating is not a single, static event. Rather, it is an iterative decision-support process that evolves alongside the project lifecycle. To establish international consistency, eliminate ambiguity among project stakeholders, and provide clear governance for stage-gate authorization, AACE International published Recommended Practice No. 18R-97: Cost Estimate Classification System — As Applied in Engineering, Procurement, and Construction for the Process Industries (complemented by the generic matrix in RP 17R-97).

For Certified Cost Professional (CCP) candidates, mastering RP 18R-97 is vital. The framework establishes clear expectations regarding estimate accuracy, required engineering deliverables, estimating methodologies, and preparation effort across all five defined estimate classes.

---

1. Primary vs. Secondary Classification Characteristics

A fundamental tenet of the AACE Cost Estimate Classification System is the strict distinction between primary and secondary characteristics. On the CCP exam, questions frequently test whether candidates understand what truly governs an estimate's class.

+---------------------------------------------------------------------------------------------------+
|                        AACE ESTIMATE CLASSIFICATION TAXONOMY                                      |
|                                                                                                   |
|   PRIMARY CHARACTERISTIC (The Sole Determinant):                                                  |
|   - Maturity Level of Project Definition Deliverables (Expressed as % of Complete Definition)    |
|                                                                                                   |
|   SECONDARY CHARACTERISTICS (Consequences & Correlates):                                          |
|   - End Usage / Purpose of the Estimate (e.g., Screening vs. Budget vs. Control)                |
|   - Estimating Methodology (Stochastic / Factored vs. Deterministic / Unit Cost)                  |
|   - Expected Accuracy Range (80% Confidence Interval, e.g., -20%/+30%)                           |
|   - Effort & Cost to Prepare (Expressed as an index relative to project scale)                    |
+---------------------------------------------------------------------------------------------------+

The Sole Primary Characteristic: Maturity of Project Definition

• The Degree of Project Definition (expressed as a percentage of total engineering and scope deliverables completed) is the only primary characteristic used to classify an estimate.
• Project definition reflects the physical and technical maturity of engineering deliverables: Block Flow Diagrams (BFDs), Process Flow Diagrams (PFDs), Piping & Instrumentation Diagrams (P&IDs), Plot Plans, Equipment Data Sheets, and Single-Line Electrical Diagrams.
• An estimate cannot be designated as a higher class (e.g., Class 1 or 2) simply because the cost estimator spent thousands of hours detailing line items or calculating prices to the nearest penny. If the underlying engineering definition is only 2% complete, the estimate remains a Class 5 estimate.

Secondary Characteristics

Secondary characteristics are directly dependent upon the primary characteristic:

1. End Usage: The intended business or project decision supported by the estimate (e.g., conceptual screening, alternative evaluation, budget authorization, commercial bid tender, or definitive project control).
2. Estimating Methodology: The calculation approach used by the estimator, progressing from stochastic (capacity factoring, parametric models, analogy) at low definition to deterministic (detailed takeoffs, itemized bills of quantities, vendor firm quotes) at high definition.
3. Expected Accuracy Range: The statistical variation (+/- percentage band) expected between the final project cost and the estimated cost, typically evaluated at an 80% confidence interval ($P_{10}$ to $P_{90}$ probability distribution).
4. Preparation Effort: The relative time, engineering hours, and financial expenditure required to produce the estimate, typically indexed against a Class 5 baseline.

---

2. Comprehensive AACE RP 18R-97 Classification Matrix

The following matrix summarizes the technical parameters of the five estimate classes defined in AACE RP 18R-97 for the process and energy industries:

---

3. Deep Dive into the Five Estimate Classes

+---------------------------------------------------------------------------------------------------+
|                           PROJECT LIFECYCLE STAGE-GATE PROGRESSION                                |
|                                                                                                   |
|   [CONCEPT] --------> [FEASIBILITY] ------> [DEVELOPMENT] ------> [EXECUTION] -----> [COMMISSION]  |
|       |                    |                     |                    |                   |       |
|       v                    v                     v                    v                   v       |
|    CLASS 5              CLASS 4               CLASS 3              CLASS 2             CLASS 1    |
|   (0% - 2%)            (1% - 15%)           (10% - 40%)          (30% - 75%)        (65% - 100%)  |
|  Screening            Study/Select           Sanction/Gate        Bid/Tender          Check/Final |
|  [-50%/+100%]        [-30%/+50%]            [-20%/+30%]          [-15%/+20%]         [-10%/+15%]  |
|                                                                                                   |
|   STOCHASTIC -------------------------------------------------------------> DETERMINISTIC         |
+---------------------------------------------------------------------------------------------------+

Class 5 Estimate (Concept Screening)

• Project Definition: 0% to 2% of total project deliverables.
• Typical Engineering Deliverables: Plant design capacity, product slate, raw material specifications, block flow diagrams (BFDs), and general site location assumptions.
• Methodology: Macro-stochastic techniques, including capacity factoring (six-tenths rule), cost-to-capacity exponential modeling, historical cost per barrel/megawatt benchmarks, and top-down parametric algorithms.
• End Usage: Initial screening to determine whether a project concept merits further capital expenditure, evaluating alternative plant locations, or portfolio prioritization.
• Accuracy Range: -20% to -50% on the low side; +30% to +100% on the high side (often generalized as -50% / +100%).

Class 4 Estimate (Feasibility Study / Concept Selection)

• Project Definition: 1% to 15% of total project deliverables.
• Typical Engineering Deliverables: Preliminary Process Flow Diagrams (PFDs) with mass/energy balances, preliminary equipment lists with rough sizing, preliminary plot plans, and generalized utility requirements.
• Methodology: Equipment-factored estimating techniques (e.g., Lang factors, Hand factors), semi-parametric models, and historical ratio curves applied to total purchased equipment costs.
• End Usage: Supporting feasibility studies, selecting between competing technical designs or technology licenses, and establishing preliminary economic metrics (NPV, IRR) to justify advancing to Front-End Engineering Design (FEED).
• Accuracy Range: -15% to -30% on the low side; +20% to +50% on the high side.

Class 3 Estimate (Budget Authorization / Sanction)

• Project Definition: 10% to 40% of total project deliverables (typically at completion of Front-End Engineering Design / FEED Phase 2).
• Typical Engineering Deliverables: Approved PFDs, preliminary Piping and Instrumentation Diagrams (P&IDs), preliminary electrical single-lines, mature equipment data sheets with budgetary vendor pricing, site civil grading plans, and defined project execution strategy.
• Methodology: Semi-detailed unit cost estimating with forced detail. Major engineered equipment items are individually priced via vendor budgetary quotations, while bulk commodities (piping, electrical, civil, structural) are quantified using standard unit ratios or forced takeoffs.
• End Usage: Serving as the definitive basis for Capital Budget Authorization (Sanction) and establishing the initial Project Control Baseline for cost and schedule management.
• Accuracy Range: -10% to -20% on the low side; +10% to +30% on the high side.

Class 2 Estimate (Control / Bid Tender)

• Project Definition: 30% to 75% of total project deliverables (detailed engineering in progress).
• Typical Engineering Deliverables: Issued for Design (IFD) P&IDs, approved plot plans, structural steel framing drawings, electrical load lists, complete piping isometric sketches, and comprehensive procurement packages.
• Methodology: Detailed deterministic unit cost line items. Bulk materials are quantified via comprehensive quantity takeoffs (QTO) from drawings and 3D models. Direct craft labor hours and wage rates are fully burdened and crew-adjusted.
• End Usage: Establishing the contractor's formal commercial tender bid, setting contractual guaranteed maximum price (GMP) baselines, and providing the control budget for project execution tracking.
• Accuracy Range: -5% to -15% on the low side; +5% to +20% on the high side.

Class 1 Estimate (Check Estimate / Definitive Bid)

• Project Definition: 65% to 100% of total project deliverables (detailed engineering substantially complete).
• Typical Engineering Deliverables: Issued for Construction (IFC) drawings, complete specifications, final vendor-certified equipment drawings, approved subcontract bids, and firm site logistics plans.
• Methodology: Complete deterministic unit cost pricing. 100% of major equipment and materials are backed by firm purchase order commitments or binding vendor bids. Subcontract work packages are priced based on firm contractor bids.
• End Usage: Final definitive check estimate to evaluate contractor lump-sum bids, baseline for commercial negotiations, resolving formal change orders, and dispute resolution/claims support.
• Accuracy Range: -3% to -10% on the low side; +3% to +15% on the high side.

---

4. Estimating Methodologies: Stochastic vs. Deterministic

A central theme in AACE cost engineering is the transition between stochastic and deterministic methodologies as project definition increases.

+---------------------------------------------------------------------------------------------------+
|                        ESTIMATING METHODOLOGY TRANSITION SPECTRUM                                 |
|                                                                                                   |
|   STOCHASTIC (Probabilistic / Top-Down)            DETERMINISTIC (Itemized / Bottom-Up)           |
|   - Capacity Factoring ($C_2 = C_1(Q_2/Q_1)^x$)    - Quantity Takeoff (QTO)                       |
|   - Parametric Cost Estimating Relationships (CERs) - Detailed Unit Price Build-Ups                |
|   - Equipment Factoring (Lang & Hand Factors)      - Firm Vendor & Subcontractor Quotes           |
|   - Historical Analogies & Cost Indices            - Crew Productivity & Wage Rate Burdens        |
|                                                                                                   |
|   [Class 5] ---------> [Class 4] ---------> [Class 3] ---------> [Class 2] ---------> [Class 1]   |
|   100% Stochastic      70% Stochastic       40% Stochastic       10% Stochastic       0% Stoch.   |
|   0% Deterministic     30% Deterministic    60% Deterministic    90% Deterministic    100% Deter. |
+---------------------------------------------------------------------------------------------------+

• Stochastic Estimating: Involves mathematical algorithms, historical ratios, and capacity-based statistical formulas that do not require itemized quantity takeoffs. Inputs are macro-level parameters (e.g., barrels per day, square footage, total equipment weight). Stochastic methods are fast and cost-effective, making them ideal for early-stage screening where scope definition is low.
• Deterministic Estimating: Involves itemized, bottom-up calculations where discrete physical quantities (linear feet of pipe, cubic yards of concrete, tons of steel) are multiplied by specific, fully burdened unit cost rates. Deterministic methods require significant engineering definition and preparation effort but yield the highest precision and detailed baseline granularity.

---

5. Statistical Nature of Expected Accuracy Ranges

In AACE RP 18R-97, the expected accuracy range is not a rigid mathematical guarantee; it represents a statistical 80% confidence interval ($P_{10}$ to $P_{90}$).

$\text{Confidence Interval} = [P_{10}, P_{90}]$

Where:

• $P_{10}$ represents the cost value with a 10% probability that the final project cost will be lower (the low-side accuracy limit).
• $P_{90}$ represents the cost value with a 90% probability that the final project cost will be lower (meaning only a 10% chance the cost will exceed this high-side limit).

Asymmetry of Accuracy Ranges (Log-Normal Distribution)

Notice that the accuracy ranges in RP 18R-97 are highly asymmetrical (e.g., -20% to +30% for Class 3, or -50% to +100% for Class 5). Cost distributions in capital construction are universally skewed to the right (log-normal distribution) because:

1. A project cannot cost less than zero (hard lower bound), but upside cost overruns have no theoretical ceiling.
2. Unidentified scope growth, technical complexity, regulatory changes, and execution delays almost always exert upward cost pressure.

Real-World Cost Engineering Calculation Example

Scenario: An engineering team submits a Class 3 estimate of $50,000,000 for a chemical plant expansion. The project definition is at 25% maturity. Based on AACE RP 18R-97, the historical accuracy range for this facility type is determined to be -15% to +25% at an 80% confidence interval.

Calculation:

• Low-Side Estimate ($P_{10}$):
  $\text{Low Bound} = \$50,000,000 \times (1 - 0.15) = \$42,500,000$
• High-Side Estimate ($P_{90}$):
  $\text{High Bound} = \$50,000,000 \times (1 + 0.25) = \$62,500,000$

Cost Engineering Interpretation: The project sanction committee should understand that there is an 80% probability that the final completed project cost will fall between $42.5M and $62.5M. Management should not establish an unyielding budget baseline at exactly $50.0M without carrying appropriate contingency.