Cost Performance Measurement and Earned Value Management

Introduction

Cost performance measurement converts project expenditure and progress data into evidence for managerial action. A project may spend less than its budget and still be performing poorly if physical progress is behind plan. Conversely, spending more than planned may be acceptable when progress is ahead and the approved forecast remains viable. Earned Value Management (EVM) integrates scope, schedule, and cost through a common measurement baseline.

EVM is not merely an accounting technique. It is a control system that requires explicit rules for defining work, assigning value, recording actual expenditure, forecasting completion, and escalating deviations. Its effectiveness depends on the quality of the work breakdown structure (WBS), the realism of the baseline, the reliability of progress evidence, and the discipline of change control.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Define planned value, earned value, and actual cost and explain their relationships.
2. Establish defensible measurement rules for physical progress and work packages.
3. Calculate and interpret cost and schedule variances and performance indices.
4. Forecast estimate at completion using alternative assumptions.
5. Integrate schedule and cost analysis, including earned schedule.
6. Diagnose performance problems and recommend corrective action.
7. Evaluate the limitations, governance requirements, and ethical risks of EVM.

1. The Measurement Logic of EVM

EVM compares three cumulative or time-phased quantities at a status date:

• Planned Value (PV): the authorized budget assigned to work scheduled to be completed by the status date. It was historically called Budgeted Cost of Work Scheduled (BCWS).
• Earned Value (EV): the authorized budget assigned to work actually completed by the status date. It was historically called Budgeted Cost of Work Performed (BCWP).
• Actual Cost (AC): the realized cost incurred for the work performed by the status date. It was historically called Actual Cost of Work Performed (ACWP).

PV and EV are measured in budgeted monetary units, not necessarily in cash. AC is measured in actual monetary units. Thus, EVM compares the budgeted value of work with the actual cost of producing it.

A fourth quantity is the Budget at Completion (BAC), the total approved performance-measurement baseline for the project or control account.

Here, PV_i and EV_i are the planned and earned values of individual work packages. The identity applies only to the approved baseline and assumes that all in-scope work is represented once and only once.

1.1 The S-curve perspective

When cumulative PV, EV, and AC are plotted over time, they commonly form S-curves. PV represents the authorized plan; EV represents the value of completed work; and AC represents expenditure. Their separation provides an early warning of cost or schedule problems. However, a visual gap is not itself a diagnosis: the analyst must determine whether the gap results from genuine performance, timing, accounting cut-off, approved change, or data error.

2. Establishing Reliable Measurement Rules

EVM is only as credible as its measurement rules. Before execution, the project team should document the following in the performance measurement baseline and related control-account plans.

Work definition and responsibility

The WBS should decompose scope into control accounts, planning packages, and work packages. Each work package should have a responsible manager, defined outputs, a budget, a schedule, and an objective completion rule. A work package should be sufficiently small to permit meaningful control but not so small that administration overwhelms management value.

Budget allocation and time phasing

The budget is distributed across activities and periods. A time-phased budget establishes PV. The project’s undistributed reserve should not be silently used to conceal overruns; authorized changes must be documented and incorporated through formal baseline control.

Progress measurement methods

Common methods include:

The selected method should reflect the nature of the work. Level-of-effort activities are appropriate for genuinely time-dependent support, but they should not be used for deliverables that can be measured objectively.

Actual-cost rules

AC should be recorded consistently with the period in which work is performed. The project control system should state how accruals, commitments, overhead, labor, materials, subcontractor invoices, and transfers are treated. If AC is delayed while EV is recorded promptly, the apparent cost performance will be misleading.

Status-date discipline

All three measures must refer to the same status date and scope baseline. A data date mismatch can create artificial variances. Corrections, reclassifications, and approved baseline changes should be traceable through an audit trail.

3. Variances and Performance Indices

Cost variance

CV is expressed in budgeted monetary units. A positive CV indicates that the budgeted value of completed work exceeds its actual cost; a negative CV indicates a cost overrun. CV = 0 indicates equality, not necessarily excellent performance.

Schedule variance

SV is also expressed in budgeted monetary units. A positive SV indicates that more budgeted work has been completed than was planned by the status date; a negative SV indicates that less work has been completed. Because both EV and PV are value measures, SV is not a direct number of days.

At project completion, SV normally returns to zero because EV and PV both equal BAC, even if the project finished late. This is a major limitation of conventional schedule variance.

Cost Performance Index

CPI is dimensionless. CPI > 1.00 indicates that each unit of actual cost has generated more than one unit of budgeted value; CPI < 1.00 indicates cost inefficiency. CPI should be interpreted with materiality thresholds and trend information rather than as an isolated decimal.

Schedule Performance Index

SPI is dimensionless. SPI > 1.00 indicates progress ahead of the time-phased plan; SPI < 1.00 indicates progress behind plan. SPI measures schedule performance in terms of budgeted work, not elapsed time. It can be distorted by inappropriate time phasing, front-loaded budgets, or work whose budget does not reflect physical difficulty.

Percentage variances

These measures express the variance relative to earned or planned value. They are useful for communication but require care when the denominator is very small or zero.

4. Worked Example: Basic EVM Diagnosis

An infrastructure project has a BAC of $1,200,000. At the end of month 6, the control account data are:

• PV = $540,000
• EV = $480,000
• AC = $600,000

Therefore:

The project has a budgeted cost overrun of $120,000 on completed work.

The project has completed $60,000 less budgeted work than planned.

The interpretation is not simply “the project is 11% late.” Rather, it has achieved approximately 89 cents of planned budgeted progress for each dollar of planned value by the status date, and it has produced 80 cents of budgeted value for each dollar spent. Management should investigate the causes, such as rework, productivity loss, material price escalation, late design information, or incorrect progress measurement.

5. Forecasting Cost at Completion

Forecasting uses current performance information to estimate the final cost. No single formula is universally correct; the selected formula must reflect the causes and persistence of variance.

Estimate at Completion based on original budget and current cost efficiency

If the current CPI is expected to continue for the remaining work:

This assumes that current cost efficiency is a reasonable predictor of future cost efficiency and that the remaining work is comparable to completed work.

EAC using remaining work and a revised estimate

If the remaining work has been re-estimated independently:

Here, ETC is the estimate to complete, expressed in actual monetary units. This approach is preferred when the original assumptions are no longer valid and a bottom-up re-estimate is available.

EAC assuming an atypical past variance

If the past cost variance is considered nonrecurring:

This assumes the remaining work will be completed at its original budget rate. It is appropriate only when the cause of the past variance has been isolated and will not recur.

EAC considering both cost and schedule efficiency

A commonly used forecasting form is:

This assumes that both cost and schedule performance will influence the cost of remaining work. It should not be applied mechanically. Schedule delay affects cost only when it creates mechanisms such as overtime, escalation, extended site overhead, inefficiency, or contractual exposure.

Variance at completion and to-complete performance

VAC is the expected budget surplus or deficit at completion.

The To-Complete Performance Index (TCPI) indicates the efficiency required on remaining work.

This version measures the efficiency required to meet the original BAC. If the approved target is the current EAC, then:

The denominator must be positive for meaningful interpretation. A high TCPI may reveal that the original target has become unrealistic, particularly when it is materially higher than the project’s established historical CPI.

Forecast example

Using the preceding example:

Thus, if the CPI persists, the projected overrun is:

The required efficiency to meet BAC is:

The project would need to earn $1.20 of budgeted value for every remaining dollar of actual cost, compared with a current CPI of 0.80. This difference should trigger a realistic reforecast and recovery analysis rather than optimistic reporting.

6. Integrating Schedule and Cost Analysis

EVM integrates schedule and cost, but it does not replace the schedule network. A negative SV may result from a non-critical activity, while a small variance on a critical path may threaten the completion date. Consequently, EVM should be reconciled with:

• critical-path analysis;
• remaining-duration and milestone forecasts;
• resource loading and productivity data;
• risk and issue registers;
• procurement and change status;
• contractual milestones and liquidated-damages exposure.

A useful diagnostic sequence is:

1. Confirm the status date, scope, baseline, and accounting cut-off.
2. Validate EV using objective physical evidence.
3. Identify the control accounts causing the largest variances.
4. Trace variances to root causes rather than symptoms.
5. Determine whether causes are one-time, systemic, or external.
6. Assess impact on critical path, quality, safety, and risk.
7. Develop corrective actions with owners and due dates.
8. Reforecast and monitor whether the action changes the trend.

A cost variance may be favorable because an invoice has not arrived, not because work is efficient. Similarly, a negative schedule variance may reflect delayed recognition of a completed milestone. Data validation precedes interpretation.

7. Earned Schedule

Because conventional SV returns to zero at completion, researchers and practitioners developed earned schedule (ES) concepts to express schedule performance in time units. Let the cumulative EV at the status date be located between two points on the cumulative PV curve. If EV lies between the planned cumulative values at periods t and t + 1, then:

where PV_t is the cumulative planned value at period t and PV_{t+1} is the cumulative planned value at the next period. The interpolation assumes that the PV curve between those points is sufficiently meaningful for estimation.

The earned-schedule variance is:

where AT is actual time elapsed, measured in the same period units. A negative result indicates that earned work corresponds to an earlier planned time than the current status date.

The earned-schedule performance index is:

Unlike conventional SPI, SPI_t can remain informative near project completion because ES does not automatically converge to actual time merely because the project is complete. Earned Schedule remains dependent on the quality and shape of the planned-value curve; it is not a substitute for schedule logic or milestone analysis.

Earned Schedule example

Suppose at the end of month 6, cumulative EV is $480,000. Planned cumulative value at month 5 is $420,000 and at month 6 is $540,000. The fraction of the interval represented by EV is:

Thus:

With AT = 6 months:

The earned work corresponds to approximately 5.5 months of planned progress, so the project is about half a month behind the time-phased plan under the interpolation assumption.

8. Illustrative Case: The Northbank Water Treatment Upgrade

Background. The Northbank municipal utility approved a 20-month treatment-plant upgrade with a BAC of $24 million. The project included civil works, process equipment, electrical integration, commissioning, and operator training. The project used monthly EVM at control-account level.

Problem. At month 10, the dashboard showed PV of $11.0 million, EV of $9.8 million, and AC of $12.25 million. Senior management initially focused on the favorable procurement variance in one account and treated the overall result as manageable.

Analysis.

A control-account drill-down showed that the largest unfavorable variance arose in civil works. The immediate causes were incomplete geotechnical information, redesign of a clarifier foundation, and rework caused by an interface error between civil and mechanical packages. AC also included an accrual reversal that temporarily distorted the prior month’s trend.

Approach. The project board required three actions: validate the accrual and restate the trend; freeze nonessential design changes until the interface review was complete; and produce a bottom-up ETC for civil works. The team also resequenced non-dependent electrical work, established a design-interface register, and introduced milestone evidence for foundation completion.

Outcome. The independent ETC produced an EAC of $27.1 million, higher than the mechanical CPI forecast of $30.0 million because the corrective plan removed several nonrecurring causes. The revised schedule forecast showed a two-month delay, but resequencing reduced the expected operational impact to five weeks. The board approved additional contingency, not a hidden baseline increase, and required monthly review of the recovery actions.

Learning points. First, aggregate indices can conceal a concentrated root cause. Second, favorable variances require the same scrutiny as unfavorable variances. Third, a bottom-up ETC may be more credible than a formula when performance drivers have changed. Fourth, baseline changes and contingency use must remain visible to preserve governance credibility.

9. Governance, Ethics, and Professional Judgment

EVM creates incentives that can be misused. Managers may select generous progress rules, split activities to earn value prematurely, defer costs, suppress bad news, or rebaseline after unfavorable performance. Such practices undermine decision quality and may constitute misrepresentation to sponsors, lenders, regulators, or customers.

Professional governance requires:

• documented and consistently applied EV techniques;
• segregation, where practical, between work execution and performance verification;
• traceable actual-cost and accrual records;
• formal authorization for baseline changes;
• transparent treatment of management reserve and contingency;
• retention of historical baselines and variance explanations;
• escalation thresholds defined before results are known.

EVM data may also contain commercially sensitive labor rates, supplier information, or personal performance information. Access should follow least-privilege principles, and dashboards should avoid presenting individual productivity data without legitimate purpose and appropriate safeguards.

10. Technology, Automation, and AI

Modern project-control platforms integrate scheduling, enterprise-resource-planning, procurement, timesheet, building-information-model, and field-reporting data. Automation can improve data timeliness, but it cannot correct an invalid baseline or subjective EV rule. Interfaces should be tested for duplicate costs, timing differences, unit conversions, and scope misclassification.

AI can assist with anomaly detection, variance narrative drafting, likely cost-to-complete estimation, and pattern recognition across control accounts. Its outputs should be treated as decision support. A responsible implementation requires explainable input data, version control, human review, protection of confidential information, and testing for systematic bias. An AI-generated forecast must never be used to conceal uncertainty or bypass approval authority.

11. Limitations and Best Practice

EVM is less effective when scope is unstable, outputs are intangible, work is highly exploratory, budgets are not time-phased, or actual costs are recorded substantially late. Indices can also be distorted by front-loaded or back-loaded budgets, level-of-effort work, large procurement deposits, accounting policies, and approved changes not yet incorporated.

Best practice is to combine EVM with milestone achievement, critical-path schedule analysis, quality indicators, risk exposure, cash-flow data, and customer acceptance. A mature control system distinguishes:

• performance variance from baseline error;
• cost variance from price variance;
• schedule variance from calendar or reporting delay;
• one-time events from recurring trends;
• forecast uncertainty from management optimism.

Thresholds should be proportional to project risk and materiality. Trend charts, control-account drill-downs, and variance-at-completion narratives are usually more informative than a single project-level CPI.

Chapter Summary

Earned Value Management integrates scope, schedule, and cost through Planned Value, Earned Value, and Actual Cost. PV represents authorized budget for scheduled work, EV represents authorized budget for completed work, and AC represents actual expenditure. CV and CPI diagnose cost performance; SV and SPI diagnose progress against the time-phased plan. EAC, VAC, and TCPI support forecasting and management decisions, but every forecast depends on explicit assumptions about future performance.

Reliable EVM requires a sound WBS, objective progress rules, consistent accounting, synchronized status dates, and formal change control. Conventional schedule variance is expressed in budgeted value and returns to zero at completion, so earned schedule can provide a more meaningful time-based perspective. EVM should be integrated with critical-path analysis, risk, quality, procurement, and milestone evidence. Ethical governance is essential because distorted progress rules or hidden baseline changes can make accurate arithmetic produce misleading conclusions.

Key Terms

Planned Value: Authorized budget for work scheduled by a status date.

Earned Value: Authorized budget assigned to work actually completed.

Actual Cost: Realized cost incurred for completed work.

Budget at Completion: Total approved budget for the performance-measurement baseline.

Cost Variance: Earned value minus actual cost.

Schedule Variance: Earned value minus planned value.

Cost Performance Index: Earned value divided by actual cost.

Schedule Performance Index: Earned value divided by planned value.

Estimate at Completion: Forecast total cost at project completion.

Estimate to Complete: Forecast cost required to finish remaining work.

Variance at Completion: Budget at completion minus estimate at completion.

TCPI: Efficiency required to achieve a specified financial target.

Earned Schedule: Time position on the planned-value curve corresponding to earned value.

Control Account: Management-control point where scope, budget, schedule, and responsibility meet.

Performance Measurement Baseline: Time-phased, authorized scope, schedule, and budget baseline.

Level of Effort: Work whose value is measured primarily through passage of time.

Accrual: Recorded cost for work performed or obligation incurred before an invoice is received.

Rebaseline: Formal approved revision of the performance baseline.

Management Reserve: Budget held for known-unknown risks and controlled outside the performance baseline.

Review Questions

Short Answer Questions

1. Distinguish Planned Value, Earned Value, and Actual Cost.
2. Why can a project with a favorable cost variance still be in difficulty?
3. What does a CPI of 0.85 indicate?
4. Why does conventional schedule variance return to zero at project completion?
5. State two conditions required for credible EVM reporting.

Long Answer Questions

1. Explain how a project team should design progress-measurement rules for different types of work packages.
2. Compare the major EAC forecasting formulas and state when each is appropriate.
3. Explain how EVM should be integrated with critical-path and risk analysis.
4. Discuss the governance and ethical risks associated with manipulating earned value.
5. Evaluate the advantages and limitations of earned schedule compared with conventional SPI.

Analytical / Application Questions

1. A project has BAC = $2,000,000, PV = $900,000, EV = $800,000, and AC = $1,000,000. Calculate CV, SV, CPI, SPI, EAC using BAC ÷ CPI, and TCPI to meet BAC. Interpret the results.
2. A project reports a favorable CPI immediately after a major supplier invoice is delayed. What data checks should the project controller perform before reporting improved performance?
3. A control account has recurring design rework, a negative CPI, and a critical-path milestone at risk. Develop an integrated corrective-action plan and explain which EVM measures should be monitored monthly.

Multiple Choice Questions with Answers

Question 1

Which quantity represents the budgeted value of work actually completed by the status date?

A. Planned Value

B. Earned Value

C. Actual Cost

D. Budget at Completion

Correct Answer: B. Earned Value

Question 2

A project has EV = $360,000 and AC = $400,000. What is its CPI?

A. 0.90

B. 1.11

C. $40,000

D. −$40,000

Correct Answer: A. 0.90

Question 3

If EV is less than PV, which conclusion is generally supported?

A. The project is behind its time-phased progress plan

B. The project is under budget

C. The project has completed early

D. The project has no schedule variance

Correct Answer: A. The project is behind its time-phased progress plan

Question 4

Which formula forecasts EAC when current cost efficiency is expected to continue?

A. EAC = AC + ETC

B. EAC = BAC ÷ CPI

C. EAC = BAC − EV

D. EAC = EV ÷ AC

Correct Answer: B. EAC = BAC ÷ CPI

Question 5

A CPI of 1.00 means that:

A. EV equals PV

B. AC equals BAC

C. EV equals AC

D. SV equals zero

Correct Answer: C. EV equals AC

Question 6

Why is conventional SV limited near project completion?

A. AC becomes zero

B. PV and EV both approach BAC, causing SV to return to zero

C. CPI cannot be calculated

D. The critical path disappears

Correct Answer: B. PV and EV both approach BAC, causing SV to return to zero

Question 7

Which progress method assigns value when defined milestones are objectively achieved?

A. Weighted milestone

B. Level of effort

C. 50/50

D. Apportioned effort

Correct Answer: A. Weighted milestone

Question 8

Using BAC = $1,000,000 and CPI = 0.80, what is the CPI-based EAC?

A. $800,000

B. $1,000,000

C. $1,250,000

D. $1,800,000

Correct Answer: C. $1,250,000

Question 9

Earned Schedule primarily expresses schedule performance in:

A. Actual monetary cost

B. Budget percentage only

C. Time units

D. Resource headcount

Correct Answer: C. Time units

Question 10

Which practice best protects EVM integrity?

A. Changing the baseline whenever performance is unfavorable

B. Using subjective progress without documentation

C. Deferring costs until invoices arrive

D. Applying documented rules and formal change control

Correct Answer: D. Applying documented rules and formal change control