In the full architecture of Earned Value Management metrics, the Estimate at Completion occupies a unique position. Every other primary metric — SPI, CPI, CV, SV — is a backward-looking measure of what has happened relative to what was planned. The EAC is forward-looking: it is a prediction of where the program will end. And because it is a prediction rather than a calculation, it carries more subjectivity, more management pressure, and more potential for manipulation than any other number in the EVMS reporting package.
Understanding the EAC problem means understanding both how EACs are supposed to be constructed and why, in practice, they so frequently tell a more optimistic story than the underlying data supports.
What an EAC Is and How It Should Be Calculated
The Estimate at Completion is the total projected cost for a program element at completion. It represents the sum of actual costs incurred to date and the estimate of what it will cost to finish the remaining work. In formal notation: EAC = AC + ETC, where AC is actual cost and ETC is the Estimate to Complete.
The ETC — the remaining work estimate — is where the judgment and the manipulation both live. There are several recognized methods for developing ETC:
CPI-based EAC. The most statistically defensible formula for EAC is derived from the program's cumulative Cost Performance Index: EAC = BAC / CPI. This formula assumes that the cost efficiency at which work has been performed to date will continue for the remaining scope. Research on completed programs consistently shows that cumulative CPI stabilizes early in a program's life — typically after 15-20% of the budget has been spent — and that it rarely improves significantly from that point. The CPI-based EAC tends to be more accurate than management-developed EACs precisely because it does not allow optimism about future performance to override observed past performance.
Management EAC. Contractors are also permitted to develop EACs based on management judgment — a "bottoms-up" estimate of the cost to complete the remaining scope. This method is appropriate when there is a documented, credible reason to believe that future performance will differ from past performance — a resolved technical problem, completed high-risk scope, or changed execution approach. It is also the method most susceptible to optimism bias and management pressure to report a favorable outcome.
Blended approaches. Many programs use hybrid methods — CPI-based estimates for well-understood scope areas, management estimates for areas with high remaining technical uncertainty, and independent government estimates for high-risk work packages.
Why EACs Get Manipulated
The EAC is the number that most directly communicates program health to senior leadership, the program office, and, on federally reported programs, to congressional oversight. A CPI-based EAC that projects a 20% overrun will generate scrutiny, recovery plan requirements, and potential contract action. A management EAC that projects a 5% overrun, for the same program with the same CPI, will not.
This asymmetry creates a predictable pressure. CAMs and program managers are not generally rewarded for delivering accurate bad-news forecasts early. They are rewarded for delivering programs that come in on budget. The rational response to a CPI that is trending toward an unacceptable EAC is to find reasons why the CPI-based projection overstates the true overrun — technical issues that have been resolved, learning curve effects that will improve future productivity, remaining scope that is structurally different from the work that drove the poor CPI.
Sometimes these justifications are legitimate. Frequently, they are not. The result is EAC optimism: a systematic pattern in which management-developed EACs are more favorable than CPI-based projections, and in which EAC growth occurs late in a program's life rather than early — when the window for effective corrective action has closed.
The Compression Pattern
The most common EAC manipulation pattern is what contract auditors call "EAC compression" — the persistent maintenance of an EAC that is significantly below the CPI-based projection, without credible documentation of why future performance will exceed past performance.
A compressed EAC has a specific visual signature in time-series EVMS data. The cumulative CPI declines over successive reporting periods. The management EAC does not rise commensurately. The variance at completion — the difference between the budget at completion and the EAC — grows more slowly than the CPI deterioration would project. The TCPI (To-Complete Performance Index) — the CPI that must be achieved on remaining work to hit the current EAC — increases beyond the values achieved on completed work, sometimes dramatically.
A TCPI that exceeds 1.1 — meaning the program needs to perform 10% more efficiently on remaining work than on completed work — is a significant surveillance concern. A TCPI above 1.2 is a program that almost certainly will not hit its current EAC. A compressed EAC with a high TCPI is not a forecast. It is a wish.
The New E20 Requirement
EIA-748 Revision E introduced Guideline E20 (Update CA EAC), which explicitly requires that Estimates at Completion be updated at the control account level on a defined schedule, with documented methodology and defensible basis. This guideline directly addresses EAC compression by pushing accountability to the control account manager level and requiring that the methodology behind each EAC be documented and defensible under surveillance review.
Under E20, a control account that has been maintaining a management EAC significantly below the CPI-based projection needs documented, auditable justification for the deviation. The narrative that "the remaining work will be more efficient" is no longer sufficient — the basis for that claim must be documented.
This is a meaningful tightening of the standard. It does not eliminate EAC manipulation, but it creates explicit audit trail requirements that make unsupported EAC optimism more visible and more difficult to sustain across surveillance cycles.
Reading the TCPI Signal
For program office personnel and oversight staff, the TCPI is the most useful single indicator of EAC credibility. It quantifies the implied performance efficiency required on remaining work and allows direct comparison against actual performance on completed work.
A program with a cumulative CPI of 0.85 and a TCPI of 0.87 has a plausible EAC — the required future performance is consistent with past performance. A program with a cumulative CPI of 0.85 and a TCPI of 1.15 has a credibility problem — it is claiming that it will perform 35% more efficiently going forward than it has to date. The burden of proof for that claim is high, and the claim is rarely supported by the evidence.
The EAC is the number that tells you where the program is going. When it is manipulated, every forward-looking management decision made on the basis of that number is made on false information. The consequences arrive late, large, and largely unavoidable.
The Forensic Intelligence Engine calculates independent CPI-based EAC projections, computes TCPI trends, and identifies EAC compression patterns — providing program offices with EAC credibility assessments that are grounded in performance data rather than management optimism.