Contingency reserve vs management reserve: where the line runs

On a contract programme, the money that is not in the work packages still exists. It sits somewhere, someone holds it, and what decides the outcome is not how much of it there is but who may commit it, and what has to be redone once it is committed. A practitioner who spends ten minutes looking the terms up will meet two contradictory definitions of the same word and conclude that one of them must be wrong. Neither is.

The three pockets in a project budget

A project budget is not an amount. It is a container with compartments, and the line between them has a name.

The budgeted work. Work packages, costed, spread over time, each with an owner. This is the only pocket in which every unit of money has an address: budgets are assigned to control accounts, and those accounts sum to the budget at completion of the measured work. The PMI Practice Standard for Earned Value Management, 2nd edition (2011), and guideline 8 of the NDIA EIA-748-D Intent Guide, Revision D (2018), describe that arrangement from two traditions.

The contingency reserve. Inside the baseline, not yet attached to any package, held against risks that have been identified. The PMBOK Guide, 6th edition (2017), defines it as time or money allocated in the schedule or cost baseline for known risks with active response strategies, and the 7th edition (2021) repeats the wording.

One point belongs here rather than later, because the rest reads wrongly without it: a contingency reserve is made to be spent, not kept (AACE International Recommended Practice No. 10S-90, Cost Engineering Terminology, revision of 11 June 2026).

The management reserve. Outside the baseline, held against work that was not foreseen at all and that still falls inside the scope of the project. The same PMBOK glossary calls it an amount of the project budget or project schedule held outside of the performance measurement baseline for management control purposes.

The cost baseline, which is that line. That glossary defines it as the approved version of the time-phased project budget, excluding any management reserves, which can be changed only through formal change control procedures. The wording has not moved since the 5th edition (2013), and it is the one technical term this article imports.

It carries a second name, which is what allows one article to speak about two worlds at once: for projects that use earned value management, the PMBOK Guide, 6th edition, states that the cost baseline is referred to as the performance measurement baseline.

One equality follows, and three independent families of documents state it in the same terms: the project budget is the cost baseline plus the management reserve. The PMBOK Guide, 6th edition, builds it that way. ISO 21508:2018, an edition withdrawn in 2026, requires the budget at completion of the performance measurement baseline plus the management reserve to equal the project budget. The NDIA, in the 2022 revision of its master definitions list, sums the same two elements into the total budget allocated to a contract.

What each pocket covers, and what neither one covers

The criterion most referentials use is whether the risk was identified. The PMBOK Guide, 1996 edition, is where the pair of terms that carried it is first attested, and that same edition already warned that the specific meaning of such modified terms often varies by application area. What a risk, an uncertainty and an issue are as events is the subject of a separate article, risk, issue, uncertainty, contingency; this one deals in pockets.

More useful is what falls in neither pocket, and three sources from three different worlds converge. The AACE terminology dictionary excludes major scope changes, extraordinary events such as major strikes and natural disasters, and escalation and currency effects. Skills & Knowledge of Cost Engineering (AACE International, 5th revised edition, 2007) lists the same families, and the Joint Cost Schedule Risk and Uncertainty Handbook (Naval Center for Cost Analysis, 2013) holds that natural disasters, mission-changing events and world events should generally be excluded from explicit uncertainty modelling.

The three do not draw the line in the same place, and the difference is worth knowing rather than smoothing over. The AACE dictionary excludes escalation outright, while chapter 9 of Skills & Knowledge excludes only inflation and currency movement that is excessive and unexpected, and treats ordinary price variation as covered. That is a limit of the split, not a further rule.

What a practitioner can act on tomorrow follows directly: new scope goes through a contract amendment, never through a reserve. The NDIA states it without softening, in guideline 14 of the EIA-748-D Intent Guide: "The budget being held in reserve must not be viewed by a customer as a source for added work scope." The wording is not a recent precaution. The same prohibition appears in DEF (AUST) 5657, the Australian cost schedule control systems criteria implementation guide of March 1994.

Sizing follows a principle, and the calculation belongs elsewhere. A contingency is the distance between the base estimate and a chosen level of confidence, which the AACE terminology dictionary and the Joint Cost Schedule Risk and Uncertainty Handbook define in the same terms. That level is a company decision rather than a result: the handbook notes that it may be mandated, or may simply be organisational practice. Contingencies also fail to add up, and the argument is an old one. Kerridge and Vervalin, in Engineering and Construction Project Management (1986), put it plainly: things going wrong on every account at once is unlikely, so the overall contingency need not be the sum of the largest individual ones, and it is held for the project as a whole. Against that stands the pattern the profession names and rejects, the remainder method, which sets contingency as whatever separates the base estimate from a budget decided in advance. AACE International Recommended Practice No. 40R-08, Contingency Estimating, General Principles (revision of 25 June 2008) classes it as never appropriate rather than as debatable.

Not one of the referentials read for this article puts a number on the contingency of a project. The methods that turn the principle into a figure are the subject of a separate article.

Who holds each pocket, who may release it, and what the release leaves behind

The contingency reserve is committed by the project manager. The PMI Practice Standard for Earned Value Management, 2nd edition, states it for its own scope, and flags on the same page that in some environments the contingency reserve is instead treated as a funding reserve owned by the customer.

The management reserve calls for an authorisation above the project manager, and the sources say so in a conditional voice worth keeping: the PMBOK Guide, 4th edition (2008), writes that the project manager may be required to obtain approval before obligating or spending it, and the PMI practice standard that its use requires special authorisation from management. The APM Body of Knowledge, 8th edition (2025), gives the most formalised chain, in three questions: was the risk identified, is the contingency sufficient if it was not, and, failing that, a change is requested. It also puts the matter plainly: contingency is not hidden money set aside to solve problems by magic, it is a management tool subject to governance.

Who exactly holds the management reserve, however, the referentials do not agree on. The DoD Earned Value Management System Interpretation Guide (2019) writes in guideline 14 that it belongs to the contractor's programme manager and not to the government. The PMI and the APM place it with management, the sponsor or a governance board. The PMBOK Guide, 7th edition, declines to settle, leaving it to the project, the sponsor, the product owner or the project management office depending on the organisation.

We read that disagreement as a question of contractual tiers, and this is our reading rather than any referential's. What is sourced is narrower: the GAO explains the difference by the reach of the contract, funds outside a contractor's control on one side and budget tied to the contract's scope on the other, and the Department of Energy publishes a one-page reference on which both tiers appear, the reserve held by the contractor and the contingency held by the department. What we add is that a project with a single contractual tier has one pocket and one holder, while a project under contract has one on each side of the relationship, so the referentials are describing different storeys of the same building. The support is the American federal case; the generalisation is ours.

Committing contingency changes nothing in the baseline: the amount was already inside it, and it moves down to a work package.

Then the movement itself. When an amount of management reserve is used to fund unforeseen work, the PMBOK Guide, 6th edition, has that amount added to the cost baseline, which requires an approved change to the baseline. The GAO describes the same crossing from the other end: the portion allocated becomes part of the performance measurement baseline, and is distributed to the appropriate control account. The rule is as old as the vocabulary, the 1996 edition of the PMBOK Guide having already required it.

The movement runs forwards only. The NDIA, in guideline 29, holds that management reserve may not be applied to completed work packages, and the PMI practice standard adds that budget is added to future baseline periods, adjusting past ones being impermissible since it could be read as revising a baseline in order to mask variances.

What makes a baseline reconstructible afterwards is not the reserve, it is the log. The NDIA expects month-end values, the origin of each amount and its application to control accounts, and a current balance, and its Guide to the Integrated Baseline Review, Revision 3 (2019), expects that record to identify changes to budgets and current balances at review time.

Three prohibitions carry the section, and they are what a practitioner remembers. A reserve does not fund additional scope, which the NDIA and the DoD both state. It does not offset a variance already incurred: the GAO forbids using it to offset or minimise existing cost variances, and the PMI practice standard warns against masking performance-related overruns. And it is not replenished out of work already done: the budget of a completed package is earned even when the actual cost came in lower, and only the budget of future work can be pulled back. The corollary was written in 1979, in DoDI 7000.10: there is no such thing as negative management reserve.

One thing no text fixes is the internal authorisation procedure. The NDIA asks only for a document, a rationale and an approval route, and refers to the contractor's own system description. The procedure is therefore written at home, and what the referentials require is that it exists and that it leaves a record.

Where these pockets come from

The management reserve is younger than project cost control, and it did not come out of PERT. The estimator's contingency existed before it, and it was removed on purpose: Robert W. Miller, in Schedule, Cost, and Profit Control with PERT (1963), records that the extreme value representing the pad, or cost contingency factor, was dropped in the DOD approach to basic PERT/COST, along with the time and cost uncertainty behind it.

In January 1970 the reserve appears, and it is not yet distinct from anything. Richard J. Lorette and Berton J. Roth, writing on the cost/schedule planning control specification, record that management may hold reserves back, and that these are one and the same thing as budget not yet handed down to the reporting structure. Their article does not use the word baseline once.

By 10 June 1977 the two are separate and carry opposite statuses. DoDI 7000.2 names and defines the performance measurement baseline, keeps the budget not yet distributed inside it, and puts the management reserve outside: the baseline equals the total allocated budget less management reserve, and the reserve is not a part of it. Seven years, and a boundary exists.

The founding criterion is six words long and has not been touched in forty-one years. DoDI 7000.2 asks contractors to "Identify management reserves and undistributed budget." The NDIA reprints it word for word in its 2006 intent guide, and again in Revision D of 2018.

The motive matters more than the dates, and it is where this article takes a position. A 1980 Department of Energy implementation guide gives it in one sentence: "In many major acquisition contracts, it may be difficult to foresee and plan all in-scope work." The same document says what the separation protects, asking that this budget be allocated to cost accounts as quickly as practicable "to maintain the integrity of the time-phased performance measurement baseline."

The unforeseen is set apart so that the measurement stays true, not so that a pocket is kept. The corollary is worth stating, because it is what goes wrong in practice: a reserve used to absorb a variance destroys precisely what it was created for.

In short

Three pockets, one line across them, and one question that decides who may cross it.

The move is small and needs no tool. On any programme under way, locate where the line runs: which amounts sit inside the measured budget, and which sit outside it. Then answer two questions in writing. Who signs to release an amount from the reserve. And where that signature is recorded, so that six months later the baseline can be reconstructed from the record rather than from memory.

Stop guessing. See the real impact.

Frequently asked questions

Q.How much contingency should a project hold?

No referential read for this article sets a figure. The five to ten per cent that circulates is credited to the PMBOK Guide and appears in no edition from 1996 to 2021. The GAO gives it as an observation, adding that it may be too little for some programmes and more than others need.

Q.What happens to contingency that is not spent?

It is not a gain. The PMI Practice Standard for Earned Value Management, 2nd edition, has any remainder removed from the measurement baseline at the end. CIRIA's 1996 guidance retires it progressively instead, as residual exposure falls, so that a surplus cannot later cover poor management.

Q.Is management reserve part of the project budget?

Yes, and not part of the baseline it is measured against. The PMBOK Guide, 6th edition, makes the project budget the cost baseline plus the management reserve. Releasing an amount from that reserve moves it across into the baseline, which is why the baseline then has to be changed.

References

• AACE International - Skills & Knowledge of Cost Engineering - 5th revised edition, with new appendices, 2007
• AACE International - AACE International Recommended Practice No. 10S-90 - Cost Engineering Terminology - Rev. 11 June 2026
• AACE International - AACE International Recommended Practice No. 40R-08 - Contingency Estimating - General Principles - Revision of 25 June 2008
• APM - APM Body of Knowledge - 8th edition, 2025
• Arthur E. Kerridge, Charles H. Vervalin - Engineering and Construction Project Management - 1986
• Australian Department of Defence - Australian Cost Schedule Control Systems Criteria Implementation Guide - DEF (AUST) 5657 - Standard, March 1994
• CIRIA - Patrick S. Godfrey - Control of Risk - A Guide to the Systematic Management of Risk from Construction - CIRIA SP125 - 1996
• Department of Energy, Office of the Controller - Cost & Schedule Control Systems Criteria for Contract Performance Measurement - Implementation Guide (DOE/CR-0015) - 1980
• DoD - DoDI 7000.2 - Performance Measurement for Selected Acquisitions - Edition of 10 June 1977
• DoD - DoDI 7000.10 - Contract Cost Performance, Funds Status and Cost/Schedule Status Reports - 1979
• DoD - DoD Earned Value Management System Interpretation Guide (EVMSIG) - 2019
• DOE - DOE EVMS Gold Card - Earned Value Management System Basics - Version of 8 March 2019
• GAO - GAO-20-195G - Cost Estimating and Assessment Guide - Best Practices for Developing and Managing Program Costs - March 2020
• ISO - ISO 21508:2018 - Earned value management in project and programme management - 2018
• Naval Center for Cost Analysis - Joint Cost Schedule Risk and Uncertainty Handbook - 2013
• NDIA - NDIA PMSC ANSI/EIA-748 Earned Value Management Systems Intent Guide - November 2006 Edition - 2006
• NDIA - Guide to the Integrated Baseline Review (IBR) - Revision 3 - 2019
• NDIA - Master Definitions List for IPMD Guides - Revision of 18 April 2022
• NDIA - EIA-748-D Intent Guide - Earned Value Management Systems - Revision D, 2018
• PMI - PMI Practice Standard for Earned Value Management - 2nd edition, 2011
• PMI - A Guide to the Project Management Body of Knowledge (PMBOK Guide) - 1996 Edition - 1996
• PMI - A Guide to the Project Management Body of Knowledge (PMBOK Guide) - 4th Edition - 2008
• PMI - A Guide to the Project Management Body of Knowledge (PMBOK Guide) - 7th Edition - 2021
• PMI - A Guide to the Project Management Body of Knowledge (PMBOK Guide) - 5th Edition - 2013
• PMI - A Guide to the Project Management Body of Knowledge (PMBOK Guide) - 6th Edition - 2017
• Robert W. Miller - Schedule, Cost, and Profit Control with PERT - A Comprehensive Guide for Program Management - 1963
• United States Air Force - Richard J. Lorette, Berton J. Roth - Cost/Schedule Planning Control Specification - 1970