PMP Cost and Finance Management: Budget, Reserves and Earned Value

Planning and managing project finance means working out how much money the project needs and when, setting aside the right reserves for risk, tracking every dollar against an agreed plan, and acting through governance when the numbers drift. Cost is not a single estimate made once: it is a budget built from estimates, held to a baseline, measured with earned value or burn rate, forecast forward and reported to the people who fund the work.

The PMP Examination Content Outline lists seven enablers for this task, and this lesson follows them:

1. Analyze project financial needs: estimate costs and aggregate them into a cost baseline and a budget.
2. Quantify risk and contingency financial allocations: size the contingency reserve from the risk analysis.
3. Plan spend tracking throughout the project life cycle: decide how cost will be measured, recorded and matched to funding.
4. Plan financial reporting: decide what each audience receives and in what form.
5. Anticipate future finance challenges: inflation, exchange rates, funding cuts and forecasts of the final cost.
6. Monitor financial variations and work with the governance process: analyze variances against thresholds and escalate through the agreed route.
7. Manage financial reserves: draw, release and request contingency and management reserve correctly.

Neighbouring tasks own related material. Identifying risks and planning responses belongs to the risk task in the Business Environment domain; this task only puts money against them. Phase-gate continue-or-stop decisions and the business case sit with the integrated plan task, and contract types with procurement.

Financial needs are analyzed by estimating the cost of the work with a technique that fits what is known, then adjusting those estimates for when the money will be spent. The choice of technique depends on how much detail exists and how accurate the result must be.

The difference between analogous and parametric is the unit. Analogous scales a whole earlier project's total; parametric applies a per-unit rate to a count. When records give a dependable cost per unit and the units are similar, parametric is both quicker than bottom-up and more accurate than analogous.

Two adjustments turn raw estimates into financial needs:

• Escalation: on multi-year work, known or forecast inflation is built into each estimate for the period in which it will be spent. Inflation is foreseeable, so it belongs in the estimates for the period in which each cost falls, not in the reserve for unidentified work.
• Life-cycle cost: when the goal is the lowest total cost to the organization, compare options over their useful life (delivery cost plus operating and support costs), not on delivery cost alone. A cheaper build with high running costs can cost more over five years.

The cost baseline is the approved, time-phased total of the work package cost estimates plus the contingency reserves; the project budget is the cost baseline plus the management reserve. The distinction decides which figure the project manager reports, measures against and asks the organization to authorize.

Two consequences follow. First, BAC is the cost baseline, so it includes contingency but never the management reserve; earned value is calculated against BAC. Second, the total the organization must authorize, including money for problems nobody has identified yet, is the budget, not the baseline. Contingency is already inside the baseline, so adding it again double-counts.

The contingency reserve is sized from the quantified risks in the risk register, most commonly by expected monetary value (EMV) or by a confidence level from Monte Carlo simulation. Identifying the risks and choosing responses happens in the risk task; here the job is to turn that analysis into money.

Expected monetary value

EMV for each risk is probability multiplied by impact. Threats carry a positive cost and opportunities a negative one, so opportunities offset threats. The contingency for the register is the sum. For example, a 15% chance of a $60,000 cost, a 10% chance of an $80,000 cost and a 20% chance of a $15,000 saving give $9,000 + $8,000 − $3,000 = $14,000. Adding up the full impacts, or ignoring the opportunities, overstates the reserve.

Residual risk after a response

When a response is planned, the cost of the response itself is funded work: it goes into the estimates like any other activity. The contingency is then sized on the residual risk, using the probability and impact that remain after the response. Holding contingency for the original exposure as well as paying for the response counts the risk twice.

Monte Carlo confidence levels

A Monte Carlo simulation of the cost model produces a distribution of possible totals, read as confidence levels: the cost the project has a 50% chance of staying within (P50), an 80% chance (P80), and so on. The contingency needed for a chosen confidence level is the cost at that level minus the deterministic (base) estimate. If the base estimate is $2.0 million and the P80 is $2.3 million, an 80% confidence target needs $300,000 of contingency. The sponsor's risk appetite picks the confidence level; the simulation supplies the number.

Contingency reserve pays for identified risks when they occur and is managed by the project manager inside the cost baseline; management reserve pays for unidentified work and is released by management through change control. Most reserve questions come down to whether the event was identified and planned for.

Reserve analysis is the recurring check that compares the reserve remaining with the risk that remains. It is how reserves stay honest: unspent contingency for closed risks goes back, and growing exposure is funded properly before it lands.

A worked BAC example: BAC is $800,000 including $50,000 of contingency. Spending $15,000 of contingency on an identified risk leaves BAC at $800,000. A $25,000 approved release of management reserve for newly discovered work raises BAC to $825,000. Contingency used for one risk is never moved to fund a different, unidentified problem; that is management reserve's job.

Spend tracking is planned in the cost management plan, the component of the project management plan that sets the rules for estimating, budgeting, measuring and controlling cost. It is different from the cost baseline (the approved numbers) and from the basis of estimates (the assumptions behind each estimate).

A cost management plan typically records:

• Units of measure and currency, and the level of precision (rounding) and accuracy range for estimates.
• Control thresholds: the variance, in percent or amount, that triggers action or escalation.
• Rules of performance measurement: the control accounts, and how earned value is credited.
• Reporting formats and frequency, and the organizational procedures that link the project to the finance system.

Earned value crediting rules

How earned value is credited matters as much as the formulas. Subjective percent complete lets optimistic reporting inflate EV, so the numbers look good while work finishes late. For short work packages (shorter than about two reporting periods), a fixed-formula rule such as 0/100 (credit only on completion) or 50/50 (half at start, half at finish) removes opinion. Longer packages use weighted milestones with objective completion criteria. The fix for unreliable EV is an objective crediting rule, because the problem is how progress is claimed, not how the numbers are calculated or reported.

Agreeing thresholds

Thresholds are agreed with the stakeholders who will act on them and written into the plan. When a finance lead and a sponsor want different thresholds, the project manager facilitates an agreement (often tiered: the project manager handles small variances, larger ones go to the sponsor), because thresholds only work when the people who act on them have accepted them.

Tracking actual spend: the S-curve, accruals and commitments

Actual cost (AC) is the cost incurred for work performed in a period, recorded when the work is done, not when the invoice is paid. Planned spend is tracked against it over time, most often on an S-curve.

The S-curve plots cumulative planned cost over time (the time-phased baseline, or planned value) and overlays cumulative actual cost and earned value. It gets its shape from slow spending at the start, heavy spending mid-project and tapering at the end. It is the standard chart for comparing planned and actual cumulative spend, because it shows both the gap and the trend in one view.

If the finance system records cost only on payment, earned value will compare credited work with cost that is missing, and CPI will look better than it is. The project manager includes accruals in AC for earned value even when the ledger lags. Commitments are not actual cost, but ignoring them hides money that is already promised, so they are tracked against funding from the moment they are signed.

Agile projects manage finance through the cost of a stable team over time: the burn rate (cost per iteration or per month) multiplied by the number of iterations gives the cost. Remaining cost is burn rate × iterations remaining, so a team costing $50,000 per iteration with six iterations left needs $300,000 more. Iterations already finished are sunk and do not count toward the remaining cost.

• Fixed budget, flexible scope. When funding is fixed, the burn rate sets how many iterations it buys, and the product owner orders the backlog so the highest-value items fit inside them. The budget is the constraint, so scope is what flexes; the team's sustainable pace is not a lever for fitting more in.
• Incremental funding. When demand or value is uncertain, fund adaptive work in increments, releasing the next tranche after each review of results (similar to a venture-capital model). This limits the money at risk and lets the sponsor stop or pivot.
• Cost per feature or per release connects spending to delivered value; story points are a team planning measure, not a currency or a value measure.

In a hybrid project, track each component the way it is planned: earned value against a cost baseline for the predictive, work-package components, and burn rate with cost per completed feature for the agile teams. Forcing story-point velocity onto hardware, or task-level work packages onto a stable agile team, produces numbers nobody can act on. The integrated report then rolls both up against the approved plan.

Earned value management (EVM) compares the value of work completed with what it cost and what was planned, then forecasts the final cost from that performance. Three inputs drive it: planned value (PV), earned value (EV) and actual cost (AC), all against the budget at completion (BAC).

Choosing the estimate at completion

The right EAC formula depends on why the variance happened:

For example, with BAC $600,000, EV $240,000 and AC $300,000, CPI is 0.8. If the cause persists, EAC = $600,000 ÷ 0.8 = $750,000; if it was a one-time event, EAC = $300,000 + $360,000 = $660,000.

Reading TCPI

A TCPI well above the current CPI means the budget target is unrealistic unless something changes. With BAC $2,000,000, EV $900,000 and AC $1,100,000, the remaining $1,100,000 of work must be done with the remaining $900,000, so TCPI is about 1.22 against a CPI of about 0.82. When no credible way to improve performance exists, the project manager reports the number honestly and takes the decision (more funding, less scope, accept the overrun) to governance, because the budget belongs to the people who approved it. Management reserve is for unidentified work, not for poor performance, and is never the project manager's to release.

Financial reporting is planned for its audience: each recipient gets the measures that answer their question, in their terms, at the agreed frequency. Variances are monitored against the thresholds in the cost management plan and handled through the governance route the plan sets.

Matching the report to the audience

• Executives and finance want the forecast: EAC and VAC ("will we finish within the money?") and a time-phased spend forecast against plan ("when are the big payments?"). Give them the forward-looking view in financial terms; detailed performance measures belong with the people who manage the work.
• One figure for "how far over or under budget will we finish" is VAC. It is the only measure that states the expected difference at completion; the other measures describe performance to date or money still to spend.
• Agile programs are often better served by spend per iteration against approved funding, shown beside value delivered (released features, outcomes) per funding period, than by earned value tables. The question funders ask is whether the money is buying value, so the report pairs spend with outcomes.

Responding to a variance

1. Analyze first. When a control account breaches its threshold, find the cause and update the forecast for that account before acting. Any response chosen before the cause is known is a guess, so analysis comes before correction, escalation or replanning.
2. Act or escalate through the agreed path. Corrective action within the project manager's authority is taken; anything beyond it goes to the sponsor or board as the plan defines. A forecast breach is escalated when it is seen, not held for the next scheduled review.
3. Change the baseline only for approved changes. An approved scope change (a client-requested feature through the change control board) updates the cost baseline. Overruns from rework, poor productivity or estimating errors stay as variances against the baseline; rebaselining them would hide the performance problem.

Underspend is not automatically a saving

Actual cost below planned value can mean efficiency or simply late work. Check earned value: if PV is $500,000, EV $350,000 and AC $340,000, CPI is about 1.03 and SPI is 0.70. The project is slightly under cost for the work it did but $150,000 of work behind, and that money is still needed to finish it. The underspend is mostly delayed work, not a saving.

Funding limit reconciliation means comparing the time-phased spend in the cost baseline with the funding the organization will release in each period, and adjusting the schedule when planned spend exceeds a period's limit. A project can be within its total budget and still be unaffordable in a given year or quarter.

• Fix a timing gap by moving work, not by adding money. If year one plans $3.1 million against a $2.6 million limit, the first move is to reschedule work, ideally activities with enough float to slip into the next funding period without moving the finish date. This changes the schedule, so it goes through change control once the baseline is approved.
• Reserves do not solve funding timing. Management reserve is for unidentified work, not for spending earlier than the funder allows. Cutting scope is also the wrong first answer when total cost already fits.
• Fast-tracking makes it worse: it pulls more work, and more spend, into the constrained period.
• Raise a funding breach before it happens. When a steady burn rate shows a team will exceed a quarterly or annual tranche, present the gap and the options (re-sequence, reduce capacity, ask for more funding) to the product owner and the funding body early. The funding body owns the money, so it decides how the gap is closed, and it can only choose well with warning.

Anticipating future finance challenges

Anticipating finance challenges means identifying the financial conditions that could change the project's cost or funding before they land, quantifying them and agreeing responses in advance. These are known exposures, so they are planned for, not left to management reserve.

• Exchange rates. A future payment in a foreign currency is a quantifiable exposure. Work with the organization's finance or treasury function to measure it and agree a response, such as a contingency allocation for the range or locking the rate. The exposure is known now, so the response is agreed now.
• Inflation and volatile prices. Build expected escalation into the estimates. For volatile inputs such as steel or fuel, track published price indices against the assumptions behind the estimate and set trigger points that start agreed responses before the costs arrive. Waiting for invoices and reporting the result as variance gives no warning.
• Funding cuts and changes in the organization's finances. When a cut is possible but not confirmed, plan now with the product owner or sponsor which scope a reduced budget would still fund, so the project can respond quickly. Planning while the cut is still uncertain turns the warning into a ready response.
• Performance trends. EAC, VAC and TCPI are forward-looking warnings: a TCPI far above CPI or a VAC that keeps growing is a finance challenge to raise now.