Program Financial Management: Tranches, Gates, Variances in ₹ Crore

Program financial management is the work of holding one approved budget for a group of related projects, releasing it to those projects in stages, controlling spend against it, and reporting the position to the people who funded it. We call it programme finance for short. The programme owns the envelope. Each project owns only its allocation.

PMI’s The Standard for Program Management (Fourth Edition, 2017) treated this as a program activity. It described it as identifying the program’s financial sources, integrating the component budgets, setting the overall budget and controlling costs. This page covers the control side of that list: tranches, gates, variances and the roll-up.

Four terms recur on this page. The envelope is the whole approved budget. A baseline is a project's approved budget within it. A variance is the gap between a forecast and its baseline, and the delegated limit is the variance size the programme manager may approve alone. The reserve is money held at programme level to cover variances.

The business case gets the envelope approved. What approvers need to see is covered in Program Business Case. This page starts after approval, when money is released against that case.

A split of ownership that we recommend:

Your governance sets the real decision rights. Program Governance Framework shows how to set decision limits for budget variances and who decides at each tier; its example uses 2% and 5% bands on a ₹40 crore baseline, a different model from the ₹1 crore limit used here. The financial rules on this page, such as the gate tests and the delegated limit, belong in your program management plan.

Release the next tranche only when the gate that closes the current one has been passed. A tranche is a block of money covering one stage of the programme. A gate is the decision point where the steering committee reviews evidence and chooses to release the next tranche, hold it, redirect it or stop.

The release test for each tranche is set when the business case is approved; this page covers running it at the gate.

The point is that money follows evidence rather than the plan. A programme that releases everything at approval has no financial lever left when a project drifts.

We run a gate on five tests. Each one has a pass and a fail:

In our illustrative programme the approved envelope is ₹105 crore: ₹90 crore of project baselines, ₹6 crore of programme-level costs and a ₹9 crore reserve held at programme level.

The reserve is 10% of the project baselines, which is our modelling assumption. For scale, the MoSPI Flash Report for July 2026 puts the revised cost of 1,775 ongoing central-sector infrastructure projects costing ₹150 crore or more at ₹37,10,642 crore against an original ₹33,70,138 crore, about 10% higher (our calculation from the report’s totals: 37,10,642 ÷ 33,70,138 ≈ 1.10).

That is a public-infrastructure list, not a corporate programme, so it does not set your reserve. It shows the scale of cost growth that a reserve and a delegated limit exist to absorb.

Tranche 1 of ₹42 crore is released at approval. Tranche 2 of ₹63 crore is released at Gate 1, which reviews tranche 1.

At Gate 1 the programme manager brings spend and forecasts. The forecast at completion is spend to date plus the estimate to finish. Our assumptions are that the programme manager holds the reserve and may approve up to ₹1 crore of reserve draws in total, counted as a running total across the programme, not per variance.

Check the working. Project baselines are 40 + 30 + 20 = 90, and adding programme costs of 6 gives 96. The reserve is 105 − 96 = 9. Forecasts of 38.5 + 33.5 + 20.5 = 92.5, plus 6 for programme costs, give 98.5. That is ₹2.5 crore above the 96 baseline, and 98.5 + 6.5 of unspent reserve = 105.

Absorb a variance at programme level when it is within the delegated limit and the reserve can cover it. Escalate it when it exceeds the limit, or when covering it would change what another project can deliver.

Count the limit as a running total of the programme manager’s reserve draws, not per request: ten draws of ₹0.9 crore would each pass a per-request test and would empty the whole ₹9 crore reserve without a committee decision.

After Project C’s ₹0.5 crore draw, ₹0.5 crore of the limit is left. The ₹1 crore limit in our model is just under 1% of the ₹105 crore envelope. Scale yours to your own envelope and set it in your governance.

Applied to Table 2:

• Project C, +₹0.5 crore. It is within the limit. C has already spent ₹0.5 crore more than its ₹6 crore release, so the reserve draw regularises that overspend. In practice, ask the programme manager to approve before spending past the release. The programme manager approves the draw, records it and reports it at the next steering review. The reserve falls from ₹9 crore to ₹8.5 crore.
• Project B, +₹3.5 crore. It is above the limit, so it is escalated with options: draw on the reserve, reduce scope, or ask for more funding. Suppose the steering committee approves the reserve draw. Because the extra cost falls in tranche 2, B's tranche 2 allocation rises from ₹18 crore to ₹21.5 crore. The reserve portion of tranche 2 falls from ₹6 crore to ₹2.5 crore before A's saving is returned (Table 3). Tranche 2 stays at ₹63 crore.
• Project A, −₹1.5 crore. We treat a forecast saving as available to the programme only when A's project manager confirms it in writing at the gate, because a forecast saving can still erode. Once confirmed, it returns to the reserve, and A's tranche 2 allocation falls to ₹20.5 crore.

The net reserve use is 0.5 + 3.5 − 1.5 = ₹2.5 crore, leaving ₹6.5 crore. Table 3 gives the same answer: tranche 1 reserve after C's draw is 3 − 0.5 = 2.5, and 2.5 + 4.0 = 6.5.

We recommend avoiding three habits:

• Netting silently. If A's saving is quietly set against B's overrun, the reserve looks bigger than it is and B's problem is hidden.
• Giving each project its own reserve. Nobody can see the total, and the money is spent first and explained later. We hold the reserve at programme level because that is where cross-project calls are made.
• Reporting by project. Every project can be green while the envelope is red.

Calculating the variance itself, including cost performance and forecasting, is covered in Earned Value Formula. Putting a ₹ exposure on each risk and adding related risks into one programme figure, the starting point for sizing a reserve, is covered in Program Risk Management.

A sponsor's roll-up should fit on one line carrying four things: forecast against envelope, reserve left, escalations open, and the tranche decision requested. Sponsors decide on the forecast, not on spend to date. At Gate 1 in our example spend is only ₹37.5 crore, while the forecast is ₹98.5 crore.

Example line from the model: Forecast ₹98.5 crore of ₹105 crore approved; ₹6.5 crore reserve unspent; one escalation decided (Project B, ₹3.5 crore); tranche 2 of ₹63 crore recommended for release.

Put project-by-project detail below the line, never instead of it. If the line cannot be written from your own records in a few minutes, the five gate tests are not yet passing.

The Project Management Institute (PMI) tests programme finance across several domains, not in one, in its PgMP® (Program Management Professional) Examination Content Outline (March 2024).

Domain I asks candidates to "define the high-level road map and financial framework to set a baseline for program definition, planning, and execution". It also asks them to estimate the financial framework "to obtain/maintain funding authorization and drive prioritization of projects within the program".

Two tasks map directly to the sheet on this page. Domain II, Task 25 asks candidates to analyse variances and trends in costs, schedule, quality and risks by comparing actual and forecast values with planned values, which is what Table 2 does at Gate 1. Domain V, Task 3 covers obtaining approval through stage gate reviews to proceed to the next phase, which is the tranche release.

Domain II, Task 27 also includes managing programme-level issues, with financial issues named among the examples. Domain V covers finance and reporting in governance standards, and lists financials among the indicators to evaluate (Domain V, Tasks 1 and 4). The outline’s knowledge lists include funding models, funding processes, budget processes and procedures, and financial closure processes.

Reading across them, we infer the pattern: set the financial framework early, hold funding authorisation, report finances to governance, and close financially. The sheet on this page covers the middle two. Financial close-out, such as final invoices and locking cost codes, sits on the close-out list in Program Closure.

PMI says it reviewed and minorly updated the March 2024 outline to align its terminology with the Fifth Edition of the Standard, published in 2024, and notes that the outline and The Standard for Program Management, Fifth Edition are not identical.

We offer a free discovery or eligibility consultation before any enrolment. It is a conversation about your goals, the PgMP® requirements and which programme suits you. It does not enrol you in anything. To ask for it, write to support@augmentconsultancy.com or call or WhatsApp +91 9816 360 116.

We mentor project managers moving into programme seats towards PMI's PgMP® credential. Our founder, Sanjeev Kumar, has more than 20 years of experience, and we have mentored or trained more than 6,000 professionals.

If you are newer to programmes, Program Management Essentials is our foundation course. It is not PgMP® exam preparation. If you are ready for the credential, the PgMP® Self-Paced Mentorship Program gives structured learning, application guidance and exam preparation at your own pace. Both are in our course store. As of November 2026, Program Management Essentials is US$99 and the PgMP® Self-Paced Mentorship Program is US$549 per enrolment; prices vary by market.

Before your next gate, run one test: can you write the sponsor’s one-line roll-up from your own records in a few minutes? If you can, send the gate pack. If you cannot, find the failing test among the five and fix it first, starting with reconciliation, because every later test depends on it.

When a variance lands, ask one question first: is the programme manager’s running total of reserve draws still under the delegated limit? If it is, approve the draw and record it that day. If not, take it to the steering committee with options, not just a request for money.

In the next 48 hours, list your last three variances and mark which ones your limit would have decided alone. If that surprises you, reset the limit before the next overrun. To test your own gate sheet with a mentor, book a free discovery call and bring it.