Construction cost-overrun statistics depend on the project population, baseline and reporting date. This 2026 review uses a clearly dated MoSPI infrastructure snapshot to explain the measures, then identifies the limits of applying it to private real estate or EPC contracts. It does not present older figures as a current nationwide construction average.
Historical data: the quantitative snapshot below is May 2024, reviewed in 2026. It is not a current all-India construction average.
MoSPI’s May 2024 report covers monitored central-sector infrastructure projects costing ₹150 crore or more. It reports 1,817 ongoing projects, including 458 with anticipated cost above original cost and 831 delayed against original completion dates. The reported aggregate cost-overrun amount is ₹5,71,080.76 crore. These are historical snapshot figures.
The count share calculated here is 458 ÷ 1,817, or approximately 25.2%. It is a share of projects in that monitored snapshot, not an average percentage cost increase. The two measures answer different questions and should not be placed on a chart as though they were comparable percentages.
The May 2024 report’s coverage is non-exhaustive and based on agency reporting. This page does not claim that snapshot is the latest available release. Use the current PAIMANA reporting portal for a fresh operational review and verify the relevant report’s definitions before making a time-series comparison.
Read the population and observation date alongside every value.
| Measure | Value | Basis |
|---|---|---|
| Ongoing monitored projects | 1,817 | May 2024 snapshot |
| Projects with cost overrun | 458 | Against original cost |
| Projects delayed | 831 | Against original completion date |
| Cost-overrun count share | 25.2% | Calculated: 458 ÷ 1,817 × 100 |
| Aggregate cost-overrun amount | ₹5,71,080.76 crore | Published snapshot amount |
No comparable nationwide averages for private real estate, infrastructure and EPC contracts are established in this review. EPC is a delivery arrangement, while infrastructure and real estate are broad project categories. A project can be both infrastructure and EPC, so these labels do not form mutually exclusive groups.
Before calculating an average, choose whether it is a simple mean of project percentages or a cost-weighted aggregate. Large projects influence the weighted figure more strongly. Report the count, original cost base and treatment of incomplete or revised records. A median can also help explain a skewed distribution, but it needs project-level data.
Keep original-sanction and latest-approved baselines separate. A revised baseline can reduce the apparent overrun without removing the increase from the original commitment. Both views may be useful, but readers need to know which one is being measured.
A cost increase can reflect quantity growth, design change, procurement rates, productivity, extended duration or revised scope. Separate these causes in a project review. A higher final amount alone does not tell you whether the original estimate was incomplete or the project’s output changed.
Monsoon exposure and labour cost variance should be assessed from the project’s location, work sequence and records. Do not assign a national percentage to either without supporting data. A weather event can affect different activities differently; a labour-rate increase is distinct from reduced output per labour hour.
Use a cause register linked to forecast movements. Record amount, evidence, confidence and owner. Allow more than one cause where appropriate, but avoid counting the same rupee under several categories. Review unresolved attribution rather than forcing every change into a convenient label.
A state ranking requires consistent coverage and treatment of projects spanning more than one state. It also needs adjustment for project mix: a portfolio of large rail projects is not directly comparable with a portfolio of smaller buildings. No state performance ranking is asserted on this page.
For a valid analysis, state whether allocation follows physical location, implementing agency or another rule. Identify multi-state projects separately or apply a documented allocation method. Show missing data and the number of projects behind each result.
Do not compare two states using figures taken from different months or different cost baselines. If the dataset changes because projects are added, completed or removed, explain that change before interpreting movement as improved or deteriorating delivery performance.
Time delay and cost overrun are related concepts but are not identical. A delayed project may remain within a particular approved cost baseline; a cost increase may occur without a reported schedule delay. Analyse the relationship using the same projects and definitions.
For an internal forecast, identify time-dependent costs such as site establishment or equipment hire and assess which period is actually extended. Keep one-off costs separate. Do not multiply a total project budget by a delay percentage and call the result the cost of delay.
Cash effects also differ from cost effects. A delayed customer receipt can create financing pressure without changing measured work value. Show the cash-flow scenario separately and document assumed payment timing. This makes the management decision more specific than a single “delay cost” number.
Software can connect budgets, commitments, actual cost and remaining work so the project team sees changes sooner. It can preserve the source of a forecast movement and assign an action owner. Those capabilities support decisions but do not establish a guaranteed reduction in overruns.
To evaluate impact, define a baseline and comparable projects or periods. Record other changes in scope, staffing and commercial conditions. A before-and-after result from one project is not automatically causal evidence or an industry benchmark.
In a BUILDX evaluation, ask to trace a forecast variance to orders, quantities and approved changes. Test partial invoices and remaining commitments to avoid double counting. Review the quality and timeliness of inputs as well as the reporting screen.
Use clear definitions so another reviewer can follow the decision.
Original sanction or latest approval, stated explicitly.
The projects and reporting month included.
Count share, mean, median or weighted cost increase.
Projects added, removed or reclassified between periods.
Illustrative example; not a customer result or statutory calculation.
For a project with original cost ₹100 crore and anticipated completion cost ₹120 crore, the increase is ₹20 crore and the overrun is 20%. If another project is ₹10 crore rising to ₹15 crore, its overrun is 50%. The simple mean is 35%, while the combined cost-weighted increase is ₹25 ÷ ₹110 ≈ 22.7%. Both are valid calculations, but they answer different questions.
Bring a representative project record and one exception to discuss the workflow with the BUILDX team.
Use these checks to prepare a review with the responsible team.
Practical answers to common questions about this topic.
Sources checked on 1 October 2026. Apply current requirements to the relevant transaction or project.
Discuss your project, current records and review responsibilities. Agree a focused demonstration and confirm the scope your team needs.