Upward revision of early public cost estimates when scrutinised at budget commitment, Australian transport projects 2001-2015

What it is. 25% of the first publicly announced cost: the average increase needed once an announcement made before any formal budget commitment was scrutinised in the budget process.

Sample. Australian transport projects completed between 2001 and 2015, from the Investment Monitor database (Grattan Institute 2016). The 25% applies to projects whose first cost was announced before a formal budget commitment, which are 32% of projects and 74% of the value of overruns; after the budget commitment their costs rose by a further 26%. The text does not give the number of projects behind the 25%.

What it is not. An owner's early public estimate, not a contractor's tender price: the client's money, and mostly before any contractor was signed. Grattan describes these as politicians' promised costs made before a budget commitment, and says they "need to be upwardly revised by an average of 25 per cent" once scrutinised: an average of revisions, with no stated sample, not a measured tender shortfall. It covers only announcements made ahead of a budget commitment and is from transport projects of $20m or more, not civil work generally. It is a stated average with no sample size. The report says early announcements are made for electoral reasons, not that the estimates were wrong because contractors priced short. The contractor's side of tender pricing has no national figure here.

Where it came from

• Terrill & Danks (Grattan Institute) (2016), Cost overruns in transport infrastructure
  Chapter 2, p.16 (text beside Figure 2.1) · source record

Considered and set aside

• Victorian Auditor-General's Office: Victoria only. Four VicRoads design-and-construct projects finished 0% to 9% over approved budget (3% on average, Figure 2C, p.12) and one public sector comparator was revised up 45%: a handful of owner-side budgets in one state, so it cannot fill the national cell.
• Ryan & Duffield: Contractors' own open-book data on 28 megaprojects: tendered margin 9% became a 7% loss. It measures the whole swing from tender to completion. The paper blames bidding competitiveness and risk transfer in words but does not split the loss between underpricing at tender and what happened in delivery, so it is not a tender-pricing figure.
• Australian Bureau of Statistics: Gives the industry's operating profit before tax as 6.5% of total income in 2024-25 (8,631 of 132,471), the margin a short tender has to eat. That is a result for the whole industry, not a measure of how short tenders were.

overrun/tender_estimate_revision_early_announcement_au · AU · F42 ·the record ·link to this figure ·dispute this figure(or email support@demiton.io)