Cost drift: what it costs a construction contractor

What it is. 23% of original contract value, final versus original.

Sample. 16 rail projects by a single contractor, AU$0.5b+, mostly WA, 2011-2014. Scope changes accounted for 99% of the cost increase; lump-sum projects that overran averaged 12.83%.

What it is not. Mostly scope variations the owner paid, not contractor losses. Median 12.3% against a mean of 23%, so a few projects dominate. Recorded as money_at_stake for that reason.

Where it came from

• Love, Zhou, Edwards, Irani & Sing (2017), Off the rails: The cost performance of infrastructure rail projects
  abstract, p.1 and Table 2 of the open-access preprint · source record

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

Open this figure's own page

What it is. 7% of contract value, final profit at completion: the 9% tendered margin plus a further 7% (a 16-point swing).

Sample. 28 completed mega projects over A$500m, Australia, commenced 2000 to 2015, road, rail, water and social infrastructure, PPP and design-and-construct, from Tier 1 contractors' open-book project data.

What it is not. Megaprojects only: the paper excluded projects under A$500m because most of them made money, so this is not the loss on a typical civil job. The sample is selected for risk (PPP and D&C, alliances excluded). Grattan 2021, which cited it, doubts it: 'The study does not explain why shareholders would tolerate such a pattern of losses.'

Our arithmetic. Table 1, Total row: tendered profit A$3,328m at 9%, so contract value about A$37.0b; final profit A$(2,681)m, about (7.2%). The paper states the same result as "its original 9% profit margin plus a further deterioration of 7%" (abstract). Its "net profit (16%)" is final minus tendered, by its own definition on p.12, so it is not the loss as a share of contract value.

Where it came from

• Ryan & Duffield (2017), Contractor Performance on Mega Projects - Avoiding the Pitfalls · free copy
  Table 1, Total row (pp.12-13); definition of net profit (p.12); abstract (p.1) · source record

History

• 2026-09-25 Read at source for the first time (previously second-hand via Grattan 2021). The paper gives the final result as a share of contract value, so the figure moves from an unquantified margin swing to a 7% loss; the grade moves from low (second-hand) to moderate (a sample selected for risk).

overrun/overrun_contractor_margin_swing · AU · F41, F42 ·the record ·link to this figure ·dispute this figure(or email support@demiton.io)

Open this figure's own page

What it is. 13.8% of the state's budgeted capital expenditure for the year, all portfolios.

Sample. The whole Queensland Government capital program, compiled by the Queensland Audit Office from state capital statements. 16.4% in 2022-23; about 15% a year on average since.

What it is not. The state's money, not a contractor's, and a spending variance for the year, not the cost growth of any project: spending can run ahead of budget because work is brought forward as well as because it costs more. The QAO restated the 2022-23 figure between reports (14% in 2023, 16.4% in 2024).

Where it came from

• Queensland Audit Office (2025), Major projects 2024 (Report 9: 2024-25) · free copy
  'Report on a page'; Chapter 2, commentary on Figure 2B, p.5 · source record

overrun/overrun_qld_capital_spend_vs_budget · AU-QLD · F41, F42 ·the record ·link to this figure ·dispute this figure(or email support@demiton.io)

Open this figure's own page

What it is. 24% of the first publicly promised cost, final versus promised.

Sample. 836 Australian transport projects of $20m or more, 2001-2016; $28 billion over in total; 34% of projects went over.

What it is not. Measured from the first promise, mostly before any contractor is signed, and paid by the owner (the taxpayer). The median overrun is 0%: a few projects dominate the mean.

Where it came from

• Terrill & Danks (Grattan Institute) (2016), Cost overruns in transport infrastructure
  PDF pp.9, 29 · source record

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

Open this figure's own page

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)

Open this figure's own page

What it is. None - no Australia-wide source read gives unrecovered input price escalation as a measured share of contract value or of overrun.

Sample. Search terms: input price escalation fixed price contract; unrecovered cost increases percent of project budget; rise and fall clause; construction input price index road bridge; contractors absorbed escalation; Australian Constructors Association Arcadis survey; Infrastructure Australia market capacity input costs. Ten sources opened. A frequency exists: the ACA and Arcadis 2023 survey of builders says all respondents absorbed some unrecoverable 2022 material escalation, and for nearly half it was more than 10% of project budgets (their own estimates, once mitigations were applied), but it is a count of respondents, not a cost. The 2024 survey chart shows absorption in bands. Rates with no unit: ABS road and bridge output prices +6.0% (year to September 2026; the June 2026 quarter alone +4.5%); BITRE's alert +1.7% (year to September 2019).

What it is not. Recorded as none on purpose: no defensible Australia-wide figure was found for unrecovered input price escalation as a share of contract value, not that none exists. The ACA and Arcadis survey is the closest measure of the cause and is people's own estimates across all construction. Its PDFs sit at a new web address from the one in the 2023 release: the original links were refused with 403 and not worked around.

Where it came from

• Australian Constructors Association and Arcadis (2023), 2023 Construction Market Sentiment Survey
  one-page infographic, 'Escalation absorption' chart; ACA release 2 May 2023 key statistics · source record

Considered and set aside

• Australian Constructors Association and Arcadis: Same survey a year on. A chart of escalation absorbed on projects during 2023 has bands (none, under 10%, 10 to 25%, over 25%) with no printed values, no sample size is stated, and it covers all construction. Reading the bar lengths would be an inference, so no figure is taken.
• Australian Constructors Association: Advocacy paper: some contractors saw supply costs rise up to 70% in a year; contractors must hold fixed prices for 12 to 18 months or even two to three years; rise and fall clauses uncommon. No measured cost to a contract. Table 1 has no printed values.
• Corinna Economic Advisory (Saul Eslake), prepared for the Civil Contractors Federation: For the 2026 oil shock: the CCF estimates 80 to 95% of civil contracts by number are fixed price; prices of bitumen, explosives and plastic pipe up 35% to 50% on anecdotal evidence. A share of contracts by count and anecdotes; the diesel scenario is filed under plant_fuel.
• Infrastructure Australia: Industry reports non-labour input price escalation of about 10 to 20% over 12 months (p.7), 51% of businesses had sought a contract variation and some build 10% variations into budgets. A sentiment range for prices, not a share of contract value lost.
• Infrastructure Australia: 68% of surveyed organisations saw non-labour input costs rise and 36% think prices have not peaked (p.37): perceptions of direction. No cost to a contract.
• Grattan Institute (Terrill, Emslie and Moran): Shows the ABS road and bridge producer price index flat 2013-16 then rising, and quotes a minister blaming market forces for a $3 billion overrun. Descriptive; no share of overrun is assigned to input prices.
• Australian Government Department of Infrastructure, Transport, Regional Development, Communications and the Arts: $32.8 billion of known cost pressures across the federal land transport program, with escalation handled ad hoc. The summary does not split by cause, and it is the owner's budget pressure.
• Australian Bureau of Statistics: Road and bridge construction output price index up 6.0% over the year to September 2026 (146.6 to 155.4; our arithmetic). A price clients are charged, a rate with no unit in the schema, not a cost to a fixed-price contractor.
• Bureau of Infrastructure, Transport and Regional Economics (BITRE): BITRE's alert on the ABS road and bridge output price index: up 1.7% over the year to September 2019 against a ten-year average of 2.4% a year. A rate, and an output price.
• Bureau of Infrastructure, Transport and Regional Economics (BITRE): Materials are 40.77% of road construction cost (Table B2, filed as context): how much is exposed, not how much prices rose against what was priced.

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

Open this figure's own page

What it is. None - no Australia-wide source read states labour hours above need, or lost productivity, as a cost or as a share of project or contract cost on a basis that can be defended.

Sample. Search terms: construction labour productivity Australia; overstaffing; hours over budget; labour cost share of cost overrun; productivity gap cost per project; Productivity Commission construction productivity; Infrastructure Australia productivity; ABS multifactor productivity construction. Nine sources opened. Rates exist but are not costs, so the schema has no unit for them: construction multifactor productivity (-0.8 growth in IA 2024; a 2022-23 fall then +2.0% in 2023-24 in IA 2025); gross value added per hour worked in housing construction (not all construction) down 12% since about 1995 against up 49% for the economy (Productivity Commission); output per construction worker down 25.4% since 2013 (RLB). Dollar totals exist only for the whole economy ($56 billion and $62 billion a year of capacity or work, two consultancies' models).

What it is not. Recorded as none on purpose: no defensible Australia-wide figure for the cost of more hours than the work needed was found, not that none exists. The Roads Australia 30% was left out because nothing shows what it measures; it is an unsourced member claim.

Where it came from

• Infrastructure Australia (2025), 2025 Infrastructure Market Capacity Report
  p.57 (summary: construction multifactor productivity rose 2.0% in 2023-24); p.61 (no consistent project-level data on subcontracting and self-performance) · source record

Considered and set aside

• Infrastructure Australia: Reports construction multifactor productivity growth of -0.8 (p.60) and that labour is about two-thirds of land transport construction cost (p.7, no source for it). A productivity growth rate and a cost share: no cost of hours above need.
• Australian Constructors Association: ABS multifactor productivity indexes for construction against other industries, and an Oxford Economics estimate that reaching the economy-wide average would unlock $56 billion of construction capacity a year. An economy-wide counterfactual of capacity, not a cost on any contract.
• Oxford Economics Australia: Estimates stagnant productivity costs Australia $62 billion a year in lost work across all construction. A modelled opportunity cost for the economy; the full report was not read and no share of a project's cost is stated.
• Rider Levett Bucknall Oceania: RLB's own series: since 2013 workers up 26.5%, hours per worker down 2.0%, output per worker down 25.4%; the Productivity Commission's 1.8% fall in construction labour productivity in 2022-23. Productivity rates for all construction, not money.
• Productivity Commission: Dwelling construction completes about half as many homes per hour as in the mid-1990s, and quality-adjusted gross value added per hour is 12% lower. Housing, and a rate, not a cost on civil work.
• Roads Australia: Members say delays and a poor working environment add 30% to total project costs. A lobby group's statement with no sample, method or definition; it mixes delay and site conditions with labour productivity, so it is not filed.
• Australian Constructors Association and Arcadis: Survey of the largest contractors: rising labour costs and stagnant productivity cited as raising costs; skilled trades up about 5% on average. Perceptions of rates of change; no cost of hours above need, no sample size.
• Australian Bureau of Statistics: Labour costs are 21.35% of total expenses for heavy and civil engineering (filed as context). The size of the labour cost base, not hours above what the work needed.
• Bureau of Infrastructure, Transport and Regional Economics (BITRE): Labour is 31.79% of road construction cost (Table B2, filed as context). A cost composition; nothing on productivity.

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

Open this figure's own page

What it is. 1.2% of the annual value of engineering construction work done by the private sector in 2025: the extra diesel cost to civil contractors of a 145 cents a litre rise in terminal gate price (national average, end of March 2026 against 27 February), if sustained for 12 months ($1.53 billion).

Sample. A scenario from a report commissioned by the Civil Contractors Federation, written in April 2026. Diesel use is the 88 million litres a month used by the federation's 1,150 member firms (Chart 10, from a CCF member survey); the price rise is the national average terminal gate price rise of 145.2 cents a litre (88.9%) between the start of the Middle East conflict at the end of February and the end of March 2026; the denominator is the report's own figure for private-sector engineering construction work done in 2025 (ABS), which the report does not state in dollars. The arithmetic matches (88 million x $1.45 = $128 million a month; x 12 = $1.53 billion), but the report's own key numbers also give 1.15 billion litres a year, which is more than 88 million x 12 (1.06 billion); the 1.2% was not recalculated. The CCF Tasmania release (considered below) says members nationally use around one billion litres a year, which fits 88 million x 12; at 1.15 billion litres the cost would be about $1.67 billion, 1.3%. The report does not state the denominator in dollars: $1.53 billion at 1.2% implies about $127.5 billion, which fits ABS engineering construction work done by private-sector contractors (public and private clients, all engineering types) rather than by civil contractors alone or for private clients alone. For scale, by 3 April the wholesale diesel price was still 127.3 cents a litre (78.0%) above pre-war, before the 30 March excise cut had fully passed through; the report notes much civil diesel is off-road and does not get the excise cut. It also reports the CCF's margin estimates (about 5% on $1 million projects, 3 to 5% up to $1 billion, 1% above $1 billion) and that the CCF estimates 80 to 95% of civil contracts by number are fixed price.

What it is not. An advocate's scenario, not a measured overrun and not an observed cost: the price rise is observed, the litres come from a member survey, and the year of sustained prices is an assumption. The report was commissioned by the Civil Contractors Federation, whose own cost calculator (considered above) puts direct fuel at about 5% of a civil project's cost, which at an 88.9% price rise would be about 4.4% of that project, so 1.2% of all private engineering work is not what a civil contractor should expect on its own work. The numerator is the extra cost of diesel to the federation's member firms only, over the whole industry's private work, and it assumes the price rise lasts a full year with nothing recovered through rise and fall, variation or a falling price; the report says some contracts do carry rise and fall but gives no share of the cost that was recovered. It is a share of a year's industry turnover, not of a contract, so it is not offered to be multiplied by a contract value. The cost is the contractor's when a contract is fixed price, but the report does not say how much of it was. It covers diesel, not other plant costs such as standby or hours above plan, and not indirect fuel in bitumen or materials.

Where it came from

• Corinna Economic Advisory (Saul Eslake), prepared for the Civil Contractors Federation (2026), The Impact of the Middle East Fuel Crisis on Australia's Civil Construction Sector
  p.9 (text after Chart 10: 'This annual figure represents 1.2%...'); p.8 ($128 million a month from 88 million litres and 145 cents); p.3 (145.2 c/L, 88.9%) · source record

Considered and set aside

• ABC News (Clare Armstrong): The CCF chief executive tells the reporter that diesel is about 7.75% of any tender and an extra $46 million on a billion-dollar project (4.6%). The article does not show how $46 million was worked out or which price rise it assumes, and it is a lobbyist's statement to the press, so it is not filed. It is a share of a tender (4.6%), not of all engineering work, so it does not sit beside the Eslake 1.2%: the two differ by about four times because the denominators and the diesel counted differ.
• Civil Contractors Federation: Gives 7.75% of contract value as the default 'fuel-impacted' proportion (5% direct fuel, 2.75% indirect) to be multiplied by the change in terminal gate price. That is a guideline cost share the user adjusts, not a measured overrun, and by the review ruling a composition share is not evidence of the cause.
• Civil Contractors Federation Tasmania: Tasmanian chapter's release: fuel up to 30% of project input costs and bitumen sealing cost up 30%. No method or sample, and Tasmania only.
• Australian Taxation Office: The off-road diesel credit fell from 52.6 to 20.6 cents a litre on 1 April 2026 after the excise cut. It is a tax rate that changes what a contractor nets, not an amount priced or an overrun.
• Bureau of Infrastructure, Transport and Regional Economics (BITRE): Fuel is 5.76% of road construction cost (Table B2) and equipment 27.44%: a cost composition, filed as context, not a measure of fuel above what was priced.

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

Open this figure's own page

What it is. None - no Australia-wide source read measures the final measured quantity against the schedule quantity.

Sample. Search terms: quantity variation; quantity growth; bill of quantities variance; schedule of rates final account; remeasurable contract; earthworks quantity variance; unit rate contract quantity overrun; Auditor-General road construction contract variations; Austroads estimated against actual quantities. Eight national or state sources were opened and none reports measured against schedule quantity. Related shares that are not this cause: Grattan found scope changes explain about 11% of the cost overrun on 51 Australian transport projects over $100 million (additions to what is built, owner-side, between announcement and completion); Love and colleagues found change orders at 11% of work value for one contractor. A Queensland Transport and Main Roads variation study (USQ, Purvis 2015) was found but could not be opened (connection refused again at review, 2026-11-02). It covers Queensland Transport and Main Roads only, so it could not fill the national cell. A further search at review for an Austroads, Auditor-General or road agency quantity outturn study found none: Baccarini's contingency study of 48 road projects and a NSW road authority estimating study compare cost against estimate, not measured against schedule quantity.

What it is not. Recorded as none on purpose: no defensible Australia-wide figure for how far measured quantities exceed the schedule on re-measurable contracts was found, not that none exists. The 11% figures above are shares of a different thing and are not offered as proxies.

Where it came from

• Terrill & Danks (Grattan Institute) (2016), Cost overruns in transport infrastructure
  Section 1.3.1, Figure 1.5 (p.13) · source record

Considered and set aside

• Love, Irani, Smith, Regan & Liu: 67 projects from one contractor; change orders were 11% of work value. A change order is an instructed variation of scope, design or error, not a difference between the schedule quantity and the quantity measured, and the paper does not separate re-measure.
• Ryan & Duffield: 28 megaprojects reported by final profit against tendered profit. No quantity measure, and the paper does not split the loss by cause.
• WT Partnership, for the Independent Pricing and Regulatory Tribunal (NSW): Compares two estimates of quantities (consultant against council) for one NSW plan before construction; earthworks differed and the council's estimate was 6% high. It is a pre-construction check, one plan, one state, and the error ran the other way.
• Victorian Auditor-General's Office: Victoria only, four projects, budget overspend of 0% to 9% attributed to land prices and geotechnical conditions. No quantity data.
• Bureau of Infrastructure, Transport and Regional Economics (BITRE): National road-works cost shares and price indexes. They say what money goes on labour, materials and equipment, with no measure of quantities built against quantities priced.
• Infrastructure Australia: National pipeline and capacity analysis. Reports input price sentiment and productivity, with no project quantity outturn.
• Infrastructure Australia: Says there is no consistent project-level data on subcontracting and self-performance, and proposes a workstream to collect core measures such as budgeted against actual cost (p.61). It holds no quantity outturn, and does not itself say budget-against-actual data is absent.
• Australian Bureau of Statistics: Industry income, expense and labour totals in dollars. No physical quantities and no job-level outturn.

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

Open this figure's own page

What it is. 28.3% of the original $6.8 billion transport allocation announced in 2020: the additional $1.926 billion Cabinet approved on 31 May 2021 to fund agreed projects at their re-baselined cost estimates, plus contingency, after some projects were rescoped.

Sample. The Auditor-General's audit of the 2020 New Zealand Upgrade Programme. 21 of the 26 announced transport projects could not be delivered within the funding allocated (5.32), a frequency, not a cost. The $1.926 billion also funded rescoped projects and a contingency fund (5.12, 5.33). The report notes that confidence in the baseline information was much lower than expected (5.30).

What it is not. An estimate against a later estimate, not against outturn, and the owner's money, not a contractor's loss. No contractor's tender price is involved: both figures are the Crown's, the first set at announcement. The addition includes a contingency fund, which makes the ratio larger than the shortfall in the original estimates, and it is net of the projects that were rescoped, which makes it smaller than the cost of delivering the original scope; the report gives the size of neither, so the direction of the net error is unknown. One programme, announced before business cases were complete, so it is not a typical New Zealand tender.

Our arithmetic. $1.926 billion (paragraph 5.33) divided by $6.8 billion (paragraph 5.32) = 0.2832, rounded to 0.283. The report's own rounded figure for the addition is $1.9 billion (5.12).

Where it came from

• Office of the Auditor-General New Zealand (2023), Making infrastructure investment decisions quickly · free copy
  Part 5, paragraphs 5.12 (p.51) and 5.29 to 5.35 (p.53) · source record

Considered and set aside

• New Zealand Infrastructure Commission / Te Waihanga: One project, and the executive summary gives no original-versus-final cost ratio. It finds the Affordability Threshold was set too low, which is the qualitative point, not a figure for how short the estimate was.
• New Zealand Infrastructure Commission / Te Waihanga: National Infrastructure Pipeline Figure 33 gives only how often reported cost went up after financial decision (51% of projects over $100 million, 28% of the rest), not by how much, so it is a frequency and not a cost.
• Atapattu, Domingo & Sutrisna: Estimators' ranking of causes of overrun, with inadequate tender documentation among the top three; no share of cost, and read from the abstract only.
• New Zealand Infrastructure Commission / Te Waihanga: States that analysis of cost overruns is out of its scope; it compares unit costs between countries.

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

Open this figure's own page

What it is. None - New Zealand national sources give how fast construction input prices moved, as index rates, but none gives what price escalation cost as a share of a fixed-price contract or of an overrun, or how much of it the contractor bore.

Sample. Searches: "Waka Kotahi NZ Transport Agency cost escalation infrastructure construction price increases cost of delivering roads report"; "New Zealand infrastructure cost escalation fixed price contracts contractors bear price risk bitumen steel diesel Te Waihanga"; "NZTA price adjustment fluctuations formula bitumen contracts cost fluctuation NZS 3910"; "Stats NZ Capital goods price index civil construction annual percentage increase road and bridge construction"; "Auditor-General New Zealand Upgrade Programme cost escalation construction costs share of increase contractors price risk"; "Te Waihanga infrastructure cost escalation share of project cost increase attributed to price escalation"; "Civil Contractors New Zealand survey members cost increases fuel bitumen contract price adjustment clauses". Index rates held here rather than as figures: Te Waihanga puts heavy and civil output price growth at 2.4% a year from 2009 to 2019 and 7.8% a year from 2020 to the March 2023 quarter; NZTA's construction index (excluding bitumen) rose 13.7% in the year to June 2022. Not opened: the NZ Transport Agency's website, the Ministry of Transport's website (the Cabinet paper on the Upgrade Programme's cost increase) and Statistics New Zealand's release pages (rendered by script) could not be read by automated request, and the Auditor-General's Upgrade Programme page returned 403; none was worked around. A frequency exists in words only: the law firm's observation that most contracts since COVID-19 do not allow escalation, with no count.

What it is not. Recorded as `none` on purpose. A rise in an input price index, however large, is not the cost of that rise on a fixed-price contract: that needs the share of work exposed, the escalation clause, and who paid. No New Zealand source read measures that. The three nearest reads (a quantity surveyor's three-contract repricing, the Agency's modelled projection, and index rates) are respectively regional, modelled and not a cost, and are not filed as a figure.

Where it came from

• New Zealand Infrastructure Commission / Te Waihanga (2023), Why do construction input costs change? The role of global and local factors · free copy
  Table 2 (Key inputs into heavy and civil construction), PDF p.11, and the text above it on p.10 (section 1.4); Table 1 on p.10; Figure 1 on p.12 · source record

Considered and set aside

• New Zealand Infrastructure Commission / Te Waihanga: Says material prices follow world commodity prices (a 1% rise lifts the New Zealand price of steel, diesel and timber by about 0.6%, and quarterly moves above 2% occurred in almost half of the 53 quarters from 2010 to 2022). Price behaviour, with no share of any contract exposed and nobody's loss.
• NZ Transport Agency Waka Kotahi: The Transport Agency's published cost indexes and bitumen series, from Statistics New Zealand inputs: the construction index (excluding bitumen) rose 13.7% in the year to June 2022, and the bitumen series stood at $1.1197 a litre in December 2022, $0.4635 a litre (about 71%) above a year earlier. Index movements for contracts that use the adjustment, with no count of contracts that do and no money paid.
• Infometrics (Brad Olsen): Reads Statistics New Zealand indexes: Capital Goods Price Index up 12.5% in the year to June 2022, transport ways up 19.1% in the year to the end of 2022, civil infrastructure about 26% above the consultancy's own trend. Economy-wide price levels quoted second-hand from the statistics, not the cost of escalation on a contract and not who paid it.
• RNZ (Phil Pennington), republished by Scoop: A news report of Transport Agency briefings that contractors on young projects carry fuel index changes for the first 12 months, and of an Agency model of a 5% to 10% rise in big project costs with oil at US$100 to US$150. A press account of documents not opened, with a modelled owner-side projection.
• Wynn Williams (law firm): States that since COVID-19 most New Zealand contracts do not allow for cost escalation, and that on a fixed price without fluctuations the contractor carries price movement. This is the mechanism, in words, with no count of contracts and no amount.
• Society of Construction Contract Practitioners: Guidance for reading fluctuation clauses under NZS 3910 contracts and for agreeing baseline indices; it holds no figure.
• Bond Construction Management Ltd for Wellington Water Ltd: A quantity surveyor's repricing of three Wellington three-waters contracts found an overall 5-6% rise since September 2021, with bitumen products 8-10% and steel 4-15% up over the quarter. The closest read to a share of a contract, but three specimen contracts in one region over three months, one consultant's estimate, and it says nothing of who bore it.
• GHD for Tauranga City Council: A forecast of roading cost indices for one council, with weights (labour 40%, plant 25%, bitumen 21%, materials 14%). Forecasts and index weights, not realised escalation on a contract.
• Massey University (School of Built Environment), Atapattu: The full thesis was read. Its survey of 92 estimators ranks "price fluctuation" with a significance index of 46% (49th of 53 causes of overrun, Table 3.2, printed p.70), noting that most projects allow price fluctuation claims; its literature review cites a 20 to 25% share of overruns from price fluctuations in a study (Adam et al. 2017) that is not a New Zealand sample and was not opened. A perception ranking and a second-hand figure from another country.
• New Zealand Infrastructure Commission / Te Waihanga: Industry-wide output, price and productivity statistics; no project-level or contract-level price exposure.
• Office of the Auditor-General New Zealand: Searched for escalation, inflation, bitumen, steel and diesel and uses none of those words; its re-baselining of the New Zealand Upgrade Programme gives no split of the increase between scope and price.

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

Open this figure's own page

What it is. None - New Zealand national sources give industry productivity indices and cost shares, not hours or labour cost against budget on civil jobs.

Sample. Searches: "Stats NZ productivity statistics industry heavy and civil engineering construction labour productivity hours worked"; "Jaffe Le Chappell Motu construction productivity heavy and civil"; "Te Waihanga cost of infrastructure construction cost drivers productivity"; "Labour Productivity in the New Zealand Construction Industry: A Thorough Investigation" (publisher page refused access, so not opened); "New Zealand infrastructure projects cost overruns study". Frequencies and indices that exist but are not costs: civil labour productivity growth of 5% (2022 edition) or 7% (2026 edition) over 2000 to 2020; annual paid hours per worker in construction down about 6% over 2000 to 2020; a Te Waihanga model estimating civil construction prices would be about 10% lower had productivity matched building (a modelled counterfactual on prices, not a share of contract).

What it is not. Recorded as `none` on purpose. Productivity statistics and a labour cost share are not a measure of hours beyond what the work needed, and presenting 22% of cost or a 5% to 7% index as a contractor's overstaffing exposure would be a misreading. No New Zealand source read reports hours or labour cost against budget on civil work.

Where it came from

• New Zealand Infrastructure Commission / Te Waihanga (2022), Economic performance of New Zealand's construction industry · free copy
  Labour productivity sections, Figures 10 to 12 (pp.16-19); firm profit margins (Figures 16 and 17); Appendix 2 on measuring labour input · source record

Considered and set aside

• New Zealand Infrastructure Commission / Te Waihanga: Gives civil labour productivity growth of 7% from 2000 to 2020 (building construction 21%), an index of output per hour for the industry, not hours over need on a job. Its earlier edition's 5% differs.
• New Zealand Infrastructure Commission / Te Waihanga: Puts direct labour at around 22% of heavy and civil industry costs (the printed table's labour cell is blank). That is the cost base exposed to hours, not an overstaffing or overrun figure.
• Motu Economic and Public Policy Research (for BRANZ): Firm-level productivity growth and dispersion for 2001 to 2012; compares firms with each other, not hours with a budget.
• NZ Institute of Economic Research (NZIER): Sector-wide productivity and competition evidence from 2013; no job-level hours or budgets.
• Atapattu, Domingo & Sutrisna: Estimators' ranking of overrun causes; the abstract names design changes, planning and tender documentation and gives no labour-hours share. Abstract only was read.
• New Zealand Infrastructure Commission / Te Waihanga: One road. It records construction progress rates falling short of the joint venture's plan, with no hours or labour cost figure.
• New Zealand Infrastructure Commission / Te Waihanga: Productivity levels are estimated for vertical (building) construction only, and its labour-share breakdown is Australian data.
• Office of the Auditor-General New Zealand: Audit of one programme's decision-making; it does not break increases into labour, plant or materials.
• New Zealand Government Procurement (MBIE): Contract-form guidance; it contains no labour or productivity figure.

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

Open this figure's own page

What it is. None - no New Zealand national source read gives plant utilisation or standby cost, fuel used against fuel priced, or the cost of fuel and plant overruns as a share of a contract or of an overrun.

Sample. Searches: "New Zealand civil contractors fuel plant utilisation standby cost study roading contractors diesel share contract price"; "Waka Kotahi NZ Transport Agency cost escalation infrastructure construction price increases"; "NZTA price adjustment fluctuations formula bitumen contracts NZS 3910"; "Civil Contractors New Zealand survey members cost increases fuel bitumen contract price adjustment"; "New Zealand infrastructure cost escalation fixed price contracts contractors bear price risk bitumen steel diesel". Not opened: the NZ Transport Agency's own website, the Ministry of Transport website and the EECA off-road equipment decarbonisation report all returned a bot-protection page to automated requests, and the Auditor-General's New Zealand Upgrade Programme page returned 403; none was worked around. Index rates, not costs: NZTA's construction index (excluding bitumen) rose 13.7% in the year to June 2022 and its bitumen series was $1.1197 a litre in December 2022, $0.4635 a litre (about 71%) above a year earlier. A model, not an outturn, and not filed: RNZ reports the Agency's early modelling at a five percent rise in big project costs at US$100 a barrel and ten percent at US$150, with no time horizon given; Te Waihanga's 2026 review repeats it, citing only that RNZ article, as annual cost inflation for projects with active fluctuation clauses. The briefings themselves are not published where they could be found. It is a modelled owner-side projection of total project cost from diesel and bitumen together, not plant hours or fuel above what was priced.

What it is not. Recorded as `none` on purpose. Sources give how large plant and fuel are in a cost structure, how fast diesel and bitumen prices moved and who carries the risk under the contract, but none measures plant hours or fuel above what was priced, or the cost of that as a share of a contract or an overrun. Cost-composition shares are context, not evidence of this cause. This register holds no ratio rather than a borrowed one.

Where it came from

• NZ Transport Agency Waka Kotahi (2026), Infrastructure and public transport contract price adjustment (NZTA cost indexes and bitumen series) · free copy
  latest-values-NEW-infrastructure-indexes.xlsx, sheet 'Construction ... excluding' (weights and escalation rates); latest-values-bitumen-price-adjustment-series.xlsx, sheet 'Cost adjustment series' · source record

Considered and set aside

• New Zealand Infrastructure Commission / Te Waihanga: Statistics New Zealand's 2020 input-output tables as tabulated by Te Waihanga: diesel is 1.7% and equipment 10.1% of the heavy and civil industry's inputs. That is how big plant and fuel are in the cost structure, now filed as context, and the paper does not compare fuel or plant hours used with those priced.
• New Zealand Infrastructure Commission / Te Waihanga: Estimates that a 1% rise in world commodity prices lifts New Zealand diesel prices by about 0.6% (p.17); a price pass-through, with no fuel use against estimate and no plant utilisation.
• NZ Transport Agency Waka Kotahi: The Transport Agency's price-adjustment index for construction other than structures weights plant and site overheads 0.35 and the bitumen series is set monthly. They record how fast prices moved for contracts that use them, not hours above what was priced, and they do not say how many contracts carry the adjustment.
• RNZ (Phil Pennington), republished by Scoop: A news report of Transport Agency briefings: diesel is seven percent of the road construction index, and on projects under a year old contractors are exposed to fuel index changes in that first 12 months. A press account of documents not opened, and a share of an index, not a measured overrun in fuel or plant.
• Society of Construction Contract Practitioners: Practitioners' guidance on recovering fuel cost rises under NZS 3910 contracts. It names the mechanisms and the records to keep and contains no figure.
• Wynn Williams (law firm): Law-firm commentary that on fixed-price contracts without fluctuation provisions the contractor generally carries a fuel price rise, and that most contracts since COVID-19 do not allow for escalation. It states who bears the risk, with no number for how much fuel cost exceeded what was priced.
• Bond Construction Management Ltd for Wellington Water Ltd: A quantity surveyor's quarter-on-quarter repricing for three Wellington pipe-renewal contracts: diesel about 13%, plant hire 3-5% and operated plant 5-7% up in the fourth quarter of 2021. Price movements for one region and one quarter, estimated by one consultant, and no hours or litres against allowance.
• GHD for Tauranga City Council: A consultancy's forecast for one council: plant other than structures is 25% of its roading output index. An index weight and a forecast, not plant use above what was priced.
• Massey University (School of Built Environment), Atapattu: The full thesis was read. Its survey of 92 estimators ranks "unavailability or failure of equipment" with a significance index of 49% (44th of 53 causes of overrun, Table 3.2, printed p.70). A perception ranking, with no share of cost, and plant hours and fuel are not separate items in it.
• Office of the Auditor-General New Zealand: The Auditor-General's report on the New Zealand Upgrade Programme's decisions was searched for fuel, diesel, bitumen, steel, escalation and inflation and uses none of those words; it gives no plant or fuel cost.

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

Open this figure's own page

What it is. None - no New Zealand national source read gives final quantities, or final account value, against the schedule of quantities tendered.

Sample. Searches: "Te Waihanga cost overruns infrastructure"; "NZ Transport Agency research report cost estimate accuracy roading outturn"; "New Zealand infrastructure projects cost overruns study Treasury major projects"; "New Zealand roading contracts quantity variations schedule of quantities remeasure final account"; "New Zealand civil construction contract variations final account over tender Waka Kotahi post implementation review"; "Office of the Auditor-General Waka Kotahi procurement contract management variations". Not opened: the NZ Transport Agency website refused automated access (its 1993 Transit New Zealand research report 23 on the reliability of capital cost estimates, the SM031 basis of payment template and the 2025 contract variation guidance draft could not be read); the publisher's site refused automated access to the full text of Atapattu et al. 2023. A frequency exists: the Pipeline shares above, and the Auditor-General's 21 of 26 projects, but neither is a cost and neither is about quantities.

What it is not. Recorded as `none` on purpose. New Zealand government guidance confirms the mechanism, that on re-measurement contracts the initial sum is adjusted as actual quantities are measured, but no source read publishes how far quantities moved. This register holds no ratio rather than a borrowed one. The Australian findings in the corpus do not apply to a New Zealand job.

Where it came from

• New Zealand Government Procurement (MBIE) (2019), Measurement Contract: Construction Procurement Guidelines (information sheet)
  Overview, p.1 (the one page of the information sheet) · source record

Considered and set aside

• New Zealand Infrastructure Commission / Te Waihanga: The National Infrastructure Pipeline shares (Figure 33) say how often reported cost rose after financial decision, 51% of projects over $100 million and 28% of the rest, not by how much, and not because of quantities.
• Office of the Auditor-General New Zealand: Audit of one programme's decisions; it reports that 21 of 26 transport projects could not be delivered within allocated funding, a frequency, and gives no split of the increase between scope, quantity and price.
• New Zealand Infrastructure Commission / Te Waihanga: One public-private partnership road. Its settlements ($400.3 million in total, $186.5 million for the consenting claim, about $50 million for a change to asphalt over one-third of the road) are variations and claims on one job, not quantity growth on a schedule.
• Atapattu, Domingo & Sutrisna: Survey of 92 estimators ranking causes of overrun; the abstract names design changes, planning and tender documentation as leading, and gives no share of cost for quantity growth. Abstract only was read.
• New Zealand Infrastructure Commission / Te Waihanga: States that cost overrun analysis is out of its scope; its cost breakdown by input is Australian data.
• New Zealand Infrastructure Commission / Te Waihanga: About input prices and wages, and the industry's cost structure; it says nothing about quantities built against quantities scheduled.
• New Zealand Infrastructure Commission / Te Waihanga: Industry-wide output, price and productivity statistics; no project-level quantities or final accounts.
• Motu Economic and Public Policy Research (for BRANZ): Firm productivity decomposition from Statistics New Zealand microdata; no project-level quantities.
• NZ Institute of Economic Research (NZIER): Sector-wide productivity and competition evidence; no project-level quantities or final accounts.

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

Open this figure's own page

What it is. 20% of budget, decision to build versus actual cost, real terms.

Sample. Roads average 20% (N=167, sd=30); all project types average 28% (N=258, sd=39). Rail, bridge, tunnel and road projects across five continents.

What it is not. Measured from the decision to build, so most of it arises BEFORE a contractor is signed, and much of it is paid by the client. This is NOT a figure for what a contractor loses, which is why it is recorded as owner_side and excluded from the headline.

Where it came from

• Flyvbjerg, Holm & Buhl (2004), What Causes Cost Overrun in Transport Infrastructure Projects?
  p.5 · source record

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

Open this figure's own page

What it is. 40.5% of the total additional cost caused by overruns, as apportioned across the parties by 52 surveyed experts.

Sample. 52 Belgian public infrastructure stakeholders, 35% of them general contractors. Public tendering authorities were assigned 49% and contractors 40.5%.

What it is not. An opinion of who pays, not a measured cost: respondents apportioned 100% of the extra cost by judgement. Contractors assigned themselves 50.7%, others assigned them 35.6%. Belgium only, and a share of the overrun, not of the contract, so it is never multiplied by a contract value.

Where it came from

• Molinari, Haezendonck, Van Rompay, Mabillard & Dooms (2025), Persisting Cost Overruns in Public Infrastructure Projects: Lessons From the Belgian Case · free copy
  Section 3.2.3.4 'Distributing the impact occurred due to cost overruns', p.18 of the accepted manuscript; survey question, Section 2.2, p.7 · source record

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

Open this figure's own page