Delivery performance
How infrastructure projects held to their budgets and their timetables. Every figure comes from the contract terms and costs recorded on each section, and a section appears only once the data supports the claim.
Measured on 1411 sections: 2 with an estimate-to-outturn comparison, 6 against the contract award, 107 delivered with a contract timetable, 64 still running.
The gaps are missing data, not clean records. A section without a recorded outturn cost or contracted duration cannot be measured, so it is left out rather than assumed on budget.
All costs restated in 2025 prices. In-progress delays measured as of Oct 2026.
Opened sections
Every section that has opened, and how far past its contract date it ran. Sort by any column.
Newest first until you sort.
What it cost
The best cost figure each section has: the outturn where one was published, otherwise the signed contract value, otherwise the pre-tender estimate. To rank them together each figure is restated into common prices and then converted to euro at that year's exchange rate. A section whose currency or price year falls outside those tables keeps its recorded cost but has no comparable figure, and sorts to the bottom.
Most expensive per kilometre first until you sort.
By contractor
Builders only, since a construction overrun belongs to whoever built the thing. Joint-venture work counts toward each partner, and name variants are resolved to one firm. Firms with a single section are left out.
| Hitachi Rail | 2 | 100 km | +89 mon=1 | – | – |
|---|---|---|---|---|---|
| ATTIKAT | 4 | 144 km | +76 mon=1 | 0% | – |
| Convensa | 2 | 41 km | +70 mon=1 | – | – |
| Webuild | 18 | 559 km | +57 mon=2 | – | – |
| Salcef | 7 | 266 km | +57 mon=2 | – | – |
| FCC Construcción | 18 | 453 km | +55 mon=6 | 0% | +12.4%n=1 |
| Euro Construct Trading '98 | 7 | 280 km | +50 mon=5 | 0% | – |
| Visio Construction Works | 2 | 48 km | +47 mon=2 | 0% | – |
| Avtomagistrali EAD | 14 | 289 km | +44 mon=3 | – | – |
| Ghella | 3 | 20 km | +39 mon=1 | – | – |
Worst median slip first until you sort.
By country
The same measures aggregated by country.
| Croatia | 70 | 1,277 km | +20 mon=5 | – | – |
|---|---|---|---|---|---|
| Greece | 164 | 3,541 km | +19 mon=15 | 20% | – |
| Bulgaria | 97 | 1,839 km | +17 mon=25 | 32% | – |
| Romania | 164 | 4,485 km | +15 mon=81 | 13% | +22.9%n=2 |
| Slovakia | 107 | 1,281 km | +14 mon=5 | 0% | – |
| Poland | 341 | 6,808 km | +12 mon=35 | 0% | – |
| Serbia | 72 | 1,461 km | +9 mon=3 | 0% | – |
| Czechia | 197 | 1,570 km | −3 mon=1 | – | – |
| Albania | 1 | 21 km | – | – | – |
| Austria | 22 | 491 km | – | – | – |
Worst median slip first until you sort.
Method
Costs carry the price year they were quoted in, and both sides of every comparison are restated into a common year before the percentage is taken, so an estimate made in 2013 is never measured directly against money spent in 2023. Costs recorded in different currencies are restated first and then converted to euro at that year's annual average exchange rate, in that order: converting first would use a rate from one year against prices from another.
Schedule slip uses the contract as the baseline rather than a press announcement, because that is the commitment the public paid for and the only baseline present for sections whose announced dates were never recorded. Negative values mean early.
Price indices: Eurostat: HICP annual data (average index and rate of change) [prc_hicp_aind], U.S. Bureau of Labor Statistics, Consumer Price Index for All Urban Consumers: All Items in U.S. City Average [CPIAUCNS], via FRED, European Central Bank Data Portal: Euro foreign exchange reference rates, annual averages [EXR, A.<CUR>.EUR.SP00.A], Eurostat: Euro/ECU exchange rates, annual data [ert_bil_eur_a], statinfo=AVG; source for RSD and cross-check for RON and USD, Council Regulation (EU) 2025/1409 amending Regulation (EC) No 2866/98: irrevocable conversion rate of 1,95583 Bulgarian lev to the euro from 1 January 2026.