What the index measures
The construction price index measures the movement of prices that clients pay for construction work, referenced to a fixed base year. The current series uses 2021 as base year with a value of 100.
The two series are regularly confused, although they measure different things. The confusion goes unnoticed because both are expressed in index points and move in similar directions.
| Series | What it measures | For updating a benchmark |
|---|---|---|
| Construction price index | prices clients pay for construction work | correct |
| Construction cost index | cost side of contractors: wages and material prices | not applicable |
| Item | Value |
|---|---|
| Base year of the current series | 2021 = 100 |
| Publication frequency | quarterly |
It must be distinguished from the Baukostenindex, the construction cost index, which reflects the cost side of contractors, meaning wages and material prices. The two diverge, because market conditions act on margin: in a period of weak workload, construction prices can rise more slowly than construction costs, and faster in a boom.
For updating a cost benchmark the construction price index is the correct series, because a benchmark derives from completed projects and therefore reflects prices paid rather than production costs.
The calculation
The formula is straightforward:
Updated value = original value × (index at valuation date ÷ index at price date)
An example sets out the calculation in full. The six lines are also the pattern for presenting it in an estimate.
| Step | Quantity | Value |
|---|---|---|
| 1 | Starting figure | 2,400 € per m² gross floor area |
| 2 | Price date of the starting figure | first quarter 2025 |
| 3 | Index at the price date | 132.6 |
| 4 | Index at the valuation date | 140.0 |
| 5 | Calculation | 2,400 × (140.0 ÷ 132.6) |
| 6 | Updated figure | 2,534 € per m² gross floor area |
Three rules apply, and the second is the most often broken. It typically costs one to two years of price movement, because a dataset's publication year sits after its price date.
| # | Rule | Why |
|---|---|---|
| 1 | Both index values from the same series and base year | otherwise the ratio is meaningless |
| 2 | Price date from the dataset, not from its year of publication | the publication year is later |
| 3 | Show the calculation as its own line, with both index values and their source | it makes the estimate auditable |
The sub-indices diverge
Alongside the overall index for new residential construction, the Federal Statistical Office publishes separate series for shell and fit-out work, and beneath those, series for individual trades. The two series move differently.
These series diverge, and the gap is wide enough to distort an estimate. It bites hardest where shell and fit-out are weighted differently in the project than in the overall index.
| Sub-series | Change on the previous year in the comparison period |
|---|---|
| Shell work | 2.5 % |
| Fit-out work | 3.8 % |
| Concrete and masonry work | low single digits |
| Roofing, earthworks and carpentry | markedly stronger |
The gap between shell and fit-out in that period is 1.3 percentage points, a little over half the shell rate. Anyone updating with the overall index spreads that difference evenly and lands wide on shell-heavy schemes.
Practical consequence: where a cost estimate is built up by work section, the corresponding sub-indices should be used rather than the overall index. A flat uplift on the total does not capture the internal shift between groups of trades and systematically understates the fit-out component.
The three per cent rule of thumb does not hold
Numerous guides assume construction prices rise by around three per cent a year. As a long-run average that is not unreasonable, but for a specific update it is unusable.
Actual rates of change vary considerably. After the sharp rises of 2021 to 2023 came a period around three per cent; in spring 2026 the rate ran at five per cent year on year, a substantial part of it within a single quarter.
Using the published index costs no more time than a flat assumption and can be evidenced to the client. There is no reason to estimate here.
Net or gross
The index is published inclusive of value added tax. That is unproblematic for updating provided you know what you are doing: since updating works with a ratio of two index values, an unchanged tax rate cancels out. A net figure can therefore be updated with the gross series, as long as the tax rate has not changed over the period.
If the rate does change between price date and valuation date, the effect must be handled separately. The system of tax basis is set out in the article on the cost scope of a benchmark.
What updating does not do
This is the most important point of the article. Updating corrects the price level, not the scope of works.
A benchmark from a building erected ten years ago describes a building with the plant standard of that time. Tightened energy requirements since, mechanical ventilation with heat recovery, heat generators running on renewable energy, charging infrastructure and building automation are not inside that benchmark. Nor does the index capture them, because it measures price movement for a constant specification.
An updated old benchmark therefore lands systematically too low, and the error grows with the distance from the price date. The separation between price and scope is thus the real limit of the method.
| What updating does | What it does not do |
|---|---|
| Corrects the price level | corrects the scope of works |
| Price movement for unchanged works | tighter energy requirements |
| ventilation with heat recovery | |
| heat generators using renewable energy | |
| charging infrastructure and building automation |
Related articles
The index values and rates of change given are positions at the date of checking. They do not replace consulting the current release of the Federal Statistical Office.
© QuostraThe actual rates of change vary considerably, by a multiple within a few years. A flat assumption is therefore right only by accident.
| Period | Rate on the previous year |
|---|---|
| 2021 to 2023 | steep increases |
| the phase that followed | around 3 % |
| spring 2026 | 5 %, a substantial part of it in a single quarter |