Germany Reports Public Procurement but Does Not Yet Measure It as a Control Instrument

AI-authored by Reinhard Brückner. Human editorial review and responsibility: Projekt Freistaat.

Since 1 October 2020, Germany has had a statutory reporting obligation for certain public contracts and concessions. The procurement statistics cover procurement procedures and concessions with an estimated net contract value of €25,000 or more. The responsible bodies must report their procedures to the Federal Statistical Office. This includes not only EU-wide tenders, but also procedures below the European thresholds. The Federal Ministry for Economic Affairs and Climate Action describes the statistics as a basis for greater transparency in public procurement.

The relationship is nevertheless incomplete: there is a reporting obligation, procedures are reported, but the data still do not provide a complete and consistently analysable picture of public procurement. Destatis itself points out that the statistical value of the data depends on the completeness and quality of the reports. Not every public purchase reaches the €25,000 reporting threshold. In addition, contracting authorities, specialist procedures and data quality are not fully organised within a common architecture. A figure for total public procurement volume therefore cannot simply be derived from the available data.

The Data Point Is Not Yet a Control Instrument

The procurement statistics are intended to bring together information on contracting authorities, procurement procedures, contract values, procedure types and contractors. In practice, however, they initially produce a collection of reports. A collection becomes a basis for control only when it is complete, comparable and can be permanently linked to other administrative data.

This is precisely where Germany’s gap lies. The reporting obligation is centrally regulated, but operational procurement remains distributed. The federal government, the Länder, municipalities and other public contracting authorities work with different procurement offices and electronic systems. The data are not generated to the same quality everywhere. Contracting authorities may also be identified differently. An authority may appear in different systems under variant spellings. Without unique identification numbers, it remains unclear whether several reports belong to the same organisation.

The time dimension is weak as well. Statistics that arrive with delays and varying degrees of completeness describe procurement after the fact. They do not control it. For an administration seeking to compare prices, identify opportunities for bundled procurement or assess the impact of procurement rules, the mere existence of a reporting obligation is not enough. The data must be structured so that they can be checked regularly and used directly.

European developments already provide a technical reference point for this. With eForms, the European Commission has standardised data fields primarily for EU-wide procurement notices. They can serve as a reference, but they do not replace a national data architecture, all below-threshold procedures or a national register architecture. Germany therefore has a standardised basis for relevant notices available. What it must do is apply suitable standards consistently throughout its administration, all the way to actual data use.

Denmark Makes Procurement More Usable Through Centralisation

Denmark illustrates one model for making procurement information more centrally accessible. The official udbud.dk portal bundles central information and refers to public tenders; it is not evidence of a fully integrated procurement database. The Danish Competition and Consumer Authority also publishes analyses and information on public procurement on a common institutional basis. This gives businesses, the administration and the public clearer access to relevant procurement data.

The difference is not that Denmark has eliminated every difficulty. There, too, the quality of public data depends on the reporting bodies and the systems they use. The qualitative difference lies in the usage architecture: information is brought together more strongly in one central location, made accessible according to uniform criteria and prepared for analysis and comparison. This can reduce search costs and increase the ability to treat procurement as a recurring process.

Germany, by contrast, has primarily created a legal obligation. Responsibility for correct and complete data remains distributed across a highly fragmented system. Destatis can process the information submitted, but it cannot by itself ensure that every contracting authority reports on time, completely and with the same technical quality. Federal law establishes the framework. Implementation lies with many bodies operating under different incentives and priorities.

The structure therefore produces precisely the result it can be expected to produce. An obligation without a binding data architecture creates data stocks, but no reliable overall view. Statistics without continuous completeness checks remain reactive. And a register without unique contracting-authority IDs can identify connections only to a limited extent. The gap is not primarily a problem of knowledge. It is an implementation problem between the rule, the report and its use.

Bauplan

The next step would be a binding public procurement register. It would not need to replace the existing procurement statistics, but should sit above them as a usable implementation layer. Every procurement subject to reporting would be recorded in a standardised data set covering at least the contracting authority, a unique contracting-authority ID, the procurement office, the procedure type, the subject of the contract, the contract value, the award date and the contractor. The European eForms should serve as a reference for EU-wide notice data, supplemented by national fields for analysing total public procurement.

The key would be to connect the obligation with verification. For every reporting body, the register would have to document receipt of the data and publish quarterly the extent to which reports had been received in full. Such verification would not be an additional statistical exercise, but a control of existing implementation. Missing reports would become visible without first creating a new substantive reporting requirement.

Implementation requires clear responsibilities. The federal government would have to define the data model, the register of contracting authorities and the verification rules on a binding basis. The Länder and municipalities would have to connect their procurement systems to these requirements. Technical integration should use standardised interfaces, so that each body does not have to maintain its own transmission channels. The central state would not implement every procurement itself. It would ensure that every procurement becomes measurable under the same rules.

One potential pitfall lies in expanding the number of data fields. A register can become unusable if it demands more information than procurement offices can reliably generate. Mandatory fields must therefore be limited, clearly defined and transferred automatically from existing procedures wherever possible. The opportunity lies in the feedback loop: the more reliable the data, the more effectively authorities can compare prices, identify joint procurement and assess the impact of procurement law.

Denmark demonstrates that centrally usable access to procurement data is organisationally possible. Germany already has the reporting obligation. What is missing is the architecture that turns reports into a verifiable and usable public infrastructure. A reporting obligation is not yet a basis for effective control.

The appropriate framework is Volume 3, “Bauplan”: the question of how a decision reliably becomes an effect. For the administration, the next step is not another statement of principle, but a register with unambiguous data, clear responsibility and regularly visible completeness.