Three weeks to answer, five years of file
A parliamentary question isn't a request for documents. It's a request for a story with evidence that holds up.
A Member of Parliament tables fifteen written questions about a subsidy scheme: when did the ministry know about the implementation problems, what internal advice was on the table, and why wasn’t the House informed earlier. The request lands at a directorate on Wednesday afternoon. Three weeks to deliver a draft answer with a supporting dossier. The owner of the file starts calling round. One colleague has moved to another Directorate-General. Another is hunting for a 2022 piece of advice referenced in a 2023 memo but nowhere to be found in the DMS. A third emails: “I think that’s still on Loes’s drive — she retired two years ago.”
Answering a parliamentary question is not the same thing as handling a Woo (FOI) request. With FOI you deliver what’s there, filtered for what may be disclosed. With a parliamentary question you build a coherent answer on the basis of what’s there, and you have to be able to demonstrate that the answer holds up. The dossier is not compiled, it is reconstructed. Which memos were on the table when the decision was taken? What advice was given at the time, and by whom? What correspondence was there with the implementing agency? Earlier parliamentary questions on the same theme — what was answered then? That information is rarely brought together in any structured way. It sits in DMSs, mailboxes, personal drives, and in the heads of the people who worked on it at the time.
Article 68 of the Dutch Constitution obliges members of government to provide the House with every piece of information requested, except where the interest of the state stands in the way. Incomplete disclosure is not a procedural error. In Dutch political life it falls into the “incident” category: a second round of questions, an emergency debate, at worst a motion. With every dossier compiled under time pressure the risk is built in that a document exists which wasn’t included — and that later surfaces through a Woo request or a parliamentary inquiry. At that point the problem is no longer the missing document. The problem is that the House was informed incompletely. And that problem (history has shown this more than once) can no longer be repaired.
We tackle this at the front end — by making the document landscape searchable before the request comes in. The platform indexes DMSs, shared drives, archive services and mail archive by content. OCR makes older scans and incoming letters searchable too. Document types — decision memo, advisory note, letter to Parliament, report — are recognised and tied to policy dossiers, with proper metadata in place. References between documents are followed on content: this decision memo points to that piece of advice, which points to that implementation report, and incidentally there’s an earlier letter to Parliament on the same topic. The step that is usually missing is built in: once a dossier is compiled, the platform sweeps the full estate for related material that isn’t in the case file. A human then decides whether those documents stay out deliberately or are added in.
The difference isn’t really speed, although it is also much faster. The difference is that the minister sends an answer whose origin can be explained — which sources were searched, which documents were included, which excluded, and on what grounds. If there’s a follow-up question, a Woo request on the same topic, or an inquiry, the full dossier is already there. And in political and administrative The Hague, the difference between assuming a dossier is complete and being able to demonstrate it is not a detail.