The tax service, and the end of the informal visit

A tax administration is judged by two things: whether it collects what is due, and whether a compliant company can go about its business without interference. Those are not the same objective and a system that optimises for only one fails at both.

What the old practice looked like

Inspections initiated by decision, with an inspector arriving on a schedule that was not published and against criteria that were not stated. That combination gives an inspector something valuable to sell and a company something to fear regardless of its records.

Interpretation varied by office, so the same transaction could be treated differently in two cities.

What changed

Risk-based selection: a system scores taxpayers against defined indicators and generates the audit plan. Removing the human choice at that point removes the market for avoiding it.

Electronic communication and the taxpayer's personal account, so that filings, correspondence and reconciliations happen without a visit in either direction.

Published individual tax rulings, so that a company can see how the authority has treated a comparable situation.

And the electronic VAT administration system, which is the single most consequential change and is discussed elsewhere.

What still causes complaint

The blocking of tax invoices by an automated risk system, which stops fraud and also catches legitimate transactions, leaving a company to unblock through an appeal process. When an automated control has a false positive rate, the appeal process is the actual service level.

The measure to watch

Not the number of audits but the proportion overturned on appeal. A high rate means the selection or the assessment is wrong, and it is the honest indicator of quality.

Fewer inspectors coming and those who do being selected by risk analysis was a measurable change on my side — the number of visits fell and their length shortened. What still draws complaint is variation in interpretation: the same transaction can be assessed differently by different inspectors. That is the measure to watch — not the penalty totals but the consistency of the answer to the same question.