Saving on Paper

In “The New Nomenklatura and the Party Cartel” we described how the apparatus grows while the parties of the centre bleed out and the fringes grow. A week later the tax on sugary drinks supplies a textbook case, and a statistic from the Finance Ministry shows that even the counting can deceive.

I. The tax whose success makes it poor

From 1 July 2027, soft drinks with at least five grams of sugar per 100 millilitres are to be taxed at 26 cents per litre, and those with eight grams or more at 32 cents. According to the coalition groups’ amendment to the 2027 Budget Accompanying Act, analysed by Christian Schlesiger for The Pioneer, introducing the tax requires 226.75 additional staff: 200 at customs, 23 at the federal IT service provider ITZBund, and 3.75 in the Finance Ministry. By 2029 the administrative cost adds up to around 90 million euros, almost 32 million in 2028 alone, against expected revenue of 925 million.

Around 3.5 cents of administration per euro of tax is not absurd for a new excise duty. The real contradiction lies elsewhere. The thresholds of five and eight grams evidently come from Britain, where such a levy has applied since April 2018. There, almost all major brands changed their recipes, largely before the levy even started. According to one study, the calorie content of the hundred leading brands fell from about 23 to about 15 billion calories a week. The Treasury had to revise its expected revenue downwards even before introduction, because manufacturers cut sugar so quickly.

For health this was a success. For the exchequer it was a shrinkage. If the German tax works in the same way, the 925 million will fall, but the customs officers will remain. The better the tax achieves its declared aim, the more expensive its collection becomes relative to its yield. The Finance Ministry does at least write that the staffing requirement is to “decrease over the years”. That can be checked in 2029.

The requirement is to be met “through vacant posts and redistribution of tasks”, the ministry says, with the addition: “as far as possible”. Either the people will then be missing elsewhere, or posts that were supposed to disappear by 2029 will remain. The coalition agreement promises eight per cent fewer staff in the federal administration by 2029; the coalition committee reaffirmed this on 2 July. For the next new tax, on cryptocurrencies, the first draft bill so far says only this under “compliance costs for the administration”: “[to follow]”.

II. Two ways of counting

The most revealing finding of the Pioneer article concerns the Federal Central Tax Office. There are two figures there. The number of budgeted posts, the “budget actual”, fell slightly from 2021 to 2026, from 2,256 to 2,211. The number of full-time staff actually working and being paid, the “working actual”, rose over the same period from 1,861 to 2,050, by ten per cent. Outwardly, then, one can say: the office is saving. But the saving happens only on paper.

This also affects our own text. The figures we relied on in “The New Nomenklatura” were budgeted posts. Budgeted posts and people actually employed can diverge, in either direction. Anyone who wants to know how large the apparatus really is must know the working actual. The Finance Ministry released it on request. As a rule it is not published.

In fairness, one must add what the figures do not show. In the tax administrations of the federal states, 107,199 full-time staff worked in 2025, 869 more than in 2021, which is less than one per cent. In the tax offices themselves, staff even fell, from 97,189 to 96,736. What grew were the IT departments, by 708 to 5,025. That looks more like restructuring than proliferation. Where proliferation does occur, it is at federal level: ten per cent more at the Federal Central Tax Office, almost nine per cent more in the operational tax administration of customs, not yet counting the sugar tax.

III. The awkward question from Spain

The Pioneer article cites an analysis by the accounting software provider BuchhaltungsButler and DataPulse Research. According to it, Germany has one full-time tax administration employee per roughly 760 inhabitants, Italy one per 1,850 and Spain one per 2,330. On its own this proves little. Fewer staff can also mean worse collection.

The comparison becomes revealing only together with a second figure. Every year the European Commission calculates how much of the VAT owed is in fact not collected. For 2023 it was 15.0 per cent in Italy, 9.7 per cent in Germany and 7.6 per cent in Spain, against an EU average of 9.5 per cent. Italy thus confirms the assumption that too few staff cost money. Spain refutes it: it manages with about a third of Germany’s staff density and still loses less VAT than Germany.

Why Germany, with so many more staff, achieves a worse result is answered by neither figure. That would be worth a study of its own. But anyone who invents new taxes should ask the question before applying for the next 227 posts: what are the Spanish doing differently?

Back to the essay: The New Nomenklatura and the Party Cartel