Where the money is

On 3 September D'Ieteren Automotive, the Belgian importer of Volkswagen, Audi, Škoda, SEAT and Porsche, told its works council that a transformation plan could cost up to 344 jobs. The reasons it gave, in VRT's report, were decarbonisation of the fleet, the mobility budget that becomes compulsory in 2027, "digitalisation accelerated by artificial intelligence" and sharper global competition. The legally required consultation began that day.

On 9 September, after the market closed, the parent group published its half-year results.

The car division's sales fell 10.8 per cent to 2.27 billion euros. Deliveries fell 8.3 per cent and market share slipped to 21.5 per cent. Adjusted profit before tax fell 66.6 per cent to 36.4 million. The group wrote 47.2 million off the goodwill of the division's own showrooms and says the trends "are not expected to improve" in the second half.

The group, meanwhile, reported what it calls a robust semester: adjusted profit before tax up 6.6 per cent to 482.4 million, outlook confirmed. The reason is Belron, the windscreen business: the group's share of its adjusted profit before tax rose 28.6 per cent to 308.2 million, on an operating margin of 23 per cent. The same day the board announced a new chief executive from December and a review of options for the minority stakes in Belron, "including a potential listing".

Now the line that answers the question a Belgian consultant asked on LinkedIn the week before: where is the money? In the first half, while its profit fell by two thirds, D'Ieteren Automotive distributed 100.4 million euros to the parent. Its net debt rose from 259.5 million to 409.0 million, "mainly driven by the distribution", in the company's own words. The distribution pushed the division's net debt up by 150 million, a year after it had paid 400 million up. At the parent, 106.4 million went out as dividend to shareholders and 29.5 million as share buybacks, on top of the extraordinary dividend of 3,974 million euros, 74 euros a share, paid in December 2024 and financed with new borrowing. The largest shareholder is the family holding, Nayarit, with 53.84 per cent of the votes at May's general meeting.

The claim this entry scores is chapter 4's: the split between productivity and pay is real and ongoing, the labour share keeps falling and AI accelerates the mechanism. The chapter gives the mechanism two channels. The power channel is what labour loses at the table. The mechanical channel is Autor's finding that the labour share falls mostly between firms: activity migrates from businesses that pay people a large share of what they make to businesses that do not.

D'Ieteren is the mechanical channel inside a single group, by analogy rather than as Autor measures it, since his firms compete in one industry and these two do not. The business with the showrooms and the mechanics is shrinking and shedding. The business that repairs windscreens, and keeps 23 cents of every euro of sales as operating profit where the car division keeps two, is growing and may be floated. The group's profit rises while the Belgian payroll falls, and nothing in the accounts has to be wrong for that to happen.

The power channel is in the other document. The roles at risk get the Renault procedure, an information and consultation process, and severance. They get no claim on whatever their leaving saves, and the cash that leaves the division already has a destination: up to the parent, out to the shareholders, some of it to buy back the shares of whoever sells. This half-year makes the point more sharply than a profitable one would. There was no surplus to share. The division went deeper into debt to pay the parent, and the parent paid the dividend anyway. The distribution does not wait for the productivity; the claim on it is settled before the year begins.

Now the case against the book, because it is strong. This is a shrinking business in a shrinking market. Belgian new-car registrations fell 2.4 per cent and the division lost share on top; a company that loses two thirds of its profit cuts costs whether or not a language model exists. AI is one reason among four in the company's own list, and the ledger cannot say how many of the 344 posts it accounts for. Nobody can yet.

So score the claim as the register states it. Its first part, that what firms make and what they pay have come apart and that the difference goes to owners, is on every page of the half-year report. Its second part, that AI accelerates the split, is on one line of a press release, written by the company, unmeasured. The verdict is confirms on the first part, and the entry says plainly that the second is unproven here. When the Renault procedure ends and the company says which functions the machines took, the ledger will score that line on its own.

Chapter 4 tells of a tea plant outside Marseille whose workers made the line 50 per cent faster by themselves, watched the surplus go to managers sent from Brussels and Rotterdam and were told the answer to overstaffing was to move the plant to Poland. They ended by owning it. Nobody at D'Ieteren Automotive is being offered the plant. They are being offered the procedure.

Who owns the machine?