The Amsterdam Dispatch: Worker in London, Entrepreneur in Amsterdam

One of the taxi drivers who took Uber's side before the Amsterdam Court of Appeal used to run fishing trips and manufacture fishing tackle. He keeps a large book of his own customers, picked a mid-range saloon to suit the clients he wants, and uses the Uber app (the court's phrase) to fill gaps in his diary. Another driver refuses any job with a long approach and cancels 41 per cent of the rides the algorithm offers him. Not dereliction. Strategy.

On 27 January 2026 the court held that these drivers are not employees — and that it could not identify a single Uber driver in the Netherlands who is.

Two maps of the same terrain

English law sorts working people into three categories. Employees get the full statutory framework, including unfair dismissal. The genuinely self-employed get almost nothing. Between them sit limb (b) workers, creatures of section 230(3)(b) of the Employment Rights Act 1996, entitled to the minimum wage, paid holiday and whistleblowing protection. The class asks only whether the work is done personally for someone who is not a client of the worker's own business; subordination is no entry requirement, which is how Bates van Winkelhof v Clyde & Co LLP [2014] UKSC 32 fitted an equity member of a City law firm inside it.

That middle category is where Uber's British drivers landed. In Uber BV v Aslam [2021] UKSC 5 the Supreme Court held that worker status is a question of statutory interpretation, not contract: the rights exist to protect people who are subordinate and dependent, and no written term can define them out, the approach built in Autoclenz Ltd v Belcher [2011] UKSC 41. Uber fixed the fare, dictated the terms, penalised ride rejection, policed performance through ratings. Drivers were workers, and their working time ran while they were logged in, in territory, ready to accept trips, at least while they were not simultaneously plying for a rival.

Dutch law has no middle category. A contract either is an arbeidsovereenkomst (employment contract) under article 7:610 of the Civil Code or it is nothing at all. Employment brings everything at once: dismissal protection, the sectoral collective agreement, back pay. The classification test comes from FNV v Deliveroo Netherlands BV (Hoge Raad, the Dutch Supreme Court, 24 March 2023, ECLI:NL:HR:2023:443): all circumstances of the case viewed in connection, nine of them listed, including the embedding of the work in the organisation, how pay is set, commercial risk. The ninth asks whether the worker behaves, or can behave, as an entrepreneur in economic life.

The litigation

FNV, the Dutch trade union federation, sued Uber in 2020 to enforce the taxi sector's collective agreement for every driver on the platform. It won at first instance: the Amsterdam cantonal judges held in September 2021 that drivers who contracted personally with Uber did so under employment contracts. Uber appealed, and a group of drivers intervened. On Uber's side.

The Court of Appeal referred questions to the Hoge Raad, which answered in FNV v Uber BV (Hoge Raad, 21 February 2025, ECLI:NL:HR:2025:319). The Deliveroo circumstances carry no ranking. Entrepreneurship weighs as heavily as any other factor, and it includes conduct entirely outside the relationship being classified: other clients, acquisition, tax treatment, reputation. It follows, the court accepted, that the same work for the same platform can be employment for one driver and self-employment for the driver in the next car.

Applying that answer, the Court of Appeal reversed (Uber BV v FNV, Gerechtshof Amsterdam, 27 January 2026, ECLI:NL:GHAMS:2026:163). The court did not dispute that several Deliveroo factors pointed towards employment: embedding in Uber's business, the take-it-or-leave-it contract, the way pay was set and skimmed. The drivers' entrepreneurship outweighed them all. The intervening drivers financed their own cars (annual vehicle costs ran to 19–32 per cent of turnover), carried their own liability and incapacity risk, kept accountants, claimed the self-employed COVID support schemes, and spread their work across competing apps and their own customer books. One pair ran six travel agencies, a window-cleaning round and a children's clothing business, and slotted Uber rides in after planning the week's fixed bookings.

The intervention was self-interest, not loyalty. The drivers told the court that choosing when to work is the essence of their businesses. The statements Uber filed put net hourly earnings at €20.60 to €27.00, and the court accepted that the intervening drivers out-earn employed taxi drivers under the collective agreement. Employee status would trade that margin, plus the entrepreneur's tax treatment, for the sectoral wage and the Dutch preventive dismissal regime: real protection, bought with the whole apparatus of subordination. For drivers like these, protection reads as constraint.

The union's fallback failed too. FNV proposed that anyone ever logged in more than fifteen hours a week be treated as an employee; the court dismissed the line as arbitrary, an assessment FNV had candidly shared. The court found that 236 drivers (online more than 32 hours a week, accepting over 90 per cent of offered rides) were probably economically dependent on Uber in fair measure. Even they got nothing: nothing was known of their individual circumstances, so no finding could be made.

One app, two verdicts

How does the same app produce workers in London and entrepreneurs in Amsterdam? Because the two courts were answering different questions. London asks whom the statute protects, and reads the relationship purposively against subordination and dependence. Amsterdam asks what contract the parties made, weighing everything with nothing pre-ranked.

Watch what that does to a single fact. A driver logged into Uber, Bolt and a local dispatch app at once is, in London, still a worker in his Uber hours: multi-apping goes to working time, not status. In Amsterdam the same fact is evidence of entrepreneurship that can tip the whole classification. Same driver, same three apps, opposite valence.

The absence of a middle category does the rest. A Dutch court that finds employment awards the full sectoral collective agreement, retroactively, fleet-wide; the all-or-nothing stakes push the analysis towards nothing. English law can split the difference at limb (b) — protection without the full employment framework. The spectrum argument made in the Industrial Law Journal last quarter, that binary status categories cannot hold the platform economy, now has its strongest exhibit, and it is not an English case.

The intervening drivers would have troubled an English tribunal too: a man who uses Uber to plug gaps between his own bookings is close to the marketing-your-own-business territory that defeats limb (b) status here. What an English tribunal would not have done is let those drivers speak for the fleet.

December is coming

Both Hoge Raad decisions declined to develop the law for the same stated reason: the legislature is already at work. The Dutch VBAR bill would recentre the test on the work itself, and Directive (EU) 2024/2831, the Platform Work Directive, must be transposed by 2 December 2026. It requires a rebuttable presumption of employment wherever the facts indicate direction and control, with the burden on the platform. France transposed on 1 March 2026. The Amsterdam judgment is best read as a portrait of the world the presumption is designed to end: the platform holds the classification high ground, and each driver must fight for status alone, on facts only the platform fully holds.

No English tribunal will be shown a Dutch judgment, and the arguments the Amsterdam court accepted are already run here daily under the "business undertaking" carve-out in section 230(3)(b). What Amsterdam supplies is the counterfactual, a system in which those arguments carry full weight, and a demonstration of where they end: a fleet reclassified around its most entrepreneurial members. The practical interest is for cross-border clients: the same operating model now produces workers in London, employees in Paris under the new presumption, and self-employed drivers in Amsterdam. Brexit means the December directive stops at Calais, so the divergence will widen before it narrows.

The court left one door open: any individual driver may yet prove an employment contract on his own facts. Each will have to do it alone.

Table of Authorities