The story of UK taxis over the last decade is not a morality tale about good cabs and bad apps. It’s a structural story. We built a market that rewards scale and liquidity on one side, and burdens local accountability on the other. The result is the great displacement: highly vetted, accessibility-capable taxis losing ground to lighter-touch pre-booked services that can roam across borders and arbitrage rules. Capping numbers won’t fix that. Changing who gets the work will.
In England, the number of licensed private hire vehicles has risen by more than 80% since 2010, while the number of taxis has remained broadly flat. Yet the quality gap has widened: nationally, only 2.2% of PHVs are wheelchair accessible, compared with over half of all taxis, and in London the contrast is absolute — 100% of taxis are wheelchair accessible, versus just 0.4% of PHVs (DfT Taxi & PHV Statistics, 2024).
This isn’t just a difference in vehicle type. It’s a difference in public purpose — between a system built around assistance and one optimised for availability.
Two tiers are the answer
The UK already has two worlds. On one side, taxis built for public trust: stronger vetting, accessibility as a design feature, regulated plying with transparent fares, and genuine local accountability. On the other, a vast pre-booked market optimised for availability and price.
The mistake has been to pretend these are just two badges on the same job. They aren’t. The taxi world is our Tier-1 public-service asset. It’s the part of the system we should be plugging directly into NHS non-emergency transport, SEND movements, social-care journeys and time-critical medical couriering. It’ll save a lot of money and improve passenger experience. The pre-booked world is Tier-2 mass demand: flexible, price-sensitive, and—done well—excellent for exactly that.
The lever for change is not new primary legislation or symbolic caps. It’s procurement and commissioning: route Tier-1 work to Tier-1 taxis on purpose, with the data standards, KPIs and safeguarding expectations that public services should insist on anyway. Do that, and you rebalance the market where it matters—in outcomes, not headcounts.
Licence the service mode
Licensing should follow mode, not mythology. Two modes are enough:
Set obligations, price controls and consumer protections by mode and apply them consistently to humans and robots alike. That keeps the regime technology-agnostic and closes the gaps where gaming thrives. Crucially, it leaves space for a deliberate policy choice: public-service contracts belong in Tier-1, delivered predominantly by taxis, with higher standards aligned to the work, not to the badge.
Where robotaxis fit (and where they don’t)
Automated Passenger Services will come, first in controlled geographies and times of day, and largely in the pre-booked mode. That’s fine—desirable, even—if they meet the same obligations for safety, data, redress and consumer protection as any pre-booked operator.
What they should not do is displace Tier-1 taxis from Tier-1 public service.
The human role is not an affectation; it is the practical answer to complexity and care. Keep robots where they are strong, keep humans where they are necessary—and let each improve the other through a common, mode-based rulebook.
A reality check on accessibility
Today, a human driver guides, reassures, and physically assists: describing surroundings, helping a passenger locate the door, fastening a seatbelt, or escorting them safely to a hospital entrance. A robotaxi can do none of these things. For those who rely on support from the kerb to the door, automation removes not a barrier but a safeguard.
In London, 100% of licensed taxis are wheelchair accessible, compared with just 0.4% of private hire vehicles. That difference is not a quirk of engineering—it reflects decades of policy choices linking human service to public trust. Robotaxis may one day supplement that work on predictable, fixed routes, but they cannot yet replace the accountability and assistance that define it.
A PHV cap is the wrong market intervention. Just procure better
The appeal of caps is understandable. The reality is they don’t bite across borders, they are blunt against algorithmic liquidity, and they often land hardest on the very drivers we claim to protect. If the problem is that high-trust work has been hollowed out, the solution is to direct that work back—with contracts that specify safeguarding, accessibility, response times, price fairness, complaints resolution, and auditable performance.
This is not theory. It is a procurement choice public bodies can make now. Map the spend. Specify Tier-1 participation and standards. Measure outcomes. Pay for reliability and care, not for badge counts.
Local accountability, national backbone
Rebalancing the system does not require ripping up local licensing. It does require a common spine: shared data standards, portable sanctions for serious breaches, and mutual enforcement so that the rules mean the same thing five miles down the road. Let regions adapt the edges—ranks, zones, demand management—while aligning the essentials. That combination—local stewardship, national interoperability—is how you stop rule-shopping without centralising everything.
Transition without drama
Change doesn’t need fireworks. Pilot in combined authorities. Grandfather fairly. Run dual systems where needed while contracts reset and data flows mature. Communicate honestly: this is not an anti-PHV crusade, nor a Luddite stand against automation. It’s a structural correction that puts assistance, accountability and value back at the centre of public-facing mobility.
What success looks like
Every percentage point that shifts Tier-1 work back toward licensed taxis improves accessibility outcomes at scale. Restoring just 10% of current PHV-type journeys to the licensed taxi fleet could double the availability of wheelchair-accessible vehicles across many English regions. That’s not theory — it’s a measurable equity gain embedded in procurement design.
The UK’s mobility debate has spent too long arguing about numbers and not long enough deciding what good work looks like — and who should do it. Robotaxis sharpen that choice. They don’t make drivers redundant; they make them indispensable where quality, assistance and accountability matter most.
Build the rules around service mode. Point the highest-trust work at the highest-standard providers. The market will do the rest—no nostalgia required.