The Real Economic Impact of Driverless Cars Won’t Fit in a Headline

How a misleading by-line still confuses us today and what economic research can teach us about measuring emergent tech.

Autonomous vehicles will add £42 billion to the UK economy by 2035 and create 38,000 skilled jobs. You’ve probably seen this figure. It appears in press releases, ministerial statements, and coverage like the BBC’s piece last week. It’s cited as though it captures the full economic transformation autonomous vehicles are due to bring.

It doesn’t.

The number has been lifted out of its original context. The language journalists use today traces back five years to a GOV.UK news article announcing the report - “UK on the cusp of a transport revolution, as self-driving vehicles set to be worth nearly £42 billion by 2035”. The briefing’s byline is misleading. Five years later, it has led to confusion.

The source is the ‘Connected and Automated Vehicles: market forecast 2020‘, a report that never aimed to measure ‘the full economic value of self-driving vehicles’. It measures something narrower: the market for selling autonomous vehicles and their components, and how much of that market UK industry might capture. What happens once these vehicles are deployed, the services, the uses, the downstream effects, is explicitly outside the report’s scope

Transformative technologies don’t reshape economies through their sales alone. Car sale numbers didn’t fully measure what cars did to the economy. If autonomous vehicles transform the economy, it will be through what they enable once deployed. Cheaper freight meaning cheaper goods. Freed time as fewer people need to drive. Released urban land as city-centre parking becomes unnecessary. These changes come from using the technology.

The £42 billion figure couldn’t attempt to measure any of this. The report’s findings were developed between February 2019 and February 2020. The technology was far less mature then. Business models hadn’t emerged. Use cases hadn’t clarified. The report did what was tractable: it measured what existing methods could measure, which was the market for vehicles and components.

Six years on, more is known. Waymo operates commercially in multiple US cities. Chinese startups are scaling autonomous last-mile logistics. Freight companies are piloting autonomous trucks. And the UK is about to join them: regulations change in the second half of 2026, with Waymo’s pilot beginning in April and Wayve following in spring.

We still can’t produce precise forecasts of what this will mean for the economy. But we can do better than a figure that measures the wrong thing. Economists who studied containerisation, the washing machine, and the horse-to-car transition developed methods for tracing how transformative technologies reshape economies. Before borrowing their tools, though, it’s worth understanding exactly what the £42 billion figure does measure, and how the authors arrived at it.

The report measures UK market share, not economic transformation

The report was commissioned to answer a specific question: how big is the global market for autonomous vehicles and their components, and how much of that market might UK industry capture?

The report measures the value of selling vehicles with Level 3 or higher autonomy. That means partial autonomy and above, not quite full passengerless autonomy. It also counts UK factories and firms supplying sensors, software, and connectivity hardware globally. It covers cars, vans, heavy goods vehicles, and buses. It includes both domestic sales and exports.

What it does not cover is what happens after those vehicles are sold. Robotaxi services. Productivity gains from freed driving time. Freight efficiency improvements. Safety benefits. Changes to where people live and work. All of this is explicitly outside the report’s scope.

How they got to the 38,000 jobs prediction

Firstly any citation of 40,000 jobs is a rounded up conclusion of 38,000 total net additional jobs by 2035.

This breaks down into ‘23,400 direct jobs in the production of [Autonomous vehicles and component technologies], with a further 14,600 indirect jobs created in the supply chain for these technologies’ (page 8)

To estimate UK jobs and market value, you need to know how much production activity will happen in the UK. That means starting with projected vehicle sales, then adjusting for trade: subtracting imports and adding exports.

For vehicle sales, the report forecasts total sales across vehicle types, then applies adoption curves. These adoption curves are S-shaped, following the standard pattern for new technology diffusion. Slow early uptake. Rapid growth as costs fall. Then saturation. The curves themselves are scenario-based assumptions, anchored to milestone years and validated through industry consultation. They weren’t derived from observed sales data. That data didn’t exist yet.

For prices, the report separates the base vehicle cost from the autonomy package. Hardware costs, sensors, compute, decline over time following a learning curve tied to cumulative production volume. Software costs stay roughly flat, reflecting ongoing licensing and development. The autonomy package cost gets marked up to estimate consumer prices.

For trade, the report uses historical patterns from similar industries as proxies. If UK manufacturers currently capture a certain share of global sensor exports, the model assumes a similar share for autonomous vehicle sensors. This is a structural choice. It projects future competitiveness from past trade patterns. The results depend heavily on whether this assumption holds and whether the UK will be as competitive in this new industry.

The 40,000 jobs figure comes from applying labour intensity ratios to the projected UK production value. These ratios, jobs per million pounds of output, are drawn from Office for National Statistics data for comparable sectors. Indirect jobs in the supply chain are estimated using ‘standard employment multipliers’.

The interesting questions are about deployment

If you’re a policymaker or a curious observer trying to understand what autonomous vehicles might actually do to the economy, you need to look somewhere else entirely. The interesting questions are about deployment: what happens when moving people and goods gets significantly cheaper. When driving time becomes usable time. When urban land currently dedicated to parking becomes available for other uses.

These are the channels through which past transport technologies reshaped economies. We can’t forecast them precisely. But we can at least identify where to look. What follows explores three areas: cost reduction and how it propagates, freed resources (time and land), and the transition costs for those who lose out.

Cheaper transport makes everything else cheaper

The core economic event may simply be this: moving people and goods becomes significantly cheaper. No driver wages. 24-hour operation. Better vehicle utilisation. Optimised routing. These cost reductions then pass on to every sector that depends on transport of goods or people. That’s nearly all of them.

Car sales didn’t capture what cars did to the economy

When cars replaced horses, the economic value went far beyond car sales figures. The most interesting developments happened when transport became more accessible, faster and more flexible. Viable suburbs. Restructured supply chains. New industries: motels, drive-throughs, out-of-town retail. Dead industries: blacksmiths and stable hands.

The £42 billion figure for autonomous vehicles may be similarly small relative to the total effect. Citing this report as the headline economic impact by 2035 is like measuring the horse-and-buggy replacement market in 1920 and calling it the economic impact of the car.

An application: Trucks that don’t sleep

The first, and among the most important changes from autonomy will come to land freight.

Long-haul trucking is constrained by driver rest requirements. Drivers can only work a certain number of hours. Trucks sit idle while drivers sleep. Autonomous trucks could run through the night. Trucks don’t need to sit empty when drivers aren’t available. The same number of trucks could move more goods. Or the same volume of goods could move with fewer trucks.

Once overnight delivery becomes cheap and reliable, that changes more than logistics costs. It changes inventory strategies. Why hold stock locally if you can get overnight replenishment? It changes warehouse locations. Proximity to customers matters less if delivery windows shrink. It changes which goods are worth shipping at all. Items with thin margins that couldn’t justify express delivery become viable.

Containerisation shows how to measure logistics innovation

Economists use input-output analysis to trace how a cost reduction in one sector flows through to all the sectors that use it as an input. If transport gets cheaper, that reduces costs for manufacturing, retail, construction, agriculture, and so on. Those sectors then become more productive or pass savings to consumers, generating further effects.

The economics of containerisation offers a useful analogy. The direct effect was simple: shipping goods across oceans became cheaper. But the indirect effects were vast. Supply chains globalised. Manufacturing shifted to lower-cost countries. Port cities that adapted thrived. Those that didn’t declined. The technology didn’t just reduce shipping costs. It restructured global trade.

A 2016 paper by Bernhofen, El-Sahli and Kneller used trade flow data to estimate containerisation’s effects. They found that trade between industrialised country pairs increased by around 17.4% for North-North (industrialised country) trade and 14.1% for World trade, but where the latter occurred 10-15 years after containerization. Less industrialised countries lacked the domestic infrastructure, railways, roads, ports, to fully exploit intermodal shipping quickly. The technology mattered. But so did the surrounding systems.

Applied to autonomous vehicles, you’d want to estimate transport cost reductions from early deployment data. Then model how those reductions flow through transport-intensive industries. The method exists. The data is starting to become available.

What could the longer-term effects look like?

If moving goods becomes radically cheaper, things that are currently “not worth shipping” become viable. Journeys that “are not worth taking” become worthwhile. What this enables is harder to predict. It might accelerate the decline of local manufacturing as more goods become worth importing.

These effect sizes depend on business models and relative costs that we can’t yet see. Honest forecasting means accepting we don’t know the specifics. But we can be confident about the channel. Cheaper transport has historically restructured economies in ways that dwarf the direct market for the transport technology itself.

Freed resources: time and land

Autonomous vehicles could release resources currently locked up. Human time spent driving. Urban land dedicated to parking. Capital tied up in underutilised vehicles.

Time recapture

UK drivers spend billions of hours annually behind the wheel. If vehicles become autonomous, that time redirects to something else. Work. Rest. Phone calls. Reading. Online shopping. Staring out the window.

Even “unproductive” uses have value. The relevant comparison isn’t “working versus not working.” It’s “cognitively occupied with driving versus free for anything else.”

Economists have a golden example: the washing machine. Studies of labour-saving household technology found that time previously spent on domestic tasks became available for paid work. This contributed to increased female labour force participation over the 20th century. Ha-Joon Chang argued that the washing machine was more economically transformative than the internet. Dozens of hours each week that was once used for handwashing, across billions of people, have been released for more productive work to be done.

The car currently demands attention. Removing that demand releases time at comparable scale. An hour-long commute that currently produces nothing except getting you to work becomes an hour available for whatever you want. Multiply that across millions of commuters. The numbers add up.

How economists measure time value

Time-use surveys measure how people spend commute time when they have a choice. Train commuters, who don’t need to watch the road, split their time between work, leisure, and rest in observable ways. This gives a baseline for what people might do with freed driving time.

Willingness-to-pay studies estimate how much people value time savings. The UK’s Department for Transport has detailed guidance, the WebTAG framework, specifying how to value time savings for transport projects. Working time is valued at wage rates. Non-working time is valued at a fraction of wages, derived from revealed preference studies about how people trade off time and money.

These methods could adapt to value released time from autonomous vehicles. There’s a question about whether released time (free for anything) is more valuable than saved time (arriving earlier). Arriving 20 minutes earlier is useful. Having 20 minutes that would otherwise be spent driving is arguably more useful. The range of possible uses is wider.

Land release

A significant fraction of urban land is parking. Surface lots. Multi-storey structures. On-street spaces. Estimates vary by city, but in Manchester around 5% of central area land is dedicated to parking.

If vehicles can drop passengers and relocate to cheaper peripheral areas, or stay in continuous use serving other passengers, demand for central parking could collapse. That land becomes available for other uses. Housing. Commercial space. Parks. Cycling infrastructure.

This is slow-moving. Land use changes take decades. They depend on planning decisions, political will, and property rights. A car park doesn’t automatically become housing just because fewer people need to park. But in cities with housing shortages, the option value of released land matters.

Policy shapes outcomes

The scale and distribution of benefits depend heavily on choices that haven’t been made.

Will robotaxis feed passengers into public transport networks or drain them? Will cities release parking land for housing or let it sit? Will displaced drivers get transition support or be left to the general labour market? Will deployment be concentrated in wealthy areas or required to serve underserved communities?

These are not technical questions with technical answers. They’re political questions. They will determine whether the gains from autonomous vehicles are broadly shared or narrowly captured. Whether the transition is managed or chaotic.

The technology creates possibilities. Policy decides which possibilities are realised. This isn’t a caveat to add at the end. It’s central to understanding what autonomous vehicles will actually do to the economy.

Transition: who loses, what changes

The previous sections focus on value creation. This one addresses the other side. Economic transformation creates losers alongside winners. Honest accounting requires including both.

Driving jobs

Taxi and private hire drivers. Delivery drivers. Truck drivers. Bus and coach drivers. These are significant employment categories. They’re often accessible work for people without advanced qualifications. The UK has around 380,000 licensed taxi and private hire drivers alone. Road freight employs hundreds of thousands more.

The transition could be gradual or abrupt. It depends on how quickly costs fall and regulations permit deployment. Geographic concentration matters. Some regions depend more heavily on driving and logistics jobs than others. The places most affected may not be the places best equipped to absorb displaced workers into other roles.

How the horse-to-car transition changed jobs

The shift from horses to cars eliminated entire occupations. Stable hands. Blacksmiths. Farriers. Carriage drivers. Horse breeders. Urban hay farmers.

Some workers moved into car-related jobs. But not one-for-one. And not the same people. A blacksmith’s skills didn’t transfer neatly to an assembly line. The transition took decades. It was absorbed through general labour market adjustment rather than active policy.

The autonomous vehicle transition may follow a similar pattern. New jobs will emerge. Vehicle operation centres. Remote assistance. Fleet management. Maintenance. But these jobs won’t necessarily go to displaced drivers. They won’t necessarily be in the same places. They won’t necessarily roll out at the same speed.

Suburban independence

If commute time becomes usable rather than dead, the cost of living further from work falls. An hour’s drive that currently means an hour of wasted time becomes an hour of reading, working, or resting. This could reinforce existing patterns of suburbanisation. Or enable new ones. Smaller towns with cheap autonomous vehicle access to cities. Rural areas that become viable for people with urban jobs.

This cuts different ways for different people. Families juggling school runs and work commutes benefit from easier logistics. Elderly people who can no longer drive safely regain independent mobility. People priced out of cities gain access to urban labour markets without urban housing costs. People who never learned to drive, or can’t afford a car, get new options.

But there are losers here too. Public transport that depends on commuter ridership may lose passengers to more convenient alternatives. Urban centres that benefit from captive commuters, people who live centrally because they have to rather than because they want to, may see demand soften. The effects depend on pricing. On service quality. On how autonomous vehicles interact with trains and buses rather than simply replacing them.

Conclusion

The £42 billion headline captures something real. It’s the potential UK share of a global market for selling autonomous vehicles and their components. That market matters. The report that produced the figure is a reasonable attempt to measure it.

But the economic transformation, if it happens, will not be measured in vehicle sales. It will be measured in what cheaper transport enables across every sector that moves people or goods. What becomes possible when driving time becomes usable time. In what cities do with land they no longer need for parking. In how we manage the transition for people whose jobs disappear.

These effects are harder to quantify. But they are certainly larger than £42 billion.