MENU
Your location:

Making disruptive technology constructive: driverless cars

    The driverless car revolution looks set to be one of the biggest changes of the 21st century. However, it will also be one of the biggest challenges, and the government must be careful in designing policies to address the issues and best unlock the potential of driverless cars. To quote Hayek, “to make competition work beneficially, a carefully thought out legal framework is required”.

    1. Safety

    Ultimately, investors are looking to develop algorithms which can extract semantic information from 3D (‘RGB-D’) video streams, allowing driverless cars to react seamlessly. By eliminating human error, this could dramatically reduce road accidents; however, safety regulations are nevertheless essential, especially when this technology is in its infancy. It may also be beneficial to initially limit when/where driverless cars can be used. Motorways, for example, though fast and dangerous, tend to be reasonably predictable.

    Similarly, cyber security should be prioritised. Without restrictions, a market in hacking, or in fake provision, would doubtless arise. As with any market founded on theft and/or dishonesty, the government should prohibit this, and quality endorsements should be obligatory. However, equally, the cost of endorsements should not bar new entrants, as this could undermine efficiency and stifle competition.

    2. Congestion

    Driverless cars could circumvent the need for car ownership, with automated cab services filling the gap where other transport modes fail. For example, high speed rail services could cover the bulk of a journey, whilst driverless cars make up the connections at either end. Where high speed rail is unavailable, driverless cars offer a possible alternative which does not require new infrastructure. On smaller roads, cars could follow individual routes to a person’s home or workplace; on motorways and larger roads, they could ‘platoon’ (drive closely in succession – connected such that they would all brake etc. simultaneously), reducing air resistance, increasing space on the road, and reducing accidents.

    However, increased congestion is also a concern. An OECD study found that, though a driverless car service complemented by an underground system could remove 90% of cars from cities, private ownership of automated vehicles without good public transport could almost double car travel. A service-based model should therefore be encouraged. Also, travel should be priced so as to encourage the use of other transport modes when congestion is high. Ideally, market mechanisms would enforce this, with driverless car services charging high prices when demand (and therefore congestion) is high. Uber’s dynamic pricing model offers potential here. However, it may also be beneficial for the government to levy congestion-based taxes on service providers; a comprehensive road-pricing scheme would achieve this. Previously, despite support by feasibility studies, road-pricing schemes in the UK have been blocked over privacy concerns – the proposed schemes, requiring satellite monitoring in individuals’ cars, were highly unpopular. Without individual car ownership, these concerns would be irrelevant, raising the possibility of resurrecting a road-pricing scheme.

    3. Drop in government revenue

    At present, speeding fines, parking tickets, road tax etc. provide income sources. To avoid falling further into fiscal deficit, the government would therefore need to find new revenue streams. Again, road-pricing could help here.

    4. Monopoly power

    The government would need to think carefully about how to encourage contestability, balancing barriers to entry against safety regulations, and access to technology against intellectual property rights. A situation where, similar to current practice, Uber provides a platform on which independent car owners can offer services through a collective ride-hailing app could be highly competitive, but the government will need to ensure that Uber’s incumbent advantage does not bar entrants, allowing them to charge providers excessively, and that providers are sufficiently endorsed[1].

    5. Unemployment

    This is less problematic than it may seem. Although there would initially be job losses for taxi and HGV drivers, a productive economy would develop other jobs (hopefully aided by retraining schemes). Also, driverless cars would likely be phased in: job losses would be gradual, enabling the job market to absorb the new intake of labour. In fact, driverless cars themselves could help by facilitating commuting, allowing workers to be more footloose without surrendering their home and community.

    6. Rents accruing to car fleet owners

    Rent-seeking behaviour is also not problematic as car fleets are not landed property[2]: they require maintenance, R&D, and coordination. It would not be the case of property owners effortlessly extracting rents.

    This is far from a comprehensive analysis; there is still much to be considered. For example, is there a need to encourage electrification of driverless cars? However, the above discussion does provide some thought on making the most of autonomous automation – hopefully, DfT will follow with more in-depth research.


    [1] TfL’s decision (22/09/2017) to revoke Uber’s licence in London indicates a tough approach to ensuring adequate safety standards and corporate responsibility from ride-hailing apps. Hopefully, if Uber does not step up to desired standards, other ride-hailing apps will fill the gap, with Gett and Citymapper already looking to innovate here.

    [2] Landed property or landed estate is a property that generates income for the owner without the owner having to do the actual work of the estate.

    Julia Bijl is a CPS Economic Research Intern. She is currently going into her final year at the University of Warwick, reading Economics. Her main interests in this field, at present, lie in transport, infrastructure and development economics. 

    Be the first to make a comment

    Centre for Policy Studies will not publish your email address or share it with anyone.

    Please note, for security reasons we read all comments before publishing.