Implementation of Driverless Cars – A case for public subsidy of private transport systems

My family had a vigorous discussion over the Christmas break on driverless car technologies and the implementation timetable and pathway (yes we are like that, and if you don’t like it don’t turn up).

While we disagreed on the timelines there was general agreement that the technology is inevitable and desirable. My view was that there is a strong case for government subsidies to implement the technology which has some similar network effects as the fax machine: who buys the first fax machine?

Now, having a driverless car has some initial advantages, even if you are the only adopter. For instance if you can read/work/sleep instead of driving it is a great time saver while reducing your chances of having an accident. However the benefits of us all having driverless cars are far greater because network benefits accumulate exponentially as the number of vehicles with the technology grows.

This means that there is a significant case for a huge publicly funded effort for implementation to maximise early adoption rates. This was reinforced for me in the last week while reading several items:

The New Killer Apps: How Large Companies Can Out-Innovate Start-Ups

Audi’s traffic light assistance helps you hit every green light

The Men Who United the States: The Amazing Stories of the Explorers, Inventors and Mavericks Who Made America

In the New Killer Apps the authors describe some of the cost savings that implementation of driverless cars in the USA including:

“The American Automobile Association studied crash data in the ninety-nine largest urban areas in the United States and estimated the total accident-related costs— including medical costs, loss of productivity, legal costs, travel delays, pain, and lost quality of life— to be roughly $ 300 billion. Adjusting those numbers to cover the entire country suggests annual costs of about $ 450 billion. Now take 90 percent off these numbers. Google claims its car could save almost 30,000 lives each year on US highways, prevent nearly two million additional injuries, and reduce accident-related expenses by at least $ 400 billion a year”

Mui, Chunka; Carroll, Paul (2013-12-02). The New Killer Apps: How Large Companies Can Out-Innovate Start-Ups (pp. 19-20). Cornerloft Press. Kindle Edition.

They also go on to postulate that there would be other savings include fuel costs due to more efficient driving, and productivity improvements due to time saving. They also state that the demand for cars would be reduced by 90% due to improved utilisation of vehicles. While it is true there would be reduced demand for cars I highly doubt it would be at this level because the reduced demand theory is largely based on the fact that we only use use our cars a small percentage of the time. I no longer have a car for this reason and use Flexicar a local car sharing service. In Australia the data indicates we only use our cars on average 4% of the time and they lie idle the rest of the time. However the figure of 90% reduction in car demand is likely to be an exaggeration due to two factors:

  1. There will be a requirements for cars at peak times that will need to be filled, meaning that at other times there will still be a large capacity underutilisation.
  2. If we increase the overall capacity utilisation of our cars then they will not last as long. If we increase average car utilisation to say 20% then we will increase the mileage of our cars 5 times. In Australia that would mean moving average distance traveled to 70,000 km per year instead of the current 14,000 ( 9208.0 – Survey of Motor Vehicle Use, Australia, 12 months ended 30 June 2012 ). That means a 5 year old car would have traveled 350,000 km so changeover rates would be much higher. (there are some interesting design issues here – designing and building cars with greater durability while still allowing technology updates for instance)

There are clearly huge savings to be made in implementation of a true driverless car system if the Google assumptions are only partly correct.

In the Audi story the article states:

“Using both live and predictive data beamed into the vehicle’s navigation unit via onboard wifi, TLA doesn’t need a single camera to tell you when the light is going to change. Local data sources provide information about traffic light patterns, and the in car system uses that data and the motion of the car to predict exactly how long it’ll be until the green light goes red”

Clearly this does not work that well unless almost everyone is on the system. If drivers ahead of you are travelling too slowly for the system or brake suddenly then it would not be of much value. Also if you were travelling slowly to match your speed against when the next light would change and behind you was a trail of angry drivers trying to pass you then it could cause more problems than it solves. This magically disappears if all cars are on the system and fuel and time efficiency are gained as well as reduced accidents.

This is what I mean by network efficiencies. There must be a tipping point at which once there are enough driverless cars on the roads that benefits start to accrue more quickly and more adoption takes place. For instance if nearly all the cars on the road were driverless and communicating with each other then travel time information would be greatly improved. However the benefits accrue to different sections of the community rather than just accruing to the user, and accrue at different time frames, and there will be many self interested parties. The following are just a few examples:

  • Reduced accident rates mean a huge reduction in physical trauma and medical costs on top of the reduction in emotional trauma. This is largely saved in the government sector both in operating costs but also in continuing demand for new hospital facilities (this is also complicated by demographic changes, growth of cities, and urban intensification).
  • Individual car owners will save money in the longer term but will have the legacy costs of their current vehicles and their financing costs which may inhibit adoption and cause political backlashes. For instance if you new car is suddenly almost worthless and you have a car loan against the asset what do you do?
  • A number of sectors will miss out on income. The government will miss out on speeding fines and drink driving fines. Panel beaters, car insurers,and car manufacturers will all suffer significant revenue losses as will taxi operators and taxi licence holders.
  • If the general public came to the conclusion that large scale adoption of driverless cars was a good thing and about to happen in the next 3 years new car sales would plummet. Who would buy a new car today if it was virtually worthless in 3 years time?

Which brings me to Simon Winchester’s fine book,The Men Who United the States. In it he describes how a young Eisenhower was part of an army project to cross the USA by road in 1919 to test the capability the road system for military transport in case of war (Lt. Col. Dwight D. Eisenhower – Transcontinental Motor Convoy, 1919).Winchester claims that this experience led to Eisenhower’s long term commitment to the National road system which was later built at the cost of hundreds of billions of dollars and changed the nature of America.

There is a similar case for a large scale public investment in the adoption of driverless cars across the world. As many of the benefits accrue to government through lower costs in the health system then there is an overriding case for the government to get involved on several levels:

  • Implementation of the necessary technology systems outside of the cars themselves which link the cars to the rest of the transport system including traffic light systems.
  • A major effort to overcome any legislative barriers and risk issues, and coordinating national approaches to the problems. As an example the implementation of all this technology is likely to result in more accurate data on causes of accidents even if the overall numbers fall significantly. There will be cases where failures in the car technology causes an accident. In that case the manufacturers are likely to be held liable for the costs in that accident through the courts. At the same time the manufacturers would not accrue any of the benefits of the large reductions in accidents flowing from the technology adoption. There is a strong case for governments sharing those costs with the manufactures to reduce the costs of implementation ( I would be against indemnifying the manufacturers as they need some skin in the game).
  • Public subsidy of the system in a similar way that we subsidise private road use and public transport systems now but at least initially for a different reason. There is likely to be significant barriers to adoption of the technology which will be tied to initial costs and social attitudes. In a networked system such as large scale of adoption of driverless cars the advantages accrue much faster with higher rates of adoption. A pure business case can be made to government subsidising the system in the initial phase to significantly reduce costs and ramp up adoption rates with the payback being more rapid reduction in government costs.

Beyond all the economic arguments the human cost of road trauma is enormous and long lasting. As someone who was hit by a car 2 years ago and was lucky to escape with some serious injuries which I have mostly recovered from I have enormous sympathy for those who have not been so lucky. I was in hospital for 10 days and had 4 anesthetics and two lots of surgery but the day I left a patient in my ward was being moved to rehab after being in hospital for over 3 months, with the prospect of never walking normally again. I was able to compete in a triathlon again last Sunday in an embarrassingly slow time but at least I could finish. My thoughts go constantly to those who have not been so lucky.

My question is where are the visionary leaders of our time who will take on the huge challenge of implementing a system that can change the lives of thousands of people over the next 50 years? Who will hold the experience of meeting a severely injured car accident victim in their head in the same way Eisenhower held in his head the difficulties of crossing the USA in 1919 and set about changing the system?

Paul Higgins

Further Links:

Large-scale trial of driverless cars to begin on public roads

The world’s first large-scale test of driverless cars will involve 100 Volvos taking to the streets of Gothenburg in 2017


U.K. town will build driverless podcar system

Milton Keynes, a town of more than 200,000 people, announced that it will begin a pilot program for a transit system that uses driverless, electric podcars starting in 2015.

The £65 million pilot project will use 100 podcars (that can hold two passenger each) which can be summoned by a smartphone. The initial test will have the podcars travel on a one mile route between the city’s train station and shopping centers and offices. Each ride will cost £2. The pilot will run for two years and continue if the test run is positive, possibly even spreading to other cities in the U.K.

Further links posted up by futurist P A Martin Börjesson:

New IHS Automotive study forecasts nearly 12 million yearly self-driving cars sales and almost 54 million in use on global highways by 2035

The Driverless City



Volvo’s first self-driving cars now being tested live on public roads in Swedish city