High-speed rail versus driverless cars

My column in The Times on British transport
priorities:

By the time HS2 is fully operational in 2033, more
than a quarter of all cars on our roads will be fully autonomous,
according to a forecast by the consultants KPMG. That may well make
fast trains less urgent, and decongested motorways more so. The
economic case for HS2 is fragile enough before taking future
driverless cars into account.

Last week on the very same day that a House of Lords committee
savaged the economic case for the HS2 railway
— costing £50 billion with contingency — another report by KPMG, for the Society of Motor
Manufacturers and Traders, estimated the potential benefits to
Britain of driverless cars at £51 billion. Per year.

The two are connected.

The Lords report, chaired by Lord Hollick, urged the government
to make incremental improvements to the existing rail network
instead, plus better links between cities in the north. It said it
had not seen evidence that the capacity constraints on the west
coast line warrant building a line to take trains that would go
faster — at 250 mph — than the fastest trains in China, France,
Germany, Italy, Japan and America.

The report also said the cost-benefit analysis for HS2, which
came out with a marginal result, relied on evidence that was “out
of date and unconvincing”, depending for example for its valuation
of non-work travel-time savings on a study done in 1994. About four
fifths of the putative benefits of HS2 are derived from the value
placed on work and non-work travel time — yet these do not even
take into account the fact that time on a train can be used
productively, especially so in the age of the mobile and wi-fi.

Pretty well the first thought in my head when I decide to take
the train for a journey of 100 miles or so, rather than drive, is
that I can use the journey time productively: one of the biggest
attractions of a train is the fact that you don’t have to drive it.
If you didn’t have to drive a car, so a road journey could be
productive (and not spent listening to, say, Test Match
Special), the attraction of the train is less; the car,
after all, can start and end at your front door. In effect, the
driverless revolution could bring the chauffeur-driven experience
to the many.

No doubt we would love to have both driverless cars and faster
trains, as well as HS3 for the north, and a bunch of new airport
runways, and lots of improvements to the road network too. But we
cannot afford everything. (Incidentally, as the Lords report says,
the benefits of HS2 will be received mostly by businessmen, so why
is the government paying for it, not business?) HS3 will cost more
per mile than HS2, so a combination of the two high-speed rail
lines is quite unaffordable. We have to prioritise.

In Britain, roads are just as full as trains. But there is no
innovation on the horizon likely to transform the demand for rail
travel as much as the driverless revolution could transform the
demand for road space.

In cities, driverless cars could cut congestion. A recent simulation at the University of Texas of
a city with driverless cars prowling for business found that
passengers need wait an average of 18 seconds for a driverless
vehicle to show up and that each shared autonomous vehicle could
replace 11 conventional cars. A study by Columbia University concluded that a
driverless vehicle fleet could cut the cost of transport by 80 per
cent compared with a personally owned vehicle driven 10,000 miles a
year — not counting the reduction in parking costs and the value of
time not spent at the wheel.

However, I suspect that on routes between cities, the advent of
driverless cars could increase congestion. When taking an
autonomous car allows you to work, motorway traffic is almost bound
to increase. An older person, for example, currently deterred by
both the difficulty of driving on the M1 through the rain, and the
effort of getting to a station and walking to a train, might today
decide to stay at home.

If in 2030 she can be collected from home and driven direct to
her destination 100 miles away while reading or snoozing, then she
is a bit more likely to make the trip. Since motorways are even
more congested than railways already, that is surely where the
problem will lie. We should be planning now for more lanes on
motorways in 2030.

Driverlessness will arrive in stages. Although Google’s and
other driverless vehicles have clocked up impressive trial journeys
without problems, and for all the experiments going on in Coventry,
Greenwich, Milton Keynes and Bristol with fully autonomous vehicles
right now, I don’t expect to be able to buy a fully driverless car
for about 15 years.

But already half of all cars are connected: I have an app that
lets me check where I left my car and how much is in its tank. It
tells me when I am wandering out of a lane and warns me of
congestion. It tells the garage if I have a fault. Some cars
already park themselves, though the “valet” system where you leave
the car at the entrance to a multi-storey, tell it to find a space
and pay is some years off.

Some cars already have adaptive cruise control, slowing down
automatically if they get too close to the car in front. Soon cars
will have “traffic jam assist” so they can take over in stop-start
traffic at slow speeds. By 2020, KPMG reckons, some cars will have
autopilot on motorways, but the driver will have to be ready to
resume control. By 2025 there should be lightweight, driverless two
-person taxis patrolling cities. Overtaking or navigating rural
lanes may take a long time to master, as will swerving for cats. By
2030 perhaps we can expect fully autonomous cars that can go
anywhere.

They will never be flawless, but nor are drivers. Insurance
needs sorting out. Yet KPMG reckons that the driverless revolution
may save up to 2,500 lives by 2030, and points out that Britain has
a technological head start in all the relevant industries, so there
is every reason to think we can become a centre of excellence in
connected and autonomous driving, and get 320,000 jobs out of
it.

Alongside this kind of stuff, I just cannot help feeling that a
very fast train, built at glacial speed (half a mile a week) over
many years of consultation, review and challenge as it punches
through Nimbyland, and at up to nine times the cost per mile of
French high-speed rail, feels like a white elephant waiting to
happen.