From twod@umich.edu Sun Oct 27 10:27:12 2002
Date: Sat, 26 Oct 2002 14:47:52 -0400
From: twod <twod@umich.edu>
To: twod@umich.edu
Subject: New Scientist | Evolution of the Car

ORIGINAL URL:
http://www.newscientist.com/hottopics/cars/article.jsp?id=22944900&sub=Socio-economic%20issues

TITLE: 
Whose roads are they anyway?

ABSTRACT:
Every day, millions of cars choke us to death in city centres the world
over. Ironically, while many people neither own nor drive one, the car
has proved so seductive that it has turned into one of the most
environmentally destructive forces of our time. One man who thinks we
can't carry on like this is Phil Goodwin of University College London, a
leading expert on traffic congestion. He wants Britain to be the envy of
the Western worldby charging motorists to drive in city centres. And
London should soon become the largest Western city to experiment with the
idea. Not, as Goodwin explains to Mick Hamer, that he expects an easy
ride . . .
(09/06/2001)
   
By Mick Hamer

Many people think the simple way to solve the congestion is to build more
roads. What do you think?

In Britain the government is planning to increase spending on building
new roads and widening others. This will cost 60 billion. But the
government's own forecasts show that the effect on the inter-urban road
network will be to reduce journey times by a second per mile by only the
end of the decade. Averaged over the whole road network, the saving is a
quarter of a second per mile each year. The normal variation in travel
times is around 20 per cent, so your journey time could already vary by
15 seconds per mile from day to day, let alone year to year. Real people
can't easily distinguish fractions of a second.

If road building won't solve congestion, what will?

The defining relationship in transport is the speed-flow curve. This
shows that the more traffic there is on a road, the slower it goes. There
are three ways out of that. One way is to build more roads to try to keep
up with increasing traffic. That has been tried for half a century or
more. A second way is to organise the road system more efficiently.
There's always scope for squeezing a bit more out of the existing network
by optimising traffic signals and one-way systems and so on. The problem
is you end up with a system that is very close to capacity, and that is a
recipe for chaos. The closer the network is to capacity, the bigger the
jams when something random goes wrong, such as a burst water main.

So what's the third way? Would it involve charging motorists?

You need to build in some redundancy, by controlling the volume of
traffic. That's been the dominant intellectual advance for the past
decade. The idea that you could solve congestion only by charging
motorists or only by improving public transport so people would
voluntarily leave their cars at home is mistaken. It's the word "only"
that undermines it all. It's like building a dam halfway across a river.

So you need both carrots and sticks?

The carrots are vote-winners but don't work on their own. And the sticks,
if you're serious enough, do work but they don't give you a long enough
political life to see them through.

They've tried pay-as-you-drive in Singapore and Oslo successfully. How
will you persuade Britons to leave their cars at home?

The traditional wisdom on road pricing was that you could prove that if
prices reflected the true costs of travelincluding the cost of congestion
and pollutionthen the system would be more efficient and everyone would
benefit whatever the government spent the revenue on. This was a
political disaster because what it did was create a medium-sized group of
people, including many motorists, who are utterly hostile to the idea
because they lose from it, and a very large group of peoplethe
majoritywho benefit from the revenue the government has to spend, but
indirectly and invisibly. So I argued that the revenue should be spent on
improving the alternatives to motoring so it increases the support for
road pricing.

If road pricing is taken up outside London, would it be introduced by
central government or by local councils?

Local councils. Unless it's done this way it simply won't happen. The
idea of road pricing has been around before. You launch into it with
great enthusiasm and the closer you get to it the more the opposition
builds up and you just kick it into another Royal Commission or study.
It's happened over and over again. I argued that local authorities should
be allowed to introduce itgiving them the incentive that they could keep
the money and spend it on public transport. Most local authorities would
be scared and only a few would go ahead. These cities would show how well
it works and this would build up public support. It's been much the same
with pedestrianisation, which I count as one of the great success stories
of the past 25 years. Every city worried whether banning cars would
destroy its city centre and create chaos, but now we all know it works. I
think it will be the same with road pricing.

What effect do you see it having in London?

At the moment the discussions are about a rather narrowly defined area
and a non-scary price level and, of course, the smaller the area and the
smaller the price, the less impact it has. But I think it's a question of
what process does it start? Which direction does it take you in? You
could imagine the thing actually working quite nicely with the increased
pricing reducing traffic levels in certain important areas and then a
radical expansion of bus priority networks. I think we've gone about as
far as we can go with the disconnected bits of bus lanes. The thing is to
put it all together and create a network that has to go, obviously, over
a much greater area than simply central London. Together with the rail
investment programmes, you could get a complete transformation.

What would you do about the other part of your strategy, improving public
transport? Commuters have a lot to put up with.

Yes, and unreliability is probably the main criticism. It may well be
preferable to have a lower average speed that you can rely on. I don't
think we've had a level playing field for about half a century. The
system has favoured car use, which has meant that traffic has grown more
than it should have done. It's time to shift the balance in favour of
public transport, cycling and walking, and a more coherent strategy for
managing the volume of road traffic. This could deliver a considerably
better outcome than the government's spending plan.

What started you off on this line of thinking?

There was a classic piece of research back in 1964 that attracted me
because the message was counterintuitive: if everyone travelled by the
slowest form of transport (bus), they get to their destination faster
than if they all travelled by the fastest method (car). Because buses
carry more people than cars, if you persuade motorists to travel by bus
you cut the number of vehicles on the road and speed up the traffic. It
was the first time I'd seen such a fascinating argument. Unexpected
results are always interesting.

So is the way we model transport all wrong?

Yes. One of the key assumptions in older transport models, for example,
was that the total volume of traffic is fixed. If you built a new road
the only thing that could change was the route that drivers chose. This
was initially only a simplifying assumption, a way of getting the
mathematics to work, but it became established thinking and was used for
assessing road schemeswith seriously distorted results.

You mean that the forecasts actually assumed that new roads wouldn't
generate traffic?

Yes. And they also meant that if you closed a road, or pedestrianised an
area, or put a bus lane in, then there would be no reduction in traffic.
We know now that both are wrong empirically. I contributed to two studies
on this. If you build more roads you get some extra traffic. The reverse
is also true: if you reduce road capacity you get less traffic. They're
obviously two sides of the same coin, but it's not yet clear whether
they're exact mirror images. The speed of or size of response may be
rather different.

Are there other false assumptions in these transport models?

All the widely used models assume that when you observe traffic levels at
a point in time the whole system has reached stable equilibrium. And when
you forecast traffic levels for 2010 or 2020, they will have reached a
new equilibrium. I don't think this is tenable. Research shows it can
take up to 10 years for people to respond fully to changes in petrol
prices, fare hikes or new services. In some senses you could say it's
never complete but, for practical purposes, the response to a major
change is greatest in the first 3 to 5 years and broadly complete after
10 years. Once you look at travel choice as a dynamic process, it emerges
that the sensitivity of the market to price changes is probably twice,
maybe more than twice as great as once thought.

Can you give an example?

Take buses. For as long as I can remember the industry has assumed that a
10 per cent increase in bus fares means you lose around 3 per cent of
your passengers shortly after the increase. So putting up the fares is
worth it because you can pocket a 7 per cent increase in revenue. But in
the long term, fares increases can lose you 8 or 9 per cent of the
passengersand in some segments of your markets you will lose more than
you gain. But these losses may be happening years later so you blame it
on something else, such as the state of the economy. And the public
transport industry, for most of our lifetimes, has eroded its own market,
year by year. That has to be reversed.

Do we need new and different models for the 21st century?

Yes. The current methods don't help you understand how to change people's
behaviour, or how long it takes for effects to build up, or in which
order policies need to be implementedoften the three most important
practical questions. Most scientific modellers are now used to asymmetric
relationships, chaos and non-linear systems. Economic models take lags
into account as a matter of course. So in the first years of the 21st
century it would be nice to bring transport methodology into the last
quarter of the 20th century.

[shim.gif]

From New Scientist Magazine   09 June 01