Superhighways, congestion, and selective statistics

The transport correspondent of the Financial Times, Robert Wright, has done an interesting post on his blog, challenging what he calls “the now-conventional wisdom among London cyclists that the 12 miles of new cycle superhighways in central London…have had no significant effect on congestion.” As he puts it:

Traffic volumes entering Central London fell 3.4 per cent between the June to September quarter in 2015 and the same quarter in 2016, part of a long-term decline that’s seen the volume of motor traffic entering central London decline by more than 20 per cent since 2000. Instead of increasing with declining traffic volumes, however, average traffic speeds in central London – the easiest available proxy for congestion – fell 3.5 per cent, to 7.8mph.

The figures (from the latest TfL streets performance report) are accurate, but they are also selective. The key indicator of congestion is in fact a quite different measure (also given in the same report) called journey time reliability. It’s the key indicator because it’s not merely a “proxy” for congestion: it measures actual congestion, tracking thousands of vehicles on ANPR cameras and comparing how long they take to make their journeys against how long they should take if the route were not congested.

And guess what? In the same quarter-on-quarter comparison, congestion in central London in fact fell and journey time reliability improved – by 2.1 percentage points in the evening peak and 0.1 percentage points in the morning peak.

screen-shot-2017-01-18-at-04-56-50

That’s actually better than the performance of several roads which are wholly outside central London and nowhere near any segregated superhighway – such as the Blackwall Tunnel approach, where JTR over the same period either improved by less than in central London, or indeed worsened.

screen-shot-2017-01-18-at-05-04-11

Which points to a second difficulty in blaming the superhighways for congestion – the difficulty in ascribing effects to particular causes. It does, though, seem unlikely that segregated cycle tracks totalling 12 miles can be causing more than a small portion of the congestion on a London main road network which totals around 1500 miles.

It seems much more likely that a rise in traffic bears more of the blame. For it is also true – but also selective – to say that central London traffic fell over these particular twelve months. Over the last three to four years as a whole, however, motor traffic in central London has increased quite sharply – by 4.5 per cent between 2013 and 2015 alone, for instance. The “long-term decline… since 2000” in motors entering central London stopped in about 2013/14.

Finally, it is not quite true to say, as Wright also does, that “private cars now account for only 18 per cent of motor traffic during weekdays in the central London congestion charging zone.” They account for 18 per cent of all traffic, but 21 per cent of motor traffic. If you add taxis (20% of all traffic; 23% of motor traffic) and private hire vehicles (12% of all traffic; 14% of motor traffic) the total comes to a rather trickier 50 per cent of all traffic, and 57% of motor traffic.

screen-shot-2017-01-18-at-05-07-17

It’s tricky because taxis and PHVs are even more inefficient users of roadspace than cars. A car, at least, is always taking somebody somewhere. Taxis and PHVs are often coming back empty from dropping someone off. Taxis are often cruising around empty looking for passengers.

The stupidity of all this really comes into focus with figures last month showing that though cars, taxis and PHVs comprise 50 per cent of the traffic (across the day) in central London, they now account for just five per cent of commuter journeys into central London.

screen-shot-2017-01-18-at-05-12-48

(The above two tables are from the latest TfL Travel in London stats, published in December.)

What this adds up to is not to indulge the silly argument that some (such as Vincent Stops, or his representative on earth, the Guardian’s Dave Hill) are trying to start between bike infrastructure and buses, but an urgent need to go much further in reducing what really clogs the central roads, inefficient motorised traffic. My old boss, Boris Johnson, asked TfL to look at making PHVs – the huge growth area of recent years – pay the congestion charge, and asked the government for powers to cap their numbers. Both these things appear to have been dropped by Sadiq Khan; certainly, there was no mention of them in his recent taxi and private hire action plan.

More broadly, as Wright says and as Johnson wrote in March, the recent rise in traffic badly demands a wider rethink of the congestion charge – the one policy instrument with a proven record of reducing congestion. It needs to be sharply increased, or made smarter. Sadiq ruled out increasing the charge in his manifesto, apart from a “toxicity” supplement on the most polluting vehicles (about 10 per cent of the total).

Khan hasn’t shown much sign so far of being able to take difficult decisions. But maybe, with his “most ambitious plan in the world” to tackle air pollution now outflanked in ambition by the plans of many other cities, he will find the will.

Advertisements

4 thoughts on “Superhighways, congestion, and selective statistics”

  1. Even if it where true that motor vehicles numbers fell despite population increases and congestion increased despite motor vehicle numbers falling, wouldn’t the logical conclusion be that it were increases in cycling that caused the drop in motor vehicles and that therefore increases in congestion would have been even worse without that modal shift?

    Like

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s