Monday, 12 August 2013

The Cobra Effect

The Cobra Effect arises when an attempted solution to a problem actually makes the problem worse.

As the story goes, the British colonial government in India had become concerned by the number of venomous cobra snakes in Delhi, and offered a bounty for every dead cobra they received.

Initially this strategy was a success, as significant numbers of snakes were killed for the reward. Eventually, however, the government became aware that a large number of "enterprising" persons had been breeding the cobras, solely for the purpose of claiming the bounty. The reward program was therefore quickly scrapped.

And so the snakes—now worthless—were set free, as a result of which the cobra menace in Delhi worsened by several orders of magnitude.

* * *

This is an instance of 'unintended consequences', a concept which was popularised in the twentieth century by the socialologist Robert K. Merton.

In a paper entitled The Unanticipated Consequences of Purposive Social Action, Merton attempted to apply a systematic analysis to the problem of unintended consequences, and suggested that possible causes include the world's inherent complexity, perverse incentives, human stupidity, self-deception, failure to account of human nature, or other cognitive or emotional biases.

Unintended consequences can be roughly grouped into three types:

i. A positive, unexpected benefit: For example, the medieval policy of setting up large hunting reserves for the nobility has allowed the preservation of many green spaces throughout London and elsewhere. Likewise, the Korean Demilitarised Zone has created a large natural habitat.

ii. A negative, unexpected detriment occurring in addition to the desired effect of the policy: For example, in 1990 the Australian state of Victoria made safety helmets mandatory for all bicycle riders, and as a result, the number of cyclists fell. Thus, whilst there was indeed a reduction in the number of head injuries—in part because there were fewer cyclists—fewer cyclists leads to fewer injuries—a health benefit model developed at Macquarie University in Sydney suggested that the decrease in exercise caused by reduced cycling as a result of the helmet law was counter-productive in terms of net health.

iii. A perverse effect contrary to what was originally intended: For example, Theobald Mathew's temperance campaign in 19th century Ireland—in which thousands of people vowed never to drink alcohol again—led those who sought to become intoxicated, without breaking the letter of their pledge, to consume diethyl ether.

* * *

Following the recent death Philippine de Gerin-Ricard, her mother told the BBC: "In situations like this, people always say they will do something. We will do something. I would like to see a deadline, so things change, so no other family goes through what I am going through. I wouldn't wish this on anyone, not even my very worst enemy."

Whitechapel Road near the junction with Vallance Road, 2013
Photo credit: Google StreetView

The cycle route which runs along the length of Whitechapel Road / Mile End Road / Bow Road is currently identified as CS2, but it used to be LCN+ Route 194, and before that, LCN Route 11. It is even marked in the LCN 2002 map as a "Cycle route, with route number".

The blue routes indicate a "Cycle route, with route number". Comparatively,
the routes coded in pink show a "Part-completed cycle route", and the
routes coded in yellow show a "Proposed cycle route."

What does it mean, a "Cycle route, with route number"? If the photo below is anything to go by, not very much at all.

Whitechapel Road near the junction with Vallance Road, 2008
Photo credit: Diamond Geezer

If this is an acceptable standard for cycle routes on the LCN, then why not just go ahead and get all of the network up and running? No seriously, why not?

"Most new ideas," said Hayden Phillips recently in The Times, "can be stopped in their tracks by deploying one, or sometimes both, of two arguments: 'the thin end of the wedge' and 'the time is not ripe'."

For the London Cycling Campaign and its supporters it's always 'the time is not ripe'. Always. For them, there is always something else much more important to be done, and this would be diminished in some way—not enhanced—if a 'network first' approach was pursued.

In February 2006, for example, the Borough Cycling Officers' Group reaffirmed their interest in a comprehensive but low-engineered cycle network. This information was passed on to TfL.

Four months later a senior TfL officer met with the LCC to discuss this proposal. They "advised that their priority is for TfL and the boroughs to complete the LCN+ by 2009/10."

This was further clarified in an email dated 31 July 2006: "I believe the reason why Simon Brammer at LCC did not ask their Planning & Engineering Committee to look at Simon Parker's proposal for a wider network was due to LCC's view that the first priority should be completion of LCN+. LCC's concern is that a wider network could (a) divert borough and TfL engineering and planning resources from LCN+ completion by 2009/10 and (b) divert financial resources."

The upshot of this was that in light of the views expressed by the London Cycling Campaign, TfL could not proceed with evaluating the case for my proposal.

* * *

It is worth taking a moment to consider the LCN+. In 2001, the LCC persuaded Ken Livingstone to abandon the original LCN. The details of this were announced in a paper entitled ‘Review of Provision for Walking, Cycling and Area Based Schemes’ and dated 5th February 2002: "There will be a change of approach in taking forward the London Cycle Network based on a slimmed-down network focused on direct, high demand, high quality routes reflecting key strategic commuter routes. The revised network will be re-branded as LCN+. It is estimated that the LCN+ network will be about one-third of the length of the planned full LCN."

Google Maps fall down if there is too much route information, and in the case of the map 
above, not all of the LCN+ routes are displayed. Actually, there is not that much 
missing—just a few bits here and there—so this map does give you a fair idea of the 
extent of the network. If you are interested to see all of the routes, please click here.  

By 2005, however, it had become obvious that unless there was a change of gear in delivering the network, it may not be completed until at least 2016—seven years behind schedule.

An investigation led by Darren Johnson remarked: "Progress of the network is not measured by how cyclists can get from one place to another along a completed route, but by the length of route completed. Over-emphasis on completing easy sections has created a piecemeal network."

"A key part of the change from LCN to LCN+," Ralph Smyth noted, "is the use of Cycle Route Inspection Meetings ('CRIMs'), bringing together engineers, cycle campaigners and other representatives, such as from bus companies. These are followed by a Cycle Route Implementation and Stakeholder Plan ('CRISP')."

As reported on, a lot of money was spent on the LCN+, but to very little practical effect. Presumably this prompted a senior LCC member to comment, with a shudder: "One of these days people are going to ask for their money back."
* * *

After a suggestion from David Hembrow that the last thirty-five years of campaigning in London have been wasted, David Arditti responded: "We have a lot that we did not have when the campaign began. Physically, one can point to some of the better pieces of infrastructure, like Camden's Somers Town Route and Bloomsbury cycle track, and Cycle Superhighway 3 in East London, and smaller features like safe bike crossings of major roads and junctions at Hyde Park Corner, Bayswater Road, Knightsbridge, Swiss Cottage, Strand, and other places, and allowed cycling paths through the Royal Parks, all features used by thousands of cyclists every day. We had none of this 35 years ago. It's all been achieved by campaigning."

Even if we accept that this little list was all achieved by campaigning—I don't, by the way, and since David has asserted it, the onus is on him to prove it, and not on me to disprove it—I acknowledge that the cycle crossing into Hyde Park from Knightsbridge was an early LCC success, and that the Somers Town route and Bloomsbury cycle track were delivered in large part thanks to the efforts of a few individuals within Camden Cycling Campaign, but not the rest, not without some proof at any rate—even if, as I was saying, we accept this list fairly reflects the fruits of the LCC's labour, it's still astonishingly mediocre.

More than this, it's just bits and pieces. More than this, it's all LCN stuff. David Arditti wasn't able to point to a single example of the LCN+ working well. I'm with David Hembrow on this one: "Many years, no progress, repeating the same things."

* * *

"The productive way," says David Arditti, "is to start by implementing small bits of the network to a high standard, to what you want as the final standard, if possible, so everyone can see the way forward, leaving gaps where there is no implementation (but the roads have not been left worse than they were before the network was defined)."

Gulp. Leave gaps in the network, was that? Well yes. If those gaps were filled (with temporary measures, say), "it would undermine the campaign because it would turn existing cyclists away from the whole idea of campaigning for separation from motor vehicles." And besides, in developing a network to the point where it functions, there would be "no real safety increase, but money is spent and political capital wasted."

As has been noted: "A perennial problem in cycle route network planning is the reliance on bright ideas and pet projects that may not have been critically evaluated for usefulness and value for money."

For David, the big issue is that attempts to develop a cycle network in London "never tackled the issue of subjective safety, and that is why they failed." A lot of his thinking about cycling in UK towns and cities appears to be informed by this premise.

For me, however, the main problem is that the cycle network was never properly planned and then made to function. Rather than develop a cycle network one piece at a time therefore—beginning with an improvement to a dangerous junction here, and then, say, an extension to an existing cycle route over there, and so on, all the way through literally thousands of changes—and doing nothing at all in between places—because, because, because—Cycling: the way ahead points out that the network "can be introduced on the basis of an overall plan (preliminary plan)".

That is to say: "All the installation measures which call for little planning may be applied without major risk of error or loss [...] Given their low cost, the small amount of extra work which they entail and the possibilities of corrections in the case of error, such measures may be adopted automatically."

It continues: "Even if the impact of such measures is not massive, it will be real."

David disputes this. He writes: "Defining the network and then putting crap implementations [i.e. 'soft' measures] on parts of it (which is in fact the CSH and generally was the LCN approach) is actually counterproductive, undermines the whole campaign, and is in no way a step forward."

* * *

Since David mentions the Cycle Superhighways, let us briefly consider them.

To see a map showing the LCN+ / CS routes combined, please click here

The map above shows all of the CS routes which were not part of the LCN+. Only one of the four routes which have thus far been delivered—CS7, from Morden as far as the Elephant & Castle—is genuinely new.

CS2, CS3, CS4, CS5, CS8 and CS9 were originally proposed as part of the LCN+, and CS2, CS3 and CS9 were originally proposed as part of the LCN. So some of these routes have been a very long time in the planning.

With the exception of CS3—which used to be LCN15 / LCN13—all of the CS routes installed to date function at a minimum level.

David says: "The counterproductive way is to try to bring the network all up in standard in one go, in one phase, but do everything badly."

Of course, I am not proposing that everything be done badly, and in suggesting otherwise, David is simply throwing shit at this strategy in the hope that some of it might stick. The idea is to do as much as possible as quickly as possible with a view to getting the network to function, and thereafter to develop it further "on the basis of priority interventions and a timetable".

* * *

After Boris Johnson promised to meet the three key tests of the LCC's Go Dutch campaign, Ashok Sinha spoke of his delight at the Mayor's commitment "to learn from the successful Dutch model". It’s fantastic news, he added.

The three key tests are:
  1. Implement three flagship Love London, Go Dutch developments on major streets and/or locations. 
  2. Make sure all planned developments on the TLRN are completed to Go Dutch standards, especially junctions. 
  3. Make sure the Cycle Superhighways programme is completed to Love London, Go Dutch standards.
If these 'tests' have anything at all to do with "the constant factors of a thoroughly understood cycling policy", then I'm a purple-sprouting broccoli.

         Spot the differences? One of the main ones is actually the build costs.
         Photo credit: Crossrider

David Arditti reports that at the recent Cycle City Expo conference in Birmingham, there was a very interesting session in which Johan Diepens of the Dutch Cycling Embassy gave a talk. More or less the last thing he said was that in planning for cycling, the critical thing is to design your network correctly. Everything else, he said, is trivial.

David interpreted this to mean that once you have decided on "your basic repertoire of designs, your units of the network, your standardised way of treating main roads and junctions, all you have to think about is an optimised network."

Assuming, for the sake of argument, that the LCN+ was "an optimised network" (in the sense that it was "big enough", and the routes were "sensible and direct enough"), it is noteworthy that the LCN+ was characterised by:
A socially inclusive cycling environment where quality standards are maintained by applying the TfL London Cycling Design Standards (LCDS)
Routes that are continuous, safe and comfortable with minimal delays.
Clear guidance using surface treatment and road markings where there is a potential conflict between cyclists and other road users.

Now, David might very well argue that the "quality standards" set by TfL were not high enough, but actually, that is completely beside the point. The point is that, according to David, you would use those standards—whatever they are—"as a bricklayer building a house from a plan works with standard bricks and blocks".

In the case of the LCN+, that clearly did not happen. And if David is suggesting that we repeat the LCN+ experiment—except with a better-designed network and much-improved standards—then, as defined by Einstein, it would be "insanity" to expect a different result.

At the beginning of 2012, I wrote to David and made the point to him that he hadn't provided me with one shred of evidence to counter Dekoster and Schollaert's claim that the level of minimum functioning is a prudent course to follow.

He replied: "What Dekoster and Schollaert meant by this "minimum level of functioning" I can't tell for certain, as I haven't talked to them. But I think this a problem of the concept getting lost in translation. They are Dutch (or Belgian), so I think what they meant was a fully segregated, or almost fully segregated, cycle network, with all the necessary links safely working. Maybe some bits a bit too narrow, too slow, indirect, or quality not great, but the required subjective safety at every point. So nothing like LCN, LCN+, the Superhighways, or anything else in the UK."

I haven't spoken to them either, but given that they were talking about where to begin, and not about where to end up, I'll bet anyone a pound to a pinch of dust that when they talk about a minimum level of functioning, they mean a minimum level of functioning.

* * *

Chris Mason from Bristol was also at the Cycle City Expo conference, and he recalls a presentation given by some consultants who had come up with a convoluted way to direct cyclists around a junction. "This had proved very complex," Chris explains, "and the consultants couldn’t understand why cyclists didn’t use the new off-road routes. Diepens chipped in with a classic piece of Dutch wisdom which amounted to—Why don’t you just make it quicker and easier for cyclists? Then you won’t need these complicated signs that nobody obeys."

That is to say, routes need to be meaningful and direct before they are anything else, otherwise they won't get used. Identifying routes which are meaningful and direct is something that can only be done during the planning phase, and making them safe and comfortable is something that can only be done during the development phase.

* * *

On 15th June 2006 Simon Brammer wrote to me as follows: "We have asked our Cycle Planning and Engineering Committee to have a look at [your] proposal, but only in relation to your recommendations for coloured signage."

         The map on the left (mine) is dated August 2005. The map on the right (LCC's)
         is dated March 2010. I put east-west routes in red because the sun rises in the
         east and sets in the west, and north-south routes in blue because blue is a cold
         colour and the North and South Poles are cold places. Very cleverly, the LCC
         took this idea, turned it on its head, and then claimed the credit.

"In terms of routes," Simon Brammer added, "there has already been significant investment in these and LCC has no plans to challenge these. This is a matter that you would need to take up with the LCN+ team."

Map showing selected LCN routes abandoned when the LCN+ was launched.
To see a map showing these routes plus the LCN+ / CS routes combined, please click here

More recently Danny Williams has written: "Let's look at Cheapside by way of an example of what's going wrong in the City."

"Would you ever let a child cycle on this road if it was open to motor-
traffic? The painted cycle sign is absolutely useless. There is no
excuse for not providing a proper continuous segregated cycle lane here."
Words by George Johnston. Photo by Mark Treasure.

* * *

Another good example of this sort of problem is Lancaster, which, as Dave Horton reports, used to be a Cycling Demonstration Town.

I left a comment on his blog to say that with a similar budget, and over a similar timescale, several German towns and cities were enabled to “introduce” an entire pro-cycling policy (network, information and promotion). (This was reported in Cycling: the way ahead, so I guess this would have been back in the 1990s some time.) I made the point that if we were to visit those same towns now, I have no doubt that their cycling environment would have come on in leaps and bounds.

Dave responded by saying that "similar seeds were sown" in Lancaster, except that, for various reasons, they "withered and died" here, but were "cultivated and thrived" in Germany.

But were they, in fact, similar seeds? No, they were not. According to Dave's own testimony: "The project’s goal to enable more short trips to be made by bike very quickly got transformed (because it is easier to do) into a focus on making the district a more attractive cycling destination (i.e. the project became about using cycling to sell the area as a tourist destination, much more than about doing the difficult things required to get local people cycling)."

Thus: "It’s certainly true that the off-road routes tend to be seen as, and to act as, the centre-pieces of our local cycling network; but they obviously don’t constitute a network."

And now, sadly, the fact that Lancaster ever was a Cycling Demonstration Town is being erased. "Why?" Dave asks. "Are we embarrassed that we might once have celebrated cycling, and imagined things could be different? Or was cycling’s celebration only a momentary blip for so long as the money lasted, and we’re now back to business- (and motoring-) as-usual?

"And who is responsible for this local institutional erasure of cycling? Central government pulled the plug prematurely on what always needed to be a long-term process. Local government was unable to develop and institute effective strategies to support cycling once centralised funding ended. Also, what seems an obvious prerequisite for long-term success—the development of broad and deep civil society support for cycling through building links at a grassroots level—was never an important objective of the Cycling Demonstration Towns project."

When I look at those routes which used to be part of the LCN+, I see similar evidence of erasure. In the case of Tottenham Court Road, the cycle route was never even made to function in both directions. Hardly surprising then that within a couple of years of the LCN+ programme drawing to a close, a gilt-edged opportunity to develop a more permanent solution ended up not being taken. This had precisely nothing to do with funding, by the way, and everything to do with planning, and the political will to see those plans through.

I had two big problems with the LCN+. Firstly, it didn't seem to me to be at all necessary to abandon two-thirds of the LCN simply in order to create a slimmed-down 'spine' network of cycle priority routes. Secondly, it was abundantly obvious that the complicated engineering works that were a necessary feature of the LCN+ would take time to deliver.

On 26th August 2004, I wrote to Ken Livingstone reminding him of this. "Whilst this process is ongoing," I continued, "it ought to be remembered that there are already 650,000 cyclists in London, and their needs are more pressing than the needs of would-be or might-do cyclists. So perhaps it would be more prudent to sort out a network that would be useful to existing cyclists first."

Nine years later, and there are surprisingly few advocates of cycling who think this is a point worth making.

* * *


* * *

Immediately following the death of Phillipine de Gerin-Ricard, Andrew Gilligan pointed out: "Changes to something as complicated as the road network have to be thought through; they cannot happen overnight, or even in the four months since the launch of the Vision. The worst outcome would be a botched instant solution which actually made things more dangerous."

The worst outcome? Worse than what? Worse than a young woman losing her life? No, come on.

Nobody is suggesting that changes can happen overnight—not even cycling's lunatic fringe—but where is the urgency? Awake, Samson, get up from thy bed! The Philistines be upon thee!

In February 2012 TfL launched their Better Junctions Review. They completed their initial review in July 2012, and based on such indicators as user feedback, cyclist numbers and collision data, they were enabled to identify their "100 priority locations". This included the Aldgate Gyratory and Whitechapel Road / Vallance Road, plus the section in between.

That was a year ago. Then what happened? According to the LCC, "Both the London Cycling Campaign and the Mayor's Cycling Commissioner, Andrew Gilligan, have rejected proposals put forward by Transport for London to improve [this junction] because they haven't been good enough."

So that's like, what? Good enough never is? I wrote to the Mayor on 4 December 2012, and quoted from a Robert Peston report (which I know the Mayor heard, since he was asked to comment about it immediately afterwards). "True risk taking," Robert Peston said, "is about analysing risks with great care but to a tight timetable. It's not about ignoring risks."

I concluded my email by quoting Churchill: "The maxim, 'Nothing but perfection', may be spelled, 'Paralysis'."

* * *

There is absolutely no dispute from this blog that the development of more permanent solutions is going to take relatively a long time. My question is: what should be done until then?

In Holland, the cycle networks are developed around a combination of main road routes (treated) and back street routes (traffic-calmed). So it's not a case of one or the other, but a combination of the both, and then let users decide which route suits them best.

In London, cyclists effectively have "no choice other than to cycle through ridiculously designed road junctions", whereas in the Netherlands, they have two choices: they can either avoid the junction altogether, or they can cycle through the junction using well-designed facilities.

Until such time as the latter option is available, what should be done? Having identified all of the "dangerous connections", do you then show them, in their place, on a cycle network map? Do you show all of the alternative routes, so people can make their own mind up about how best to get from A to B? Do you invite experts to come and speak to you, and to address you at your conferences, and then do the exact opposite of what they say? No? So what do you do then?

No comments:

Post a Comment