How Energy Performance Certificate data could be really useful, but isn't

One of the frustrating things about discussing housing in this country is that we have historically lacked some key data which would allow us to compare ourselves more accurately with other countries. For example, there are no good statistics on the size of homes we're building now. It is commonly accepted that we build very small homes compared to other new countries, but the only some of the evidence for this is very out of date (see this Policy Exchange report from 2005 which cites these EU statistics from 2002 which cite English House Condition Survey data from 1996). It may well be true that we are building small homes at the moment but we just don't have good enough data to say for sure. [Update: I completely forgot about RIBA's excellent research on this very topic. Thanks to Rebecca for reminding me. So the data gap isn't quite as large I thought, though much of the below still applies.]

Relatedly, we can't really compare our house prices with those in other countries because the simplest consistent comparison, price per square foot or square metre, is not available to us. This matters to people who make housing policy, but it also matters to people thinking of moving house between different countries.

There is a solution in sight, however. The law requires an Energy Performance Certificate to be produced for every house that is sold or rented out. An EPC is drawn up by an expert after looking over the house, and captures key information about the energy efficiency of the house. But it also captures other information, notably the type of house, its size in square metres, and its exact address. There are now about 7 million domestic EPCs, all held on a single register and as of today searchable by address. I just looked up the EPC for a house down the road from me, which is the same kind of Victorian mid-terrace as I share with friends. It pretty much confirms what I thought, which is that our house retains heat about as well as a sieve.

To get back to my point though, what this means is that we've got a huge and growing database of home sizes. And because the Land Registry has recently started releasing its data on house prices, again with the exact address provided, it should be possible to link the two datasets together to calculate the average price per square metre in different parts of the country, for new as well as old houses. More sophisticated analysis could also reveal the extent people are willing to pay for for more energy efficient homes.

I don't know whether anyone in government is working on this. As far as I can see they're not, and that wouldn't surprise me as the key department (Communities and Local Government) is these days shedding statisticians and generally doing less analytical work.

But it should be possible for academics and laypeople to analyse the data in this way. The problem is that the government has decided that EPC data should only be available in bulk to certain organisations and only if they are prepared to stump up the money for it. The costs range from 1p to 10p per record depending on how much detail you want, but in any case this quickly mounts up if you want any kind of comprehensive database at local or regional level.

The government says these prices are to cover the costs of disseminating the data. Maybe that's fair enough and maybe it isn't, but it does mean that the kind of useful analysis I've described above can't be performed by anyone outside central government. So if the CLG are determined to ration access to the EPC data by price I think it should really be doing its own analysis and making the most of this data on our behalf.

10% of Inner London gets to work by bike

A while back I posted an analysis of Census data showing the trend in the proportion of people who cycle to work in Inner and Outer London from 1971 to 2001. We won't have equivalent figures from the 2011 Census for several months, but we can use another source of information on travel to work, the Labour Force Survey (which I mentioned yesterday when talking about trends in car travel).

The chart below shows the proportion of LFS respondents in Inner and Outer London who reported using bicycles as their main mode of transport to work between 2004 and 2011. One important thing to note is that in each year the data is from the October to December quarter of the survey only, so differences in autumn weather from year to year will affect the results somewhat. Also, as with the car data yesterday these figures are estimates based on relatively small samples of people, and so they have largish confidence intervals around them. This means that there are few if any statistically significant year-to-year changes - but for both Inner and Outer London the trend over the whole period is fairly clear.

In late 2004 around 6% of Inner London workers commuted by bike, rising to around 10% in late 2011. For reference, the share who travelled to work by car fell from 20% to 16% over the same period. A few more years of this and they'll be level pegging.

As we have seen before, cycling is much less common in Outer London, but at least it now seems to be rising, from a 2% share in 2004 to 3.5% in 2011. Not shown is the trend in the rest of the UK, but it's basically flat at around 3% throughout.

[Note: The LFS data here was downloaded from the Economic and Social Data Service]

People are using their cars less (but still quite a lot)

There has been much talk here and in other countries about the apparent declines in personal car travel seen in the last few years. Generally these are relatively small declines compared to the huge increases in car travel seen in previous decades, but if per capita car travel has peaked and is falling then that would at least be noteworthy from a social point of view, and possibly quite important for transport policy - see for example this letter from the head of the Transport Planning Society arguing that the Department for Transport's London traffic forecasts are out of whack.

The chart below shows my calculation of daily car travel in miles per person in Britain from 1949 to 2010, derived from these DfT statistics on car travel and population data from the Census and from ONS mid-year estimates.

According to these figures, per capita daily car travel peaked in 2004 at 11.7 miles and has been trending slowly downwards since then. A couple of caveats are probably in order at this point: we are in a recession, which usually reduces travel, and while these figures are per person, strong population growth could in future increase total car travel even if the per capita average continues falling. Finally, the reduction in car traffic is slightly offset by a rise in light van traffic over the same period.

We've also got data for London, though only going back to 1993. The chart below uses population data from ONS but published on the London datastore. The trend here is quite different, already flatlining in 1993 and falling fairly consistently from the turn of the millennium. While car travel per capita has fallen across London, the drop is particularly large in Inner London, down 28% over the period. The average Inner London now travels a shade under four miles a day by car, compared to just over six for Outer Londoners and eleven for the average Briton.

Lastly, much of the talk in the US is about whether car travel is particularly falling among younger people. To try and get a feel for this I looked at data from the Labour Force Survey on the main mode of transport people use for getting to work. Obviously the caveat here is that commuting is only a subset of all travel, but it's the best we can do for now. The chart below shows the proportion of people in broad age groups who reported travelling to work by car in 2004 (the earliest year I could find) and in 2011.

It's really important to emphasise here that this is based on a sample survey, so the estimates have confidence intervals around them (the little black lines). This means that in most cases the change is not statistically significant - including the apparent increase in car use among 16-19 year olds. The only age bands in which there was a clear, statistically significant change over the period were 25-29, 30-34, 35-39 and 50-54 year olds, all of whom were less likely to drive to work in 2011 than in 2004. So there's evidence of a fall in car commuting, but mainly by the 'young-ish' rather than the young (only around half of whom commute by car anyway).

What women want (when it comes to cycling facilities)

About 70% of cycle trips in Britain are made by men. But in countries where lots more people cycle, like Germany or the Netherlands, the gender split is roughly equal.  You can see the same relationship when comparing areas within countries, so it seems like a fairly robust link. Where cycling is a marginal activity it is mostly dominated by men, but where it is a normal part of everyday life women cycle just as much.

As this article in Scientific American from 2009 says, women are therefore a sort of "indicator species" for bike-friendly cities. But why? Well,

First, studies across disciplines as disparate as criminology and child ­rearing have shown that women are more averse to risk than men. In the cycling arena, that risk aversion translates into increased demand for safe bike infrastructure as a prerequisite for riding. Women also do most of the child care and household shopping, which means these bike routes need to be organized around practical urban destinations to make a difference.

So on average women seem to put a premium on safer cycle routes (that go somewhere useful). And the available evidence says this translates into a clear preference for facilities with a higher degree of separation from general traffic. When Jan Garrard and others carried out a survey of cycle use across a range of locations and infrastructure types in Melbourne, they found that "Consistent with gender differences in risk aversion, female commuter cyclists preferred to use routes with maximum separation from motorized traffic".

Similarly, when John Pucher and others looked at who uses which facilities in New York, they concluded:

Women in all five boroughs clearly prefer off-street paths ... the average percentage of women cyclists on paths is about three times greater than the percentage of women using on-street facilities such as bike lanes or simply bike routes on shared traffic lanes ... In short, the greater the physical separation from motor vehicle traffic, the higher the women’s share of cyclists.  

So the evidence seems fairly clear, and on that basis I think it would be good to see people arguing for more high-quality, separate cycling routes as a contribution towards greater gender equality.

But at the same time one probably shouldn't overstate the gender divide here. After all, it's not as if the proportion of men who cycle regularly in the UK is particularly high: it's probably fairer to say that current cycling facilities suit a tiny minority of men and an even smaller minority of women. Basically, it is risk-averse people (i.e. most people) who are discriminated against by our current cycling infrastructure, and if we had much safer cycling conditions we would see far more men cycling too. So while cycling infrastructure is a feminist issue in that the current setup particularly discriminates against women, it's also one of the many issues where feminist solutions would also benefit men in general.

Long-run trend in commuting into central London

[Cross-posted to London Transport Data]

The first statistics on commuting into central London were collected in the 1850s (of which more later), but the first figures comparable to the present date from around a century later. The chart below shows the trend since 1956 in the number of people (in thousands) measured as entering central during the weekday morning peak, broken down by whether they used rail (national rail, London Underground or TfL), bus or private transport (car, coach, taxi, cycle and motorcycle). NB, walking isn't included.



The number of morning commuters peaked at about 1.25 million in 1962 and then fell through most of the next twenty years. The pattern over the last thirty years is dominated by peaks and troughs linked to London's economic performance, with notable booms and busts in the late 1980s, early 2000s and in 2007-08. Rail is the dominant mode throughout this period, even more so in recent years, reaching 79% of the total in 2010. In fact the more interesting changes happened on the road and only really show up when you leave rail out. Detailed data on road traffic only starts in 1969 but the chart below interpolates back to estimates from 1961 to show the broad modal split of road commuting over a nearly 50-year span. It shows buses and cars twice swapping places as the dominant mode of transport for commuters, with bus ridership sliding throughout the 60s and 70s before shooting up again in the early 2000s. Interestingly, this latter shift seems to have started before the introduction of the congestion charge: the number of car commuters into central London fell by nearly a quarter between 2000 and 2002, before the C-charge was introduced in 2003.



The most notable trends in the last decade have been the continuing fall in the car share of commuting, and the rise in cycling. The chart below shows cycling's share of road commuting into central London since 1969. In the early 1970s cycling accounted for just 1% of road commuting (and therefore a much smaller share of total commuting), but by 2010 this had risen to 12%. Given the combined motorcycle/cycle figure in 1961 was 13%, it seems fairly plausible that cycling now accounts for a higher share of central London commuters than at any point in the past. Also, if current trends continue (a big if) it won't be long before more people are coming into central London on two-wheelers than in cars.



I mentioned at the start that these kind of statistics were first collected in the 1850s. This refers to a survey by Charles Pearson, who hired 'traffic-takers' to stand 'at all the principal entrances to the city of London, to take their station from eight o’clock in the morning till eight o’clock at night' and count the number of persons and vehicles leaving or entering the City over the twelve-hour period. The City was a much larger part of 'London' in the 1850s than it is now, and Pearson measured somewhat different flows and used a different methodology, but his results, shown in the table below, are still fascinating.  

Estimated number of persons and vehicles going into and out of the City daily in 1854, counting them all both ways.

Omnibus	88,000
Other vehicles	52,000
River steamers	30,000
Via Fenchurch St and London Bridge rail	54,000
Foot passengers	400,000

Railways were still in their infancy and there was no Tube yet, but the most striking result here is how many people walked into Central London. That's not so surprising, as London was much smaller and denser than it is today so most people would have been within an hour's walk of the City. It's frustrating that we don't have comparable figures on walking today (at least, not that we could easily find) but as the city is so much more spread out you would expect walking's share to be much lower, though still significant.

Recent academic research on cycling safety

The annual conference of the Transportation Research Board in the US features a huge number of papers and presentations across the whole spectrum of transport studies, and in recent years has included a good few about cycling safety. I looked through the papers from the 2011 conference and the abstracts for the upcoming 2012 one and found some interesting stuff, summarised below.

First, from the 2011 conference:

Bicycle Route Choice Model Developed from Revealed-Preference GPS Data (Broach et al)
Using an analysis of bike-mounted GPS data, finds that "cyclists are sensitive to the effects of distance, turn frequency, slope, intersection control, and traffic volumes. In addition, cyclists appear to place relatively high value on off-street bike paths, enhanced neighborhood bikeways (bicycle boulevards), and bridge facilities."

Disaggregate Exposure Measures and Injury Frequency Models of Cyclist Safety at Signalized Intersections (Miranda-Moreno et al)
Analysing data on cycle collisions and traffic flows, finds that "cyclist collisions were sensitive to changes in both cyclist and motor vehicle flows. A 10% increase in bicycle flow was associated with a 4.4% increase in the frequency of cyclist injuries. A 10% increase in the total number of motor vehicles that passed through the intersection would result in a 3.4% increase in cyclist injury occurrence."

The impact of bicycle lane characteristics on bicyclists' exposure to traffic-related particulate matter (Kendrick et al)
This one's about whether on-street or segregated cycling facilities expose cyclists to more air pollution. "Ultrafine particle exposure concentrations are compared in two settings: (a) a traditional bicycle lane adjacent to the vehicular traffic lanes and (b) a cycle track design with a parking lane separating bicyclists from vehicular traffic lanes. Traffic measurements were made alongside air quality measurements. It was observed that the cycle track design mitigates ultrafine particle exposure concentrations for cyclists."

Cyclist safety on bicycle boulevards and parallel arterial routes in Berkeley, California (Minikel)
Compared cycle collision rates on main arterial routes with 'bicycle boulevards', traffic-calmed side streets signed and improved for cyclist use. Found that "collision rates on Berkeley's bicycle boulevards are two to eight times lower than those on parallel, adjacent arterial routes. The difference in collision rate is highly statistically significant, unlikely to be caused by any bias in the collision and count data, and cannot be easily explained away by self-selection or safety in numbers."

And now here are a couple of papers due to presented in the 2012 conference this January:

Exploring Factors Influencing Bicyclists' Perception of Comfort on Bicycle Facilities (Li et al)
Detailed survey of cyclists in Nanjing finds that "the environmental factors significantly influencing bicyclists' perception of comfort included the width of path, presence of grade, presence of bus stop, physical separation from pedestrians, surrounding land use, and bicycle flow rate. For on-street bicycle lanes, the contributing factors associated with perception of comfort included the width of bicycle lane, width of curb lane, presence of grade, presence of bus stop, amount of occupied car parking spaces, bicycle flow rate, motor vehicle flow rate, and rate of use of electric bicycles. The results suggested that bicyclists perceived a higher average comfort on physically separated bicycle paths as compared to on-street bicycle lanes."

Safety of Urban Cycle Tracks: Review of the Literature (Thomas)
Review of studies into cycle tracks from various countries concludes that "one-way cycle tracks are generally safer than two-way and that, when effective intersection treatments are employed, constructing cycle tracks on busy streets reduces collisions and injuries. The evidence also suggests that, when controlling for exposure and including all collision types, building one-way cycle tracks reduces injury severity even when such intersection treatments are not employed."

I think if there's an overall conclusion to draw from these it's that cycling facilities which don't require cyclists to mix with motor vehicles are likely to be significantly safer, for reasons of air quality as well as reduced collision risk (which I think most people accept anyway).

Contributory factors in cyclist casualties

Mark King of the Guardian writes:

DfT statistics reveal that the biggest single contributory factor in cycle deaths is the cyclist failing to look properly (25% of fatalities), followed by failing to judge the other person's path or speed (10%), the cyclist entering the road from the pavement (8%), and careless or reckless behaviour (8%).

This gives the impression that errors by cyclists themselves are the dominant factors contributing to cyclist deaths. And that would be the wrong impression, because (a) the figures quoted refer to all accidents, not just fatal ones, and (b) by definition they exclude any actions of motorists or other non-cyclists. The rest of this post goes into these reasons in more detail.

The statistics Mark King is referring to are from DfT's 2010 road casualties report here. Table RAS50005 in that report shows 'Contributory factors reported for vehicles in reported accidents' and under 'Pedal cycles' the top 'Contributory factor attributed to vehicle' is indeed 'Failed to look properly' at 25% of the total number of cases. But these figures don't mean what Mark King thinks they mean.

First of all, it should be fairly clear that this table doesn't just deal with fatal accidents. There are 11,929 incidents reported under the 'Pedal cycle' category, and I think most people know that there weren't 11,929 cycling fatalities in 2010. So that's one rather important mistake.

The second issue is more subtle but is important to understand. As the DfT report makes clear in the paragraph directly above the table in question, these figures are about the contributory factors attributed to vehicles involved in accidents, not the factors behind each accident. Put another way, the table describes what vehicles of each type did, not what caused casualties of each type. So the 168,990 figure at the end of the 'Car' column indicates the number of cars involved in reported collisions, not the number of car driver/passenger casualties (the fact that the figure is labelled as 'Number of vehicles' and not 'Number of casualties' is a clue).

So the figures quoted don't show all the factors contributing to cyclist casualties, but only what cyclists did in collisions involving cyclists, which will include a number of collisions where there were no cyclist casualties at all (e.g. involving pedestrians).

Unfortunately DfT don't seem to publish statistics showing the contributory factors to cycling casualties. Fortunately, the Transport Research Laboratory published a report in 2009 which does just this (in fact, Mark King mentions it in his article).

Briefly, the TRL report lists 430 collisions where a cyclist was killed (between 2005 and 2007), of which 354 were reported as involving a collision with another vehicle (figure 7-2). In 183 of these fatal collisions there were contributory factors attributed to cyclists, of which 31% (which equates to 57) involved the cyclist failing to look properly (table 7-4). 57 is 13% of 430, so you can probably say that cyclists failing to look properly was reported as a contributory factor in 13% of cyclist fatalities, not 25% as Mark King says. It's also worth noting that figure 7-6 shows that drivers failing to look properly was reported as a contributory factor in an almost identical number of cases (44% of the 131 cycle fatalities where a contributory factor was attributed to drivers). As the TRL report says, "attribution is split fairly equally between the cyclist and driver/rider of the motorised vehicle".

Finally, it is really important to understand that these figures are just based on what police officers subjectively reported when they were called to the scene of a collision. As the TRL report states very clearly, "this does not necessarily mean that the cyclist or other road user was in fact responsible". The media and the rest of us should bear that in mind when using these figures.

Some estimated cycling casualty rates for London and other cities (a work in progress)

Update! I've added figures for Helsinki (thanks Tommi) and Vienna (from here). Details are in the accompanying spreadsheet.

It would be very interesting and useful to have some statistics allowing us to compare the safety of cycling between different cities around the world, but this kind of thing is very hard to come by. As far as I can see it's nobody's job to collect statistics from cities in different countries, so at present nobody does it. In some places, individual city governments/authorities/mayoralties do collect and publish statistics, but only some of these are comparable with other cities. This post is the start of an attempt to collect together figures which we can use to compare London to other cities in terms of cycling casualty rates.

Please note, these figures are just initial estimates, cobbled together from available sources, some of which may not be correct. I would very much welcome any better information, and especially any similar information for other cities. If there's something here you think is wrong, instead of getting annoyed why not send me a correction?

The table and chart below show some summary information on distance cycled, cycle casualties, and the cycling casualty rate (per distance cycled) in Amsterdam, Copenhagen, Berlin and London (and now Helsinki and Vienna). This spreadsheet gives more detail on the sources, which where possible are official statistics but in a couple of cases (e.g. the total distance cycled in Amsterdam per day) are not.

	Amsterdam 2009	Copenhagen 2010	Vienna 2008	Berlin 2009	Helsinki 2008	London 2010
KM cycled per day (m km)	2.2	1.21	-	4.82	0.82	1.73
KM cycled per year (hundred millions of km)	8.03	4.05	2.76	17.58	2.98	6.32
KSI per year	104	92	71	538	170	467
KSI per 100m km per year	13	23	26	30.61	57	74
Million km per KSI	7.7	4.4	3.9	3.27	1.8	1.4

For the casualty rate I've used the number of cyclists killed or seriously injured per 100 million kilometers cycled per day in each city. I know the casualty rate per km isn't very good for comparing the relative risks of different modes of transport where the average trip length is very different, but it should be fine for comparing cycling casualty rates between cities. I guess you could also calculate a casualty rate per trip, but then you get into how to define a 'trip', which isn't as straightforward as you might think.

I should probably say that this certainly isn't an attempt to make London look like a dangerous place to cycle. While I think it could certainly improve on that score, I'm sure there are many, many cities with worse casualty rates. But most of them probably don't collect any robust, comparable statistics. There is a selection bias of sorts at work here, in that it is generally those places which have fairly strong cycling cultures which collect the best statistics on cycling.

I'd love to add some more cities, but comparable data is hard to come by. I looked around for data from Paris and found absolutely nothing, while I couldn't find any statistics on distance cycled in New York. Any help on these two or on any other cities would be really appreciated.

Another important caveat: In each case I've taken the city borders as read. Measuring the extent of a city is notoriously difficult, everyone does it a bit differently, and in most cases statistics are just based on administrative borders rather than any 'functional area'. So it maybe that with these statistics each city is defined differently, for example with some including suburbs and some not. I know that the London area used here includes a lot of suburbs, but I don't know whether the other cities do. I'd be interested to hear from anyone who does and who thinks this might significantly affect the answers.

Cycling trends in London since 2001

A couple of weeks ago I wrote about how cycling to work became more common in Inner London and less common in Outer London between 1971 and 2001. This provoked an excellent post on the 'cycle of decline' in Outer London from David Arditti (see also here).

Now I'm going to look at what information we have on more recent cycling trends. The results of the 2011 Census won't be out for another year or thereabouts, so instead I'm drawing on the Department for Transport's National Road Traffic Survey data, which you can access here. DfT count the average daily flow of traffic on major roads throughout the country on a yearly basis, and there's more detail on how it works here. I used some of this data for a previous post on trends in cycling at Kings Cross.

DfT's data includes traffic counts at nearly 1,900 count points on major roads in London between 1999 and 2010, with a few gaps here and there. The usual major categories of motor vehicle are counted, as are cyclists. And by averaging all the counts in a particular borough you can get an idea of the trends at borough level, with the caveat that the data doesn't cover minor roads.

The first chart below shows the trend in DfT's average daily count of cycles between 2001 and 2010 in Central London (Westminster and the City of London), the rest of Inner London and Outer London.

So what we see is an increasing divergence between the inner city and the outer suburbs over the last decade. The average cycle count increased by about three-quarters in Central London and doubled in Inner London, but rose by only about a third (122 to 159) in Outer London.

The chart below (click to embiggen) gives some borough-level data, showing the percentage change in average daily cyclists count by borough, with the Inner London boroughs in blue and the Outer London in red.

Update: I have added another chart, this time showing cycling's share of total traffic in each borough in 2001 and 2010. This shows that in the case of some boroughs (like Barnet) which have seen fairly high proportional increases the actual level of cycling (compared to overall traffic) is still quite low.

Again there's a stark contrast between Inner and Outer London boroughs, though cycling increased in nearly every borough over the period. Barnet and Greenwich saw relatively high increases in cycling for Outer London boroughs, while Westminster had a relatively low increase for a central borough (contrast with the City).

My post on Census data included a chart which identified a fairly close relationship between increases in car use and declines in cycling. The chart below plots the percentage point change in cars' share of average traffic flows versus the change in cycling's share, at borough level, and this time the relationship looks even stronger, with areas where cycling increased the most also seeing the biggest falls in car traffic (note, 'car' traffic here doesn't include goods vehicles). Those boroughs in the top left corner (with both big decreases in car traffic and big increases in cycling) are City, Lambeth, Southwark, Islington and Camden. The two Inner London boroughs where the share of car traffic has risen are Newham and Tower Hamlets.

Finally, here are a couple of charts which show the change in the share of cycling and car traffic at each count point, arranged by distance in kilometres from the centre of London (Trafalgar Square, to be exact). There's a good bit of noise at the level of individual count points, but the overall patterns are still fairly clear - the roads closest to Central London generally saw the largest increase in cycling's share of total traffic and the biggest drop in car traffic's share.

The urbanisation of cycling in London

I've just published a post over at LondonTransportData about what I think are some very interesting trends in the proportion of people who cycle to work in London. The data is from four successive Censuses, 1971 to 2001, and at London level it looks like this:

Not all that interesting, but go below regional level and you get this.

What we're seeing here is that whereas cycling to work was twice as common in Outer London as in Inner London in 1971, by 2001 it was the other way around. Rates of cycling to work fell for 40 years in Outer London and rose in Inner London.

Here's the pattern at borough level:

As I say on LTD, two things stand out here. First, boroughs with relatively high cycling rates in 1971 tended to see decreases over the next 30 years. Secondly, Inner London boroughs nearly all saw an increase while Outer London boroughs nearly all saw a fall. The only Inner London borough that didn’t increase its cycling rate over this period was Newham.

On the basis of this data it looks like cycling in London changed from a mostly suburban to a mostly urban phenomenon over the course of a few decades, although it's worth bearing in mind that it remained very much a minority pursuit in both areas throughout the period. It would be nice to know why this change happened, and I'd be interested to hear any theories anybody has, bearing in mind that it all happened the introduction of the congestion charge, and mostly before we had much in the way of red routes, or 20mph zones.

Here a are a few theories on what might have caused this turnaround:

Traffic: There was less motor vehicle traffic in both Inner and Outer London in 1971 than in 2001. Perhaps as a result cycling conditions in the suburbs were relatively benign, as there was more space on the roads than in the inner city. But as traffic increased, two things happened: in Inner London, rising congestion made commuting by car painfully slow, so the bike became a more attractive option for some. But in Outer London, increasing traffic just increased the collision threat faced by cyclists, and most of them stopped as a result.

The Census data provides some tentative support for this hypothesis: as the chart below shows, car commuting increased much more in Outer London over the 40 year period, and across London there was a strong negative relationship between the change in car commuting and the change in cycle commuting.

Infrastructure: Maybe it was all about changes in the provision of cycling facilities in Inner and Outer London, with cycle lanes and paths being added in Inner London and taken away in Outer. Or maybe there's a link with the first theory, in that the lack of infrastructure in Outer London didn't matter when there wasn't much traffic but became a deterrent when there was.

Labour market: Maybe it was the pattern of commuting that changed, with longer journeys less suited to the bike becoming the norm in Outer London over time. I'm sure there's data out there on trends in commute lengths, but it might be hard to disentangle cause from effect.

Of course, it could be a mix of the three, or something else altogether. But I think anyone interested in growing cycling in London needs to think about why this happened. I'd also love to hear whether there's any evidence of a similar pattern in other cities.