The following is an investigation of the flawed evidence of the effects of LTNs on pollution and active travel

There are five main pieces of evidence constantly quoted for LTNs leading to traffic evaporation, reduction in pollution and increased active travel by Better Streets for Enfield, London Living Streets and various other activist groups. They are:

1. The counts from Walthamstow Village before and after this LTN was implemented.
2. The Kings Report on the benefits of air pollution reduction in Waltham Forest Borough
3. The Cairns report studying 60+ cases of traffic displacement worldwide
4. Rachel Aldred's papers on increased active travel when living in, or near an LTN
5. The Walthamstow Village Final Report

Let’s look at each one in detail.

1. The Walthamstow Village counts.

These showed a substantial reduction in counts inside the LTN, as expected, after it was introduced and an increase in counts on perimeter roads also as expected. The claim is there were approximately 14000 less counts per day inside the LTN and an increase of only 4000 counts per day on the perimeter roads therefore the claim is 10,000 counts have disappeared. 10,000 as a proportion of the total traffic counts (inside plus outside) of 64000 gives a traffic evaporation of around 16% as quoted by groups such as Better Streets For Enfield and London Living Streets. This is STATISTICAL NONSENSE on four levels:

 

1. These are counts and not cars. A car travelling through the LTN was counted on every road it passed down so a car passing through this particular LTN (if you look at a map of it) would have typically been counted on 2 or normally 3 roads. There was only one count point on each perimeter road. This means you cannot simply subtract one from the other and call them all cars, they are counts not cars. Typically a good estimate is a car passing through the LTN would average around 2.8 counts. As some cars would have travelled, after LTN closure, on two of the perimeter roads a car outside the LTN averages around 1.6 counts. Unless you know every cars pre and post LTN routes you cannot be accurate but these reasonable estimates would reduce the 16% evaporation to around 9%, still significant but...
   
   
2. A car entering the LTN before it was closed off may well have been counted on the perimeter road before it entered the LTN. When it could no longer take this route, all its counts from inside the LTN would have disappeared but when it stayed on the perimeter road as extra traffic it would NOT have been counted as extra traffic because it had already been counted on this road before the LTN was introduced. ie 2 to 3 counts will have disappeared for these cars from inside the LTN and no addition counts will have appeared on the perimeter roads. It's like cars were just evaporating, except they didn't.
   
   
3. Only the three immediate perimeter roads were counted for extra traffic. The counts themselves showed increased traffic and congestion on these roads yet no effort was made to assess the traffic stress on other roads, (Wood St. in particular), as cars took alternative routes. It is obvious to all who are experiencing the effect of the Fox Lane LTN that as the immediate perimeter roads become backed up that some traffic takes a wider berth. Winchmore Hill Rd, Wynchgate, Morton Way, Broomfield Lane, Hoppers Road and many roads further afield have all shown considerable increased traffic, none of which was counted before the LTN and likely not to be counted at the end of the trial. The same happened in Walthamstow Village but because this additional traffic was not counted it is included in the percentages as traffic that just evaporated, except it didn't.
   
   
4. Some of the traffic that was originally counted on the perimeter roads pre LTN, that didn't even enter the LTN at any stage, may well have diverted to other routes once these perimeter roads became congested. So they were counted pre LTN on perimeter roads and not counted post LTN so, hey presto they just evaporated, except they didn't.

 

None of the above proves there was not traffic evaporation. (i) shows statistically it was far less than often quoted and (ii) to (iv) shows it was probably a lot less, if anything of any significance at all. The latter statement is backed up by the figures for the borough overall. The increase in car traffic flow for the Waltham Forest borough between 2015 and 2018 was 11.9% (416 mill miles to 469million miles from Dept. of Transport stats file TRA 8902). 2015 is when the Walthamstow Village LTN was introduced and by 2018 this and other LTNs in Waltham Forest were fully bedded in. This is not very consistent with car evaporation. Clearly there may be many other factors at play here when you look at the borough as a whole, but you might expect these to affect neighbouring boroughs as well, yet Redbridge had 6.8% increase, Enfield 3.7% , Harringay 5.9% and Newham 9.6% over this same period.

 

In addition, the British Heart Foundation in a recent survey placed Waltham Forest as the second worst authoritative region for air pollution in the country, again not consistent with traffic evaporation.

2. The Waltham Forest Kings Report

This is often quoted as a scientific paper, it is not. It is a commissioned report that Waltham Forest paid Kings College London to produce by running figures supplied by the Enjoy Waltham Forest Team through modelling using the Londonair maps. They then converted these benefits into increased life expectancy following changes to pollution levels between 2013 and 2018 and projected to 2020. Part of the paper was to model typical walking and cycling trips through the borough and show how the improvement in pollution levels between 2013 and 2020 had led to increased life expectancy. Had it been a scientific study it would have done a control experiment by running similar walking and cycling routes in other Outer London boroughs. It produced pollution maps showing how NOx and Suspended Particulate Matter had reduced in the borough between 2013 and 2020. It adjusted these models slightly for some new 20mph speed limits and some road closures, and reduced further the pollution levels in the models since Waltham Forest supplied figures suggesting an increase in active travel to school.

 

However if you zoom out on the Londonair maps you will see the improvement in all the air pollutants was London wide. Had they done control experiments they would have got almost identical results for these control groups in any Outer London Borough. According to Londonair the London wide decrease in air pollution was almost certainly down to each successive new diesel engine model being much cleaner and better filtered than the previous generation, together with a small input latterly from the Inner London ULEZ. The extra reductions from the 20 mph may have had a small input but any tiny input from car reduction from the school run would be negligible and as we have seen from the figures above, those figures supplied for the modelling were almost certainly wrong as Waltham Forests traffic went up significantly more than its neighbouring boroughs during this time period. Given that, the life expectancy increase claimed are wrong and would have been bettered by the other boroughs who didn't have LTNs. The report did however allow Waltham Forest, and various pro LTN groups to produce pretty pollution maps of Waltham Forest and infer that the better air quality and improved life expectancy was due to their Mini Holland measures, which it wasn't. Note again the British Heart Foundation survey result above.

3. The Cairns Report

The Cairns paper looked at over 70 cases where closures to roads or road adjustments had led to traffic displacement to try to discover what happens to this displaced traffic and how much of it simply disappears. In the end they settled on 63 cases and claimed a median average of around 11% traffic evaporation. A median average is where you line up your 63 examples in order of percentage traffic disappearance and the one in the middle, in this case the 32nd in the list is your median. Now let’s look at some of these 63 cases.

 

1. (i) In 28 of the 63 cases (44%) no counting of traffic was done on the surrounding area, they just used the figure for the decrease in traffic in the area/roads closed or restricted with no account for displaced traffic.
   
   
2. In some cases there was only partial counting of displaced traffic. For example the 1 month closure of Tower Bridge. They used figures that took into account extra traffic on four others bridges into the City of London but not The Rotherhithe Tunnel, the closest crossing to Tower Bridge on the east side.
   
   
3. Some examples were double counted. In 9 cases the same event is counted twice, just over a different time period. Some schemes are included for effect on traffic after 6 months and included again, as a different one of the 63 examples, for the effect after 1 year. It may be useful looking at the change in figures over time but to include both sets of figures in an overall average, particularly when finding a median, is basically double counting. As there was more double counting of schemes above the median this is raising the median figure.
   
   
4. There were several examples used which included natural disasters (Earthquakes, bridge collapses etc.) but no account was taken in these for the economic effect of the disaster that may have affected traffic flows. For example the Hanshin-Awaji earthquake (Kobe) where the main elevated road collapsed. Traffic flows when it was restored were less than half before the earthquake. No mention of the economic effect caused by the destruction of 400,000 properties and the major destruction of much of the port. There was also the fact there was a reluctance among many to use the restored elevated highway they had seen topple like a house of cards just a year earlier.
   
   
5. One example was a pre-planned ONE DAY closure of the A13 Road (Rotterdam) on a Saturday. Is this really an example of traffic displacement/ evaporation comparable to any permanent proposal ? Oh and a second of the 63 examples was a pre planned closure of the same road a week later. A third was, yes you've guessed it, the same road closed the week after that. These count as three of the 63 examples. They shouldn't be in there at all but any decent scientist would have averaged these into one example. And guess what, they were all above the median average for traffic evaporation. Just removing these three cases and the Kobe earthquake immediately reduces the median to 8.4%.
   
   
6. Eleven of the 63 examples showed an OVERALL INCREASE in net traffic with more additional traffic being found on the alternative routes than had gone missing from the closed/ restricted route. This is suggesting the road changes were causing more people to take to their cars, or more likely there is a huge margin of error in all these studies.
   
   
7. From above at least 41 of the 63 examples have a question mark against them, of the remainder 11 showed an increase in net traffic.
   
   
8. Finally there is a huge question mark over how the percentages used were calculated. The figures used in the paper were the reduction of traffic observed as a percentage of the traffic that had originally passed through the area now closed or restricted, NOT as a percentage of the whole traffic. If you take for example a case in Leeds which was exactly on the median average quoted, its traffic reduction as a percentage of the original traffic within the restricted area comes out at 10.6%, the figure used by Cairns. If you represent it as a percentage of the traffic in the whole measured area, including the roads traffic was diverted onto, it reduces to 2.5%.

 

In a further example, if the Walthamstow Village scheme highlighted above, with its flawed statistical calculations, had been included in the Cairns paper they would have taken it at face value, trusted the local officers to have correctly identified ALL the possible alternative routes traffic would likely to have been diverted onto (which they clearly didn't) and then expressed the percentage traffic evaporation not as 16% shown above but 39%.

 

lot of commitment would most likely attract individuals who were very pro or very anti the changes being made about which they were being surveyed. This 24% figure is for the base rate data in the initial 2016 survey. By the time of the first year follow up survey there had been an average drop out rate of about 50% . Except that the drop out rate was very uneven. 58% of the original mailshot recruitment dropped out while only 38% of the cycling database dropped out. So by the time you could do any meaningful measurements conservatively over 28% of your sample population came from cycling groups in an area where 3% of the population cycle 3 times a week or more (Outer London cycling rates from Sport England as cited by Rachel in this same paper). Although some adjustments were made for the drop out rates, is this a representative sample population ? In particular is this a representative sample when your survey is about changes in active travel?

 

The survey was also very under representative of the BAME community with only one third of the expected number in the surveys sample population compared to the general London population. The mailshot group and the Oystercard group would have been expected to reach the 33% of Londoners who identify as BAME. However, of cyclists in London, 86% of male and 94% of female cyclists are white. So the massively over large representation of cyclists in the survey may also have led to the BAME imbalance.

4. Rachel Aldred's paper on increased active travel

This was what is known as a longitudinal survey where the same group of people are followed over a period of time to assess any change in their behaviour. Unlike a simple public survey which might take 15 minutes to complete, these individuals were asked to effectively keep a diary of walking, cycling and car use over a period of around 5 weeks in May-June 2016 and then the same period for the following 3 years after Mini Holland style road changes had taken place. The author obviously had difficulties in finding a large enough sample group who would make such a commitment. As such, her final paper was littered with "p values" greater than 0.05. The "p value" is a statistical test you carry out on the data to rule out something known as the Null Hypothesis. It's generally accepted that if p is greater than 0.05 it fails this test. If you can't rule out the Null Hypothesis then any proof you are trying to demonstrate, any changes in the data, could be there by random chance.

 

But this is not the only problem with the paper. In any paper involved in surveying people, the results are only meaningful if your sample groups are a true representation of the population in the area you are assessing. Normally you have a sample base of a number of individuals from which you then generate your sample population on a random basis, whilst also ensuring it is a balanced similarly to the total population of the area. In this case the response to the initial mailshot was only around one percent. It's not surprising nearly 99% did not respond given the commitment involved. As the sample numbers weren't large enough the authors then turned to two Transport for London databases containing people who had responded to surveys before and were willing to do so again. One was through the Oystercard database, which you would expect to be a fair representation of random Londoners. The other was a TfL database of cycling groups !

 

As a consequence 20% of her total sample were from the Tfl cycling database. In addition there were another 10% that came from the two TfL databases where they could not identify which of the two databases they had come from, (why I don't know as they sent out the emails recruiting them from the databases). Assuming this unknown section was in proportion to the Oystercard/cycling database ratio this would add another 4% from the "unknowns" that were actually from the cycling database. So at least 24% of her initial sample were from cycling databases. In addition a proportion of the initial mailshot may also fall into this group as the response to such a survey that required a

5. The Walthamstow Village Final Report

£27 Million pounds was spent on about half a square kilometre in the Walthamstow Village scheme with much of the money being spent on road closures and traffic calming. There were other areas of improvement such as street lighting, tree planting and general visual improvements. Many figures and quotes have been taken from this report by groups such as BSfE and London Living Streets suggesting it was a tremendous success. I have already dealt with the claims of traffic evaporation above but I urge all councillors to read this report in full. In doing so note how biased it has presented the data, allowing for pro LTN groups to cherry pick information and give a good impression on what was effectively a failed scheme. I will list some of the many examples:

(i) You will often see quoted only 1.7% of those surveyed in the final report said they would scrap the scheme. This was NOT asked as a survey question. It comes under the category of "Residents comments: Additional adjustments that could be made to the scheme". So 1.7% of those who responded (25 people from 1483 sampled) were moved enough to actually write "you should scrap the scheme" or something similar as a comment. In the same section, "I'm happy with the changes" got a figure of 0.6%. Yet this 1.7% figure appears throughout the report including in the executive summary as if it had been asked as a question on the survey. The 0.6% figure, strangely, didn't get a mention.

1. To the question "Perceived change in regular journey quality" the report says "Approximately 64% of respondents felt the quality of their journey had stayed the same or improved". Now actually 52% had said it had stayed the same and 12% said it had improved. If you are spending £27 Million shouldn't this 52% be regarded as failure as no improvement has been achieved for them despite the huge amount of money spent ? 498 people actually said their journey quality was WORSE compared to 166 who said it had improved, exactly three times as many. This data could have, and should have been, reported as "Approximately 88% of respondents felt the quality of their journey had stayed the same or decreased".
   
   
2. You will often see the figures of a 19% increase in walking reported and a 28% increase in cycling ( Actually these should be 16% and 25% because they didn't subtract the people who said they walked or cycled less). However the figures for car use stayed EXACTLY THE SAME. If people were walking and cycling more it wasn't as an alternative to using their car, so no likely traffic evaporation from this source. The figure for increased cycling looks impressive but its from a very small base. Only 4% initially said they were regular cyclists so this may have risen to 5%. However, there was no indication as to whether any of this small additional cycling was new cyclists or all done by existing cyclists cycling more. As there was no change in car usage it would suggest the latter.
   
   
3. From 1,352 respondents almost an identical number said they perceived a decrease in traffic volume (34.5%) as said they perceived an increase (33.8%). A difference of 13 people.
   
   From 1,389 respondents 27.6% said their overall perception of their street had increased, 27.5% said it had decreased. A difference of 2 people.
   
   The logical conclusion is that the benefits for some were to the detriment of others, with no overall improvement. Sound familiar ?
   
   
4. They asked businesses for their overall opinion of the scheme and reported this as "54% either having a positive response or not expressing an opinion" where the figures show 47% had a poor opinion and 20% had a good opinion. They called this a mixed response.
   
   
5. There was a generally better received opinion of other improvements, especially around the commercial centre, and for spending on street lighting, foliage and general improvements of the physical appearance of the area. The visitors to the area were also surveyed and seemed generally to have a much better opinion of the improvements than the residents. When asked how the visitors journeyed to Walthamstow Village 1% said they walked, 0% cycled. When asked if they had changed their mode of transport to the village since the improvements, only 2% said they had, although obviously none to cycling and few, if any, to walking.
   
   
6. In the summary of Addition Adjustments it states that "55% of residents said they would not change anything". This could be the biggest deceit of all.

 

It appears that if you left this box for suggested additional improvements blank you were classed in the category “No/Nothing/None”. If you didn't comment it appears that you were classed in the same group as someone who actually wrote "I wouldn't change a thing."

 

Now I have no knowledge of whether our councillors have independently read all these reports and knew everything above already or whether they have simply believed what others have told them about the conclusions and are sitting there wondering why the traffic isn't evaporating as it should. Indeed there may be many at BSfE or London Living Streets that have been taken in by the "spin" from Waltham Forest, and the work of the full time lobbyists and non-independent statisticians.

 

I urge any of the decision makers, or indeed any members of BSfE or LLS to at least to read the final report from Walthamstow Village, and see what they really got for their £27million. I mean the FULL report, not just Waltham Forest’s summaries. Look at the actual figures and then and how they reported it.

 

Now from a personal point of view I am not against properly configured cycle lanes. Cycling is a great form of transport for those who can and wish to cycle and they deserve to be protected as made safe as possible. I am also not against promoting active travel in any sense. The benefits to everyone in reduced pollution, less congestion and the benefit to the individual of the exercise is obvious. It should be promoted by education and encouragement. But it is clear from Rachel Aldred’s “LTNs For All” paper that LTNs are designed with a carrot and stick approach as she states on page 12. The “carrot” supplies quiet roads for some individuals to encourage walking and cycling, and the “stick” is to make journeys by car longer in both distance and time, therefore less convenient so you are more likely to walk or cycle. The added congestion is a necessary part of the “stick”. This means that LTNs CANNOT BE ENVIRONMENTALLY FRIENDLY which can be demonstrated with some simple maths…..

TfL say around 35% of car trips are 2km or less. They agree a proportion of these would not be replaced due to disability, age, heavy loads being transported, young children etc. and suggest 22% of all car journeys could be reasonably walked or cycled. So for every 100 cars 22 of these car trips could reasonably be removed. A 2km car journey in Outer London according to TfL journey times is about 4 minutes driving. If you were amazingly successful and removed 20 of these car trips then you have removed 20 x 4 = 80 car minutes from the road. But of the original 100 cars 80 are remaining on the road. If you delay these 80 cars for just one minute with longer journeys and congestion, you wipe out any pollution gain you have made from removing nearly all short journeys. Indeed even just a one minute delay will have increased pollution because you have delayed not only the remaining 80 cars but all the buses, lorries, LGVs and taxis. Remember THIS IS THE CASE IF LTNs DO REMOVE CAR JOURNEYS AT A LEVEL NEVER SEEN BEFORE. If they don't, and the evidence above says they wont come anywhere near it, the pollution situation can only be much worse.

 

Reducing car journeys by education and persuasion would definitely reduce pollution. Closing large tranches of road space and forcing vehicles into longer journeys can only increase pollution.

Conclusion

In conclusion, all the evidence presented by various groups for the effects on traffic from LTNs used to justify the trial is deeply flawed. But you don't really need the figures and examples above to see the results. You can see this on our streets in the already “bedded in” schemes. Increasing congestion, increasing frustration, increasing pollution and a divided community.

Peter Payne

References: