Thames Valley Speed Cameras

An independent report        

Comparing the 3 years before and after (the usual method to present camera statistics) there were:

BEFORE: ... 122 collisions in the 3 years before
AFTER: ...... 98 collisions in the 3 years after


This is a drop of 20%.


Figure 1 clearly shows that the entire 20% drop occurred 2 full years before the cameras were installed but TVSRP claim 44%, not 20%.


TVSRPs document "windsor-and-maidenhead" states that the 44% drop was (p1): "Based on the three years before the cameras were installed compared with the most recent three years (2006 – 2008)" and also states (p1): "The majority of the fixed cameras have been in place for over 10 years ...". And they are right. In fact the average time over which the 44% drop occurred is 12.3 years therefore most of this reduction occurred at some of the older sites over 6 years AFTER the cameras were installed.


The cameras cannot have caused a drop in collisions 2 years before they were installed and, if they have failed to cause any reduction for 6 years, they cannot be credited with a drop that materialises at some of the older sites around 12 years later.


Figure 1 shows there was virtually no change in the number of collisions after the point when speed cameras were installed but notice that TVSRP don't claim that their speed cameras actually "caused" the 44% drop. We must conclude that the 29 speed cameras in RBW&M have not produced any reduction in the number of collisions whatsoever.


Figure 2 The above "injury collisions" consist of those that resulted in fatal, serious and slight injuries as recorded by the Police but over 80% of these collisions involve only slight injuries. If the collisions where people were killed or seriously injured are analysed separately, we find that they went up.


But there are 2 significant effects that influenced the number of serious injuries:

1) The threshold to classify an injury as serious changed in 1999 in Thames Valley.
2) There was a trend of lower serious injuries on the vast majority of roads where there were no cameras.

The influence of both of these effects can be compensated for by calculating the percentage of area-wide KSIs that occurred at camera sites, see figure 3.


Figure 3 The number of collisions that resulted in fatal or serious injuries increased by 27% after the cameras were installed (compensation applied to correct for the area-wide trend and Police reporting change in 1999).


Figure 4 There were 3 deaths in the 3 years before, and 6 deaths in the 6 years after. That's 1 death per year both before and after cameras. The numbers are too low to be significant but there is no evidence that lives have been saved.




The evidence suggests that the 29 speed cameras in RBW&M have not reduced the number of collisions, have not prevented any serious injuries and have not saved any lives.




This report evaluates all the active cameras so does not include site TV8007 in Ascot because it was decommissioned following nearly 9 years of operation. This site has data for 5 years before, the 9 years of operation, and then 4 years after decommission.


There were no fatalities in this entire period. There were no serious injuries before, 7 during operation, and 2 after decommission. There were 17 total injuries before, 44 during operation, and 13 after decommission.


That works out as these averages per year:

  before     with camera     camera removed

Serious        0                0.8                   0.5

total          3.4               4.9                   3.3


Serious injuries went up when this camera was installed, and reduced when it was removed.


Total collisions went up when this camera was installed, and fell back down again when removed.


Had this camera been included in this report, the RBW&M speed cameras would show an increase in total collisions and an even greater increase in serious injuries.




This report investigates the effect of the speed cameras in RBW&M on the number and severity of road traffic collisions.




The evidence suggests that the 29 speed cameras in RBW&M have not reduced the number of collisions, have not prevented any serious injuries and have not saved any lives.




The results of this report raise sufficient concern that the author considers it essential for RBW&M to only operate their speed cameras within a scientific trial in order to evaluate the effect of their speed cameras. This trial should take the form of a Randomised Controlled Trial that is rigorous, independently organised and independently supervised.




From 2000 to 2011, the speed cameras in Thames Valley (Oxfordshire, Buckinghamshire and Berkshire) were operated by Thames Valley Safer Roads Partnership (TVSRP). Like all partnerships, TVSRP were initially funded by the income made from speed camera fines but this stopped in March 2007. The government then took the speeding fines and added millions of pounds of taxpayers money to create a road safety grant of £110m per year to be distributed to councils. The councils could then choose how to spend this to improve safety on their roads.


This transformed the partnerships into something akin to a business where TVSRP sold a range of road safety services, with speed cameras as their core operation, to any council that wished to spend our money on speed cameras. TVSRP received additional income from the courses that offenders attended (speed awareness courses etc).


On 20th October 2009, the Councillors for the Royal Borough of Windsor and Maidenhead (RBW&M) met to decide whether to continue to purchase the services of TVSRP costing £200,000 per year of the councils budget.


To help sell their services to RBW&M, TVSRP produced this document "windsor-and-maidenhead" which states (p1): "The majority of the fixed cameras have been in place for over 10 years and the casualty reduction record is impressive with a 44% drop in recorded injury collisions since the cameras were installed".


So should we be "impressed" with this?


The 44% drop is at all 13 fixed camera sites in RBW&M. Each of these 13 sites has 2 fixed camera locations (Gatsos) and 3 of these sites are also visited by mobile cameras. That's a total of 29 speed cameras. For all these camera sites combined, TVSRPs database allows a comparison of a maximum of 3 years before the cameras and 6 years after.

Supporting information for all 13 active fixed speed camera sites in RBW&M:


Note: TVSRP have corrected around 5000 errors in their database, this report uses the new corrected and verified database. TVSRP agree that the above graphs are correct.


rbwm_data.xls ..... spreadsheet with the data for this report.

rbwm_graphs.ppt ..... powerpoint with graphs of all 13 active fixed sites for all collisions and KSI collisions.

rbwm_raw_data.xls ..... the database extract for this report (the original raw data) so the reader can check the results of this entire report for themselves.


Note: This report displays the same 3 year period before and after cameras as used by TVSRP therefore it is directly comparable with reports issued by TVSRP to analyse the performance of the cameras.


Other information: my report of the TVSRP public meeting in Maidenhead, rbwm_tvsrp_meeting.doc.


Speed cameras in RBW&M have raised over £1.7m in the last 5 years, but this still does not cover the cost of running them. Taxpayers must pay the shortfall of around £50,000 per year, RBWM_speed_camera_income.xlsx