Assessing Speed Camera Benefit Cost-Effective Road Accident Reduction or Flawed Analysis, Wishful Thinking and False Claims? Let’s look at the facts:-

Assessing Speed Camera Benefit

Cost-Effective Road Accident Reduction

or

Flawed Analysis, Wishful Thinking and False Claims?

Let’s look at the facts:-

Idris Francis [email protected]

01730 829 416 07717 222 459 December 2013

Presenter’s Background

1957 State Scholarship in Pure and Applied Mathematics and Physics, Llandysul

1960 1st Class Honours B.Sc. in Electrical Engineering, University College Swansea and Institution of Electrical Engineers Prize for Wales

1960-62 Postgraduate research, Eng. Dept, University of Cambridge

1963-65 Circuit Design Engineer, Leo Computers, London NW10

1964-96 Founder and Director, Flight Link Control Ltd, specialising in remote control of moving objects and later joysticks controls.

1985-94 included writing computer relational database software for company accounts, production, costing and sales

1992 Queens Award for Export Achievement, primarily based on an innovative joystick control which became the world standard for electric wheelchairs

2000-date Many thousands of hours studying road casualty data and trends, claims for speed cameras benefit and the data on which they are based.

"When you have worked out the answer, you must always ask yourself - does it make sense?“

The advice of Thomas Davies, my wonderful mathematics teacher in Cardiganshire in the 1950's, that I have followed ever since and found invaluable.

Applying that simple test to the Alice in Wonderland world of speed cameras shows immediately and in many ways that claims of camera benefit make no sense whatever.

Contents

Basic Data

Accident and casualty parameters and trends before cameras, and now where there are no cameras

Partnership claims of camera benefit far exceeding what is ever possible

Close to forty adverse camera effects, many applying across the country not just at sites

Seriously flawed analysis that lead to claims of far greater benefit than would ever be possible

What really happens at sites that would qualify for cameras have none

How to use that information properly to assess camera benefit, if any

Effectiveness and cost-effectiveness, in isolation and compared to other measures

Much exaggerated DfT estimates if the values of accidents prevented

Misleading legal advice provided by police and partnerships.

This presentation ( including more detail) and other files are on the CD and may be copied freely.

“We will make speeding as socially unacceptable asdrink-driving”

One widely publicised objective of the CameraHypothecation Scheme, 2000

Have they succeeded? Well, as one journalist pointedout recently, they have achieved the precise opposite –the huge number of fines, resulting in 10m peoplehaving 3 points, 2m having 6 and 1m 9, the vastmajority for modest breaches of limits involving nodanger, have now made speeding tickets so much the“norm” as to be socially acceptable and a commontalking point! Once again the law of unintendedconsequences bites back!

Years

Casualties

1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008

10^2.6

10^2.8

1000

10^3.2

10^3.4

10^3.6

10^3.8

10000

UK Fatalities per 100bn veh km, Log

SI per Veh km(no scale)

Logarithmic Scale (Constant % fall each year would show as falling straight line)

The area at the bottom right hand of the graph, between the dotted blue line - the long term trend - and the black line, actual fatalities, now represents in excess of 10,000 more fatalities than would have been expected in 1993

Trend line

Fatal per veh km(no scale)

The SI data from about 1996, provided as usual by the police, is now known to be bogus, as hospital records show no such fall.

NOTE Based on motor vehicle traffic. Growth of all traffic has been higher.

Years

Casualties

1975 1980 1985 1990 1995 2000 2005 20100

1000

2000

3000

4000

5000

6000

7000

8000

GB FALLING FATALITY and SI TRENDS

GB FatalitiesSpeed Cameras

Hypothecation Scheme

World BustBrown Boom

ERM Bust

GB Serious Injuries (multiply by 12)

Years

Cas ualtie s

1994 1996 1998 2000 2002 2004 2006 2008 2010 20120

500

1000

1500

2000

WALES KSI

"For core fixed and mobile sites a comparison between the baseline figure and 2008-2010 casualty data across Wales shows a 71.49%* reduction in killed or seriously injured (KSI) casualties"From GoSafe web site.

However as most cameras were installed in the early 2000's theirbaseline data will be in 1999/000/01 roughly, and the 98% of Welshroads that had no cameras saw reductions of roughly 30% in that time.It is essential that estimates of camera benefit allow for the effects of trend, especially over the longer term.

3 year average 1,235

3 year average 1,813

Main Causes of Long Term Downward Trends in Fatal and Serious Injuries

* Improvements of all kinds in vehicle design - better brakes, tyres, steering, road-holding, seat belts, air bags, ABS, crumple zones, stability systems

* Better roads and road surfaces, more motorways

* Slowing traffic growth, no longer rising, for the first time since WW2

* Better and quicker medical and other help at accidents and later

* Fewer pedestrian casualties as car ownership widens

* Falling reporting levels of non-fatal injuries (down 25% in recent years)

* and others.

“The reduction in non-fatal road traffic injuriesreported in police statistics probably representsan increase in under-reporting of these injuries,or a reduction in minor injuries, or both.

Hospital statistics show that there has been noappreciable reduction in injuries that areserious enough to warrant hospital admission”

BMJ, doi:10.1136/bmj.38883.593831.4F (published23 June 2006)

“29. Up to this point we have accepted the assurances of theGovernment that its casualty data were robust and that goodprogress was being made on bringing down the number of peoplekilled or seriously injured. Given the significant yet unexplaineddivergence in the trends for deaths and serious injuries, and giventhe growing body of evidence of changes in the reporting rates, wecan no longer conclude that good progress is being made oncasualty reduction. Indeed, we are worried that Ministers are notchallenging their officials sufficiently and that policy-makers andpractitioners are being lulled into a false sense of security.”

Transport Select Committee Report “Ending the Scandal ofComplacency” October 2008

Adverse Effects of Speed Cameras

The file "Dangers" on the CD contains a long but far from complete list of the adverse effects of speed cameras including, sudden braking, tailgating etc. all of which can lead to crashes not only at camera sites but also across the country (because drivers do not necessarily know whether there will be a camera in the next few hundred yards).

Following protests from Safe Speed and others, the DfT issued an Invitation to Tended for an investigation of these adverse effects but when asked by the presented a year later what progress had been made, replied that it had been cancelled because there be none! Fully documented on the CD – see “Adverse effects Investigation Cancelled.

This despite not infrequent media reports of deaths and injuries caused directly by the presence (real or imaginary) of cameras, including cars and motorcycles braking so harshly that they leave the road and hit immovable objects such as trees, walls etc.

It is clearly impossible to quantify these effects with any accuracy, but it is simply not acceptable that they are ignored not only by the DfT but also by camera partnerships.

Significance of the Graphs and Patterns

The DfT publishes a great deal of data but it is not possible to understand what is happening without seeing it in graph form.

The patterns we have seen so far and others to follow show that while it is a matter of almost pure chance whether an accident happens at a particular place and time, in aggregate accident and casualty numbers had do contain patterns which allow us to understand not only what has happened and what will happen, but also what would have happened at camera sites had cameras not been installed – which is of course the only way to assess camera effect.

Those patterns have been hiding in plain sight for many years, yet understanding and analysing them turns out to be not only almost childishly simple, involving only simple arithmetic and said graphs with no need for complex and questionable mathematical formulae, computer models or assumptions that may well be invalid.

First though, some basic information on the real significance of speeds above limits in accident causation and how Partnerships and others claim far greater benefit than would ever be possible:

DfT Table RAS50007 2011 Killed Seriously injuredContributory factor No % No %

Road environment contributed 158 9 2,409 12Vehicle defects 55 3 450 2

Injudicious action 498 28 4,604 23Exceeding speed limit 242 14 1,378 7Travelling too fast for conditions 226 13 1,759 9Driver/rider error or reaction 1,185 68 13,395 66Poor turn or manoeuvre 202 12 2,842 14Failed to look properly 433 25 6,882 34Failed to judge path or speed 200 11 3,186 16Swerved 116 34 4,190 21Impairment or distraction 426 24 3,152 15Impaired by alcohol 166 9 1,386 7Behaviour or inexperience 467 27 5,247 26Careless, reckless or in 285 16 3,533 17Pedestrian only 302 17 3,779 19Total number 1,752 100 20,396 100

Putting Speeding in Context Stats19 causation data shows that speeding was involved in:

* 5.1% of all collisions* 7.6% of KSI collisions* 13.8% of fatal collisions

• No speeding was involved in:* 99% of child pedestrian injuries* 94% of all injuries * 95% of cyclist fatalities* 99% of collisions with cyclists •96% of motorcycle accidents

* Speed cameras cover 1% of rural road length and 3% of urban

It would therefore be literally impossible for speed cameras to achieve the 30/40/50/60/70% reductions often claimed for them, even if they eliminated all speeding – which they do far from achieve. (Indeed, at many sites, speeds actually rise.)

Deaths Each Day in Britain (approximate)

All Causes...........................................................1,800Avoidable Hospital Deaths (infection, medicalerrors, neglect etc. ...............................................200Suicides....................................................................10Falls at home.............................................................7Road Deaths, all kinds ..............................................6As above, involving speeding ...................................1Primarily caused by speeding...................................0.5As above, on the 2% of roads with cameras ...........0.05 (0.003%)

Might a visitor from another planet wonder why we are spending £100m a year trying to reduce 0.003% of deaths in this country, when the same money could save vastly more lives spent in other more cost-effective ways? Like mops, buckets and disinfectant? When being in a hospital bed is several hundred times per hour, more likely to result in accidental death than being in a car at 70mph on a motorway ?

GoSafe’s Claim of £45,354,695 Benefit in 20011/12

The table on Pg. 17 of the report, showing a (ludicrously precise) total camera benefit of £45,354,695 carries a foot-note, “The savings are based on the reduction in 2011 in the number of casualties in Wales compared with the 1994/1998 average, at current live camera sites".

That however is to ignore that over the same period KSI fell across Wales by 38%, despite 98% of roads having no cameras, and that the same would surely have happened at camera sites had they not had cameras.

In any case none of those cameras could have had any effect on KSI in from 1999 to their installation and (as maximum benefit must be reached within months of installation, no further % reductions afterwards’

It is not clear what overall % reduction was used to calculate that figure but the 38% which would have happened anyway must surely account for a major part if it – perhaps 75% or £30m. Much the same applies to slight accidents, reducing benefit by a further £3m or so.

Even that much reduced benefit is far too high however because the DfT’s valuations of accidents are themselves preposterous – see later.

More detail on the longer version of this presentation on the CD

Misleading Statements in GoSafe’s 2011/12 Report

Paragraph 4 of the Introduction and again in the Executive Summary shows reductions in various casualties at camera sites, with the implication that they were due to the cameras. However most or all of those cameras would have been installed some years before and whatever effect they might have had would have levelled off after a few months or a year or so at most, making it impossible for those further reductions to have been due to the cameras. (That is not of course to say that whatever % reductions they achieved in their early years reversed, only that they did not and could not continue to increase many years later).

"PERFORMANCE.....at camera sites is encouraging: by the end of 2011, KSI casualties had been reduced by 50.77%* compared with the 1994-1998 annual average.

Four significant figure results from data which not accurate two! More seriously, the 50.77% fall in KSI at sites seems impressive until adjusted for the 38% fall on the 98% of roads that had no cameras, the same applies to the 61.27% fall at sites compared to 54% overall. The difference in the first case, of 23 KSI, might or might not be statistically significant bearing in mind variable reporting levels, the 2.5 child KSI in the second case is certainly not.

More Discrepancies in the GoSafe Report for 2011/12

Further, as above, the effect of any camera necessarily starts when it is installed and reaches its maximum within a few months or at most a year - so any further reductions years after installation, right up to 2009-11 would have been nothing whatever to do with the cameras.

For both reasons those reductions are seriously misleading.

"Camera sites in 2011 showed a 69.64% reduction in the number of people killed or seriously injured".

That statement is confusing and possibly in error, first because it fails to specify what baseline was used and secondly because, being well above the 50.77% previously mentioned, the figure implies a baseline well before 1994-98, which would of course be absurd.

"Motorists are continuing to break the law at camera sites and this demonstrates the continued need for educational messages to motorists“

Might it not, equally or more so, demonstrate that the cameras are failing to reduce speeding? And if cameras don't what chance have educational messages?

VALUE OF PARTNERSHIP WORK

Here "misleading" is inadequate a word to describe serious misrepresentation extending to many millions of pounds. When false claims such as this are used, as they often are, to seek and secure new funding, it amounts in the presenter's view to fraudulent misrepresentation or any one of several related offences, including breach of duty of care. The clue is the footnote stating that"The savings are based on the reduction in 2011 in the number of casualties in Wales compared with the 1994/1998 average, at current live camera sites".

As previously pointed out, 38 of the 50.77% fall in KSI over that period was due to long term trend, not to camera effect. Accordingly, the value of KSI benefit achieved should have been reduced by a factor of 50.77/12.77 = 4 to 1 and the £38m total should have been of the order of £9.5m, a reduction of £28.5m. The data for slight injuries is not to hand but the £7m quoted should probably be halved. making a total discrepancy of about £32m and leaving a notional benefit of about £13m. Even that however is absurd because the DfT values used are themselves wildly exaggerated - a matter which the presenter intends to take up with a Select Committee - in that the figures for "lost output" do not exist, they are fantasy - others always take the place of those unable to work, to meet that demand.

Assessment Report 256 - Statistics on the Scottish Safety Camera Programmehttp://www.statisticsauthority.gov.uk/search/index.html?kw=|Scottish%20Government and on the CD

Claims similar to those of the Wales partnership were summarily dismissed by the above Report which pointed out (in words more acerbic than the presenter would have thought possible in a public document) that the claims:

1/ Totally ignored long term trends

2/ Totally ignored regression to the mean (see later)

3/ Risked being seen as biased because it its authors were employed by the Partnership – whose newly appointed head statistician, then resigned.

And more of the same. Anyone wishing to understand properly the ways in which camera benefits have been wildly exaggerated not only in Scotland but across Britain would do well to read the UK Statistics Authority’s assessment. And the same applies to Safer Roads Humber:

Safer Roads Humber Annual Safety Camera Progress Report

April 2010 – March 2011

59% reduction in the number of people killed or seriously injured at core safety camera sites in the Humberside Partnership area

42% reduction in the number of injury collisions at core safety camera sites in the Humberside Partnership area

£73,223,760 saving in terms of killed or seriously injured

9% reduction in the average speed and a 11% reduction in the 85th percentile speed

32% reduction in the number of vehicles exceeding the speed limit at camera sites.

Safer Roads Humber annual report 2010- 2011

Published March 2012

The partnership has now been operating safety cameras for eight years and the annual report gives details of the partnerships performance at core safety camera sites……..

Figures from the report show that, in the eight years since safety camera enforcement began, there has been a 59 per cent reduction in the number of people killed or seriously injured at the core safety camera sites. In real terms there are 411 people alive and well today that would have been killed or seriously injured if safety cameras had not been introduced.

Years

Casualties

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 20110

10

20

30

40

50

60

70

80

90

100

110

120

GB v Humberside Falling KSI Trends

GB KSI

Both scaled to 100% average in 2001

42% GB Fall

Humberside

33% HumberFall

For obvious reasons, casualty trends in each police area tend to be similar to national trends as shown here..However local data being smaller is more volatile,again as shown here..

For that reason what is significant in this sort of comparison is not whether one graph is above or below the other - that depends very much on the arbitrary choice of when the figures were scaled to match but how they change over time.

For the same reason of volatility percentage falls vary considerably depending on quite when the comparisons start and end

Years

Casualties

0

10

20

30

40

50

60

70

80

90

100

110

120

GB v Humber Camera Sites, KSI

Scaled to 100% average in 1999/00/01

55% Fall at 63 camera sites

But the cameras not switched on until April to Aug 2003 so they could not have been responsible for the big drop in 2002. There was a small rise in 2003 when they were switched on and no change in 2004!

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

SRH Core camera sites

100%

Impressive at first glance?

Years

Casualties

0

10

20

30

40

50

60

70

80

90

100

110

120

GB v Humber Camera Sites, KSI

GB KSI, 98% of road length having no cameras

Both scaled to 100% average in 1999/00/01

42% GB Fall

55% Fall at 63 camera sitesBut the cameras not switched on until April to Aug 2003 so they could not have

been responsible for the big drop in 2002. There was a small rise in 2003 when they were switched on and no change in 2004!

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

**

SRH Core camera sites

100%

Impressive at first glance?

KSI

0

20

40

60

80

100

120

Humbers ide KSI Comparis ons

* Operational Apr 2003

71 actualsites

100%

3 year Selection Period1999/00/01

254 SITES 3+ KSI

Years Start at vertical lines identified by dates

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

SRH's 63 camera sites averaged 3 KSI in 3 years in the selection period but there were another 189 sites that would have qualified at the same timeon that basis, that did not have cameras installed. The graph shows how little difference there wasbetween their results.

So far we have seen that, for the 63 sites at which enforcement started between April and August 2003::

much or all of the observed reduction at those sites would have happened anyway, without cameras, as it did across the 98% of roads across the country which have none

it is clearly impossible, even if speeding were eliminated entirely for that alone to bring about reductions of more than 5% or so in KSI, let alone the 59% claimed in the Partnership Report

reductions in speed and speeding have been minimal (which is why the Partnership prefers to use % not mph to make the numbers seem larger)

a large part of the observed reduction happened in 2002 before the cameras went live between April and August 2003.

in 2003 and 2004, after switch-on, KSI at those sites hardly changed

it is not credible that cameras, having failed nothing for 16 to 21 months could then cause the steep fall in 2005 .

since 2005 the trend has been slightly upwards in contrast to elsewhere

We now turn to much the most important part of this Presentation, how properly to assess camera effect (if any)

To do that however we first need to understand the seriously flaws in the simplistic method first introduced in the equally flawed Eight Area Trial of 2000/01 (full details on request, available early December 2013), and indeed how those defects have been increased over time by Partnerships continuing to claim further % reductions many years after cameras achieved their maximum effect.

Casualties

0

20

40

60

80

100

120

140

Why Camera Effectiveness Assessment is Nonsense

Camera effectiveness is claimed by many Reports, and indeed stated explicitly in the 2005 Handbook for Partnerships, to be the fall in casualties from the average in the 3 year site selection period before camera installation up to date.

Claimed as Camera Effect

However this makes the assumptions, invalid as we have seen, that:

(a) casualties would continue indefinitely at the same level were cameras not installed,

(b) the 3 year average accurately represents that long term level. rather than being an unusally high level which led to the site being selected in the first place.

(c) having reduced casualties almost immediately by far more than are ever caused by speeding in the first place, cameras then continue to reduce them further over time.

Casualties

0

20

40

60

80

100

120

140

Why Camera Effectiveness Assessment is Nonsense

However, given that sites are chosen for unusually high casualty numbers,and given long term trends,observed falls are clearly primarily or indeed entirely to the two effects shown below:

Long term trend

Regression to the Mean(Return to Normal)

A new analysis of when and where some 7 million injury accidents happened from 1985 to 2011 confirms beyond rational doubt that this is indeed what happens,on average, across the country, and that results at camera sites are, again on average,indistinguishable from similar sites that have no cameras.

That claims for camera effecftiveness have been based on incompetent- or, worse,deliberately misleading - analysis has long been clear to critics. The new analysis however confirms that they provide no identifiable benefit and certainly none that remotely justify either the expense or the "war on motorists".

The New and Better Finney FTP Method

The fundamental blunder in speed camera assessment from the beginning has been to ignore or downplay the significance of trend and regression to the mean to be able to claim credit for the cameras for the observed reductions.

A case in point was the 4th Report of 2006 which admitted in the fine print that regression to the mean accounted for 75% of the observed falls in KSI after trend, leaving only 25% for camera effect’ - then chose to announce their findings without adjusting it.

As recently as April 2013 the current Minister, Stephen Hammond, wrote to the presenter that:

"while it would be desirable to include some explicit allowance for regression to mean, no reliable method had yet [in 2005] been established for doing so”

That might have been true then, but is no longer.

Following a meeting I obtained with the DfT in September to protest at their view, they now accept Dave Finney’s method ( (confirmed by the presenter’s analysis of 220,00 qualifying sites that never had cameras what Finney found at 75) is a valid way of assessing camera data and supposed benefit.

The file “DfT endorses new method” on the CD provides full details, what follows is only a brief summary.

Mr. Finney and I would be happy to work with the Welsh authorities to implement use this analysis on Welsh data.

YEARS

KSI

0

10

20

30

40

50

60

70

80

90

100

110

120

130

FINNEY FOUR TIME PERIOD METHOD OF CAMERA ANALYSIS

SiteSelect

RTM Fall

3 year AssessmentPeriod

Pre Selection Baseline

Dummy data with no long-trend

Method1/ Ignore data in SSP, almost always unusually high2/ Ignore data in RTM period, fall is back to normal, before camera3/ Also ignore data in year before SSP as likely to be higher than nornal4/ Use pre-SSP, normally the same as in RTM, as baseline5/ For each site make a judgement about most representative pre-SSP data6/ Use data in assessment period to assess fall from pre-SSP7/ Ignore data 3 years or more after installation, no further reductions possible8/ Camera effect is pre-SSP level to assessment level

Gap

The Importance of accurate data, particularly timing of accident selection and camera installation

Accurate timing of data is essential if the Finney Method is to provide accurate assessment of camera effect. Its results for 75 Thames Valley sites (no camera benefit whatever) were based on meticulously researched data that showed that at every single site the high levels of the site selection periods fell immediately (in the first quarter) to the same level as the pre-selection period and that planning and logistical delays invariably meant that cameras were not installed less than a year later.

It also found that the Thames Valley Partnership - like Scotland and Humber - had been claiming as camera benefit KSI and PIC reductions that had (a) happened before camera installation (b) were clearly due to regression to mean and perhaps (c) (as Scotland and Humber certainly had) further reductions well after camera affect (if any)

would have stabilised.

Summary of Finney Method

1/ Ignore data in SSP, almost always unusually high

2/ Ignore data in RTM period, fall is back to normal, before camera

3/ Ignore data in year before SSP, likely to be higher than normal

4/ Ignore data 3 years or more after installation, no further reductions possible

5/ Use pre-SSP data, normally the same as in RTM data, as baseline

6/ For each site choose the most representative pre-SSP period

7/ Use data in assessment period to assess fall from pre-SSP level

8/ Camera effect is pre-SSP level to assessment level

9/ When completed analysis shows, as for Thames Valley, that cameras provide no meaningful benefit, dismantle them and sell them for scrap

The presenter’s contribution to this method has been to confirm Finney’s findings on regression to the mean and its timing across the country and at all speed limits.

The analysis used some 4m accident records to identify 220,000 examples of sites that would have qualified for cameras but which for the most part did not receive them.

The following graphs, only a few examples of the thousands that can be produced from Excel spreadsheets of the KSI data, show to a compelling degree how consistent these changes are, in terms of the close match of the post-RTM to pre-SSP levels and the instantaneous fall after the SSP, well before any camera would be installed in response to the SSP data being found to justify it.

It is now possible for the first time to assess camera benefit reasonably accurately and beyond dispute and put an end to years of serious misrepresentation by the authorities.

How to identify 220,00 examples of sites that would have qualified for cameras

This turned out to be surprisingly easy because police Stats19 records of reported accidents include not only when but where they happened, to within 10m as defined by two 5 or 6 digit grid numbers.

It is therefore a simple exercise (indeed a child of 10 in possession of card indexes and a very great deal of patience could do it) to assign every accident to its correct place in a new database of accidents that happened in any specified area.

The first 3 digits of the Easting and the Northing codes conveniently specify 1 km sq areas, approximating to normal camera sites, though it almost equally easy to specify smaller areas (for instance in congested regions where there are many roads and many accidents). Or indeed to select particular areas on the basis of their roads' characteristics.

Having done that and assembled a database of some 300,000 sites in which at least one KSI accident had occurred over 25 years it is then a simple matter to select from them all examples showing at least 3 (or 4, or 5) KSI in any 3 year period, thus approximating to the threshold for camera installation – and then produce graphs of their results, as follows.

Effectiveness and cost-effectiveness, in isolation and compared to other measures

In 2006 the DfT gave Transcom figures purporting to show that cameras were “the most cost-effective measures yet devised”. The figures were so obviously wrong that the presenter forced the DfT to admit that, on the simple basis of their comparison, they are in fact 9 times less cost- effective than vehicle activated signs – though an independent accountants report confirmed that they are about 50 times less cost-effective, to the extent that there could not be any possible justification for their continued use. Both documents are on the CD, and the fully detailed record of the false DfT analysis, initial denial and then admission is available at www.fightbackwithfacts.com/cameras-versus-activated-signs/ and http://www.fightbackwithfacts.com/how-the-figures-were-skewed/

And even that was on the assumption that they achieved the sort of accident reductions that we can now show they most certainly do not!

In summary, it is obvious, and should have been obvious from thebeginning, that spending £100m a year to tackle such a tiny % of theproblem, and only on 3% of our roads never did make any sense atall.

No tweaking of the figures – claiming that RTM is less significant thanclaimed here, that speeding is more significant than the official figuressuggest, can make any difference to that basic point.

One example of the hopelessness of the camera objective is that, evenif we accept the 4th Year Report’s wildly inflated claim of 100 livessaved, that’s 1 life per £50,000 pa camera every 50 years! Would theNHS save only 1 life every 50 years, for £2.5m? And even thatignores the casualty statistics showing worsening, not improving,trends which must be at least in part due to the many adverse effects ofcameras.

So are there better, more cost-effective ways of reducingcasualties? The DfT certainly thought so when they changed thefunding system:

Cost Effectiveness Comparison Cameras, Signs, Humps v 20mphzones

The DfT figures shown in the 2006 Select Committee Report also coveredroad humps and 20mph zones. After correcting the bizarrediscrepancies in the camera data and the method of analysis, andassuming the same valuation of accidents avoided, the comparisonbetween all 4 methods is:

Camera Sign Hump 20mph ZonePIA Before 29 (5yrs) 31 (10yrs) 10 (5yrs) 10.5 (1yr)

PIA After 18 0 7 2.4PIA Reduction 11 31 3 8.1PIA Reduction per annum 2.2 3.1 0.6 8.1PIA Reduction per annum % 38 100 30 77Installation Costs £40k £14k £10k £200kOperating Costs pa* £52k** £ 2k*** £1k £20kRatio of Benefit/Costs 38/52 100/2 30/1 77/20Relative Benefit/Cost 1 68* 41 5.3*based on 10 year cost amortisation ** changed from 80 to 1 due to same values of accidents avoided in same year.

Misleading legal advice provided by police and partnerships.

Across the country virtually all police and camera partnerships provide legal advice on their web sites and paperwork. Much of that advice is incorrect or seriously misleading.

This is simply not acceptable – it is not the job of the police or partnerships to give legal advice, and I am not aware of any other type of offence for which they do.

But if they do, then they should certainly not give misleading advice that not only risks perverting the course of justice but also, to my certain knowledge, has done so on many occasions, usually by persuading innocent defendants to plead accept penalties.

Documents on the CD cover the range of incorrect advice across the country and also the more limited number of examples for Wales.

END