By on July 12, 2010

The TaxPayers’ Alliance and Drivers’ Alliance last week calculated that UK speed cameras issued £87,368,227 (US $131,256,380) worth of tickets in fiscal 2009 without any demonstrable safety benefit. Since speed cameras were first installed on British roads in 1991, the roads became more dangerous than they would have been without photo enforcement, according to the report.

“The road casualty rate has declined at a slower rate since speed cameras were introduced in the early 1990s,” the study explained. “Using the road casualty rate from 1978-1990 it can be estimated that 1,555,244 more road casualties have occurred from 1991-2007 than would have if the 1978-1990 trend had continued.”

Such figures should see a continuous steady decline because advances in automotive technology including anti-lock brakes, stability control, crumple zones and airbags have made vehicles significantly safer over time. Those who are injured in an accident are also more likely to survive as medical treatments and trauma care likewise advance.

In terms of the number of fatal accidents per billion passenger miles traveled, the casualty rate fell from 773 in 1979 to 331 in 2007. The pre-camera fall, however, was far more impressive. Had cameras never been installed, the group projected the casualty rate would have been 128 in 2007.

Automated ticketing machines in the UK are operated by speed camera partnerships representing local city and county officials teaming up with police. Beginning in 2007, the revenue generated by each partnership was sent directly to the national treasury with money reallocated back to the partnership through road safety grants. This setup was designed to remove the impression that the cameras were merely being used to raise revenue. Newly elected Prime Minister David Cameron now pledges to stop grants for the purposes of adding new fixed speed cameras.

This was the first count of tickets that included magistrates’ courts along with the national figures. Partnerships in England and Wales accounted for the greatest number of tickets — £65,748,850 (US $98,721,524). Another pound;19,214,594 (US $28,850,691) went through magistrates courts in England and Wales. Scotland issued £1,641,630 (US $2,464,905) in automated citations while Northern Ireland collected £763,153 (US $1,145,909).

View a copy of the report in a 200k PDF file at the source link below.

Source: PDF File Research Notes No. 3 — speeding fines (TaxPayers Alliance and Drivers Alliance, 7/12/2010)

Get the latest TTAC e-Newsletter!

19 Comments on “UK: Speed Camera Ticketing Slowed Advance in Road Safety...”

  • avatar
    M 1

    While I’m no fan of automated tax coll— er, fines — it’s a little difficult to accept this conclusion without some explanation of how or why speed cameras would cause this change.

    Rear-end collisions from drivers nailing the brakes when they think they’ve been caught? More wandering as people take their eyes off the road trying to spot the cameras? Maybe more traffic on smaller and unfamiliar roads by folks trying to avoid known cameras? Throw us a bone here.

    Also, it looks to me like the trend resumed around 2000…

  • avatar

    To have any validity, this claim should show that the rate of decline in similar jurisdictions without a massive buildup of speed cameras continued on a straight line. I’m not sure if that control should be in Europe, Australia/NZ, or the USA/Canada, but without it this is meaningless.

  • avatar

    Actually, if you go to the group’s website and read the study it appears to be done fairly rigorously. If you believe in such statistical analysis, this one is fairly convincing. In fact, they even point out (in footnote 6, page 9) that a change in statistical record-keeping in 1993 should have driven the casualty rate downward significantly. Instead, there does appear to be a trend toward the unsafe from the expected trend. Could an alternate hypothesis be suggested? Sure- this study isn’t bulletproof by any stretch of the imagination. I thought that maybe airbags were the answer, slanting the 80s data with a safety improvement rate that could never be sustained after the phasing in of airbags was complete. That would largely negate this study- however, per the anonymous and unaccountable editors at Wikipedia, airbags in Europe were not common and certainly not the majority of cars on the road until well into the 90s, so if anything, the speeding camera era should have benefited from the reduced mortality of the airbag phase in- but no, that isn’t what we see. Despite the increased utilization of airbags and the newer calculation method for passenger-Km traveled, both of which should have skewed the curve towards a safer line than the projection, we see the opposite. Something caused the improvement to stop- I think they might be onto something, actually.

  • avatar

    A similar conclusion to this study was shown somewhere some months ago. Related to Australian roads.

  • avatar

    Beginning in 2007, the revenue generated by each partnership was sent directly to the national treasury with money reallocated back to the partnership through road safety grants. This setup was designed to remove the impression that the cameras were merely being used to raise revenue.

    W. T. F.

    Removes the impression but not the reality.

  • avatar

    If it weren’t for those darned speed cameras casualty rates would have fallen to zero deaths per km by about 2012, and would continue into the negatives where new people spontaneously appear in places where there used to be accidents.

    There are reasons to dislike speed cameras, but this article is based on a ridiculous linear extrapolation on a downward trending statistic that HAS to level off at some point.

    • 0 avatar

      They tested the linear fit of two segments of years against each other for a difference in slope- but no extrapolation of any linear model outside of the actual range used to calculate it was used in their analysis- just in their silly graph. Instead, they proved a slope difference which was true between two different eras (not using that extrapolation line from the graph, at all). As you and others correctly point out- this was inevitable. Having proven that the inevitable happened, they then tried to spin it as proof that traffic cameras are preventing progress in safety. Their conclusions are suspect, granted, but their statistical methods are not as bad as folks are suggesting.

    • 0 avatar

      Most things from this submitter are sketchy at best. This one is just obviously bogus.

  • avatar
    DC Bruce

    To the skeptics, I would suggest that the proponents of automated ticketing devices have justified their installation on the basis of their alleged effect in improving road safety. That they have demonstrably failed to do so should be a sufficient refutation of this argument. That they have actually degraded road safety (assuming this to be the case) is just “gravy.”

    The political reality is that, once any source of government revenue is established, it is extremely difficult to eliminate — because the funds it generates must be replaced.

    I spent the weekend of July 4 in Los Angeles. A sure sign of California’s fiscal crisis is the posting at various red light camera-equipped intersections of signs warning that the penalty for getting your picture taken is something in excess of $400. Wow!

    Of course these gadgets, and enforcement of extremely low BAC against drivers through the use of “sobriety checkpoints” all illustrate that the costs of striving for 100% legal compliance greatly outweigh the benefits.

  • avatar

    The projection model is so obviously wrong to begin with: it projects a negative casualty rate by 2010.

    Only a retarded person would use a linear model to predict the drop of the casualty rate. An exponential decay model would be more convincing.

    • 0 avatar

      There is a very valid point in your objection. There would have been a mathematically mandatory slow down in improvements at some point. No one could argue that, and that is why I said that another hypothesis could be put forth to explain the results. However, the analysis was between the models of each era, not the projection of one model into the other model’s era. The graphic is misleading, but their analysis is not as stupid as the graph would suggest. In the actual paper, they model each era and test the models for a difference. They conclude, essentially, that a slope change has occurred- only then do they make a rather dramatic leap to the conclusion that they were aiming at the whole time. In my opinion, you have provided a very plausible explanation of the observations that serves as a challenge to their conclusions- but I must disagree that their statistical methods were retarded.

    • 0 avatar

      Although I’d refrain from using the word “retarded” to describe a person who inappropriately used a linear model in a statistical analysis (I’ve actually been in email correspondence with the analyst and though she’s very stubborn about her findings, she’s actually been very polite, civil and courteous, and certainly not retarded), your point is quite right. The extrapolation is very silly indeed.

      Furthermore, their data source goes back as far as 1952, so why they’ve only displayed data from 1978 is anyone’s guess, though the more cynical among us might suggest that this is the cutoff point that makes the data look most flattering to their desired conclusions. From about 1962 onwards, accident rates have followed an exponential decay pattern (more or less) right through to the late 80s. There’s a bit of discontinuity at 1990 but a) not as much as claimed in the report, b) not as significant as the break in 2000 and c) even if that break were real, it occurred way before the majority of speed cameras were installed.

      I’ve done the analysis, and written my take on it here:

  • avatar

    I like how they just draw a line on that graph as the point where speed cameras were introduced. Undoubtedly, the number of speed cameras was rising that whole time. To do it properly, they need to note how many speed cameras were operating at each point in the graph (or what percentage of road was being covered).

    I don’t like speed cameras, but this is not even close to convincing analysis.

  • avatar

    Anytime I see something like this, I’m remembered of the phrase that one of my professors beat into me:

    “Correlation is not proof of causation”.

    He also liked to remind us of the dangers of extrapolation:

    “I gained 5 pounds over the Christmas holiday, so in 5 years I’m going to weigh 1500 pounds!”

  • avatar

    > I would suggest that the proponents of automated ticketing
    > devices have justified their installation on the basis of their
    > alleged effect in improving road safety.

    Ah, so. Did the casualty rate increase or decrease after the cameras were installed?

    The real question is: What is the cause of the decrease? For that matter, what was the cause of the previous decrease?

    Likely the decreases in fatalities are due to an increase in the number of safety devices — safety glass, seat belts, airbags, crumple zones… and maybe speed cameras. Certainly, were I a proponent of automated ticketing devices I’d argue that speed cameras were the latest in a long line of devices which has steadily improved driver safety over the years. (I’m not — but certainly not because I believe they have made driving more dangerous as postulated here)

    Correlation is not proof of causation, indeed.

  • avatar

    Money people. It is all about money. Change the penalty to community service and would the cameras exist? Probably not. All this “safety argument” is really just spin. Everyone will argue one way or the other (just like politics) over the non-issue (safety) while they cotinue to suck money out of the population.

  • avatar
    Dr Strangelove

    The y axis reads “casualties per passenger kilometre”. WTF – so there are several hundred people killed for each passenger traveling 1 km? Sounds more like the charge of the light brigade than road traffic.

    Also, “Using the road casualty rate from 1978-1990 it can be estimated that 1,555,244 more road casualties have occurred from 1991-2007 than would have if the 1978-1990 trend had continued.”

    WTF again – 1.5 million incremental casualties? There is some serious innumeracy here. Several aspiring journalists should have gone to school a little longer.

    • 0 avatar

      They are defining casualty as death or serious injury, not just deaths. Also, the units are 10*9 kM, not kM- their fault for not being clear in the graphs (it was clear in the text of their paper, but still- their mistake- good catch). The graphs really are awful.

      An average difference of 140 casualties per billion kM traveled would yield 98,000 per year in the UK (they roughly drive 700 billion kM per year). The claim made was for 16 years, so 98,000 times 16 would give you 1.57 million incremental casualties. I’m using a steady average, not an increasing gap, but still- the math supports a number around a million and a half if you believe in the model in the first place (and therein lies the real problem).

      I think we are witnessing a questionable or leading premise and calling it bad math. I think the model is the problem, but the execution thereafter looks pretty solid to me. I do think this analysis is misleading, and even worse, I think it is intentional. If the journalists involved take your advice and head back to school, I suggest an ethics class be added to the curriculum.

    • 0 avatar
      Dr Strangelove

      The death toll on British roads is something like 3000 per year, which adjusted for population is similar to most countries that actually school their new drivers (coupla years back UK had actually the lowest numbers, don’t know if that still applies).

      So it’s all hopeless nonsense.

Read all comments

Back to TopLeave a Reply

You must be logged in to post a comment.

Recent Comments

  • 18726543: My first car (the hand-me-down, not the first car I ever bought) was a 1989 Taurus. It was purchased used...
  • Peter Gazis: Mike A. NOT A COMMUNIST. Communists are atheists. Martin Luther King was a Baptist Minister.
  • DenverMike: The economic world places Mexico in Central America even though no such continent exists. Or that Mexico...
  • la834: Actually that C/D test went from 15-inchers (base wheel on Golf) to 19 (optional on Golf R). My ’07 Golf...
  • EX35: No. In my experience, nissans vary depending on model, but most seem to hold up better than the Kias I’ve...

New Car Research

Get a Free Dealer Quote

Who We Are

  • Adam Tonge
  • Bozi Tatarevic
  • Corey Lewis
  • Mark Baruth
  • Ronnie Schreiber