CBS SF Bay Area published an article about the debate surrounding CA Assembly Bill 612, which would increase the duration of yellow signals for intersections with red-light cameras.

An opponent of the bill argues that it's dangerous to lengthen yellow lights:

“This bill would put millions of California drivers at risk for increased crashes,” said Richard Retting, a consultant for Redflex. The company supplies most red light cameras in California. . . . Retting points to a study by the American Academy of Sciences. “That study shows that when yellow timing is increased, crashes go up,” he said.

Redflex offers a rationale for this claim in their letter to Sen. Mark DeSaulnier opposing AB 612:

Mr. Retting concludes that, "Providing excessive yellow signal timing as mandated under Assembly Bill No. 612 violates established engineering practice. This would encourage drivers to enter intersections further into the yellow phase and could disrupt the flow of vehicles from intersection to intersection, ultimately creating a risk of increased crashes."

The bill's sponsor counters that the claim is not credible, noting bias and counterevidence:

Jay Beeber with Safer Streets L.A. disagrees. “He is a scientist who has been bought and paid for by the red light camera companies,” he said.

Beeber, who sponsored the bill, told KPIX 5 that data in the study is too limited. “Georgia increased yellow light times by a full second. They had a reduction in red light running and have had no ill effects,” he said.

Unfortunately, the article does not provide any verifiable evidence for or against the opposition's claim. Neither the article nor the letter to the senator actually cite the “study by the American Academy of Sciences,” so I cannot judge its validity. I personally find the claim dubious, but I could be convinced by solid evidence. What are the facts here?

  • It might be worth treating all four cases of (with/without citation camera) * (short/long light timing), as the cameras are often claimed to cause a change in driver behavior as well. Jul 2, 2013 at 14:16
  • In at least one US city there are camera-enforced stop lights that alternate red/green/long-yellow/red/green/short-yellow/red. Might that pattern increase wrecks and yet distort data if each light is described with a single average yellow length?
    – Paul
    Jul 7, 2013 at 8:56
  • One option I have seen is to have a period after the yellow where the lights are red in all directions. This is a "hidden" way to make the intersections safer, in that drivers are less likely to learn the "trick" and adjust to it. People drive by habit, but adapt to variations that they notice. So changing yellow light timing will result in changed behavior at that site. But red is red and still illegal.
    – user29285
    Oct 8, 2015 at 23:45

2 Answers 2


There is some evidence to suggest that increasing yellow light timing creates fewer accidents. Here (PDF) is a study on yellow light timing. They measured the number of red light violations at six intersections in Philadelphia, Pennsylvania over the course of a year, before and after increasing the yellow light duration by one second. Here are a couple excerpts taken from their results:

After yellow signal timing changes, violation rates at the experimental sites declined at each location, with reductions ranging from 21 to 63 percent. (p6)

Additionally, at different comparison sites, they note:

After accounting for these changes in driver behavior at the comparison sites, the changes to yellow signal timing at the experimental sites were associated with a 36 percent decline in the odds of a red light violation... (p6)

So, I would say the evidence points towards a correlation between fewer accidents and longer timings. Additionally, there's a logical argument here: A longer yellow light timing means more warning when a light is about to turn red, means a better chance that people will stop in time.

I have no idea why increasing yellow light times could possibly increase accidents; that is a logical leap which does not make sense to me. The argument given has a logical flaw: While yes, they would be entering later in the yellow light phase, they would still be entering with the same amount of time to spare.

Edit: Some more details. Apparently I need to read the paper more thoroughly! Thanks to OP for pointing these out.

There's a quote from the same person (Retting) who argues against the longer yellow-light periods, arguing for them, cited in the paper; a classical hipocrisy:

A study of modified traffic signal change interval timing at urban intersections reported that injury crashes were reduced by 12 percent at experimental sites relative to control sites (Retting et al., 2002). (pg. 9)

In addition, from the same paper:

Stimpson et al. (1980) reported that increases in yellow signal timing duration of 1.3 seconds significantly reduced potential intersection conflicts. (pg. 9)

It would seem this argument has been rather destroyed.

  • 2
    That shows a reduction in red light violations but not accidents. Does the study also address that? I can't check – the link is broken. Jul 2, 2013 at 5:31
  • 1
    @BraddSzonye if there are fewer violations, given the rate of accidents per violation stays the same, there will be fewer accidents.
    – jwenting
    Jul 2, 2013 at 5:34
  • 3
    I agree that there's probably a correlation – but it's not a given that the rate of accidents per violation would stay the same. Jul 2, 2013 at 5:36
  • @Bradd I fixed the link; the paper doesn't specifically address accidents, but it's implied.
    – user13259
    Jul 2, 2013 at 5:53
  • 2
    Driving habits are well known to vary from place to place. If drivers in California are known to habitually run yellow lights, it might be that a longer yellow light in PA would have a positive effect, but in CA a negative one. This is still a good answer though; it does address the question, even if there's a chance it doesn't address the specific legislation.
    – Flimzy
    Jul 2, 2013 at 18:35

Richard Retting is perhaps the most skillful salesman for red light camera companies we have ever seen. He is employed by the camera company Brexford and has worked for others. He has worked for the IIHS whose member insurance companies make money with surcharges in some states when their policyholders get camera tickets.

It is interesting that anyone concerned with their reputation would agree to work with Redflex after they were thrown out of Chicago in a $2 million dollar bribery scandal. And please note that it takes a LOT to get thrown out of Chicago for bribery and corruption.

There is a long list of studies and investigative reporters articles on our website showing drastic reductions in violation rates with longer yellows, rates that stay down over time. Increasing the yellows by 0.7 to 1.0 seconds typically drops violation rates by 60% to 90%. Anyone objecting to huge reductions in violations with longer yellows is likely in the camera revenue stream.

James C. Walker, Life Member-National Motorists Association, www.motorists.org

  • You're not really answering the question asked. It's not about breaking law, but about HAZARD.
    – Wertilq
    Jul 4, 2013 at 20:30

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .