As Houston’s city council considers what to do with the current red light camera program, two choices seem to be on the table. Opponents want the program discontinued or at least curtailed. Proponents want the program continued or expanded.
A new Baker Institute study, “Report on the effects of Houston-area red light monitoring cameras,” suggests a possible compromise: Increase the number of intersections with two cameras by eliminating selected intersections with only one camera. Our data shows the installation of a second camera reduces collisions not only at the approaches with cameras, but also reduces collisions at approaches without a red light camera.
Like the authors of a Texas Transportation Institute report released earlier this month, our research concluded that red light cameras in the city of Houston are effective at reducing intersection collisions. Moreover, we find that the effect of red light cameras on collisions is distinct and independent of other sources of variation in intersection collisions.
Our study, which was funded by Rice University, used data collected in Houston from 2003 through 2009 by the Texas Department of Transportation. After filtering out weather, time and geography-related causes of collisions by using intersection and approach collision data, we made predictions about the number of collisions that should occur under different sets of circumstances.
Houston, also, initially installed only one camera per intersection. A second camera was installed in 20 intersections after one year. This difference allows us to test the effect of one and more-than-one camera on intersection collisions.
Our model suggests that Houston’s program was responsible for a reduction in traffic accidents of approximately 35% between September 2006 and December 2009. In this time, there were 1,413 collisions at intersection approaches (streets entering the intersection) with cameras. Our model predicts 1,438 collisions for these same intersections: a deviation of only 1.8% from what actually occurred. Absent any cameras, we estimate that collisions at these intersections would have totaled 2,230.
Although we find cameras are associated with a reduction in collisions, the effects of the cameras depend crucially on how long the cameras are in place and the number of cameras installed at an intersection. Intersections with one camera see reductions in collisions only on the approach with the camera: We do not observe a reduction in collisions at any other approach in the same intersection.
The installation of a second camera reduces collisions not only at the approaches with cameras, but also reduces collisions at approaches without a red light camera. This spillover effect is only observed at intersections with two cameras. We attribute this effect to the placement of the second camera. Second cameras were placed adjacent to approaches with existing cameras. This effectively targets side-impact collisions, the type of collision mostly frequently caused by red light running.
At present, there are 50 intersections in the city of Houston with 70 cameras. Twenty of these have two cameras and 30 only one. Our model would predict that consolidating the 30 single-camera intersections into 15 two-camera intersections would reduce collisions.
- Download “Report on the effects of Houston-area red light monitoring cameras” from the Baker Institute website.
Robert Stein, Ph.D., is the fellow in urban politics at the James A. Baker III Institute for Public Policy. He is also Rice University’s Lena Gohlman Fox Professor of Political Science and a nationally recognized political analyst and expert on elections.