You're misrepresenting the results of the study. The study recommended the continuation of the program. The rear-end collisions were less severe than types of collisions it reduced, it was a safety improvement overall.

From the findings section of the actual study (emphasis mine):

>Quantitative studies conducted in this project demonstrate significant safety benefits of the current RLC program. As a result, it is appropriate to recommend continuation of the program. Most of the intersections have experienced an improvement in safety, particularly in terms of severe angle and turn crashes, albeit with an accompanying increase in less severe rear-end crashes. The safety benefits extend beyond the immediate vicinity of the RLC intersections, evidenced by a significant spillover effect. However, some intersections appear to experience no significant safety impact. Recognizing that crashes are the result of complex interactions amongst many factors, and subject to considerable randomness, these deviations should be used as opportunities for detailed investigation and learning to design and deploy more effective automated enforcement programs

http://www.transportation.northwestern.edu/docs/research/RLC...

Similar benefits have been demonstrated in other cities. Doesn't mean it will work in every city, but there is evidence it works in Chicago.