2.1 Automated Enforcement

The use of automated enforcement systems to address speeding and red-light running are in use across the United States. Many States have prohibitions in their laws to prevent the use of automated enforcement technology; others have enabling legislation and/or parameters on the use of the technology; and others still have no legislation that addresses the technology’s use.

Automated speed enforcement (commonly referred to as “photo radar”) and red light camera systems should be used as a component of a broader traffic safety and speed management program supported by a demonstrated need through problem identification. These systems should be used to support traditional enforcement efforts, or be deployed in locations where enforcement may be unsafe or impractical for LEOs to make traffic stops.

Automated enforcement systems function by capturing violations, recording relevant data about the violations, and recording images of the violator vehicles. Red light camera systems employ sensors linked to a camera and data collection equipment. Vehicles that enter an intersection against red signal lights are detected; the cameras capture series of images (and with some systems, video) to depict the violations. Sensors provide additional violation data, such as the vehicle speed, the time the light had been red at the point the vehicle entered the intersection, and temporal information. Images and violation data are reviewed at a later time, and when appropriate a traffic citation is issued and mailed to the registered owner of the vehicle. Some States involve driver liability to determine responsibility for violations. This approach requires a more involved process in which approaching and receding images are captured, and include an image of the driver. Review and processing of citations in such States is more intensive, and places a higher burden on the State to identify the driver for a conviction or finding of responsibility. Many States operating in this manner apply penalty points against the license of the driver.

Other jurisdictions use a registered owner liability approach to enforcement. The processes for this approach are generally more limited and are not reliant on charging the actual driver of the vehicle. This approach places the burden on the registered owner, regardless of who was driving the vehicle to resolve the citation. In many cases, the only defenses would be in cases where it can be demonstrated the vehicle had changed ownership, was stolen, or an error occurred in processing the citation.

Guidance documents have been produced by the FHWA and NHTSA for the use of red light camera systems and automated speed enforcement. Red-Light Camera Systems Operational Guidelines (FHWA, 2005) provides information on red light camera program costs, effectiveness, implementation, and other issues, Eccles et al. (2012), and NHTSA and FHWA (2008) released automated enforcement program and operational guides with information on identifying problems and setting up and maintaining an effective and transparent, community-supported enforcement program using speed or red light cameras.

In 2011 and 2012 Miller et al. (2016) surveyed agencies with current or recently discontinued automated speed enforcement programs. They found that States and agencies varied greatly in the legislation and the technologies used for automated enforcement. These differences influenced the amount and type of data collected, types of fixed and mobile camera units used, and enforcement duration and schedule. Almost 63% of the 90 responding agencies reported not being aware of the NHTSA Speed Enforcement Camera Systems Operational Guidelines (NHTSA, 2008). Half of the surveyed agencies reported past or upcoming plans for evaluation of crash-related effectiveness of the program; however, 62% of agencies regularly reviewed speed data and 63% regularly reviewed crash data to determine locations for deployment of automated enforcement.

Use: Red light camera systems are used extensively in other industrialized countries and were first employed in the United Sates in 1993 (National Campaign to Stop Red Light Running, 2002). As of September 2019 red light camera systems were being used in 341 communities in 22 States and the District of Columbia (GHSA, 2019; IIHS, 2019b). As of 2018 speed cameras were being used in approximately 137 jurisdictions in 14 States and the District of Columbia (GHSA, 2019; IIHS, 2019c). Speed cameras also are used extensively in other countries (Speed Camera Database, 2019; WHO, 2004).

Effectiveness: The effectiveness of red light camera systems has been studied previously and mixed results with respect to crash type and experience were found. It should be noted that red light camera technology does not cause traffic crashes. While the presence of a red light camera system has reflected increased numbers of lower impact rear end crashes at intersections where the systems are installed resulting from drivers stopping for the red light, research has also found a reduction in more dangerous offset and right angle crashes at intersections with red light cameras (Aeron-Thomas & Hess, 2006; Decina et al., 2007; MacCubbin et al., 2001; McGee & Eccles, 2003; Retting et al., 2003; Washington & Shin, 2005; WHO, 2004; IIHS, 2017). The best-controlled studies have found that intersections with high total volumes, higher entering volumes on the main road, longer green (through) cycle lengths, protected left turn phases, and higher publicity may also increase the safety and cost benefits of red light camera enforcement (Council et al., 2005; Washington & Shin, 2005). Additional studies may provide greater insight into whether or not such crashes persist where the technology is in place for longer periods of time. The effect of warning signs, public education, and familiarity with the presence of the system in the fullness of time is not clear.

Warning signs for drivers indicate the presence of automated enforcement systems in the community, and the approaches where the technology is deployed. Washington and Shin (2005) recommended the use of warning signs as they enable drivers to come into compliance before crashes or enforcement events occur, and provide fair warning to drivers of potential enforcement action in general. The researchers also caution that less expensive engineering solutions should be sought before implementing camera programs.

The use of speed cameras can contribute to reductions in speed and crash experience. Decina et al. (2007) reviewed 13 safety impact studies of automated speed enforcement internationally, including one study from a U.S. jurisdiction. The best-controlled studies suggest injury crash reductions relating to the introduction of speed cameras are likely to be in the range of 20 to 25% at conspicuous, fixed camera sites. Similarly, in South Australia, injury crash data for 35 safety camera intersections for 5 years before and after the speed camera installation showed an estimated reduction of up to 21% (Kloeden et al., 2018). Wilson et al. (2010) reviewed 28 studies of automated enforcement from U.S. sites and found reductions of 8 to 49% for all crashes and reductions of 11 to 44% for crashes related to serious injuries and fatalities. Covert, mobile enforcement programs also result in significant crash reductions area-wide (Thomas et al., 2008). Crash-based studies from the United States have reported positive safety benefits of crash and speed reductions from mobile camera enforcement on 14 urban arterials in Charlotte, North Carolina (Cunningham et al., 2008), and from fixed camera enforcement on an urban Arizona freeway (Shin et al., 2009). In Great Britain the effects of fixed speed cameras on crashes were estimated by examining data from before and after camera installations at 2,500 locations. Researchers estimated that installing another 1,000 cameras could prevent approximately 1,130 collisions and approximately 330 serious injuries (Tang, 2017). In France 2,756 speed cameras were installed from 2003 to 2010. A program evaluation estimated that the cameras prevented around 15,000 road traffic deaths during that time (WHO, 2017).

The Shin et al. (2009) study examined effects of a fixed camera enforcement program applied to a 6.5-mile urban freeway section through Scottsdale, Arizona. The speed limit on the freeway was 65 mph; the enforcement trigger was set to 76 mph. Total target crashes were reduced by an estimated 44 to 54%, injury crashes by 28 to 48%, and property damage only crashes by 46 to 56% during the 9-month program period. Since analyses found low speeding detection rates during peak travel times, the target crashes (speeding-related crashes) were considered to be those that occurred during non-peak flow periods (weekends, holidays, and non- peak weekdays hours). In addition to the crash reductions, average speed was decreased by about 9 mph and speed variance also decreased around the enforced zones. Another positive finding from this study was that all types of crashes appeared to be reduced, with the possible exception of rear-end crashes, for which effects were non-significant. Thus, there were no obvious trade-offs of decreases in some crash types at the expense of increases in others. The program effects should be considered short-term. There was also very limited examination of spillover effects, including the possibility of traffic or crash diversion to other routes.

Speed cameras were also installed on Interstate 10, west of central Phoenix, and were supported by mobile (vehicle mounted) speed camera units. In 2009 and 2010, a political determination was made to discontinue the speed camera program. Among the factors impacting the decision was the fatal shooting of the operator of a mobile speed camera in his vehicle that created concerns for the safety of field personnel. Additionally, a change in administration in the State shifted the view of automated enforcement in general, and on the freeways around Phoenix, in particular. However, there are local jurisdictions in Arizona that have retained their automated enforcement systems, and continue to operate speed enforcement and red light camera programs.

Pilot project evaluations of speed camera use in the United States have also obtained promising speed reductions from fixed speed cameras in low-speed school zones in Portland, Oregon (Freedman et al., 2006), and low-speed limit residential streets and school zones in Montgomery County, Maryland (Retting et al., 2008). In the latter case, speed reductions attributed to spillover from the automated enforcement program were also observed on unenforced comparison streets. In an update to the original study by Retting et al. (2008), Hu and McCartt (2016) evaluated speed data from 18 of the 20 original speed cameras and data from 9 of the 10 control sites. Between the 6 months before and 7.5 years after the start of the speed camera program, mean speeds decreased by 13% at the camera sites, 5% at the spillover sites, and by 4% at the unenforced comparison sites. The percentage of vehicles exceeding the speed limit by more than 10 mph decreased by 64% at camera sites, by 39% at spillover sites, and by 43% at unenforced comparison sites.

The percentage of speeders was also substantially reduced when police-operated photo radar enforcement vans were present in a work zone on a non-interstate highway in Portland, Oregon, but there was no carry-over when the enforcement was not present (Joerger, 2010). Given that there was no evidence of any accompanying signs or publicity, there was, however, no reason to expect carry-over outside of the enforced periods. Crash and injury outcomes were not evaluated in these studies.

The use of fixed speed cameras has also been evaluated in Norway. Hoye (2015) investigated the effects of speed cameras on injury crashes and the number killed or severely injured (KSI) on short, medium, and long road lengths downstream of camera sites from 2000 to 2010. Short road lengths were 100 m upstream to 100 m downstream of the camera site, medium road lengths were 100 m upstream to 1 km downstream of the camera site, and long road lengths were 100 m upstream to 3 km downstream of the camera site. There was a 22% reduction in injury crashes on road sections of medium length, but no significant reductions for short or long road lengths. Additional speed cameras installed in 2004 or later furthered the reduction in injury crashes and KSI with 9% and 39% reductions respectively on long road lengths, and 32% and 49% reductions respectively on medium road lengths.

Costs: Costs will be based on equipment choices, operational and administrative characteristics of the program, and specific negotiations with vendors. Cameras may be purchased, leased, or installed and maintained by contractors for a negotiated fee (NHTSA & FHWA, 2008). Most jurisdictions contract with private vendors to install and maintain the cameras and, to process images and violations. A substantial portion of the fines from red-light citations is generally used to cover program costs (Washington & Shin, 2005). Operating costs of automated enforcement systems vary based on the nature of the system, administrative costs, and negotiated fees to vendors providing services to a jurisdiction. Many systems are turnkey operations in which a vendor provides all the equipment, vehicles, and support services necessary to collect violation data and issue a citation. The cost for this service may be based on a fixed monthly fee, or on a negotiated fee for issued or paid citations.

Costs to communities or States for the installation of fixed equipment can vary based on the type of system, the number of devices in use, and the type of sensors being employed to collect violation data. Jurisdictions must make the return on investment decisions for accepting these costs based on their determination of need, risk versus mobility assessment, and budgetary projections and constraints.

Fixed speed camera costs may not be similar to those for red light camera programs, based on volume of activity and violations they generate. An economic analysis estimated the total cost savings of the Scottsdale freeway fixed speed enforcement were from $16.5 to $17.1 million per year, considering only camera installation and operational cost estimates and crash cost impacts (other potential economic impacts were not considered) (Shin et al., 2009). Chen (2005) provides an extensive analysis of the costs and benefits of the British Columbia, Canada, mobile speed camera program and estimated a societal savings of C$114 million and a savings of over C$38 million for the Insurance Corporation of British Columbia that funded the program. Gains et al. (2004) reported a 4:1 overall societal cost to benefit ratio of operating the national (fixed) speed camera program in the United Kingdom based on 33% reductions in personal injury crashes at camera sites and a 40% reduction in the number of people killed and seriously injured. Also, Tang (2017) estimated net benefits of installing 1,000 cameras to be around £21 million in Great Britain based on data from 2,500 fixed cameras crashes.

Time to implement: Once any necessary legislation is enacted, automated enforcement programs generally require up to 9 months to plan, publicize, and implement.

Other issues:

• Laws: Many jurisdictions using automated enforcement are in States with laws authorizing its use. Some States permit automated enforcement without a specific State law. Others prohibit or restrict some forms of automated enforcement (GHSA, 2018a; IIHS, 2019a). In yet others there is no specific statute, and it cannot be inferred from case law whether the State allows automated enforcement. As of December 2018, nine States permit the use of speed cameras under at least some circumstances, 13 States have laws that prohibit speed cameras, and 28 States have no laws addressing speed camera use (GHSA, 2018a). See NCUTLO (2000) for a model automated enforcement law.
• Public acceptance: Public surveys typically show strong support for red light cameras and somewhat weaker support for speed cameras (NHTSA, 2004). A 2011 nationally representative survey of drivers found that 86% thought automated speed cameras would be acceptable to enforce speed limits in school zones. Significant majorities also thought they would be acceptable at high-crash locations (84%), in construction zones (74%), and in areas that would be hazardous for police officers to stop vehicles (70%) or would cause congestion (63%). Thirty-five percent thought automated camera enforcement of speeds is acceptable on all roads (Schroeder et al., 2013). Support appears highest in jurisdictions that have implemented red-light or speed cameras. A survey of District of Columbia residents found 76% favored speed cameras, with even higher support among non-drivers (Cicchino et al., 2014). A larger majority of 87% favored the use of red light cameras. Interestingly, support was lower for measures not currently in use, including photo-enforcement of stop signs (50%) and yielding at crosswalks (47%). Again, support was higher among non-drivers for these measures. However, efforts to institute automated enforcement often are opposed by people who believe that speed or red light cameras intrude on individual privacy or are an inappropriate extension of law enforcement authority. They also may be opposed if they are viewed as revenue generators rather than methods for improving safety. Drivers responding to the NHTSA survey, although indicating support generally for automated enforcement in certain types of locations or conditions, were also more likely to somewhat agree or strongly agree with the statement that speed cameras are used to generate revenue (70%) than with the statement that speed cameras are used to prevent accidents (55%) (Schroeder et al., 2013). Such concerns should be carefully and openly addressed in any automated enforcement program. FHWA recommends, for example, that per citation payment arrangements to private contractors should be avoided to reduce the appearance of conflicts of interest (FHWA, 2005). A case study from the Portland, Oregon, RLC program indicates that the vendor payment structure is a blended contract. The vendor receives a fixed amount per intersection to install and operate the cameras (the city picks the sites) and a monthly amount based on the number of citations that are issued (Eccles et al., 2012). The marginal amount decreases with more citations issued. The current payment structure is $27 per citation for the first 500 paid citations in a month, $20 for citations 501 to 700, and $18 for each paid citation over 700. Two research papers have discussed how Australia and the United Kingdom have dealt with the opponents of and controversies associated with speed cameras and expanded programs at the same time (Delaney et al., 2003; Delaney et al., 2005). Also see Eccles et al. (2012) for more in-depth description of best practices for speed camera programs and case study examples of sustained programs.
• Legality: State courts have consistently supported the constitutionality of automated enforcement (Poole, 2012).
• Covert versus overt enforcement: Covert, mobile speed camera enforcement programs may provide a more generalized deterrent effect and may have the added benefit that drivers are less likely to know precisely when and where cameras are operating. Drivers may therefore be less likely to adapt to cameras by taking alternate routes or speeding up after passing cameras, but data are lacking to confirm this idea (Thomas et al., 2008). Public acceptance may be somewhat harder to gain with more covert forms of enforcement (NHTSA & FHWA, 2008). Fixed, or signed, conspicuous mobile enforcement may also be more noticeable and achieve more rapid site-specific speed and crash reductions at high crash locations. However, the use of general signs in jurisdictions with automated enforcement (not at specifically enforced zones), media, and other program publicity about the need for speed enforcement may help to overcome the idea that covert enforcement is unfair, and promote the perception that enforcement is widespread, enhancing deterrence effects. Based on lessons learned abroad, a mix of conspicuous and covert forms of enforcement may be most effective. See Belin et al. (2010) for a comparison of Australian covert and Swedish fixed, overt systems. NHTSA and FHWA’s operational guidelines document outlines other considerations of overt and covert speed enforcement and signing strategies (NHTSA & FHWA, 2008).
• Halo effects: More research is needed to shed light on spillover effects (positive or negative) of automated speed enforcement programs of varying characteristics. While fixed cameras may yield more dramatic decreases in crashes at the treated sites (which, however, are often sites with high crash frequencies that are likely to decrease in subsequent years) than mobile enforcement, there is little reason to expect that there would be a significant positive spillover effect. In fact, some studies have detected crash migration related to conspicuous, fixed camera enforcement (Decina et al., 2007). There is also a possibility of negative spillover resulting from mobile camera enforcement, but signing and random deployment practices may reduce that possibility (Thomas et al., 2008).
• Average speed (over distance) enforcement: A review of the evidence to date suggests that enforcement (using cameras and camera sites) of average motorist speed over distance is associated with reductions in average and 85th percentile speeds, and the proportion of speeding vehicles (e.g., Ilgaz & Saltan, 2017). Such systems have the potential to reduce speed variability and improve traffic flow characteristics, and may help to avoid negative halo effects such as crash migration to downstream sites that fixed or overt mobile enforcement sometimes experience (Soole et al., 2013). In Italy section control was implemented in December 2005, and by 2014, a total of 320 camera sites were operational. A program evaluation showed decreases in mean speed and speed variability. For example, on urban motorways, mean speed decreased by 10% and the number of crashes decreased by 32% (Montella et al., 2015).
• Enforcement threshold: Victoria, Australia, has had success with a program that tightened enforcement tolerances as part of an overall speed management package that included automated and other enforcement, publicity, and penalty restructuring (D’Elia et al., 2007). An experiment in Finland also found that lowering the enforcement threshold of fixed, speed camera enforcement on a rural, two-lane road from 20 km/h (12.4 mph) above the limit to 4 km/h (2.5 mph) above the limit (advertised as zero tolerance) and publicity of the measure reduced mean speeds by 2.5 km/h (1.6 mph) and speed variance by 1.1 km/h (0.7 mph) in comparison with a similar, camera-enforced corridor where the threshold was not reduced (Luoma et al., 2012). The percentage of vehicles exceeding the speed limit was reduced from 23% to 10%, so deterrence of speeding was increased without increasing the processed citations (police or administrative burden). The speed effect of the reduced threshold was within the range of effect of the initial implementation of the automated camera enforcement.
• Implementation Considerations: Ontario, Canada, offers suggestions for municipalities that are considering initiating a red light camera program based on the lessons learned during 13 years of red light camera program operations. As of 2014 there were over 190 camera operating sites in South and Central Ontario, spanning seven municipalities according to Solomon et al. (2014), who offer suggestions for improving the effectiveness of these programs covering aspects related to planning, implementation, performance, evaluation, and supporting policy.

---