Safety Impacts and Other Implications of Raised Speed Limits on High-Speed Roads (2006)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

36 3 Project Survey Results and Analysis Two surveys were carried out and analyzed during the study. The first was an internet-based survey of state DOTs, intended to obtain information from them on a variety of issues related to speed limits on high-speed roads. Section 3.1 discusses this survey, including its execution and results, as well as the conclusions that were drew from it. The second was a limited telephone survey of state police or highway patrol agencies, intended to learn about agency responses to the NMSL repeal and, more generally, about their decision- making practices regarding the allocation of traffic enforcement resources. Section 3.3 discusses this survey. As will be seen, one of the questions in the survey of state DOTs concerned studies that were performed around the time of the NMSL repeal to plan for and assess the impacts of the speed limit changes. A number of respondents provided references to studies that their Department had carried out or commissioned. These were obtained from the DOT and reviewed by the project. The reviews are also provided in Section 3.2 below. 3.1 Survey of State DOTs The description of Task 2 in the project Scope of Work reads: “Conduct a survey to collect data on the experiences of state DOTs that have raised speed limits. This survey should be designed to collect information including, but not limited to, the following: • range and magnitude of speed limit changes in the states; • mileage and types of highways with raised speed limits; • design and traffic engineering procedures used to determine if and where speed limits should be changed; • legal limitations and other factors in the decision to raise speed limits; • published and unpublished information summarizing the DOT’s experience since raising speed limits; • availability of before and after data on volume, speed, number of accidents, and accident rates; and • willingness of the DOT to provide those data for the study.” Section 3.1.1 below describes the steps that the study team carried out to prepare and execute the survey; Section 3.1.2 presents summaries of and extracts from the responses that were received; and Section 3.1.3 develops the conclusions that the team drew from these responses. A complete copy of the survey instrument is provided in Appendix A.

37 3.1.1 Survey Preparation and Execution 3.1.1.1 Preparation The study team prepared a draft survey questionnaire that covered the data elements specifically identified in the project SOW, as well as others that were of interest to the study. In developing the questionnaire, a number of possible structures were considered. In the end, the team decided on a fairly simple approach that attempted: • to establish the location and characteristics of speed limit increases on high-speed roads since the repeal of the NMSL in 1995; • to identify available data, reports and studies relating to the speed limit changes and their impacts; • to identify other available sources of data on traffic volumes and traffic safety; • to understand the protocols and procedures for decisions relating to highway patrol deployments for speed limit enforcement; and • to understand the protocols and procedures for decisions relating to speed limit increases, either across an entire class of highway facilities (e.g. rural Interstates) or on individual high- speed roads or road sections. A basic issue had to be considered in designing the questionnaire: the answers to many of the questions potentially involved considerable quantities of data. A questionnaire that attempted to collect such data might be excessively unwieldy and, moreover, might dissuade respondents from completing the survey. In general, therefore, it was decided that the questionnaire would not directly ask for such data; rather, it would ascertain if the information existed and, if so, would ask how to obtain it at a later time, if necessary. 3.1.1.2 Execution It was decided to administer the questionnaire via an Internet web site accessible using a standard web browser. Suitable survey respondents were identified in each state DOT with the help of the TRB State Representatives. The designated person was then contacted and asked to contact the survey team to obtain a password with which to access the survey web site. In order to ensure the largest possible response rate, team members repeatedly contacted the designated state DOT personnel over the course of approximately six months, using both email and voice mail for this purpose. The web site home page offered basic information about the purpose and use of the survey instrument, and asks respondents to identify themselves. The questions were divided into sections, roughly corresponding to the types of data identified in the project SOW. In a particular web session, a respondent could choose to answer questions in any or all of the sections. Respondents could save a partially completed survey and return to it later. When the entire survey was complete, the respondent was thanked and the survey permanently saved. (A Microsoft Word file version of the survey was also made available to respondents who preferred to work offline; the completed questionnaire was then emailed to the study team.)

38 3.1.2 Presentation of Survey Results Responses were eventually received from 33 states. The following table identifies the responding states and the main survey respondent in each. Alabama Mr. Timothy Taylor Alaska Mr. Kurt Smith Arizona Mr. David Duffy Arkansas Mr. Mike Selig California Mr. Craig Copelan Connecticut Mr. Robert Uricchio Florida Mr. Patrick A. Brady Georgia Mr. Keith Golden Idaho Mr. Lance Johnson Illinois Mr. L. W. Gregg Indiana Mr. John L. Nagle Iowa Mr. Tim Crouch Kansas Ms. Linda Voss Kentucky Mr. Duane Thomas Louisiana Mr. Peter Allain Maryland Mr. Manu Shah Massachusetts Mr. Richard F. Wilson Michigan Mr. Leo Arens Minnesota Mr. Dan Brannan Missouri Mr. John Schaefer Nebraska Mr. Randall D. Peters Nevada Mr. Michael Lawson New York Mr. David Woodin North Dakota Mr. Allan A. Coulin Oklahoma Mr. Red Miller Oregon Mr. Steve Reed Rhode Island Mr. Frank Corrao III South Carolina Mr. Don Turner Texas Mr. Darren McDaniel Virginia Mr. Curtis Meyers Washington State Mr. Ezekiel W. Lyen West Virginia Mr. Roger L. Russell Wisconsin Mr. John Corbin Some of the respondents provided detailed answers to the survey questions, while others were relatively brief. This section presents their answers, summarizing all the various responses that were received, and in some cases highlighting interesting comments from individual respondents. The presentation follows the organization of the survey itself, and each survey question is discussed in turn. It should be noted that the survey was not intended to be a statistically rigorous effort. Since it is not known why some states responded and others did not, the responses may be biased in some

39 unknown way.13 On the other hand, since responses were received from two-thirds of the states, with no part of the country particularly under- or over-sampled, the results can be considered a reasonable indication of conditions and practices in a substantial majority of the states. Nonetheless, the possibility of a bias should be kept in mind when examining and interpreting the survey results. 3.1.2.1 Answers to Questions in Part A: Speed Limit Change Data A-1) Did your Department raise posted speed limits on any high-speed road sections following the repeal of the National Maximum Speed Limit (NSML) in 1995? Among the states responding, only Oregon did not raise posted speed limits after the repeal of the NSML in 1995. A-2) Has your Department studied the traffic impacts (for example, on speeds, highway safety, volumes and composition, route choice, etc.) of these speed limit changes? A number of states carried out some form of study of the traffic impacts of the speed limit changes following the NMSL repeal. Studies considered changes in speeds, or crash statistics, or both. In some cases, the studies were carried out by state universities or research institutions commissioned by the DOT. Several states that carried out such traffic studies provided references to reports from those studies. In these cases, the project contacted the DOTs to obtain the reports. The reports obtained in this way are reviewed in Section 3.2 below. A-3) Has your Department studied other impacts of the speed limit changes? Examples might include impacts on environmental factors (air quality and/or noise), business and commercial activities, or other areas. A few of the reports identified in question A-2 discuss potential non-safety impacts of speed limit changes, but none mentions a specific study of these impacts. The review of these reports in Section 3.2 identifies these discussions. A-4) Has your department studied the overall benefits and costs associated with the changes? No such studies were mentioned. 3.1.2.2 Questions in Part B: Related Data B-1) What traffic data (such as volume and composition, speeds, number of accidents and accident rates) does your Department collect and maintain on a regular basis? All responding states carry out regular traffic monitoring activities including traffic volume and composition, and crash counts and rates. (The traffic count data might be the responsibility of a different section of the DOT than the crash data, however.) The collection of speed data is more variable between states, with some states collecting it routinely at certain locations; others collecting it at random locations on a spot basis; and yet others collecting it on the occasion of audits or assessments of individual roadways. 13 We were however equally persistent with all non-respondents in our efforts to persuade them to participate.

40 B-2) Does your Department operate any instrumented highways (roadway facilities with a high density of traffic sensors and detectors collecting and recording data at short time intervals on an ongoing basis)? California, Connecticut, Georgia, Illinois, Kentucky, Louisiana, Maryland, Massachusetts, Michigan, Minnesota, New York, Rhode Island, Texas and Washington State reported operating instrumented highways that deploy a significantly greater density of data collection capabilities than is typical of a standard statewide traffic count program. 3.1.2.3 Questions in Part C: Speed Limit Enforcement Decisions C-1) What role, if any, does your State DOT play in determining the levels and location of highway patrol deployments for speed limit enforcement on high-speed roads? No state DOT reported having a direct role in determining speed enforcement activities: this is the responsibility of the State Police, Highway Patrol or similar agency. A number of state DOTs reported that they regularly provide police with data on crash locations and characteristics, and on speed limit compliance. Several responses described specific arrangements that are of interest. The Washington State DOT meets regularly with the State Police to discuss issues of mutual concern, including speed limit enforcement. The West Virginia DOT makes recommendations to the State Police about locations meriting enhanced enforcement activities. In Missouri, the Division of Highway Safety has recently been transferred from the Department of Public Safety to the Department of Transportation. Among other functions, the Department of Highway Safety provides assistance to state and local law enforcement agencies for targeted enforcement. One of the anticipated benefits of the move to MoDOT is more efficient sharing of speed and crash data, which should improve the planning of enforcement activities. Speed limit enforcement in work zones is a special case. A number of respondents described close coordination between their DOT and the police for the patrolling of work zones, including arrangements for compensating the police for this activity. C-2) If the Department is involved in such decisions, how does it decide where and how intensively speed limits should be enforced? In most cases, as noted above, the DOT’s role consists at most of providing speed and crash data to support decisions made by the State Police or another similar agency. A few states provided additional details. New York: “State Police and NYSDOT are involved in an Aggressive Driving Campaign and NYSDOT has identified high accident locations where aggressive driving has led to a significant accident problem.” Virginia: “VDOT coordinates with police for their presence to enforce work zone speed laws on larger projects involving comprehensive Congestion Management plan; VDOT also works with law

41 and public safety agencies to develop incident management components including speed enforcement.” Washington State: “The WSP utilizes a number of statistical analyses from a variety of resources to make deployment and enforcement decisions. Collision data indicating the causation for collisions, speed, DUI, etc., are factors that drive these decisions. Enforcement practices are based on data driven decision making based on strict analysis of the WSDOT Speed Report for example.” West Virginia: “In cooperation with the WV State Police, the DOT installs aerial enforcement markings at select locations on Interstate and other four lane highways. Our only other involvement is to recommend to the state police that they provide enhanced enforcement at certain locations identified as having a higher than average percentage of motorists violating the speed limit.” C-3) Were there changes in your State’s enforcement policy following the repeal of the National Maximum Speed Limit in 1995? Of the 33 respondents, 15 stated that there had been no change in enforcement policy. Ten respondents either indicated that they didn’t know, or referred us to the State Police. Four respondents did not answer the question. Four respondents indicated that there had been a change in enforcement policy; their responses are reproduced below: New York: “Not sure, but initially State Police tried to prevent the unofficial tolerance level (about 7 to 8 MPH) from rising [after the NMSL repeal]. Speeds were already 10 to 15 mph above the 55 MPH speed limit and they did not want to see such a uniform shift occur when the speed limit became 65 MPH. So the enforcement tolerance initially became tighter, but as resources have become more difficult, the tolerance has loosened. I have had troopers tell me that you will get a 7 or 8 mph tolerance in a 65 MPH zone because judges want to see tickets that are non- disputable. Somebody could dispute that the radar or speedometer is off a few MPH, but if you are doing +15 MPH over the limit, then they got you.” North Dakota: “Enforcement policy changed on four-lane divided highways.” Oklahoma: “Only one change regarding speed limits occurred, which established a zero tolerance policy; and tickets for less than 9 miles over the speed limit were not recorded on the DMV record.” Washington State: “Strict enforcement of the new 70 mph speed limit [following the NMSL repeal] was instituted. Media resources were solicited to provide the public with information on the change and the enforcement decisions.” C-4) What are the levels of traffic fine for different degrees of speeding? Are there other penalties as well (e.g. driver’s license revocation)?

42 A number of states provided detailed descriptions of the system of speeding fines and penalties in place, or references to their on-line statutes. Many respondents suggested that the survey personnel contact the State Police or similar agency for information, or provided references to their on-line statutes. More detailed responses were received from several states. C-5) What is the legal Blood Alcohol Content (BAC) in your State, and how and when has it changed in the last decade? Most states reported that a BAC limit of 0.08% has been instituted relatively recently, although in a few cases the 0.08% limit has been in force for several years or more. A number of the states that still have a BAC limit of 0.10% in effect reported that the lower 0.08% limit is currently under consideration by the legislature. C-6) Does your State have graduated driver’s licenses and, if so, how and when did these arise? What sorts of restrictions on young drivers are in place? Most respondents reported having graduated driver’s licenses for young drivers and, in most cases, these were initiated in the past few years. (The respondent from Maryland pointed out that it was the first state to institute graduated licenses, in 1978.) A number of these respondents provided detailed information on the restrictions that are entailed; these typically involve limitations on driving after dark, on the number of passengers, and the like. Connecticut was alone in reporting no restrictions on young drivers. Other respondents referred us to the state police or similar agency. C-7) Please suggest someone we might contact in another State agency (Department of Public Safety, State Police, etc.) for further information on speed limit enforcement decisions. Most states recommended that the survey personnel contact the State Police or related agencies for further information on speed limit enforcement decisions. Responses usually provided information on specific contact persons in those agencies. Section 3.3 below presents the results obtained from a follow-on survey of State Police agencies. 3.1.2.4 Questions in Part D: Speed Limit Change Decisions D-1) Please describe how speed limits are determined for high-speed roadways in your State. In determining speed limits, how much importance is given to design speeds versus observed uncongested operating speeds? In general, most respondents reported that maximum speed limits on particular road classes are fixed by statute. Decisions about speed zones on particular road sections are generally based on engineering studies that may be initiated directly by the DOT, or at the request of local governments, police or (in some cases) citizens’ or other groups. Such studies typically rely heavily on (i) the section’s crash history, when safety is a particular issue in determining the speed; and (ii) observations of the 85th percentile operating speeds under free-flow conditions, in accordance with recommendations of the Manual on Uniform Traffic Control Devices (MUTCD) or its state equivalent(s). A number of other factors also may play important, but secondary, roles. On newly opened roads, the initial speed limit is generally determined using the design speed, but the limit may later be revised as operational experience is accumulated.

43 Arizona stated that a section’s design speed did not affect its speed limit determination. Iowa explicitly mentioned design speed as a factor in determining speed limits, but emphasized that it is secondary to the observed speed. Massachusetts mentioned that a section’s design speed was only used as an upper bound in setting its speed limit. Oregon stated that the design speed was used on new highways to determine a temporary speed limit, but that the permanent limit was based on observations of the 85th percentile speed after the road had been in service for some time. California, Michigan, and Minnesota reported having datasets for comparing speed limits, design speeds, and operating speeds. Some of the more detailed responses to this question are reproduced below: Illinois: “Maximum speed limits are statutory and determined by legislative action. Altered speed limits are set on the basis of an engineering and traffic investigation based principally on the 85th percentile speed with some adjustment factors such as the crash rate, access points, pedestrian activity and parking. Taken together, such adjustment factors may justify an altered speed limit up to 9 miles per hour less than the 85th percentile speed.” Nevada: “Special traffic studies are conducted considering the criteria specified in the MUTCD. A recommendation is then made by the Traffic Information Division to the Chief Traffic Engineer who considers other criteria unique to the study area before forwarding his recommendation the Director, who makes a final determination based on the recommendations and his personal judgment. Design speed is viewed as the upper limit regardless of operating speeds, but operating speeds can suggest a lower limit than the design speed.” D-2) Please describe in detail the process by which your Department decides to modify (raise or lower) the posted speed limit on high-speed roads, either for individual road sections or for an entire class of facility (e.g. rural or urban interstates, other limited access facilities, other high-speed roadway). The following comment is a typical response to this question: North Dakota: “Typically speed studies are conducted upon receiving a request to alter a speed zone. Based on the 85th percentile speed we try to set a realistic speed limit within 5 mph of this speed within legislative limits. Other considerations include horizontal or vertical alignment, crash rates/types, roadway design, access control, adjacent development and political.” Other factors that were cited as considerations in deciding about a possible speed limit change include crash history, sight distance, traffic volume, “field conditions”, “engineering judgment” and the like. A number of states simply referred us to the MUCTD or their local equivalent for detailed information on the process used to make decisions about modifying speed limits. D-3) Has the Department established any written rules or guidelines to be followed when making these decisions?

44 Arizona, California, Florida, Georgia, Idaho, Illinois, Kansas, Kentucky, Louisiana, Maryland, Massachusetts, Michigan, Minnesota, New York, Oklahoma, Oregon, Texas, Virginia and West Virginia referred to specific written manuals, rules or policy guidelines for making decisions about changing the posted speed limits. D-4) Please describe other factors (legal limitations, public opinion, interest groups, political considerations, etc.) that play a role in making decisions about raising speed limits. A number of respondents described mechanisms in place in their states by which public opinions about speed limits can formally be expressed: for example, provisions by which citizen or other groups can request speed studies. Comments on the influence of political or other factors in speed limit decisions were few and generally rather circumspect. A sample of the responses is given below. Arizona: “While we try to minimize the above factors they do occasionally play a role.” California: “Public opinion is a high priority when changing speeds on high speed facilities; California Vehicle Code section 22345.5, which stipulates that proposed speed limit changes must be coordinated with the California Highway Patrol and that county or city boards of supervisors may arrange public hearings to discuss proposed speed limit changes, ensures this. In addition, when the freeway speeds for cars and trucks were considered for changing after the repeal of the National Maximum Speed Limit, a committee was put together by the management of Caltrans to evaluate the proposed changes and to determine how best to proceed with the existing differential speed limit of 55 mph for heavy trucks. Committee membership included representatives of the Highway Patrol, trucking and insurance industries, local government and various safety interests who collectively made the decision to retain the differential speed limit on California freeways.” Iowa: “The Iowa legislature has debated the speed limit issue every year since the repeal of the national maximum speed limit. So, yes legal limitations and political considerations are considered in the speed limit decision. In the past, we have also sought local/public opinion by informing the county/city officials of a planned increase in the speed limit for individual sections of road. In a couple of situations, the locals were opposed to the increase and it did not happen.” Minnesota: “It would take 3 pages to answer these factors’ individual impacts but it safe to assume that they do impact the decision making process based upon the intensity of each issue.” Missouri: “According to state statute, [the DOT’s] study is the only deciding factor in raising/lowering speed limits. We may consider requests to change speed limits.” Nevada:

45 “All of the above [i.e. legal limitations, public opinion, interest groups, political considerations, etc.] factor into the Director's final decision, but are not considered as part of the special studies that are conducted.” 3.1.2.5 Comments From Part E: General Comments E-1) We welcome any observations that you may have about the impacts of speed limit changes on high-speed roads in your State. As an example, you might have comments on the following questions: • Overall, has the repeal of the NMSL affected traffic safety in your State? • Have speed limit changes on high-speed roads influenced driver behavior and/or traffic safety on other road classes as well? • Have truck route choices changed since the NMSL repeal? Are some portions of your State’s roadway network either safer or less safe because of this? • Has the elimination of NMSL-related speed enforcement mandates changed the focus of highway patrol activities? Has this had an effect on traffic safety? • Have changes in speed limits on high-speed roads impacted the environment? the business community? public opinion? other impacts? • Any other issues that you would like to raise? This question elicited useful and interesting comments from a number of respondents. There were several brief anecdotal comments to the effect that the NMSL repeal either (i) had no effect on traffic safety, or (ii) had an initial negative effect that became negligible after a few years as drivers adjusted to the changed speed situation. Some of the more illuminating responses that were received are reproduced below. In an attempt to ensure candid responses, the instructions for this question in the survey promised that comments on this question would remain anonymous. Accordingly, any readily-identifying information that was included in the replies has been redacted. Respondent 1: Overall, has the repeal of the NMSL affected traffic safety in your State? “I will provide this summary that I had prepared back in 1998. On December 18, 1995 the speed limit was raised from 55 mph to 65 mph on approximately 2,500 miles of freeways. On January 8, 1996 the maximum speed limit on nearly 1,300 miles of rural freeways was increased from 65 mph to 70 mph. Safety: “The fatal accident rate (accidents/MVM), for all freeways where the speed limit was changed, increased by about 4.4%. This change is based on accident data for two years before and two years after the speed limit change. “The fatal plus injury accident rate (accidents/MVM), for all freeways where the speed limit was changed, increased by about 2.3%. This change is also based on accident data for two years before and two years after the speed limit change.

46 “The fatal accident rate for all freeways statewide (includes freeways not having a speed limit change) increased by 1.5%. There was no significant change in the fatal plus injury accident rate for all freeways statewide. These results are based on a comparison of accident data for 1994- 1995 with accident data for 1996-1997. “The percentage of total accidents where speed was the primary cause increased slightly from 44.8% before the speed limit change to 44.9% after the speed limit change. These percentages are based on accident data for a representative sample of locations where the speed limit was changed and on similar 18-month before and after periods. In comparison, the 1997 statewide percentage of total accidents related to speeding was approximately 42.6%. “The percentage of total accidents where the influence of alcohol was the primary cause decreased from 6.4% before the speed limit change to 5.5% after the speed limit change. These percentages are also based on accident data for a representative sample of locations where the speed limit was changed and on similar 18-month before and after periods. In comparison, the 1997 statewide percentage of total accidents involving the influence of alcohol was approximately 6.1%. Operations: “The average 85th percentile speed of vehicles within the 65-mph zones changed from 67.1 mph before to 68.8 mph after the speed limit increase, a 1.7-mph or 2.5% increase. “The average 85th percentile speed of vehicles within the 70-mph zones changed from 70.3 mph to 72.3 mph after the speed limit increase, a 2.0-mph or 2.8% increase. Conclusions: “The average 85th percentile speeds have increased a very small amount (1.7 mph and 2.0 mph) since the initiation of the new higher speed limits on [our] freeways. “Safety data indicates a 4.4% increase in the fatal accident rate and a 2.3% increase in the fatal plus injury accident rate for all freeways having a speed limit change. These changes are not significant, considering that the accident rates for the after period are both lower than the 1996- 1997 statewide accident rates for all freeways (6.0% lower for the fatal accident rate and 16.5% lower for the fatal plus injury accident rate). “The percentages of total accidents, where speed and the influence of alcohol were the primary causes, did not change significantly after the speed limit increase. The percentages of these types of accidents were approximately the same as the statewide percentages for 1997. “Subsequently reports prepared by [a state] university which evaluated the before and after impact of the change in speed limits using a predictive model determined that collision rates did not decline as rapidly on freeways where the speed limit had been increased as those where the speed limit was retained at 55.” Have speed limit changes on high-speed roads influenced driver behavior and/or traffic safety on other road classes as well?

47 “Unknown.” Have truck route choices changed since the NMSL repeal? “Unknown.” Respondent 2: “When the NMSL was repealed, the […] Department of Transportation, along with the […] State Police, organized a task force to study the impact of the abolishment of the national maximum speed limits. This study included engineering traffic investigations, as required by state law. These investigations used data from automatic speed monitoring stations which were then in place throughout the state and applied widely accepted speed limit determination methodologies. The study justified retention of the 55-mile-per-hour speed limit for the conventional state and local two-lane highway system and the 65-mile-per-hour speed limit for the rural freeway system. This decision was supported by the […] General Assembly […]. This act established statutory maximum speed limits of 55 miles per hour on both the state and local highway systems and a maximum speed limit of 65 miles per hour on both the rural freeway and tollway system. “The task force did, however, recommend increasing the speed limit to 65 miles per hour on some rural, four-lane divided, high-type highways and on some portions of rural freeways which fell under the federal guideline for urban areas, but which operate as rural highways. A review of the accident rates on these systems showed that the speeds could safely be increased with little, if any, impact on either the number or severity of accidents. “Altogether, the speed limit was raised on less than 250 miles of highway (less than one percent by mileage). This included 126 miles of Interstate and other freeways and 118 miles of 4-lane divided, high-type highways. “We have no evidence that the speed limit changes on high-speed roads influenced driver behavior and/or traffic safety on other road classes or roads, no evidence that truck route choices changed, and no evidence of any effect on traffic safety or the environment.” Respondent 3: “Drivers’ average speeds have slowly been increasing.” Respondent 4: “The traffic safety in [our state] has been affected by the repeal of the NMSL. There is continued pressure to increase the speed limit and our experience on those sections that have had increases has indicated an increase in the severity and numbers of crashes. Even if the increased speed limits do not lead to additional crashes, those crashes that do occur will be more severe.

48 “We continue to see slight increases in speeds on our two lane highways and our interstate highways. I don't know if that is related to the higher speed limits in other states or more a reflection of today's society. “With increasing speed limits, there are a couple other issues that surface and should be addressed. Should trucks be allowed to drive the faster speeds? This inevitably comes up during legislative debates, should there be a separate lower truck speed limit. Past research is limited and the results vary. The other issue is the minimum speed limit. We currently on have a minimum speed limit of 40 mph on the Interstate system. It was 40 mph when the maximum was 55 mph. Now the maximum is 65 mph and the minimum has not changed. If the maximum is increased again, should the minimum speed limit be increase to try and keep the variance in speeds at a minimum level?” Respondent 5: “…As I recall crashes either stayed the same or were reduced on 93 percent of our highways when we raised the speed limit.” Respondent 6: “We’ve experienced a slight increase in speeds.” Respondent 7: “We have no unique data to quantify any of these issues.” Respondent 8: “The NMSL never was about safety, it was about energy conservation. Pick whatever study you like and unscientific people will distort the results to fit their argument. I believe any legislatively mandated speed limit is inappropriate. Different roadways have different ‘characteristics’ and those, along with prevailing speeds, should determine the appropriate limits. “The repeal of legislatively mandated limits was a good thing and a long time coming. States have the responsibility for all matters concerning their roadways and ought to have the authority to manage them as they see fit. The NMSL was yet one more example of creeping Federalism, as is the .08 BAC, federally mandated seat belt laws, federally mandated revocation of drivers licenses for any drug related offense, the LCV size and weight freeze, and on and on. The elimination of any and all of them is in keeping with the fundamental principals of the Constitution as articulated in the 10th amendment which is that ‘The powers not delegated to the United States by the Constitution, nor prohibited by it to the States, are reserved to the States respectively, or to the people.’ Amen to that.” Respondent 9:

49 “1. Repeal has had minimal effect on traffic safety: On highways with 65 MPH all accident rates are down, number of fatalities are down, but total number of accidents and injuries are up. “2. Yes, speeds are up on other highways as well. “3. Trucks probably are sticking more to the interstates with 65 MPH as it allows them to get to their destinations faster and with less fear of a ticket. “4. Not sure - Need to ask State Police. “5. People seemed to be satisfied with the 65 MPH speed limit based on the volume and type of letters we receive on the subject. Most letters now ask why can't you add highway XX, rather than I wish you did not increase the speed limit to 65 MPH.” Respondent 10: “The elimination of NMSL related speed enforcement mandates, have not been appreciable. “Public opinion seems to be favorable towards the increase in speed limits in [our state]. “Some truck traffic may have changed to include roadways with higher speed limits. The Patrol does not have data to substantiate this. “Traffic fatalities have not risen in [our state] since the speed limit increases in 1996. In some years after 1996, traffic fatalities were substantially lower.” Respondent 11: “[Our] drivers tend to drive up to 10 miles per hour over any posted speed limit. Raising the speed limit to the 85% usually results in motorist driving faster than the 85% speed. The key to maintaining safe speeds on our roads is enforcement and there is very little enforcement of motorists speeding up to 10 miles per hour over speed limit.” Respondent 12: “[Our] maximum speed limits are legislated. The lower speed limits are determined by the 85th percentile speed. Environmental Speed Limits have been enacted in [three cities]. The lower speeds did not have much of an impact on the 85th percentile speeds.” Respondent 13: “Since the repeal of NSML in 1995, most states have changed speed limits on their interstate highway systems in our nation. Due to there not being any policy, procedure or standard, the state agencies made their decisions on their own. This situation has brought lots of controversies or concerns regarding the effects of such changes on safety, operation, roadway design,

50 enforcement, environment and productivity, etc. In addition, no reliable research reports have been found. “Therefore, it is strongly recommended that NCHRP prepare a guideline for the changes of speed limits based on engineering and statistical data analysis.” Respondent 14: “When we raised speed limits, we set them to reflect the actual operating speeds on the affected roadways. We have not identified any significant impacts on route choices, safety or other issues.” 3.1.3 Conclusions From the Survey of State DOTs The results of the survey of state DOTs provided the study team with a considerable amount of useful information. Although the survey was not exhaustive, many of the responses shed light on hypotheses about speed limit change impacts that were of interest to the study. This section highlights a number of the most relevant and interesting conclusions that were drawn from survey results. 3.1.3.1 Studies of the Traffic Impacts of the NMSL Repeal In general comments, respondents felt that average travel speeds had increased following the repeal, but most noted that this was part of a general trend established over a long time period and affecting most road classes. With one exception, survey respondents intuitively felt that impacts of the NMSL repeal on traffic safety had been either insignificant or non-existent. One respondent did cite an increase in crash numbers on sections with raised speed limits, but did not provide data to quantify the effect. The same respondent noted that higher speeds could be expected to increase the severity of the crashes that do occur. 3.1.3.2 Studies of the Non-traffic Impacts of the NMSL Repeal No state that responded to the survey has studied the non-traffic impacts of the NMSL repeal, with one partial exception: the NYSDOT study of post-NMSL impacts did consider effects on the trucking industry. 3.1.3.3 Impacts of the NMSL Repeal on Traffic Enforcement State DOTs that responded to the survey reported that they have no direct responsibility for decisions about traffic enforcement; their role is generally limited to providing the State Police or similar agency with data on locations with high crash rates or speeds. The respondents made very few references to systematic changes in enforcement policies post- NMSL. A number of states stated that they became less tolerant of speeding than they had been

51 during the period of the NMSL. No information obtained so far supports the hypothesis of a systematic shift away from speed enforcement on high-speed roads to other types of policing activity following the repeal. Interestingly, Washington State specifically noted that a policy of strict speed enforcement was implemented specifically on roads with limits that were raised to 70 mi/h. 3.1.3.4 Systemwide Impacts of the NMSL Repeal The project Scope of Work asks for an analysis of the “systemwide effects” of speed limit changes, mentioning trip generation and diversion as specific examples. None of the respondents referred to trip generation impacts in any way. Two respondents speculated that some truckers might have shifted to Interstates with higher speed limits following the repeal, in order to save travel time and reduce their exposure to police patrols. Other respondents simply replied that no data was available to investigate this potential effect. Some researchers have speculated that speed limit changes on high-speed roads could have “spillover” effects on other road classes as, for example, drivers change their tripmaking (particularly route choice) decisions or police change their speed enforcement strategies. Two respondents explicitly mentioned that there was no evidence to support such a hypothesis, while the others did not refer to the idea. It seems likely that such effects, if present at all, are small and difficult to detect. 3.1.3.5 The Speed Limit Change Decision Process Little information was received concerning decisions that affect the maximum speed limit for entire classes of roadway facilities; the responses tended to focus much more on decisions affecting individual facilities or roadway sections. The practical details of the speed limit change decision process vary somewhat from state to state, but all responding states rely on observed 85th percentile free flow speeds as one of the principal factors used in setting speed limits, with consideration also given to any of a number of secondary factors. The FHWA’s Manual on Uniform Traffic Control Devices (MUTCD) is the standard engineering reference used in setting speed limits, although a number of states have developed local adaptations of the MUTCD and/or speed limit policy guidelines. Most states reported that a roadway section’s design speed was not considered when setting speed limits, although a few respondents mentioned using design speed as a secondary factor, or as the basis for setting the initial speed limit on a newly opened road. NCHRP project 15-18 focused on relationships between design speeds, operating speeds and speed limit decisions. The final report of the project (Fitzpatrick et al. 2002) may be consulted for further information on this topic.

52 3.1.3.6 Other Insights from the Survey One respondent mentioned the question of different speed limits for trucks as an issue worth considering. The same respondent suggested that the impact of changes to minimum speed limits should also be studied. When the NMSL was in effect, the national minimum and maximum speeds on Interstates were 40 mi/h and 55 mi/h, respectively. With the NMSL repeal, many Interstates have maximum speed limits of 70 mi/h or more, while the minimum speed has generally remained at 40 mi/h. Has this increase in the range of legal speeds produced traffic or other impacts? The literature appears to be silent on this question. Other respondents emphasized the importance of speed limit enforcement and enforcement policies in determining actual speeds. One suggested that drivers use a posted speed limit as a reference and then choose to drive at a somewhat higher speed; strict enforcement is needed to ensure that the limit is actually respected. 3.2 Review of State Studies of the NMSL Repeal Question A-2 of the survey of state DOTs identified studies of the NMSL repeal that had been conducted by the Departments. The project obtained and reviewed these studies. The following subsections present and discuss the reports so obtained. 3.2.1 Arkansas In Arkansas, the State Highway Commission is responsible for establishing speed limits. Following the NMSL repeal, the Commission met in July and August 1996 to consider increasing speed limits on freeways with rural and suburban characteristics, and on rural expressways with high-type partial access control. The recommended new speed limits were as follows: Rural freeways: cars: 70 mi/h; trucks: 65 mi/h Suburban freeways: 65 mi/h Urban freeways: 60 mi/h Rural expressways: 65 mi/h, except where review indicates need for a lower limit These recommendations were based on a Speed Limit Study for Arkansas Highways (1996) prepared earlier by the Arkansas State Highway and Transportation Department. In addition to its recommendations, this study noted briefly the increased mobility, efficiency, time savings and cost savings that might derive from increased speed limits. Following implementation of some of the recommended speed limit changes, a Study of the Safety Impacts of Increased Speed Limits in Arkansas and a Review of Proposed Speed Limit Increases on Additional Routes (November 1997) carried out a simple before-after comparison of fatal crash counts and fatalities one year before and after post-NMSL speed limit increases. Possible changes in traffic volumes and other safety-related factors were not controlled for. The study found a 5% increase in fatal crashes and a 15% increase in fatalities on all routes with

53 raised limits. The changes varied considerably by facility type. Fatal crashes and fatalities decreased by 12% and 10% respectively on rural freeways, but increased by 89% and 145% on suburban freeways. In 1998, the State Highway and Transportation Department completed a follow-up study, Speed Limit Study for Arkansas Highways – 1998 Update. This study built on the fatal crash and fatality data developed by the 1997 study for different roadway classes following the changes suggested in 1996. It generally recommended full implementation of the post-NMSL limits, consistent with safety and speed data. It noted that the recommended limits would provide a smooth transition between speed limits on urban, suburban and rural freeways. In the particular case of rural expressways, it recognized their variability regarding degree of access control and connectivity to other similar facilities, and recommended lower speed limits for specific highways. 3.2.2 Iowa The 1995 NMSL repeal led to the preparation of a Report on Speed Limits and Safety for Iowa Highways, published by the Iowa Task Force on Speed Limits in January 1996. The report contained a thorough and thoughtful review of possible approaches to post-NMSL speed zoning policy in the light of Iowa and national speed and traffic safety trends. However, it did not make specific recommendations regarding speed limits, preferring to compile, analyze and present the relevant facts, and leaving speed limit policy decisions to the State Legislature and other policy makers. In 1996, the Legislature authorized the Iowa DOT to increase speed limits to 65 mi/h on certain divided, multi-lane highways. Following an initial review, 248 miles of highway had their speed limits so increased. By January 2001, a total of 680 miles on 28 sections of Iowa’s rural freeways and expressways had had their limits increased to 65 mi/h following a review of section design characteristics and crash history, and a field inspection. The Task Force on Speed Limits published in January 1997 a Report on Results of Speed Limit Changes After Repeal of the National Maximum Speed Limit. Subsequently, it has published annual update reports on speed limits; the latest available report is for 2002. These annual update reports track a number of speed, travel, safety and enforcement variables. They also contain comparisons of Iowa’s highway safety experience with that of neighboring states. The reports examine trends in the percentile distribution of speeds. They document a steady upward trend in 85th percentile operating speeds in the years following the implementation of the new speed limits. They also note an irregular downward trend in the pace speed (the 10 mi/h speed range containing the highest number of vehicles), which suggests increasing speed variance. However, it is possible that this trend reflects the effects of increasing congestion (which tends to accentuate the differences between peak and off-peak travel conditions) more than those of the speed limit change (which may or may not increase the variability of individual vehicle speeds). The reports also highlight a roughly upward trend in the percentage of vehicles exceeding the 65 mi/h speed limit. However, a much larger fraction of Iowa drivers complies with the 65 mi/h limit than did with the 55 mi/h limit during the NMSL period.

54 With respect to rural expressways and freeways for which the speed limit was increased, crash rates were found to increase between mid-1996 and the end of 1997. These increases were found across all considered crash categories, including fatal crashes, fatal and injury crashes, all crashes, fatalities, fatalities and major injuries, and other (less severe) injuries. All crash rate categories increased by at least 20%, with fatal crashes and fatalities increasing by 497% and 587% respectively. Crash rates on non-interstate freeways with full access control were roughly two to three times higher than rates on the Interstate system. The number of speeding citations issued by the State Patrol did not change significantly in the period 1993-1999; however, the number decreased in 2000 and 2001. 3.2.3 Kansas Speed limit statutes in Kansas were changed in March 1996, following the NMSL repeal. The revised statues authorized limits of 70 mi/h (up from 65 mi/h) on most rural multi-lane divided highways, and of 65 mi/h (up from a variety of lower limits) on most urban interstate and two- lane rural highways. By June 1996, new speed limit signs had been posted on most highways in the state. A study conducted for the Kansas Department of Transportation (KDOT) by researchers at Kansas State University (Najjar et al. 2000) investigated possible changes in speeds and crash rates as a result of the speed limit changes, applying before-after analysis methods to both rural interstate and rural two-lane highways. The speed analysis concentrated on the 85th percentile speed, using t-tests to compare section speeds in the approximately one-year periods before and after the speed limit change. On rural interstates, the research found (at a 95% confidence level) that the 85th percentile speed increased by 3 mi/h following a 5 mi/h speed limit increase. The safety analysis distinguished rural interstates, urban interstates and two-lane rural highways, and considered total crash rate, fatal crash rate and fatality rates using a crash database maintained by the KDOT Bureau of Transportation Planning. Investigation of possible rate changes was conducted using a three-step sequential analysis methodology consisting of: • a statistical (t-test) comparison of monthly crash rates of each type in the periods 1993-1995 vs. 1997-1998 (the year 1996, in which the speed limit changes took effect, was intentionally omitted); • an examination of time series plots of crash rates in the period 1993-1998, to determine if any significant change in rates could be visually detected; and • a final determination, considered conclusive if the preceding two steps agreed regarding the presence or absence of a significant change, and inconclusive otherwise. This methodology determined that there was no significant change in the three considered crash rates on either rural or urban interstates following the speed limit change. On the other hand, the changes in these rates on two-lane rural highways were found to be significant. These latter

55 represent a group with heterogeneous engineering and traffic characteristics, and saw post- NMSL speed limit increases ranging from 5 to 20 mi/h. Closer examination of the results revealed that about 7% of the rural two-lane highway sections accounted for most of the observed increase in crash and fatal crash rates. 3.2.4 Louisiana In Louisiana the speed limit on rural interstates was raised from 65 to 70 mi/h on August 15, 1997; limits on urban interstates remained unchanged. Schneider (undated) carried out an analysis of the impact of the speed limit change on the number and severity distribution of crashes, distinguishing between fatalities, injuries and PDO crashes. A comparison of roadway crash performance in 1996 and 1998 identified a 37% increase in fatal crash counts on interstates, albeit with no overall increase in fatal crashes when all roadways were considered. Comparing the same two years, there was a 9% decrease in injury crashes across all roadways, but a 1% increase on interstates; similarly, there was a 2% increase in PDO crashes on all roadways, but a 14% increase on interstates. Louisiana has specific safety issues that most other states do not have. The presence of swamps in much of southern Louisiana results in significant numbers of elevated bridges and highways. These are particularly prone to crashes because of the limited maneuvering room and the prevalence of fog due to the nearby water. The analysis found considerably greater increases in all crash rates on elevated interstate facilities compared to other interstate facilities. The report did not examine VMT and crash rates, but is noteworthy for considering a variety of other safety-related factors including, among other things, the age and gender distribution of drivers and their crash involvement, the distribution of crashes by day of week and time of day, the use of seatbelts and the effects of weather conditions. The study also attempted to ascertain the change in fuel consumption costs associated with the speed limit change. Changes in VMT and taxed fuel consumption were very small between 1996 and 1998. The report notes an overall decrease of 0.2% in fuel economy (mi/gallon) during this period, but notes that the changing vehicle fleet (increased numbers of SUVs and light trucks) makes it impossible to impute this change to the speed limit change. Considering the effect of speed limit change on travel time costs, the study notes that approximately 85% of travel in Louisiana occurs on urban interstates and roads with speed limits of 55 mi/h or less. About 25% of VMT is produced on interstates, of which 60% (i.e. 15% of total VMT) occurs in rural areas where the speed limit was increased. The report concludes that the change in travel time and associated costs would be most strongly felt by long distance travel, such as commercial trucking and vacation travel. 3.2.5 Michigan Following the NMSL repeal, the Michigan Legislature directed the Michigan Department of Transportation (MDOT) and the Michigan State Police (MSP) to designate 500 miles of rural

56 freeway for which the speed limit would be increased from 65 to 70 mi/h, and to study the impact of the speed limit change on vehicle speeds and crashes over a six-month period. Taylor and Maleck (1996) carried out this study. They found increases in the 50th and 85th percentile speeds of less than 2 mi/h at some locations, with most locations exhibiting increases of less than 1 mi/h. The lags inherent in reporting crash data did not allow an analysis of the crash impacts of the speed limit change. The legislature then authorized the MDOT to raise the speed limit on an additional 1000 miles of rural freeway on January 1, 1997. (Truck speeds remained at 55 mi/h during this time.) The study of the speed limit change impacts was expanded to include the additional freeway sections. Following a series of reports on the shorter-term impacts of the changes, Taylor (2000) issued a comprehensive study of the speed limit change impacts over the period 1997-1999, comparing traffic and crash data for this period with those for the three-year period preceding the speed limit change. It was found that, over the considered period, total crash counts increased by 10.5%, severe crash counts increased by 4.5%, but fatal crash counts decreased by 9.3% on the freeways experiencing the speed limit changes. Lack of comprehensive traffic volume data made it impossible to compute crash rates and identify rate changes. However, a comparison of the increase in total crash counts with the estimated growth in VMT during the same period suggested that the total number of crashes increased by less than the overall traffic growth rate. 3.2.6 New Jersey The 1987 NMSL relaxation had little effect in New Jersey because it concerned only rural interstates, of which New Jersey has very little mileage. The 1995 NMSL repeal, on the other hand, led the New Jersey Legislature in late 1997 to raise the 55 mi/h speed limit to 65 mi/h on portions of the highway network including interstates and other highways with similar design and access control. A default limit of 65 mi/h was also established for the New Jersey Turnpike, the Garden State Parkway and the Atlantic City Expressway. It was decided that limits would be raised on approximately 400 miles of highway, and that these would be monitored over an 18-month period to determine the impacts of the speed limit change, in order to develop a policy regarding speed limit changes on other portions of the highway network. Impacts monitored during the 18-month period were to include travel speeds, safety performance, enforcement experience and environmental impacts (air quality and noise). Approximately 475 miles of roadway were selected for this purpose in May, 1998. A number of criteria were applied in the selection process, including minimum section length (10 miles), minimum design speed (65 mi/h), adequate spacing of access ramps, and absence of significant recurring congestion (to avoid creating unsafe driving situations by the speed limit increase). The application of these criteria had the effect of concentrating the selected segments in rural and suburban environments.

57 In advance of the speed limit change implementation, a more aggressive traffic fines schedule for violations of the 65 mi/h limit was developed and put in place. “Before” traffic volume and speed measurements were also carried out on and near all the highway sections designated for changed limits, as well as on a sample of highways for which the speed limit would not change. NJDOT measured all sections’ traffic volumes and speeds at least once every three months during the 18-month period. Throughout the period, the State Police collected data on crashes and traffic law violations. Average speeds were found to change by minimal amounts on the sections with changed speed limits. Typical changes were less than ±2 mi/h. The only exception was on the New Jersey Turnpike, for which average speed increases of 3-4 mi/h were found. It was noted that other factors, such as enforcement policy changes and public outreach efforts, probably affected the observed speed changes. Air quality and environmental noise level changes were determined using standard models based on measured traffic volumes and travel speeds, and taking account of possible travel pattern (routing) responses to the speed limit change. It was determined that the speed limit change resulted in increases of 0.20%, 0.90% and 1.15% in traffic-related emissions of volatile organic compounds (VOCs), nitrous oxides (NOx) and carbon monoxide (CO), respectively. These changes were deemed nominal. Similarly, it was determined that the noise level change resulting from the observed (small) travel speed changes would not be perceptible in the noise environment adjacent to the highways. Fatal crashes and fatalities decreased on the concerned sections by 7.9% and 9.6%, respectively, in the 18 months following the speed limit change compared to a similar period before the change. Total crashes increased by 18.3% on these sections, while the number of crashes with injuries and number of injuries increased by 9.4% and 5.9%, respectively. It was found that 55 mi/h zones adjacent to the 65 mi/h zones exhibited slightly greater increases in total crash counts. The study noted that, in the period between 1984 and 1996, crash rates on New Jersey highways varied by up to 12% per year. Consequently, in the 18-month duration of the impact study, it was not possible to determine conclusively whether the observed changes were due to normal fluctuations in traffic rates, or were an effect of the speed limit changes themselves. 3.2.7 New Mexico Following the NMSL repeal, the New Mexico Legislature set the maximum permissible state speed limit at 75 mi/h, effective May 15, 1996. Beginning on that date, the New Mexico State Highway and Transportation Department began posting new 75 mi/h speed limits on many rural highways formerly having 65 mi/h limits. The specific highways had been identified in engineering studies conducted earlier, and included the state’s three rural interstates. Criteria considered in these studies included design speed, pavement condition, level of traffic congestion, and existing travel speeds. In conjunction with the speed limit changes, New Mexico put in place an ongoing effort to study the effects of the higher speed limits through the analysis of speed monitoring and traffic crash data.

58 Davis (1998) provides an analysis of speed monitoring data collected several months before and after the implementation of the speed limit change, and of crash data over a period from two years before to one year after the speed limit change. It was found that, on two of the state’s rural interstates, average and 85th percentile speeds increased by over 2 mi/h, and the percentage of vehicles exceeding 80 mi/h almost doubled. Speed increases on the remaining rural interstate were less than 1 mi/h, and the percentage of vehicles traveling over 80 mi/h increased very slightly. This facility carries a significant fraction of heavy vehicle traffic, which may account for some of the difference. On the two rural interstates where travel speeds increased, towaway crashes increased by 29%, injury counts by 31%, incapacitating injury counts by 44% and fatalities by 50%. On the interstate with much smaller speed increases, there was a small but statistically insignificant decrease in crash severity. The increases in crash and injury counts were much larger than the increase in traffic volumes on the same facilities during the study period, so that the crash and injury rates increased significantly. The increase in incapacitating injuries affected out-of-state drivers much more than New Mexico drivers, suggesting that increased vacation travel may be responsible for the increase in injuries. Injury occurrence tends to peak during the summer travel season, corroborating the hypothesis. Observed seat belt usage rates by New Mexico drivers are among the highest in the nation, and the difference in seat belt usage may account for some of the difference in the injury increase between New Mexico and out-of-state residents. On the two most affected rural interstates, multiple vehicle crashes accounted for 63% of the increase in incapacitating injuries and, of these, 84% occurred between vehicles traveling in the same direction (primarily through side swipes and rear-end collisions). This suggests that speed variance may have become a problem on these facilities following the speed limit increase. Many sections of New Mexico’s rural National Highway System experienced speed limit increases (typically to 65 mi/h, although some were posted at 60 mi/h and a few remained at 55 mi/h). Taken as a whole, this network is very heterogeneous, and there was no apparent overall effect of the speed limit changes on either speeds or crash performance at the level of this system. 3.2.8 New York In June 1995, legislation was approved in New York State to allow for a 65 mi/h speed limit on approximately 1,200 miles of rural interstates and other highways with similar design and usage characteristics. Since New York State had chosen to not change speed limits at the time of the 1987 NMSL relaxation, limits on these facilities had remained at 55 mi/h and the speed limit increase was uniformly 10 mi/h. At the time of the speed limit changes, the New York State Police, with authority for speed enforcement on the affected facilities, undertook a number of measures to enhance compliance with the new limits. These included dissemination of public service announcements and

59 informational literature on speeding consequences, establishment of dedicated highway patrols and speed enforcement details, and increases in the amount of speed enforcement equipment available to patrolling vehicles. The legislation also required the New York State Department of Transportation (NYSDOT), in conjunction with the New York State Thruway Authority (NYSTA) to prepare and submit a report on the impacts of the speed limit increases. Impacts on crash performance, travel speeds, traffic volumes, traffic mix (commercial vehicle percentage), speed limit compliance, as well as on the trucking industry and general public were specifically to be identified. In 1999, NYSDOT and NYSTA, together with the State Police and Department of Motor Vehicles, published their Report on the 65 MPH Speed Limit in New York State. The impact of the speed limit changes on crash counts and rates was performed using a before- after analysis covering three years before and three years after the speed limit change. It was found that total, fatal and injury crash rates decreased by 4%, 29% and 5%, respectively, on the roads with increased limits. In the period just prior to the speed limit change through December 1998, it was found that average travel speeds on affected NYSDOT facilities increased from 64 to 67 mi/h. (Systematic “before” speed data was not available for NYSTA facilities.) The 85th percentile speed increased from 69 to 74 mi/h, and the percentage of traffic exceeding the speed limit by more than 10 mi/h dropped from 40% to 11%. Between 1994 and 1998, traffic volumes on affected NYSDOT facilities increased by 13% on average. On these facilities there was an average increase of 8% in the percentage of commercial vehicles, although the specific composition varied considerably between facilities. Statewide, the VMT on rural interstates increased by approximately 11% over the same period. Total volumes on affected NYSTA facilities increased by 6%-21% over this period, with major road sections typically seeing much smaller increases in the percentage of commercial vehicles. The report notes that NYSTA volumes might have been affected by independent factors such as the inauguration of the E-Z pass system. The number of speeding tickets issued by the State Police decreased by 0.4% between the three year “before” and the three year “after “ periods. Ticketing records suggest the presence of “speed creep” (a tendency for ticketed speeds to increase over time following a speed limit change) of 0.6-0.9 mi/h/year, but the report notes that the observation period is too short to draw definitive conclusions regarding such trends. As part of the report preparation, discussions were held with the New York State Motor Truck Association (NYSMTA) regarding the specific impacts of the speed limit changes on truckers and the trucking industry generally. The NYSMTA did not have an official position regarding the 65 mi/h speed limit, and it was not aware of any studies of the speed limit change impacts on the trucking industry. However, the Association reported that the general consensus of its members was that increasing the speed limit to 65 mi/h in New York State had little impact on the trucking industry. The Association favors a single speed limit for all vehicles, rather than differential limits as practiced by some states.

60 Finally, the report estimated that roughly 4.4 million vehicle-hours per year were saved in New York State as a result of the increased limits. This figure was developed using average speed data and traffic volumes on the affected facilities. 3.2.9 Texas With the repeal of the NMSL, the Texas speed limit law of 1963 went back into effect for the first time since 1974. This law requires the Texas Department of Transportation (TxDOT) to post passenger vehicle daytime speed limits of 70 mi/h and nighttime limits of 65 mi/h on all state roadways outside urban areas, unless engineering and traffic studies demonstrate that lower speed limits are warranted. Truck speed limits on these facilities are set at 60 mi/h during the daytime and 55 mi/h at night. TxDOT conducted studies of the state highway system to identify sections that should have speed limits lower than the maximum allowed. Except where the need for such reduced limits was identified, speed limits were generally increased from 65 to 70 mi/h on rural interstates, and from 55 to 70 mi/h on other facilities affected by the NMSL repeal (urban interstates, rural multi-lane divided highways, urban multi-lane divided highways, rural multi-lane undivided highways, and rural two-lane US and state highways). Speed limit change impact studies were also conducted. An initial study by Pezoldt et al. (1997) reported on impacts of the speed limit changes after nine months of operational experience, focusing on rural and urban interstates. A later study by Griffin et al. (1998) extended the analysis to include data through early 1997, and to encompass non-interstate as well as interstate facility types. The 1998 study carried out three main types of analysis: a comparison of measured vehicle speeds in the periods before and after the speed limit change, a longitudinal analysis of injury (i.e. non-PDO) crashes, and an analysis of confounding factors that may have affected highway safety performance during the same period as the speed limit changes. Speed measurements were available from 30 permanent speed monitoring sites on highways for which the speed limits were raised to 70 mi/h in 1996, and generally covered the period from 1991 through 1997. These do not necessarily constitute a statistically representative sample of speeds on all highways with raised limits (for example, urban facilities are very under- represented), but can serve to indicate the nature of any speed changes that may have occurred. Moreover, the monitoring equipment provides 24-hour averages over all vehicles, and so does not distinguish daytime/nighttime or car/truck speeds. It was found that, following the speed limit change, average speeds on the rural interstates in the speed monitoring sample rose by roughly 3 mi/h, and by 8, 4, 7 and 5 mi/h, respectively, on the urban interstates, rural US highways, urban US highways, and rural state highways in the sample. Crash data used in the analysis were derived from the Texas Department of Public Safety Accident Files. Crashes are classified as K, A, B or C according to the most severe injury sustained: fatality, incapacitating injury, non-incapacitating injury or possible injury, respectively. Aggregate crash classes were constructed by combining the above classes: K (fatal crash); KA (crash with fatal or incapacitating injury); KAB (crash with fatal, incapacitating or

61 non-incapacitating injury); and KABC (crash with any kind of injury). PDO crashes were not considered in the analysis. Twenty-four time series equations were estimated, corresponding to the four aggregate crash categories and six highway categories. Each model was estimated based on the corresponding monthly crash counts prior to the speed limit change, and then used to predict expected post-change crash counts under the assumption that prior trends would continue without modification (i.e., that the speed limit change had no effect). Actual crash counts were then compared with the predicted counts, and a statistically significant difference between the two was taken as an indication that the speed limit change did in fact have an impact on safety. This time series analysis approach is known as Winter’s additive method. Note that, in the application here, changes in VMT are not explicitly accounted for, although the effects of a steady traffic growth trend would be captured by the approach. The analysis revealed that KABC crashes on rural interstates increased by a significant 16% in the 15 months following the speed limit change; changes in particular categories of crashes were not statistically significant, due in part to the smaller number of observations. On urban interstates, KA crashes increased by 75%, KAB crashes by 49% and KABC crashes by 28%; all these increases were statistically significant. On non-interstate rural multi-lane divided facilities, significant increases in all of the crash categories were observed following the speed limit change. Non-interstate urban multi-lane divided facilities gave mixed results by crash category. KAB and KABC crashes on non-interstate rural multi-lane undivided highways increased significantly, by 16% and 9% respectively, but no significant change was found in K or KA crashes. Finally, KA, KAB and KABC crashes on rural, two-lane US and State highways increased significantly following the speed limit change, but no significant change in fatal (K) crashes was detected. Factors other than the speed limit increases were analyzed to determine if they might have contributed directly or indirectly to the observed changes in crash counts. These included the prevalence in reported crashes of DWI, truck involvement, speeding over the limit, driving at a speed unsafe for conditions, number of vehicles involved, darkness, wet road surface conditions, and snowy/icy road surface conditions. The involvement of these factors in injury crashes in comparable months in the periods before and after the speed limit change was examined. Based on investigating police officer reports, there was no indication that crash-involved drivers were any more likely to be intoxicated after the speed limit increase than before. Similarly, there was no indication that trucks were more likely to be involved in crashes after than before. A smaller percentage of drivers were speeding over the limit following the change (as would be expected), but the number of drivers who were traveling at a speed unsafe for conditions was roughly the same before and after the change. Crashes involving multiple vehicles, during hours of darkness or on wet surfaces were all roughly comparable before and after the speed limit change. On the other hand, the prevalence of crashes in snowy/icy conditions increased following the change. It was not possible to determine the extent to which this increase resulted from unusually high exposure to snow and ice during the analysis period (and so would have happened independently of the speed limit change), and to what extent it resulted from the interactions between snowy/icy conditions and higher speed limits.

62 When the effect of snowy/icy conditions was accounted for in a re-analysis of the safety impacts, the formerly significant increase in KABC crashes on rural interstates and on rural multi-lane undivided highways became insignificant, and the estimated increases in other crash and highway categories, while remaining statistically significant, were reduced by zero to eight percentage points. 3.3 Survey of State Police Agencies The project also conducted a limited a telephone-based survey of state police highway traffic law enforcement policies and decision processes, in order to determine how these may have changed following the NMSL repeal. The following sections describe the preparation and execution of the survey, and present the results that were obtained. 3.3.1 Survey Preparation and Execution The telephone interviews were structured around a set of questions that were prepared to cover the topics of interest. However, the detailed conduct of the individual interviews was intentionally left flexible, in order to maximize the opportunities to elicit useful information and comments from the interviewees. The initial contacts in this effort were with the police representatives identified by respondents in Section C of the survey of state DOTs. These individuals sometimes directed the survey personnel to someone else in the organization better able to answer the questions. It also turned out in a number of cases that the personnel recommended in the DOT surveys were no longer available. In these cases, project survey personnel attempted to identify and contact an alternative respondent in the state police agency, but this often proved to be a very inefficient and time-consuming process. After a certain number of interviews had been conducted, the survey personnel found that each additional interview was simply repeating and confirming information that had been obtained in earlier ones. At that point, it was felt that the likelihood that an additional interview might produce new and useful information had become quite small. In view of the difficulty of identifying respondents in new agencies and of the project’s resource constraints, it was decided to stop the survey. 3.3.2 Presentation of Survey Results Surveys were successfully conducted with police officials in 18 states. The following table identifies the responding states and the main survey respondent in each. Arkansas Lt. Ray Coston California Mr. John Keller, California Highway Patrol Florida Major Ernesto Duarte Georgia Mr. Nigel Lange, Georgia State Patrol Idaho Glen Schwartz, Idaho State Police Illinois Sgt. Dianne Vanderkooy

63 Indiana Major Thomas Melville Iowa Mr. Bob Thompson, Governor’s Traffic Safety Bureau Maryland Sgt. Moore Minnesota Major Mike Asleson Missouri Captain Terry Moore Nebraska Major Anderson New York Major Jon Van Steenburg / Lt. Jon Tibbits North Dakota Major Mark Nelson Oklahoma Mr. Brandon Kopepasah, Dept. of Public Safety Texas Lt. Taylor and Major Gonzalez Virginia Mr. Cox, Virginia State Police Washington State Lt. Vasser The discussions held with these respondents revealed a number of common features: • The representatives interviewed stated that the primary goal of their agencies was the reduction of traffic fatalities and crashes. Speed limit enforcement decisions as well as allocations of patrols to other functions are made with this as the number one priority. • Most respondents were not aware of any intentional change in their states’ speed limit enforcement practices as a result of the NMSL repeal. • Several respondents noted that the most significant change in enforcement practices after the repeal of the NMSL was a decrease in the speeding enforcement tolerance (i.e., drivers were formerly allowed to go some amount over the speed limit before getting a ticket, but this threshold was reduced as speed limits and driving speeds increased following the NMSL repeal). • Where there were changes in enforcement patrol allocations after the repeal, these were generally the result of analysis of crash and fatality data, indicating the geographic areas or enforcement functions in which greater resources were needed. • In many state agencies, the individuals who were involved in speed limit enforcement decision-making at the time of the NMSL repeal have retired or left the agency, and no institutional memory remains of the decision-making process that was followed at that time. In addition to these common features, a number of individual responses provided interesting perspectives and insights. California acknowledged that its speed limit enforcement decision-making was affected by the NMSL and its repeal. During the time that the NMSL was in effect, California had argued for less Federal control over speed limits. In addition to the philosophical point that setting speed limits is a state’s right, California’s argument was based on concerns about the NMSL impact on driver behavior (since most drivers were disobeying the speed limit, it decreased the overall respect for traffic laws), as well as about how it might distort police manpower allocation. Specifically, California, like many other states, was concerned about Federal legislation authorizing sanctions (including reduced access to Federal construction funds) against states that inadequately enforced the NMSL. The threat of these sanctions could, for example, lead a governor to order the police to issue more speeding tickets in order to ensure that the state

64 received its share of construction funds. After the NMSL repeal (including the relaxation of speed limits on rural interstates in 1987), there was no longer an incentive for an allocation of resources away from other types of traffic safety enforcement or other types of facilities. It should also be recognized that enforcement policies and practices tend to change over time due to the natural development and improvement of decision-making methods, quite independently of the NMSL enactment or its repeal. For example, Missouri noted that its speed limit enforcement decision-making process has changed since 1995, but not because of the NMSL repeal. Rather, this process has evolved to take advantage of the availability of better traffic safety statistics. Moreover, a Police Allocation Model developed at Northwestern University is now used to help in this process. Similarly, in Washington State there was a change in enforcement-related decision-making in the late 1990s, but this change also does not appear to have been a direct result of the NMSL repeal. The new process in Washington makes systematic use of the traffic safety statistics maintained by the state DOT, together with time and activity reports completed by patrolling officers, to focus resources towards reducing crashes and fatalities caused by speed, DUI, seatbelt non-use and aggressive driving. Interestingly, Minnesota has an anti-quota law that prevents law enforcement officials from giving police officers instructions on how many tickets to issue, and so limits the extent to which police officials at the state level can establish broad speed limit enforcement policies The survey personnel also attempted to identify the extent to which political factors may influence speed limit enforcement allocations. In some states, this influence is quite open and transparent. For example, in Florida it was reported that police sometimes receive specific speed limit enforcement requests from local lawmakers. The police generally attempt to be responsive to such requests. However, if a major enforcement effort is not warranted, they tend to periodically revisit the locations in question, rather than mount a sustained enforcement program. In New York, the governor’s Traffic Safety Committee can become involved in the process of determining speed limit enforcement allocations. 3.3.3 Conclusions from the Survey of State Police Agencies More generally, to the extent that police agency budgets influence the feasibility of different enforcement efforts, and legislatures determine agency budgets, the agencies cannot be said to be totally isolated from political influences. Comments from Texas indicated this to be the case there. Washington State indicated that because the State Police Chief is appointed by the governor, and thus holds a somewhat political position, the decisions made in monthly executive staff meetings on enforcement might incorporate political input. Oklahoma indicated that some political influence results from the fact that the Legislature places a statewide cap on speed limits (of 75 mi/h), and then allows the municipalities or counties to set their individual speed limits. Overall, however, political influences on enforcement decisions seemed to be slight enough not to warrant significant comment by the representatives contacted for the survey.

65 3.4 Summary This chapter has presented the design and results of surveys of State DOTs and State Police agencies conducted by the study. The survey of State DOTs also identified a number of documents that were prepared around the time of the NMSL repeal, to plan for and analyze the impacts of changed speed limits, and these reports were also reviewed in this chapter. The results obtained and presented here are, for the most part, of a qualitative rather than quantitative nature. They do not contribute directly to the development of statistical models of the impacts of speed limits on speed choices, crash occurrence and/or crash severity. Nonetheless, the information gathered and insights gained from the surveys contributed to and guided the development of a number of the project’s conclusions and recommendations.