Time for Answers

by Fred Nix

Part III of a series

Regulators put great store in improving on-road safety by developing elaborate truck safety regulations. Canadian jurisdictions, for instance, have spent more than a decade implementing the National Safety Code. And regulators and road agencies on both sides of the border have ambitious goals to improve road safety. The new US Federal Motor Carrier Safety Administration has announced a goal of reducing truck-related fatalities by 50 percent by 2010 and the number of persons injured by 20 percent by 2008. In Canada, regulators have set a target of reducing by 20 percent the number of persons killed or seriously injured in collisions related to commercial vehicles by 2010.

Collision rates provide an objective measure of highway safety indicating how one jurisdiction is performing compared to another or how one jurisdiction is performing over time. What they do not reveal, however, is why one jurisdiction has more or fewer collisions than another. One factor that presumably explains the variances in highway safety among the different jurisdictions is the large number of truck safety regulations. Yet, in similar fashion to the situations reported with roadside inspections and truck crashes in the first two installments of this series, little effort, particularly in Canada, is being put into actually trying to understand the effectiveness of these regulations. Questions that need to be asked about the puzzling data coming back from the field are not being asked.

Consider, for example, earlier work which compared collision rates between Canada and the US. Four rates — (1) trucks involved in collisions per capita, (2) trucks involved in collisions as percent of all vehicles involved in collisions, (3) trucks involved in collisions per registered truck, and (4) trucks involved in collisions per kilometre of travel – show little difference between the two countries. For example, the total collision involvement rate per 100,000 people in 1996 was 150.7 in Canada and 148.9 in the US.

This is puzzling for several reasons:

1. Canada has heavier trucks than the US and some allege that safety is compromised as truck weight increases. Canada’s national 62.5 tonnes is considerably heavier than the US 80,000 lb;

2. Canadian roads are probably not as good as those in the US. There is more travel on two-lane roads and collision rates are much lower on high-class, divided freeways than they are on two-lane roads;

3. Canada presumably has worse weather (snow and ice);

4. Canada does not have mandatory drug and alcohol testing while the US has had such regulations in place for over a decade;

5. Canada is behind the US in some areas of safety regulations (eg, assigning safety-ratings or measuring relative levels of safety performance);

6. Canada allows longer hours of work for truck drivers than does the US;

7. Canada probably has a worse record on safety audits as not all provinces are doing such audits vigorously.

Despite these differences, however, the apparent truck collision rates between the two countries are not much different. It seems odd that all these regulatory, road design and climatic differences seem to have so little impact. All of this suggests that the issue of how truck safety regulations affect on-highway safety needs revisiting.

Let’s first examine the makeup of currently available data. Transport Canada collects truck collision information from the provinces and territories and there now are consistent statistics from all jurisdictions for the period 1991 to 1998. (see Table 1 on page 20). The quality of the information has improved. For the first time there are estimates of kilometres of travel, important in providing a true picture of truck activity in Canada. There is also a more reasonable estimate of the size of the Canadian fleet.

However, even with these new estimates on the number of trucks and vehicle kilometres of travel (vkt), there are still problems. The first is that at the time of writing only two quarterly surveys on kilometres of travel – fourth quarter of 1999 and first quarter of 2000 – were available. To provide an initial look at what collision rates in Canada may be, these data are used as follows: (1) the average number of “in-scope” trucks for the two quarters are used to estimate vehicle populations, and (2) total vkt for the two quarters is doubled to give an estimate of annual travel. (Results are shown in Table 2.) But if travel peaks in the summer, the estimate of annual vkt may be lower than the true number. There is indication that this is, in fact, the case. Statistics Canada used to produce a quarterly survey of for-hire carriers that included estimates of total travel. In 1993, of the 6.1 billion kilometres of travel, 48.7 percent occurred in the first and fourth quarters. Presumably, then, collision rates developed here are overstated as – assuming for-hire truck activity is indicative of all truck activity – the denominator is two to three percent smaller than it should be. However, the use of vkt information for the period October 1999 to March 2000 along with collision information for the year 1998 presumably has the opposite effect. That is, with a growing economy, it is likely that truck vkt is increasing by two to three percent per year.

The second problem is that both the fleet numbers and the vkt are estimates and some numbers shown for vkt on Table 2 are not reliable. There is either missing data (for one of the two categories of trucks used in the survey or for one of the two quarters) or the estimates have coefficients of variation above 25 percent. The only estimates of vkt that may be considered reliable (if doubling six months of information to estimate annual activity is acceptable) are those for Qubec, Ontario, Alberta, British Columbia and Canada.

Third, there are still questions about the reliability of the numbers on truck populations. The jurisdiction where the problem appears most severe is Qubec and this may be because it is the only province that does not register trucks on a gross weight basis. (The new survey classifies trucks on the basis of registered weight.) Consider the following three numbers: (1) the registration agency (SAAQ) itself showed 100,718 registered trucks in the province in December 1996, (2) a 1995 estimate, based on R.L. Polk data obtained by Transport Canada, showed 125,408 trucks (at the time, it was thought that the difference between 100,718 and 125,408 was accounted for by farm trucks), and (3) the “in scope” trucks — registered trucks that, on the basis of Statistics Canada’s new survey, exist during the survey period — for the first quarter of 2000 numbered 66,201. There is quite a difference in these three numbers and there has to be some suspicion about the accuracy of Statistics Canada’s latest ones.

Comparing truck collisions (see Table 3) among the Canadian jurisdictions raises questions. Admittedly it is suspected there is a problem with the numbers for Qubec in columns one and two (collision rates well over double the national average?). There are no problems, however, with the numbers in columns three and four (comparing collisions per capita between 1991 and 1998) and, in both cases, the rate for Qubec is significantly higher than it is in other jurisdictions. In the case of Alberta, it appears odd that the performance indicated in the first two columns is relatively good (better than the national average) but that the performance in the last three columns appears so poor. Part of the explanation is the relative level of truck activity. There are 4.7 trucks per 100 people in Alberta compared to a national average of only 1.9 and annual truck travel per capita amounts to 2,559 kilometres compared to a national average of only 897 kilometres.

Table 4 contains new data comparing Canada and the US. In considering road safety, it is the total occurrences of collisions that are important. However, because there may be different definitions used, comparisons of total trucks involved in collisions on both sides of the border may be slightly ina
ccurate. Therefore, fatal and injury collision rates are also shown.

There are qualifications about the accuracy of the information shown. Important ones about Canadian data have already been mentioned and may make the rates shown in the last two lines (per registered truck and per kilometre) slightly higher than they should be (the denominators may be too small). In the US, information on trucks involved in fatal collisions is considered accurate. But the numbers for total collisions are an estimate and information on total trucks and total vkt may not be entirely accurate. What is not known, and this may be serious, is whether the definition of a collision in both countries is similar. Presumably, definitions of “fatal” and “injury” are the same but the precise definition of “property only” is not known in either country.

As in the previous table, then, only cautious observations are possible. The most obvious is that there is more truck activity in the US than in Canada. This higher relative level of truck activity probably accounts for part of the higher collision rates in the US when population is used as the denominator. The second observation is that, while the absolute number of collisions has been falling in Canada, the opposite is true in the US. Presumably, this has something to do with the faster pace of economic activity in the US during the 1990s. The third and potentially most significant observation is that this new data in the last two lines (collision rates per truck and per kilometre) seem to indicate a somewhat higher rate of collisions in Canada than in the US. Admittedly, given the possible weaknesses in the data, this has to be labelled a tentative observation.

Many factors contribute to the cause or severity of a collision — the road design; the design of the vehicle; the load-securing devices and the characteristics of the freight; the weather; the skill of the driver, the size, weight or configuration of the truck; traffic levels; compliance with safety regulations; etc. The phenomenon is so complex that it is difficult to sort out cause and effect. This gives regulators a problem when it comes to designing or implementing new or revised regulations, or enforcing existing ones. They are put in the position of having to say how many collisions will be avoided, or how many lives will be saved, if, for example, the specifications for rear bumper guards on trailers are changed or if allowable hours of work for truck drivers change.

It is difficult for regulators to set such goals when the understanding of how regulations affect road safety is so weak. The available research in this area is notably limited. In Canada there has been no recent research on this subject at all. And while much of the existing research confirms that the safety regulations evaluated are effective, there is other research suggesting the opposite.

For example, a study in Illinois looked at all trucks stopped for an inspection in the state over a number of years (almost a million trucks), and then the crashes those vehicles were involved in during a 12-month period after the inspection. There was no significant statistical relationship between those trucks found with an out-of-service defect and crash rates. One would suspect that carriers that allow vehicles or drivers out on the highway with defects serious enough to be classified as out-of-service would tend to have more crashes than other carriers. Apparently not. And this possibly raises a question about the point of roadside inspections.

The most important truck safety regulations in Canada are those of the National Safety Code. And the most important standard in the code is NSC#14 (safety ratings) as this is where, in effect, a carrier’s compliance with other safety regulations is judged. Provinces and territories are at various stages of implementing NSC #14 and the federal government is planning to amend current legislation such that any jurisdiction that has not implemented this safety-rating regime will have its power to issue operating rights to extraprovincial carriers removed. The problem is that provinces and territories are all implementing NSC #14 in somewhat different fashions with the result that there is no nationally consistent safety-rating regime emerging. And this is where the lack of research into safety regulations becomes critical. In assessing the different procedures now being developed, it is almost impossible to determine which safety-rating regime is more or less effective than any other. It may be that it will take some years to make this determination. But, such a determination can only be made in those jurisdictions where procedures are established to assess the effectiveness of the regulations.

Even if all jurisdictions were implementing a completely uniform safety-rating regime, there would still be a need for research into the effectiveness of these regulations. The tax-paying public and the industry, which bears much of the cost of this regulatory burden, have the right to know if the money is being spent wisely.


Have your say


This is a moderated forum. Comments will no longer be published unless they are accompanied by a first and last name and a verifiable email address. (Today's Trucking will not publish or share the email address.) Profane language and content deemed to be libelous, racist, or threatening in nature will not be published under any circumstances.

*