SPECIAL ISSUE IN HONOR OF CLAIRE LABERGE-NADEAU

ACADEMIC ARTICLES

Lateral damage and point of impact in intersection crashes: implications for injury
par Mary L. Chipman, Gerald Lebovic, Ediriweera Desapriya and John Gane

Les collisions en 1987 et en 1999 : comparaisons entre les personnes utilisatrices du téléphone mobile en 1999 et les toujours non-utilisatrices
par Urs Maag, Claire Laberge-Nadeau, Jean-François Angers, François Bellavance, Louis François Poirier, Denise Desjardins et Stéphane Messier

Modélisation et estimation des effets individuels et d’entreprise avec des données de panel : une application aux parcs de véhicules
par Jean-François Angers, Denise Desjardins, Georges Dionne et François Guertin

A Profile of Adolescents who Attend Driver Education for the Insurance Discount: Are Insurers Rewarding Bad Risks?
by Pierro Hirsch and Urs Maag

Le contact visuel à la traversée d’une intersection par les piétons à Montréal et à Toronto
par Jean-Pierre Thouez, Jacques Bergeron, André Rannou, Robert Bourbeau et Yves Bussière

Étude par simulation mathématique de l’efficacité d’un dispositif de retenue pour enfant selon la configuration d’installation
par Marius-Dorin Surcel et Michel Gou

Current Events
under the responsibility of Rémi Moreau

1.  Séisme majeur au Pakistan
2.  Le bilan des pertes dues à l’ouragan Katrina est révisé
3.  Un troisième trimestre désastreux pour les assureurs américains sur le plan des 
     catastrophes naturelles
4.  Les investisseurs américains s’inquiètent des coûts liés au risque climatique
5.  Le pouvoir d’emprunt du Fonds national contre l’inondation aux États-Unis est revu à la
     hausse
6.  Les réassureurs, réunis à Baden-Baden, ont réclamé un durcissement du marché,
     mais non nécessairement des hausses tarifaires drastiques
7.  Munich Re est protégé par un marché dit Catastrophe bonds en cas de tempêtes
     européennes à venir 
8.  Swiss Re deviendrait le plus gros réassureur de la planète
9.  Le dédommagement des clients américains de Marsh
10. Les assureurs chinois peuvent investir dans des opérations étrangères
11. Le marché de Londres est en régression
12. Les collectivités territoriales françaises durement touchées par les violences urbaines
     qui ont enflammé l’Hexagone au début de novembre
13. La tempête du verglas de janvier 1998 sévit encore
14. Nouvelle directive européenne sur la réassurance
15. Le Sénat américain proposerait une loi en 2006 créant un fonds lié à l’asbestose
16. Un rapport sénatorial sur le coût de l’amiante en France
17. Une étude du Lloyd’s révèle que les entreprises internationales prennent de plus en
     plus conscience de la gestion des risques
18. Standard Life, le numéro un européen des mutuelles, serait démutualisée en 2006
19. Les méga-catastrophes du vingt et unième siècle (pertes de vie et/ou pertes
     assurables et économiques)
20. Le sida dans le monde en 2005 : le cap du 40 millions est atteint
21. Le groupe Shell a confié à CMGL la gestion de son portefeuille de sinistres de
     responsabilité civile et recours
22. La garantie de l’État sur les risques terroristes
23. Le naufrage du Prestige, trois ans après
24. Le RIMS s’intéresse au Enterprise Risk Management (ERM)
25. La première victime humaine de la grippe aviaire en Chine et un deuxième foyer 
     d’infection dans l’Ouest canadien
26. La Table de la sécurité routière

The Internet Surfer Page
The Société de l’assurance automobile du Québec

Lateral damage and point of impact in intersection crashes: implications for injury
by Mary L. Chipman, Gerald Lebovic, Ediriweera Desapriya and John Gane

Crashes in intersections may result in damage to vehicles and injury to occupants in many different ways. One vehicle hitting the side of another (T-type) is the classic side impact; however, we have argued in the past that other crash configurations (e.g., L-type) may subject occupants to similar risks because both vehicles may sustain lateral damage. To test this assumption, we examined crash data from police reports of 4032 intersection right-angle crashes (IRC), collected by the Insurance Corporation of British Columbia for 2002. We compared the risk and types of injury in target and bullet vehicles for T-type crashes, L-type crashes by front and rear fender involvement and for all other IRC crashes. There were 787 T-type crashes (impact into either side of target vehicle), compared to 798 L-type crashes (impact into front fender) and 350 L-type (impact into rear fender). Overall, injury risk was 23.5%. Proportions injured were very similar for occupants of target and bullet vehicles in T-type crashes (OR = 0.996; 95% ci 0.80 to 1.24.); for L-type crashes, the proportions were 23.2% for front and/or front fender involvement and 15.0% for crashes involving the rear fender of one vehicle and the front or front fender of the other (OR = 1.71; 95% ci 1.40 - 2.10). Apart from rear fender crashes, proportions injured were very similar (P > 0.05). Other factors, notably weather, lighting, land use and vehicle damage differed significantly by crash type, and were strongly associated with injury risk. Since rear-fender crashes are a small proportion of IRC crashes, this suggests that it is not necessary to subdivide crashes by configuration in IRC crashes.

Keywords: Lateral impact, vehicle damage, injury.

Les collisions en 1987 et en 1999 : comparaisons entre les personnes utilisatrices du téléphone mobile en 1999 et les toujours non-utilisatrices
par Urs Maag, Claire Laberge-Nadeau, Jean-François Angers, François Bellavance, Louis François Poirier, Denise Desjardins et Stéphane Messier

The cell phone when used at the wheel may be a distraction that could be harmful to road safety. A research project on 12 691 drivers users and 23 387 drivers non-users has shown that the users, men as well as women, have relative risks for collisions and for injury collisions that are 38 % higher that for non-users (Laberge-Nadeau et al., 2003).

Two questions arise : (i) Are the two cohorts comparable in terms of collision rates before using the cell phone? (ii) How to respond to the argument that cell phones pose no risks since their use is increasing but collision rates diminish?

Among the 36 078 respondents of 1999 there were 18 707 with a driver’s permit in 1987 : 5 107 men became users, 7 779 remained non-users, 1 365 and 4 438 respectively for the women. Collision rates in 1987 for the two groups are not significantly different, neither for all collisions, nor for injury collisions. Between 1987 and 1999 the collision rates for non users declined significantly more than for the ones who became users; this result holds for men and for women.

In conclusion : (i) On the basis of the collisions in 1987, one can assert that the cohort of male users and the one of female users are comparable, although slightly less safe for the younger users (16-34 years old), to the cohort of male non-users and of female non-users respectively before the use of a cell phone. (ii) The collision rates diminish differently for the two cohorts; they decrease more for the non-users than for the users, both for men and women.

Keywords : Cell phone, distracted driving, collision, relative risk.

Modélisation et estimation des effets individuels et d’entreprise avec des données de panel : une application aux parcs de véhicules
par Jean-François Angers, Denise Desjardins, Georges Dionne et François Guertin

In this article, we propose a detailed analysis of the modeling and estimation of accident distributions of vehicles belonging to a fleet. This analysis uses panel data to account simultaneously for vehicle as well as fleet effects. The distribution of accidents can be affected by both observable and non-observable factors. Non-observable factors are modeled as random effects.

Keywords: Insurance for fleet of vehicles, vehicle effect, fleet effect, panel data, econometric modeling, estimation.

A Profile of Adolescents who Attend Driver Education for the Insurance Discount: Are Insurers Rewarding Bad Risks?
by Pierro Hirsch and Urs Maag

Problem: The overrepresentation of adolescent drivers in crashes is a robust phenomenon. Driver education (DE) is a popular countermeasure, and in most North American jurisdictions, insurers grant automobile insurance premium discounts to DE graduates. However, over the past 20 years, evaluations have consistently demonstrated that DE does not reduce, and may even increase, crash risk among adolescent novice drivers. Providing premium reduction incentives to DE graduates may possibly increase crash risk in two ways. One, by reducing the overall cost of licensing and car ownership, insurance may increase driving exposure. Two, insurance may increase morale hazard, a careless attitude toward prevention. The human and financial losses resulting from adolescent crashes are a serious problem for public health and for insurers. Insurance is also known to increase moral hazard, a tendency to make dishonest claims – losses due to fraud are a significant problem for insurers. Therefore, the DE insurance discount may not be optimally efficient for reducing insurers’ losses or for improving the public health. One approach to investigating the effects of the DE insurance discount is to study the characteristics and the driving records of adolescents who are insurance-motivated, i.e. those who attend DE partly or entirely for the insurance discount.

Method: A cohort of 1,804 novice drivers 16- to 19-years of age of both sexes completed an extensive questionnaire on learning methods, including motivation to attend or not to attend DE, risk taking, and lifestyles. Questionnaire data were linked on an individual basis with government records on exam performance, violations, and crashes. Among the participants who attended DE (N = 1,536), the importance of the insurance discount in their motivation to attend DE was studied in relation to violation and crash records during the first 450 days of unsupervised driving and explanatory variables from the questionnaire.

Results: Insurance-motivated participants, compared to those who were not motivated by the insurance discount, were more likely to have: greater violation risk, more tolerant attitudes towards speeding and risk taking, and less financial support from family for all licensing and driving related expenses. Insurance motivation was also associated with the likelihood of presenting fraudulent DE certificates and expressing a willingness to defraud insurance companies.

Discussion: Increased violation and crash risk associated with insurance-motivation may possibly be due to greater morale hazard. The data also indicate that insurance-motivation may be associated with greater moral hazard and potential future losses for insurers. Alternative methods for insuring adolescent drivers are suggested with the aim of decreasing insurance losses and injury risk by attempting to decrease both morale and moral hazard.

Keywords: Adolescent drivers, driver education, crashes, insurance, morale hazard, moral hazard, risk taking.

Le contact visuel à la traversée d’une intersection par les piétons à Montréal et à Toronto
par Jean-Pierre Thouez, Jacques Bergeron, André Rannou, Robert Bourbeau et Yves Bussière

Visual contact is one of the indications of interaction between a pedestrian and a motorist. It is also an important function with respect to cognitive information when a visual attention pedestrian prepares to enter a public crosswalk at an urban intersection.

This research presents the results of field data collected in Montreal and Toronto in the spring of 2001, which analyzed the behaviour of pedestrians at signalized crosswalks. The results showed that visual contact is more frequent in Montreal than in Toronto having regard to gender and age groups. There were also significant differences in behaviour of pedestrians as they began to cross and during the crossing of intersection. Logistic regression analysis was used with visual contact/no contact as the dependent variable. With regard to the behavioural model there was 1,63 times greater chance in Montreal and 2,12 in Toronto of observing visual contact when the pedestrian did not respect the rules of crossing an intersection. It is significant, however, that in Montreal there is lower probability of observing visual contact when the pedestrian do not cross at a pedestrian crossing. In Toronto there is a higher probability of visual contact if the pedestrian is not strictly observing the rules for crossing the intersection.

With regarded to the environmental model a pedestrian in Montreal has more chances of making visual contact when he/she crosses intersections in residential areas, streets with three or more lanes, in the presence of a standard stoplight, and at intersection located on the periphery of the city center. In Toronto, we found the last two environmental variables included in the Montreal model. In addition, the automobile flow appeared in the Toronto model. The discussion underlines the importance of the notion of attention with regard to visual contact between pedestrian and driver in an environmental context.

Keywords: Pedestrian behaviour, field data, crosswalk, urban milieu.

Étude par simulation mathématique de l’efficacité d’un dispositif de retenue pour enfant selon la configuration d’installation
par Marius-Dorin Surcel et Michel Gou

This project uses dynamic simulation to assess the effectiveness of child restraint systems for various installation configurations in the cases of both frontal and side impact. For this purpose a model was developed using the madymo software (essentially the part that uses the multibody method) while finite element modeling was used to model the wings of the child restraint system and the vehicle body to provide a better representation of the contact between the child dummy, the restraining device and the structure of the vehicle. The model was evaluated for a side impact but, in order to perform more thorough verification, frontal collisions were also considered. The study shows that isofix lower anchorages provide a better protection for the head while the results for lower flexible anchorages and vehicle belt installations although generally similar are not as good. The top tether reduces head injury criteria, especially when vehicle safety belts or lower flexible anchorages are used, but has practically no influence when the child restraint system is installed using isofix anchorages. The model will also allow to measure the effect of inadequate installation, such as caused by the play introduced in the child restraint system due to the use of winter clothing.

Keywords : Side impact, child restraint system, ISOFIX, MADYMO, modelling.