TRAFFIC VIOLATIONS

Traffic Violations

Abstract

This paper separates from the empirical accident literature by using register data of approximately 9 million insurance contracts, the aim being to establish the role of age and gender on traffic violations. Data contains input variables (characteristics of the vehicle owner, the vehicle and annual mileage) and output variables (claims). In addition, information about traffic violations of individuals is added. The main advantage is that data contains information of both accident involved and accident free vehicle owners. We estimate three probit models of different traffic violations; speeding, other traffic offences and convictions. Individuals are more likely to have speeding tickets up to their 30's and thereafter the probability decreases with age. The probability of having on-the-spot-fines for traffic offences decreases with age. The probability of having convictions increase until the middle-age, and decrease for older vehicle owners. The analysis confirms previous findings in that males and young males in particular, tend to be riskier compared to women. In the second part of the analysis we include more explanatory variables (characteristics of the vehicle, annual mileage, at-fault claims and insurance coverage). Our findings suggest that individuals with traffic violations are more likely to report at-fault claims (accidents).

Table of Contents

Introduction4

Literature Review5

Statistical approach6

Part I: establishing the effect of age and gender6

Part II: Including endogenous variables6

Data7

Descriptive Statistics8

Results10

The relation between traffic violations, age and gender10

Age and gender subgroups12

Conclusions15

References17

Traffic Violations of Men and Women

Introduction

Traffic accidents lead to death and disability, which result in significant costs to society and the individual. A number of factors contribute to an accident such as vehicle design, speed, and road design and driver impairment. The accident literature has shown that traffic violations are one of the best predictors of crashes (Parker et al.; 1995, Forward; 2006, 2008). There are furthermore demographic differences in both violations and crash rates.

Males and young males in particular, tend to have more aggressive driving behavior than women. This manifests in a greater frequency of violations of traffic regulations. On the other hand females, and older drivers, seem to commit more errors than male and young drivers (Åberg and Rimmö; 1998). Driven distance is emphasized as another accident predictor since exposure increases with number of kilometers. Identifying accident predictors and reducing risk is highly relevant for insurers and coincide with the public interest of reducing the number of accidents. The economic relevancy is reflected not at least in the pricing of risk by actuaries, since insurance rates are based on different groups, partly based on driver and vehicle characteristics, correlated with ex post risk (claims).

This paper extends previous literature that has sought to identify accident predictors by using a rich dataset of insurance contracts. There are at least two advantages with the present data. First, it consists of approximately 9 million contracts including characteristics of the vehicle owner and vehicle, annual mileage, and reported accidents (claims). Second, our data contains register data on traffic violations of individuals that both reported an accident and individuals that did ...