Risk Homeostasis

A Theory about 

Risk Taking Behaviour

Common Misunderstandings

The choice of denominator for the calculation of accident rates

The finding that drivers tend to move more slowly in road stretches with a history of an elevated accident risk per km driven[1] brings up the issue of how to calculate the rate of accidents: per km of driver exposure, per time unit of driver exposure, or by others still, such as divided by the total length of roads in a given jurisdiction, or per the number of registered vehicles. Here the two most common of these are considered: per unit distance travelled and per time unit of driving (per hour or per capita per year). Making a clear distinction between these two is crucial for an appreciation of what has happened in the past, as well as for an understanding of the tenets of risk homeostasis theory.

What has been said here does not mean that efforts at reducing the accident rate per km driven serve no meaningful purpose. When successful, they allow drivers to drive a greater distance per death or injury or damage, thereby reaping all the attending advantages. The reduction of road accidents per km driven is generally welcomed by ministries of transport, but agencies involved in public health still face the same number of people in hospital beds or morgues. The implication is that road safety researchers and practitioners much face the question: “Do we want to achieve more mobility per death or fewer deaths?”

Statistics show that over periods extending the best part of the 20th century, the accident rate per km driven has dropped significantly. In the US, for example, the accident rate per km driven dropped by a factor 11 or so, while the annual traffic death rate per citizen, while showing major annual variations, was about the same in 1996 as it had been in 1923 (and the numbers of km driven per citizen rose by about the same factor, i.e., eleven; The National Safety Council, various years).[2] Other data, pertaining to the province of Ontario in Canada, show that in a period of sustained economic growth between 1955 and 1972, the accident rate per km driven dropped, while it rose per capita. Yet, one critical comment on risk homeostasis argued that it says that risk per km driven cannot be reduced,[3] while another confused the two denominators altogether, and treated them as if they were of the same nature.[4]

Instead, what the notion of risk homeostasis in traffic does say is this. When roads are made more forgiving, (e.g., through more and wider lanes, medians and shoulders and guardrails) and cars more crashworthy (e.g., through padded dashboards, collapsible steering columns and shatterproof glass), or drivers’ vehicle handling skills are improved, drivers will respond by driving more and faster or choose some other adjustment action, and thereby attempt to maintain the equilibrium between the degree of risk experienced and the degree of risk accepted.

Some 20 years ago, airbags were advertised by car manufacturers as providing drivers with near-immortality in a crash, and governments in some parts of the world have included them in the required vehicle manufacturing standards.

What has been the effect on safety per head of population? There are occasional reports showing that a car occupant survived a crash thanks to the airbag. There are also reports that show that show that car occupants died due to an inflating airbag. So, what is the over-all effect?  Research by economists in the State of Virginia in the US compared the accident experience of passenger cars with and without airbags and came to the conclusion “that insurance industry-generated data reveal that injury claims increase following adoption of an air bag system” and that “Virginia State Police accident reports indicate that air-bag-equipped cars tend to be driven more aggressively and that aggressiveness appears to offset the effect of the air bag for the driver and increase the risk of death to others."[5] This statement is similar to the one made by another economist whose California data “lead to the conclusion that, controlling for other influences, as driving conditions become safer, accidents become relatively more injurious for not-at-fault road users than for at-fault drivers, the former  including pedestrians and bicyclists.[6]

[1] See Wilde, G.J.S. (2014).Target Risk 3, Sections 3.1, 3.2 and 3.3.

[2] See Wilde, G.J.S. (2014). Target Risk 3, Sections 5.1 and 5.2.

[3] Huguenin, R.D. (1982). Zur Problematik von Risikohomöostasetheorien in der Verkehrspsychologie. Zeitschrift für Verkehrssicherheit, 28, 180-187.

[4] Evans, L. (1985). Human behavior feedback and traffic safety. Human Factors, 27, 555-576.


[5] Peterson, S., Hoffer, G. and Millner, E. (1995). Are drivers of air-bag-equipped cars more aggressive? A test of the offsetting behavior hypothesis. Journal of Law and Economics, 38, 251-264.

[6] Traynor, T.L. (1994). The effects of varying safety conditions on the external costs of driving. Eastern Economic Journal, 20, 45-60, p. 59.