Risk Homeostasis

A Theory about 

Risk Taking Behaviour

Common Misunderstandings

2. Homeostasis is a process, not an outcome

In 1929, Walter Cannon,[1] an American physiologist, proposed the term “homeostasis” as a label for the dynamic process that had been discovered some 70 years earlier by the French physician Claude Bernard.[2] Cannon showed considerable wisdom in calling it homeo-stasis, not iso-stasis. “Homeo” means “like, matching, agreeing” while “iso” means “same, equal, identical,” as in isobar, isotope and isotherm. An isotherm is not a homeotherm, and isosceles is not homeosceles. Encyclopedias explain the difference between isostatic and homeostatic. Isostatic has to do with a state of sameness, homeostatic with a mechanism that keeps the output at a desired level. Homeostasis, therefore, should not be viewed as a process that keeps the output the same, at an invariably fixed level. As was stressed by researchers at Harvard University another 70 years later:


Bernard‘s and Cannon’s teachings that the integrity of higher forms of life relies on maintenance of a constant internal milieu have been central to modern physiological theory. Unfortunately, the teaching of [the concept of homeostasis] to successive generations of medical students has led to an overly simple perception being embedded in the collective medical consciousness. Cannon never suggested that every physiological variable is tightly regulated within limits, nor did he indicate that even the most well-regulated variables were maintained at an absolute constant level.

It is interesting to note that Cannon (1929), in the article in which he first outlined his concept of homeostasis, specifically pointed out that even the most tightly regulated variables may oscillate. He accordingly defined homeostasis as the process which regulates a physiological variable within certain limits, but that the variable may oscillate between those limits, and the limits themselves may change in response to some special demand.[3]


The “overly simple perception,” alas, is not limited to people in the medical profession. Some researchers and practitioners in the field of safety and public health seem to suffer from the same affliction.

You may have noticed that I often belabour the point that the target level of risk is not fixed once and forever—nor is the accident rate per capita. But I think there are good grounds for reiterating it. The fact is that some published critiques of risk homeostasis theory have mistakenly interpreted it as stating that the target level of risk, and thus the accident loss, is immutably fixed.[4],[5] This misinterpretation explains why one critic made the perhaps amusing, but no less erroneous, quip by referring to the theory as “the law of the conservation of misery.” Risk homeostasis does not imply a law of “the conservation of accidents,”[6] just as homeostasis of body temperature or blood pressure does not imply invariant body temperature (think of fever) or invariant blood pressure (think of rest versus exertion).

Homeostasis is a process, not an outcome, let alone an invariant outcome. Expressions such as “partial homeostasis,” “exact homeostasis,”[7] “incomplete homeostasis” and similar ones that have cropped up in well over 25 years of “the great risk homeostasis debate” make very little sense. The use of such expressions betrays two basic misunderstandings of the nature of homeostasis. For one thing, these expressions are mistaken because they refer to outcome, not to process. For another, they misinterpret the outcome as something that should be fixed and invariant. Some people have referred to risk homeostasis theory saying that it implies constancy of risk.[8],[9],[10] They’ve even mislabeled it “the constant risk hypothesis.”[11] We stress that the main hope for developing interventions that are capable of reducing the accident loss per person is located in the very pliability of the target level of risk, as has been demonstrated in a great number of empirical studies under real-life conditions.[12]

[1]Cannon, W.B. (1929). Organization for physiological homeostasis. Psychological Review, 9, 399-431.

[2]Bernard, C. (1859). Leçons sur les propriétés physiologiques et les altérations pathologiques des liquides de l’organisme. Paris: J.B. Ballière.

[3]Moore-Ede, M.C., Sulzman, F.M. and Fuller, C.A. (1982). The clocks that time us: physiology of the circadian timing system. Cambridge, Mass.: Harvard University Press.

[4]Evans, L. (1986). Risk homeostasis theory and traffic accident data. Human Factors, 27, 555-576.

[5]McKenna, F.P. (1987). Behavioural compensation and safety. Safety Science, 9, 107-121.

[6]Michon, J.A. (1979). Personal communication. Department of Psychology, University of Groningen NL.

[7]Janssen, W.H. and Tenkink, R. (1988). Risk homeostasis and its critics: time for an agreement. Ergonomics, 31, 429-433.

[8]Oppe, S. (1988). The concept of risk: a decision theoretic approach. Ergonomics, 31, 435-440.

[9]Rumar, K. (1988). Collective risk but individual safety. Ergonomics, 31, 507-518.

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

[11]Veling, I.H. (1984). A laboratory test of the constant risk hypothesis. Acta Psychologica, 55, 281-294.

[12] Wilde, G.J.S. (2001) Target Risk 2. Toronto: PDE Publications, Chapter 11.