Is risk homeostasis real?

Is risk homeostasis real? There are studies that both confirm and deny its existence. Many of these studies are biased, but research that tries to eliminate the bias seems to indicate that risk homeostasis is real.

Risk homeostasis is the theory that people like a certain level of risk in their lives, and that if you eliminate one form of risk, people will tend to compensate by taking additional risks. If this really happens, it has significant public policy implications. Wearing seatbelts, for example, might make people feel safer so that they compensate by taking other risks. If the dangers from these additional risks exceeds the danger from not wearing a seatbelt, then we may actually be better off by not requiring people to wear seatbelts than by requiring people to wear them.

In the case of seatbelts, it can be difficult to get an unbiased answer as to what’s best. Some researchers, particularly those with Libertarian leanings, seem to want to prove that making people wear seatbelts doesn’t make sense. You can see an example of this here. Other researchers, like those working for insurance companies or government health and safety organizations, seem to want to prove that making people wear seatbelts is a good idea. You can see an example of this here.

Supporters of either side in this debate may selectively cite statistics that support their position while ignoring those that don’t. Because of this, there are studies that have shown that requiring drivers to wear seatbelts is both a good idea and not a good idea, which makes it difficult for policy-makers to make an intelligent and informed decision about what’s best to do.

There are also areas of information security in which we might expect to see the effects of risk homeostasis if the effect is actually real. We might expect people running anti-virus software, or example, to be less careful when opening emails or attachments to emails if they believe that their anti-virus software is protecting them.

Jeremy Jackson, while a student at Simon Frasier University, decided to test whether or not risk homeostasis is real. To do this he carefully created an experiment that tried to eliminate all of the sources of potential bias that earlier experiments might have had. His experiment tried to test the accuracy of the original study that claimed that drivers compensate for the additional safety that wearing seatbelts provides by driving more recklessly. You can get his full findings here.

The bottom line is that the effects of risk homeostasis seem real, so that we can expect to see people compensate for decreased risks in one area by taking additional risks in another area. It’s still not clear, however, whether or not this compensating for decreased risks in one area leads to an overall gain or loss.

It might be the case, for example, that if you eliminate a risk that’s causing $10 million in loss that you get people taking compensating risks that only cause $5 million in loss. If that’s the case, then trying to eliminate the first risk makes sense. If the compansating risks cause $15 million in loss, however, then you’re better off not trying to eliminate the first risk. Because it’s impossible to tell which of these will happen for any particular risk, the best way to manage any given risk needs to be looked at on a case-by-case basis.

  • Ben Lewis-Evans

    As a researcher examining theories of driver behaviour I was interested to see the thesis to which you linked and the claim it proves Risk Homeostasis is true.
    I have to disagree however. The research while it does show behavioural adaptation (the term used to typically describe a reaction to a safety invention which is opposite to the intended effect – although it can be used in a positive fashion too), it does not show the complete compensation predicted by Risk Homeostasis Theory.
    Furthermore the manipulation used in the linked Thesis is very artificial and biases the results towards the production of behavioural adaptation. By making the penalties within the experiment very clear and very certain the researcher has created a situation quite different from real world driving. As one of the problems with real world driving is that the penalties associated with risky driving are very rarely encountered, and certainly are not certain in their outcome.
    The “rules” of behavioural adaptation state that it will occur if:
    1. The driver is aware of a change
    2. That the driver thinks the change will effect them
    3. That the driver has a reason to change their behaviour
    4. That the environment he driver is within allows for the change
    As you can see, the linked research clearly sets off rule 1 and 2 right from the start. Whereas in real world driving many of the changes which occur in the world around drivers may go unnoticed or be perceived as unimportant.
    Finally the linked research purely examines the conscious level of decision making and behaviour production (through the use of explicit manipulations made clear to the participants). Much of driving however is handled by highly automated unconscious skills, which most likely also play a role in behavioural adapation.
    The thesis is interesting, but (mores the pity because it would make my life easier) does not prove real world evidence of Risk Homeostasis.


  • Vic Napier

    Although there is theoretical support for risk homeostasis from the behaviors of large groups, from the standpoint of individual psychology risk homeostasis has very hard to scientifically define and measure. That may be changing, however, as the result of new research in neuropsychology. fMRI experiments hint at prefrontal-striatal-midbrain activity that may be constantly at work modeling future behaviors. Given a propensity for an individual to maintain a particular level of risk – something supported by the DR4R gene that codes for sensation seeking – and a tantalizing possibility emerges. It just might be possible that the brain is constantly modeling the interaction between future behaviors and events in order to achieve the most desirable outcome, and part of that outcome would be to maintain a particular level of risk.


Leave a Reply

Your email address will not be published. Required fields are marked *