I recently renewed my car insurance, and one of the main questions was "How many kilometers do you drive a year?". I initially thought that the insurance price for collisions should be directly proportional to this number.
But in reality, the function should not be linear, since by not driving I am changing the density of drivers on the road at any one time, thus reducing the likelihood of collisions occurring per car hour. It may take a monopoly (i.e. a single insurer) to be able to pass on such savings in a non-statistical manner.