> You _do_ still have the problem of people dying as a result of private industry, but the rate at which people die per mile is an order of magnitude lower if y

> You _do_ still have the problem of people dying as a result of private industry, but the rate at which people die per mile is an order of magnitude lower if you take the bus versus a car. These numbers don't reflect the absolute difference in safety between these modes of transportation. Imagine a world in which buses replace cars as the primary mode of transportation. You've now got many, many more buses on the road. So, if there is a collision, rather than probably being car-on-car or car-on-bus, it's most likely to be bus-on-bus. That will raise the fatalities per billion passenger miles in two ways: (1) The amount of energy in a bus-on-bus collision will be quite a bit larger than when a car hits a bus or a car hits a car. (2) When there is a fatal collision, the number of fatalities is likely to be higher. (It takes no stretch of the imagination to imagine if one person on a bus dies in a collision, his fellow passengers are more likely to also die.) You also have second-order effects, like vastly reduced visibility for bus-drivers, pedestrians, bicyclists... > Forcing manufacturers to deal with the externalities we're discussing would also incentivize them to encourage safe driving, no? No. Not if they don't actually have any control or responsibility over the externality, like in the case of drunk driving, which was my original point. The only things a manufacturer could do in that case is fight for prohibition or install a breathalyzer in every car. The former was tried and turned out to be pretty bad for society. The latter probably would destroy the industry and/or be trivially worked around by the dedicated drunk.