My mother tells me that where she lives there are cameras that will catch you if you don’t come to a complete stop at the octagonal sign. Your license plates will be photographed and you will be sent a bill in the mail. The fine is close to $500. That’s a lot more than I remember it.
Quiz: suppose the technology improves for detecting whether a violation has taken place. Should the fine increase, decrease or stay constant?
Cheap Talk 
7 comments
Comments feed for this article
October 6, 2009 at 10:22 pm
Alan
In my opinion it depends on how the fine was arrived at initially. If the level of the fine was decided by picking an expected value of the fine then the fine should be lowered. If some E(fine) is a proper deterrent then technology increasing the chance one gets caught should allow the fine to be lowered and maintain the same expected value. (detterent effect trumps other concerns)
If however the fine was arrived at by picking a ‘reasonable amount’ for the fine then it should stay constant. (fairness concerns trump other concerns)
Maximizing revenue from fines might also be a concern. Raising it until some point, but I’m having trouble thinking of what the binding constraint is for fines, which I suppose may explain the $500 fine for rolling through a stop sign.
October 7, 2009 at 12:14 am
Alex
Alan, I would think that if the objective is raising revenue then the fine has to be lowered.
Lets say that the fine was initially $500 before the cameras were installed. Your probability of being caught and fined $500 depends on the probability of a cop being there to pull you over which is a low probability at most intersections (I’m excluding the possibilities of the cop letting you off and contesting the ticket in court). So lets say the probability of getting caught before cameras was 5% and therefore your expected fine for running a stop sign is $25. That’s enough for me to decide to stop but I can see a certain segment of the population valuing rolling stops over $25 and thus generating a revenue stream for the government.
Cameras lets say increase the probability of being caught to about 90% accounting for a 10% chance that the camera isn’t working. Then the expected fine for running the stop sign is $450. Under these conditions we would see movie stars, professional athletes, and CEOs who really like running stop signs paying the fines but these are few and far between and will likely generate little revenue for the government.
Of course, this analysis assumes that decisions made at a stop sign are always rational and made with complete information and they are not. Often people run stop signs because they don’t see them. Or if they aren’t from around the area they might not know about the cameras or exactly what the fines are.
October 7, 2009 at 1:42 am
samson
Running stop signs are a negative externality. The socially optimal number of stop signs ignored is almost certainly greater than zero. Whatever this optimal number is, it does not change with the detection technology, since the determinants of the social costs and benefits of running stop signs has not been altered.
The private benefits of running stop signs can be represented as a demand curve over the number of run stop signs. The social marginal cost function (presumably increasing) would be then determine the socially optimal number of stop sign violations. The optimal deterrent in a world of risk-neutral people is to set the expected fine to equal the private marginal benefit of the marginal transgressor. Since E[fine] = Pr(detection)*fine, increasing the probability of detection should induce a lower fine, in order to remain at the social optimum, as Alan pointed out.
If the govt has no concern of social optimality, but only wishes to raise revenue, then we have a classic monopolist problem with zero marginal cost. Set the E[fine] to equal the price at which the demand curve has unit elasticity.
Now if people are risk-averse, it is possible to show that the result is still true, i.e. the fine should be decreased, but what isn’t immediate is that E[fine] should be the same. I think this would depend on the distribution of degree of risk-aversion and pvt benefit from running stop signs.
October 7, 2009 at 8:25 am
Daniel Reeves
What Alex said: it should go down in order to keep the *expected* cost — probability of detection times the fine — of running the stop sign constant.
October 7, 2009 at 8:58 am
gappy
I agree with Samson. Moreover, the marginal cost of an infraction is greatly reduced with cameras, and that difference in cost is transferred to the driver.
October 7, 2009 at 4:53 pm
mobile
If history is any guide, it will increase. And after that, the fine for vandalizing official police cameras will also increase.
October 7, 2009 at 10:17 pm
jeff
I would add one thing to the excellent analysis given above. Prior to the technological advance there may have been a binding constraint and this can reverse some conclusions.
The best example comes from a consideration of type I and type II errors. We can imagine that the fine is set to internalize externalities as best as possible subject to *not imposing high costs on the innocent.* This constraint is binding when the technology is poor.
If the technology for detecting a violation is poor then there will often be false positives (type II errors.) That is, non-violators will be ensnared. A fine set so high as to deter violations may then impose an unacceptably high expected cost on the innocent. This binding constraint would keep the fine lower than the first-best.
Then when the technology improves, false positives become rarer and the constraint is slack enabling the fine to be *increased* in the direction of the first-best.
Taking this one step further, a cynical interpretation of the $500 fine is that the camera can be trusted to fine only the guilty. Before the cameras, a $500 would be unnaceptable not because the police are any worse at spotting offenders but because the police could not be trusted if the fine were so high.