Saturday, 21 January 2017

Auckland Transport vs Fare Evaders

One of the courses I took last year was Economics 303: Law and Economics. I think you might characterise this as being one the "policy economics" papers along with Public Economics, Energy and Resource Economics and Environmental Economics but I have not taken any of those other three. What I would say is that the course looks at several aspects of law from an economist's point of view, i.e. legal topics in economics theory. One of those topics was crime. Fare evasion is a crime.

The basic principle of all the economics I've done is that people are rational. When it comes to crime, then, a criminal commits an offence because their marginal utility (satisfaction) from offending exceeds the marginal cost of offending. In some sense, then, a criminal is a producer and the product they supply is crime. That's not an entirely apt analogy (what is the demand?) but you get the idea.

As I have said many times, economics is about scarcity, choice and allocation. In terms of crime this manifests largely in terms of the optimal level of deterrence. Thus, what one needs to do is have some sort of model of the costs of apprehension/enforcement and a model of the net harm caused by crime (we have to treat the offender's utility as the private benefit... crime is now more a production process with negative externalities). Naturally, then, the objective is to minimise the social loss experienced by crime. To do this, what is actually considered is the expected utility to the criminal. In this understanding, criminals offend when the expected utility from crime exceeds that from their next best alternative use of their resources. This expected utility takes into account good and bad possible states... these are states society can influence.

Let's recap:
  • scarcity -- criminals have only so many resources with only so many uses; government has only so many resources with only so many uses
  • choice -- criminals choose to offend based on expected utility; government chooses resources to minimise the social loss of crime
  • allocation -- more resources will go to crime if  the expected utility from crime is more often relatively higher; government will allocate resources based on their understanding of what reduces crime's costs the most
At least, if everyone is rational.

So, how does this all work? Well, what does expected utility actually mean? Very simply, it is an expectation (like in statistics) and that means it basically weights the utility and costs of an action/procedure against the chance they happen. So, if, for instance, we assume that there's a 10% chance of getting $10 and a 90% chance of getting $20, the expected value is .1*10 + .9*20 = $19. In terms of 303, specifically, the model we used dealt with the probabilities of good and bad wealth states. And, in fact, everything largely depends on whether or not a criminal is risk averse, risk neutral or risk loving. That is, how does their risk profile affect their expected utility? I think the best way to explain this is by example.

Let us assume that John is a fare evader (a type of crime). We will assume that his utility curve has the form: U = w2, where w represents wealth from fare evader. We will assume, in very simple terms, that the ideal state from John's fare evader is wg = 10 and the bad state is wb = 5. These numbers aren't very realistic, I know, but they're sufficient to explain. Finally, we'll assume the probability of the good state is P(wg) = .8 and the probability of the bad state P(wb) = .2. What is the expected wealth? (We'll denote this W.)

W = .8*10 + .2*5 = 8 + 1 = 9

What is John's utility from the expected wealth?

U(W) = 92 = 81

What is John's Expected Utility?

E(U) = .8*100 + .2*25 = 80 + 5 = 85

That is, John is risk loving. In other words, he prefers to play the lottery (i.e. be subjected to the chance created by uncertain wealth states) than receive for sure the expected wealth from the lottery. Risk averse people are the other way around and risk neutral people aren't fussed either way. There is an interesting consequence of these profiles for government, though. Let us imagine that the cost associated with the bad state doubles. That is, wb = 0 now.

W = .8*10 + .2*0 = 8 + 0 = 8

U(W) = 82 = 64

E(U) = .8*100 + .2*0= 80 + 0 = 80

What if the chance of the bad state doubles? That is, P(wb) = .4.

W = .6*10 + .4*5 = 6 + 2 = 8

U(W) = 82 = 64

E(U) = .6*100 + .4*25 = 60 + 10 = 70

Implication? If you increase the likelihood of the punishment, then John is less likely to fare evade than if you increase the punishment associated with fare evading. The problem is, as noted in this video, there are costs associated with increasing the probability of capture. (Notice how it also discusses far evading?) And, of course, the criminals need to be made aware that they are more likely to be caught... this also costs money. There are fewer costs associated with the less effective (assuming, as I think is fair, fare evaders are risk lovers) raising of levels of punishment. What to do?

Took This Photo Myself
Apparently, Auckland Transport has gone for the more effective option, even though it costs more money. I cannot really speculate as to why they've taken this particular route, but it may matter more to AT to make sure it has the most accurate data possible when it comes to passenger numbers and station use than anything else. Also, because only one behaviour is being targeted there are fewer choices that need to be made (the risk profiles for, say, car thieves and cat burglars might be quite different, even though both are thieves at core) and there has been news about the cost of fines going up over the last few years. Possibly, the reason why the probability of capture has increased five fold is that AT expects the increased number of fines to pay for the increase wage costs for inspectors. Or maybe they're lying. Who knows? Point is, I saw this ad and I had taken a course that suggested a line of reasoning behind the decision to run it.

The question of fare evasion is a superficially obvious one: catching a ride on a train you haven't paid to use is stealing. This is similar to arguments deployed against downloading and streaming but there are important distinctions. For one, when you get below the superficial level of theft is wrong, you realise that we're really talking about (literal) free riding. Auckland Transport recently confirmed that Westfield train station would be closing in March 2017 (why they are so keen on introducing changes in the busiest period of the year I do not know) and this brought up the question of why Te Mahia train station would be remaining open with its even lower patronage figures in the comments section of Transportblog. The answer that came up was the question of fare evasion: perhaps there were actually substantially more users of Te Mahia than get recorded in the data? This was controversial because, after all, how do you measure this? But the reasoning is sound: when it comes to provision of social services it is really important to have accurate data. If it is true that fare evaders nearly caused Te Mahia to be train station-less there is a clear illustration of the potential for the interests of the butcher, baker and candle-stick maker to harm the social interest. In other words, it is entirely right to dedicate effort towards catching fare evaders.

Whether we're talking about fare evasion or more traditionally serious crimes (I hope you can see that fare evasion can have wide-ranging consequences), I have no idea how practically useful the paradigm of understanding crime as taught in Economics 303 is. I can say, though, is that what we have here is a beautiful illustration of a point of education: insight into phenomena that one encounters in the course of life. It may not be the world's most charismatic topic, but fare evasion has thus proven an ample demonstration of how education unfolds as one progresses. It has also shown the truth of my characterisation of economics and, possibly, this post has offered insight into the information loading of videos versus text.

No comments:

Post a Comment