Why rational choice is useful, as explained in introduction to economics

I was teaching a class in ‘The Economic Way of Thinking Yesterday’ and we did a little thought experiment to try to get an intuitive understanding of why rational choice is helpful for policy makers. The course is for public policy students, most of whom have politics backgrounds. The topic of this thought experiment was something along the lines of ‘how can we decide where to allocate government resources’? One of the main definitions of economics is ‘the study of the efficient allocation of scarce resources’. 

What I was trying to show is not only the problem that economics tries to nut out, but also why a crude but mathematical approach (as opposed to high theory about say, inequality) might be helpful. I was amazed by how well it all went, so I thought I’d write it up.

The discussion question was the following (note that we had discussed costs and benefits in an earlier question, here I made everything $100million so it is just about benefit):

Imagine the government has $100 million to spend on any of the four following programs. Assume the government is benevolent and effective, but do not assume that any of these programs already exists:

a.      Disability pensions

b.      Tax cuts

c.       Public art

d.      The Military

What should the government spend the money on?

What ought the government to spend the money on?

The difference between ‘should’ and ’ought’ is that the former implies a positive statement while the latter implies a normative judgement. We investigated how economics engages with these issues throughout the thought experiment.

I led the class in a discussion. First I asked them to rank their own preferences in secret from 1–4. I then asked whether anybody thought one or more of these items was not worth spending a cent on, or more clearly, was worthless. Nobody raised their hand in any of the three classes, so I think it’s safe to say that most people think these four items are of some value.

I then got information about people’s first preference. For the first class of 22, it was something like the following. Other class have different breakdowns, but I’m using this one because the split between tax cuts and pensions (which are diametrically opposed) was so helpful:

Item	1st pref.
Pensions	9
Tax Cuts	8
Art	1
Military	4

We had a quick look to see whether a simple voting mechanism—majority rules—would deliver an outcome. Here it would not (9 < {0.5*22}). We discuss more complex voting systems below.

We then turned to the public sphere for a solution. I asked people with first preferences for each of the items to provide some reasons for why they think their item should take priority. Some of the volunteered reasons are given below:

Pensions:

·         The most vulnerable in society need to be protected

·         Government taxation and spending has a redistributive role

·         The disabled are in a sense and to a degree reliant on the society (I asked a follow up question here: ‘but are they reliant on the state?’)

Tax Cuts:

·         Stimulates economic activity

·         Always popular for the government

·         Libertarian arguments

·         More money in people’s pockets to spend on what they want (i.e. they know how to satisfy their preferences better than the government)

Arts:

·         Non-rivalrous good

Military:

·         The government’s first priority is to provide defence and law & order

·         The military provides stability (though too much military can tip the other way)

·         Need to protect national security and resources like fisheries

I asked the students whether on the basis of these reasons anyone had changed their preferences. In one of the classes someone had—more in favour of tax cuts. The power of dialogue! But there was minimal change in preferences. I noted that economics tends to take preferences as given, especially in the sense that preferences will not change between now and when the relevant decision is going to be made. This is a powerful simplifying assumption, and one that we have just demonstrated is often not so silly. Given that almost everyone’s preferences were still the same after dialogue, we were no closer to a decision about where to allocate resources than we were before the dialogue.

Next we considered more complex voting systems. A system that would work here is where you do head-to-head majority rules votes until all-but one item is eliminated. I can’t be bothered doing the math, but I suspect that pensions would win there. However, it is possible for that style of voting to lead to perverse outcomes depending on what head-to-heads take place first. We instead explored a crude run-off system:

Item	1st pref.	2nd pref.	3rd pref.	4th pref.
Pensions	9*	6	5	0
Tax Cuts	8	8*	5	3
Art	1	3	3	15*
Military	4	5	9*	4

At this point, I mentioned the median voter theorem. I had earlier mentioned public choice theory as an application of the economic way of thinking to politics, and most of the students have politics backgrounds, so I thought this would be interesting to them. I drew it up using a bike paths situation. Canberra has a lot of bike paths because a lot of people like to ride, but it also has a lot of roads because people like to drive. Those who ride wants lots of bike paths and those who drive would rather the monies are spent on roads. There are also some intermediate types who mostly ride but still want roads for when they drive and vice versa. How might this play out in a stylised situation where we can vote (single-issue) on between 0 and 5 bike paths and the population is 10?

0 XX

1 XXX    

2 XX

3 X

4 X

5 X

In this situation, how many paths will be built? The students’ intuition kicks in pretty hard and they realise quickly that the answer is 2 because that’s the half way point—the median. In a single-issue majority-rules vote, coalitions will form and the winner will be determined by who tips the middle. This is the median voter theorem, which uses some simple mathematical modelling to provide an insight that allows us to make predictions in the case of more complex phenomena. Economics does this a lot. More generally, what economics does it to start by making a lot of simplifying assumptions until a basic model can be built. It then proceeds to gradually relax assumptions so that more complexity can be taken into consideration. 

Another important point emerges here, which is that the median is welfare maximising across the whole of society. The extremes (all roads or all paths) get something, but it is mostly the centre that dictates things. Having to care about the welfare of everyone in society and not just the majority makes things quite complicated when it comes to distributing resources.  

Coming back to our example, the median voter would, I suspect, be somewhere around the dominant preference spread, but again I can’t be bothered doing the math: pensions (1st), cuts (2nd), military (3rd), and art (4th). In none of the classes did such a person exist! It is quite possible that even a preference run off voting style would not result in a welfare maximising outcome.

At this point, I introduced the notion of a social planner. Where governments need to get re-elected and so care mostly about just getting a majority vote, the social planner (or technocrat) is interested in something a little different, namely, maximising welfare for everyone. When we invoke the social planner, we assume that they are benevolent, unbiased and omniscient, a ludicrous assumption, but a very helpful one for modelling purposes.

To get a sense for the welfare involved in each item, I asked the students to assume that a first preference was worth 4 points of utility, the second 3 points and so on. This got us the following:  

Item	1st pref.	2nd pref.	3rd pref.	4th pref.	Total ben.
Pensions	9 x 4 = 36	6 x 3 = 18	5 x 2 = 10	0 x 1 = 0	64
Tax Cuts	8 x 4 = 32	8 x 3 = 24	5 x 2 = 10	3 x 1 = 3	69
Art	1 x 4 = 4	3 x 3 = 9	3 x 2 = 6	15 x 1 = 15	34
Military	4 x 4 = 16	5 x 3 = 15	9 x 2 = 18	4 x 1 = 4	53

Tax cuts come out on top as providing the most benefit in terms of satisfying people’s relative preferences. In this sense, choosing tax cuts is welfare maximising. I mentioned here that it is critical to understand that what economics cares about is welfare or utility, not surplus, output, growth or anything like that. Those things are contingently tied to welfare, but it is utility that is the fundamental quantum. It is for this reason that some economists (notably the Chicago school) think there is a moral imperative to ignore equity considerations and go for the most Pareto efficient distribution of resources. This is a classical utilitarian way of conceiving of social welfare.  

Note: Pareto efficiency is where nobody can be made better off (in terms of utility) without making somebody else worse off. Perfect markets bring about Pareto-efficient outcomes.

I turned to illustrate this point with one of the most common mechanisms used in economics: the humble pie. If we, as a class, where making pie, would it be more important to make the pie as big as possible or to ensure that everyone gets a fair share of it?

Of course, people say that both are important. I point out that these are the equity and efficiency criterions. These are crude but powerful criteria upon which a social planner could decide how to allocate resources. Economic models, for the most part, show you how to make the pie as big as possible and how much the pie shrinks as you make the slices of it more even. I mentioned the disincentive effects of even the most efficient forms of taxation at this point as an illustration. Economics as a discipline passes no judgement upon what the optimal trade off is, but it can illustrate places where trade-offs can be minimised so that you get lots of bang for your buck (e.g. cash transfers to poor farmers rather than rice subsidies, or correcting market failures for public goods). This is very helpful for policy makers. How much equity a society should have is left to the public to decide and communicate to politicians, who in turn impose these values upon technocratic social planners. 

So economics is often very crude, but it is also very powerful and something worth having in your analytical arsenal.