Just a little bit of Macro First

I don't have any easy answers to the "Micro first" vs "Macro first" debate when teaching Introduction to Economics. But this is what I do, and it seems to work a bit. And I have probably taught Intro about 30 times in my life. That experience should count for something.

1. We (Carleton University) are one of the last holdouts that refuses to split Intro Economics into two half courses. We are being "difficult". But I refuse to say someone has had an introduction to either micro or macro unless they have had an introduction to both. You need to be able to see it from both sides, or you can't really see it at all.

2. We start out with micro, BUT..........................

I start out with a Production Possibilities Frontier. Apples and Bananas.

You can teach a helluva lot with just a PPF. And not just opportunity costs. What is the relation between the slope of the PPF and the relative price of apples and bananas? Why is the PPF curved, and what does it mean? (And unless you can answer that question, you do not understand why supply curves slope up.) And what does it mean if the economy is inside the PPF?

Next I teach comparative advantage and trade. For that you need two PPFs, with different slopes. You can call them "Farmer" and "Rancher", or you can call them "Canada" and "US".

I first assume barter. The Canadian BIB students know they can make a profit by "buying low and selling high". So they fill their canoe with Canadian apples, cross Lake Ontario, swap their apples for US bananas at the US relative price, then paddle home, swap their bananas for apples at the Canadian relative price, and find out they have more apples than when they started (minus what they ate while paddling). Which is their Profit! So other traders follow them, so Canadian and US relative prices converge. Standard stuff.

Then I introduce Money, and The Nightmare Scenario. Cue scary music. What happens if the dollar price of both apples and bananas is lower in the US than in Canada? Will free trade with those dastardly clever cheap productive Americans cause mass unemployment in Canada???

Then I talk about exchange rates. Whose "dollar" are we talking about? Yes, if the exchange rate is too high, The Nightmare Scenario would indeed happen.

But we don't blame free trade for The Nightmare Scenario; we blame the Bank of Canada, for keeping the exchange rate too high. The Bank of Canada should print money, to make the exchange rate drop, until Canada can export apples.

The alternative, under fixed exchange rates, is to wait for mass unemployment to cause Canadian wages and prices to drop, until Canada can export apples. Which might take some time.

Then I go on to all the regular micro stuff, like demand and supply curves. (Because demand and supply don't make any sense unless you understand why people trade. Which is why Mankiw does trade theory before demand and supply.)

I'm not saying it works perfectly, but it seems to work. The students understand that if the Bank of Canada gets it wrong, the economy will be inside the PPF, and all that opportunity cost stuff might not work. And it only takes about 20 minutes extra to teach it, on top of the regular micro stuff. It's totally unrigourous of course, but so what.

It probably works better in a smallish open economy like Canada; I'm not sure how I would do it if I were teaching in the US.

Comments

[Frances Woolley]

Nick, the way I see it, macro first v. micro first is basically just a way of framing a much deeper, more fundamental question: what are the core, paradigmatic assumptions of our discipline? Micro first is basically a way of saying "the core paradigm of economics is the theory of rational choice. Everything follows from that. Until you know that, you know nothing."

Your trade example, although you call it macro, is firmly within that micro rational choice paradigm - it's a story about people making choices, and the consequences of those choices.

It seems that those arguing for a different first year are really taking a shot at the centrality of that rational choice model. For example, here's Noah Smith, http://www.bloombergview.com/articles/2015-11-24/most-of-what-you-learned-in-econ-101-is-wrong

"We now have an academic economics profession focused on examining evidence and an Econ 101 curriculum that focuses on telling pleasant but often useless fables. Econ education needs to get with the times."

Rational choice theory is harsh. It's nice to believe that given a few gentle nudges - labelling trash containers "landfill" rather than "waste", for example - people will do the right thing. It seems cruel and callous to think that people just basically do what's in their self-interest, and the best way to get people to change their behaviour is to change incentives.

Yes, there are lots of ways in which people's behaviour departs from the simple rational choice model. But those exceptions shouldn't blind us to the model's overwhelming power. Even the example Noah Smith expounds at length to discredit standard intro - that minimum wages don't have much effect on employment - is a no brainer to explain in a simple intro framework. If the demand for labour is elastic, minimum wages will have a big effect. If the demand for labour is inelastic, minimum wages will have a small effect. So it turns out either that the demand for labour is pretty inelastic in the short run (not surprising) or that employers have some degree of monopsony power (also standard intro level stuff). That we draw the diagrams with elastic labour demand curves and competitive market equilibrium may say something about the ideological position of people who write best selling intro textbooks, but it doesn't really get at the fundamental validity of the rational choice approach.

[Gene Callahan]

" It seems cruel and callous to think that people just basically do what's in their self-interest..."

In pure rational choice theory, this is simply a tautology, not a "harsh" fact of life. It means nothing more than "what we see people doing, that's what they think it is in their interest to do. So Mother Teresa no less than Michael Milken acted in her "self-interest" in terms of the pure theory of choice.

To posit that people act mostly for "material gain" or something of the sort is an additional assumption on top of the model, and is unnecessary for rational choice theory per se.

[jonathan]

I wouldn't call that approach "starting with macro". I also think it's fairly standard (my own first micro class started with PPF, then comparative advantage). I don't think we directly addressed the "all goods cheaper" fallacy, but I agree that this would be a useful exercise.

The only thing I disagree with is teaching about fixed exchange rates. To make sense of that, you need some frictions in international adjustment -- either sticky prices, non-traded goods, non-integrated capital markets, etc. I think that it's more important to help students understand the frictionless benchmark first.

[jonathan]

I wouldn't call that approach "starting with macro". I also think it's fairly standard (my own first micro class started with PPF, then comparative advantage). I don't think we directly addressed the "all goods cheaper" fallacy, but I agree that this would be a useful exercise.

The only thing I disagree with is teaching about fixed exchange rates. To make sense of that, you need some frictions in international adjustment -- either sticky prices, non-traded goods, non-integrated capital markets, etc. I think that it's more important to help students understand the frictionless benchmark first.

[jonathan]

(Feel free to delete the double post and this one -- I got an error on my first post attempt, and so hit post again.)

[Nick Rowe]

Frances and jonathan: My post wasn't clear enough. The only bit that I consider really *Macro* is when I added money to the Ricardian trade model. The rest of it is very standard, and I just put it in there for context, so readers could see at what point I went off onto a macro theme.

But that does raise the question, that I ducked in my post: what the hell is Macro anyway? Because General Equilibrium theory is about "the economy as a whole", but it's as least as much micro as macro. But *Money*, I claim, is a macro topic. Money is what makes macro different from micro. I should do a post on this.

Frances: Yep, I can't imagine teaching Intro without teaching some sort of theoretical explanation. And concentrating on explaining exactly what the "economic perspective" is, which includes rational choice, and supply and demand, etc., seems essential too.

Funny things about minimum wages and elasticity is: a lot of people who say minimum wages won't hurt employment (much) also say immigration won't hurt wages (much). I think that's right. But the first says the labour demand curve is very inelastic, and the second says its very elastic. Having some sort of theoretical perspective does discipline you a bit about how you approach the data and policy.

jonathan: "The only thing I disagree with is teaching about fixed exchange rates. To make sense of that, you need some frictions in international adjustment -- either sticky prices, non-traded goods, non-integrated capital markets, etc. I think that it's more important to help students understand the frictionless benchmark first."

I disagree. Fixed exchange rates is just one possible monetary policy. It's not a friction. But the sticky prices or wages are a friction.

[jonathan]

Nick: Sorry, I wasn't clear. I'm fine with teaching about exchange rate regimes, and how they make no real difference under flexible prices (just determine whether P or e adjust).

My disagreement is with discussing unemployment (operating inside the PPF). For this you need sticky prices or some other friction that prevents the standard monetary adjustment mechanism from operating, which I think is confusing at that point. I think at that stage you want to be explaining the classical benchmark, rather than introducing Keynesian mechanisms. That's for macro!

[Nick Rowe]

jonathan: Ah. I misunderstood you. But I still disagree. First year students worry about aggregate demand, even though they can't articulate that worry. And it's hard to teach them micro if they have this uneasy feeling about aggregate demand at the back of their minds. So I think we need to say something about it, to face The Nightmare Scenario, and say something about what causes it and how it might be fixed, early on in the course. Even if that means a lot of handwaving.

[Frank Restly]

Nick,

At what point between micro (barter) and macro (money) does time become a precious resource?

Not sure this is relevant but in Engineering and Physics there is the jump from 2 dimensional problems to 3 dimensional problems. There is also the jump from problems involving static (non-moving) objects to problems involving dynamic (moving) objects. The way it was taught for me was start with 2D static objects, move to 3D static objects, then move to 2D and 3D dynamics. The reason it is done this way is because of the math - statics can be done with algebra and trigonometry primarily. Dynamics requires calculus.

And so economics has been broken down into small (micro) and big (macro). In line with physics, economics could be of the static and dynamic varieties. Static economics would be economics without trade or barter (no moving goods). Dynamic economics would be economics of trade.

The second axis that economics could be broken (similar to 2D and 3D physics) would be time as a fixed / infinite resource. Rather than money as a discriminating factor between micro and macro, I believe time is the discriminating factor. Individuals have a fixed lifetime to make individual decisions and they know it (micro decision making). The human population has a seemingly unlimited lifetime to arrive at a group consensus (macro decision making).

Consider the following course layout:
Economics IA - Static Economics, Infinite Life Agents And Goods (Algebra and Trig required)
Economics IB - Static Economics, Fixed Life Agents And Goods (Algebra and Trig required)
Economics II - Dynamic Economics, Fixed Life Agents and Goods (Calculus required)

[Nick Rowe]

Frank: it is easy to do a (simple) static 2D model of trade. The Edgeworth box is one example. Or simple supply and demand curves is a second.

[Frank Restly]

Nick,

I wasn't sure the analogy was appropriate (physics versus economics) but I thought I would put it out there.

Reading what I could find on the Edgeworth box:
https://en.wikipedia.org/wiki/Edgeworth_box

The Edgeworth box is a division of resources description, not necessarily a trade description. We can say that given a certain pre-ordained division of resources, Octavio and Abby may each try to jump to a better indifference curve. But it does not establish why Octavio and Abby try to move in the first place (is it gravity pulling them?) and what might impede them from doing so (friction?).

[Frank Restly]

If we all live forever and goods don't wear out, rot, or turn to dust then it really doesn't matter how long it takes to jump from one indifference curve to the next and what might be in our way when we try to do so. If, we will all die someday and our apples will eventually be a pile of mush, then it matters quite a bit.

[John Hawkins]

Frank, true dynamics (in the way you are describing) in macro is very lacking in its current state. Anything concrete would not be part of an intro course, it would be part of a Ph.D. seminar.

[Frank Restly]

John,

I don't understand why that is the case - is it the math? Engineering students ended up with 3 semesters of calculus and an course dedicated to differential equations by the end of their 2nd undergraduate year when I went to school. And most had taken introductory calculus in high school as well.

[rsj]

Frances,

The question is not whether rational choice theory is harsh, but whether it is relevant. This is the point that Noah was making -- sure you may get the direction of sign right, but the goodness of fit may be very poor. If a theory does a poor job of explaining something, then perhaps it should be de-emphasized. The specific example of employment is relevant because while the sign is mostly right, the goodness of fit is just horrible. Other affects appear to be much more important for understanding minimum wage changes on employment, in which case don't we owe it to students to teach the most important factors at play rather than factors that just tell a nice story?

Consider the case of psychology. A typical psychological first year curriculum would include a bestiary of various disorders, for example. They could, if they wanted to, try to reduce human mental health to a single grand mathematical model which for purposes of mathematical convenience is assumed to be maximization of convex functions. Then they could start rolling out mathematical theories of personality disorders in terms of a least action principle. This would be a barbaric and wasteful exercise. Yet in consumer choice theory such violence to reality is considered OK. Why? It's really surprising how it's OK to have these 19th century Grand Unified Theories of the Mind in economics when all the other social sciences consider such grandiose claims to be both unprofessional and discredited.

Consumer choice has to be much more complex than psychology. Certainly the firms that do marketing and set prices don't tend to focus on consumer choice theory when it comes to maximizing their profits.

Noah suggests that comparative statics be given the level of attention it deserves -- e.g. treated as a niche hypothesis -- which is taught along with many other hypothesis in a type of bestiary, with no attempts to form a Grand Theory Of Consumer Choice. The central core of the curriculum would be mathematical tools for data analysis and hypothesis testing. I find it intriguing.

[Frances Woolley]

rsj: "This is the point that Noah was making -- sure you may get the direction of sign right, but the goodness of fit may be very poor."

And this is the point that I disagree with. Humans generally speaking consider themselves fundamentally different from, say, chimpanzees, or other animals, whose behaviour can be readily described by maximize utility=f(consumption, leisure, sex, social status). Or perhaps even maximize utility=f(number of surviving offspring). Because this is the nature of cognition - we look for patterns - commonalities in random objects, differences in things that are apparently similar. Someone from another planet observing our species would probably see us as just another animal, also striving to maximize consumption, leisure, sex and social status.

Sure, there are lots of fun and exciting things that we can do around the edges to nudge human behaviour one way or another. For example, we can show that people eat more when they're given bigger plates, less when they're given smaller plates, remarkable amounts of disgusting stale popcorn when they're not thinking about what they're doing.

But what that kind of analysis fails to explain - and what that so-called "niche hypothesis" of rational choice theory can explain perfectly well - is why people get out the small plates when food is in short supply. Why every known society creates rituals around dinner to harness and control people's impulses.

Psychology cannot, in fact, do this, because it's inherently an inductive discipline. See this post on http://worthwhile.typepad.com/worthwhile_canadian_initi/2011/09/the-behavioural-economists-dilemma-induction-versus-deduction.html.

[Avon Barksdale]

I agree completely with Frances Woolley. She is absolutely right on this issue.

To understand anything in nature – and that includes economics – you need to discover the laws. There is an awful lot of pseudo-science nonsense in the social sciences, but economics is not one of them. Economics does have laws: 1) Humans are rationally self interested and organize their lives accordingly, 2) Prices adjust until the amount demanded equals the amount supplied.

At times, we see these laws break down, but they usually don't. Of course they are an approximation to reality, but they work really well. Students need to understand that organizing structure of these laws from day one, and that is the micro foundation of economics. In physics, we start with Newton's Laws of Motion. They too are an approximation to reality, but if you're not too small, not too big, and not going too fast, Newton's Laws are perfectly valid.

Noah is talking about collecting data and going through the motions of science without really doing science. You need to understand the laws or you won't know how to questions of data, or how to use data to ask deeper questions. Everything and nothing will seem a pattern.

I recommend Richard Feynman's take on the social sciences. https://www.youtube.com/watch?v=IaO69CF5mbY He was one of the greatest minds in human history. Economics is a serious science precisely because it has laws to organize its structure. Make sure students understand that.

[rsj]

Frances,

chimpanzees, or other animals, whose behaviour can be readily described by maximize utility=f(consumption, leisure, sex, social status).

This is not true. Please point me to the relevant research documents that describe these universal "laws" governing the behavior of animals. Seriously, why is making up these Grand "Laws" OK in economics when every other social science field has called Bullshit on this.

[rsj]

...And it's worth pointing out that something as simple as the weather, for example, is extremely complex and cannot be accurately modeled based on first principles because even though the basic underlying laws are in this case known, solving those differential equations is too complex and they exhibit chaos -- e.g. knowing the laws doesn't tell you much about outcomes because of extreme sensitivity on initial conditions. But for something as complex as why we (or animals) make the choices that we do, we don't even know the laws.

Agent based Models of Macro economics have similar characteristics -- e.g. just one agent in the economy changing their behavior can cause the entire economy to go in a different direction.

But of course human behavior (or animal behavior, or even the behavior of a single celled protozoa) is much more complex than the weather.

There are now super computers trying to model a 100 neuron organism and they can maybe kinda make some progress, but it's not at all clear on how accurate the model is. The animal wiggles a bit like you'd expect. Yet you talk confidently of how you can describe the behavior of a Chimpanzee. It really does feel like I've stepped back into the 1850s when I read this stuff.

Why is it OK to lie so much in economics and make such grandiose claims about human behavior? I have to think because while other social sciences have tended to either abandon making such claims unless they are corroborated by data, no such compulsion for accuracy has been felt here, until recently. Noah is championing this as a revival of economics that cares about promoting knowledge of the world rather than advancement of a particular ideology, but there is an old guard for whom philosophizing about how the world works is good enough, as long as the math checks out and the story sounds plausible.

[rsj]

..and as a simple example, it's not the case that the dopamine levels in your brain are going to be some function of current consumption -- the brain has only a finite number of receptors, so there would be a "maximum" utility that could be reached. But of course it's all relative, based on your own experience, how much pleasure you get, or if you get any pleasure, from consumption. And that type of chemistry is extremely complex with not nearly enough known about it. Let alone creating a single number for "social status". Utility -- a concept rejected by Adam Smith as too simplistic to describe behavior -- cannot map to something like dopamine because it has to be a single valued function. But if you were even to modify your function to be a bit more realistic: u_n = f(c_n, u_(n-1)), u < u_max, u > u_min, u changes in discrete intervals, then your "laws" start looking a lot closer to chaotic laws that don't help you make grand claims about consumer choice.

[Nick Rowe]

There was a late frost this Spring. This Fall, I noticed that local apples weren't as cheap as they normally are. I think I can explain the relation between those two facts with a simple supply and demand diagram. I want my first year students to understand that explanation.

They can learn statistics and econometrics in other courses. Frances teaches student how to formulate a hypothesis and test it against the data. Those skills are useful too. But those skills are more than just observing and measuring correlations.

[Tel]

FW: chimpanzees, or other animals, whose behaviour can be readily described by maximize utility=f(consumption, leisure, sex, social status).

rsj: This is not true. Please point me to the relevant research documents that describe these universal "laws" governing the behavior of animals. Seriously, why is making up these Grand "Laws" OK in economics when every other social science field has called Bullshit on this.

If you want to read interesting studies on apes, try Frans De Waal (be cautious, he happily mixes his own presumptions with his observations, but I guess we all do to some extent, you will have to sort that out yourself).

However, there's a more basic philosophical way to look at this which avoids the need for research papers, let's suppose you observe the chimp has a choice of eating bananas, or apples, or peanuts and you observe him eating the bananas. There's only one logical conclusion you can make; which is the chimp must have "wanted" to eat those bananas. If the chimp had eaten the apples instead you would have concluded the chimp "wanted" to eat the apples.

Indeed, if the chimp grabs a banana and sticks it into his ear, you would never the less have no choice than to conclude this is also utility maximizing behaviour... thus giving rise to the definition of "wanted to do" in an empirical sense, is that given some choice this is what happened. In the cases where the is no choice, of course, you cannot make any behavioural observation at all.

I'll point out a very specific phrase used by Frances Woolley: "whose behaviour can be readily described"... this is quite accurate, of course all behaviour in all circumstances can be readily described after the fact, as intentional utility maximizing behaviour, given the constraints of the situation, and the given individual cognitive capability. You see, this is always true in a tautological sense. If you were to attempt to describe it as utility minimizing behaviour, then how would that work?

What Frances Woolley did NOT say is that chimpanzee behaviour could be reliably predicted ... because prediction is much more difficult, you need to know not merely that the creature is attempting to maximize utility, but you need to know the entire calculation it internally uses to decide that utility, including any random "tie-breaker" factors, and this modelling would need to include learning and adaptation over the medium to longer term. Very difficult indeed.

[Tel]

Why is it OK to lie so much in economics and make such grandiose claims about human behavior? I have to think because while other social sciences have tended to either abandon making such claims unless they are corroborated by data, no such compulsion for accuracy has been felt here, until recently.

While Keynes was busy screwing up the world economy (making bureaucrats more powerful than they had ever dreamed), and leaving behind the legacy that we now must deal with, another man became quite bitter and angry about this rising tide of economic tomfoolery, and he wrote along these lines:

Economics is haunted by more fallacies than any other study known to man. This is no accident. The inherent difficulties of the subject would be great enough in any case, but they are multiplied a thousand fold by a factor that is insignificant in, say, physics, mathematics or medicine—the special pleading of selfish interests. While every group has certain economic interests identical with those of all groups, every group has also, as we shall see, interests antagonistic to those of all other groups. While certain public policies would in the long run benefit everybody, other policies would benefit one group only at the expense of all other groups. The group that would benefit by such policies, having such a direct interest in them, will argue for them plausibly and persistently. It will hire the best buyable minds to devote their whole time to presenting its case. And it will finally either convince the general public that its case is sound, or so befuddle it that clear thinking on the subject becomes next to impossible.

In addition to these endless pleadings of self-interest, there is a second main factor that spawns new economic fallacies every day. This is the persistent tendency of men to see only the immediate effects of a given policy, or its effects only on a special group, and to neglect to inquire what the long-run effects of that policy will be not only on that special group but on all groups. It is the fallacy of overlooking secondary consequences.

I'm sure this political influence on Science doesn't just happen in Economics. There's an excellent article "Edward Krug Flunks Political Science" by William Anderson, which describes how the process works (actually Ed Krug also describes these events from his own perspective, you can buy Krug's book for a few bucks on Kindle, and yeah he does also sound pretty bitter about the situation, understandably so from my perspective).

[Avon Barksdale]

Nick,

Get to the laws or you will be on the wrong side of Feynman: https://www.youtube.com/watch?v=IaO69CF5mbY

[nivedita]

@Avon, I think Feynman's point was that social scientist's haven't discovered the laws yet.

The two laws you set out for example, I think are of much more limited applicability than Newton's laws of motion.

Everyone recognizes that much of their own behavior is not consistent with rational self-interest. I drink more coffee and eat more food than is healthy for me, and I know that, and if someone asks me I will say that my coffee consumption and food consumption are irrational and against my self-interest. I don't actually change them though.

If you look at the prices around you, it is not often that they adjust until quantity demanded equals quantity supplied -- in fact, most often we see prices that don't adjust at all, in daily life, haggling has gone out of fashion. Indeed, this happens most readily in situations where it is forced to happen, the prototypical examples being futures or stock markets.

[rsj]

There was a late frost this Spring. This Fall, I noticed that local apples weren't as cheap as they normally are. I think I can explain the relation between those two facts with a simple supply and demand diagram. I want my first year students to understand that explanation.

The first problem is that if you try to teach this as a general "law" to your students, then they will think that an increase in immigration drives down wages, which it doesn't, that long wait lists for iphones will drive up prices, which they don't, etc. E.g. you will be teaching them how to think wrongly about the relationship between prices and availability of products. There is a difference between finding some common themes that sometimes apply and sometimes don't -- and you don't really know whether they apply or not until you do some studies. Universal laws are very scarce and simply not available in micro-economics. So go ahead and teach them that, as long as you also teach them that these types of effects are seldom the dominant effect that goes into determinations of price and availability. In which case, why not teach a whole host of such rules of thumbs? E.g. quantities of products generally don't run out unless a supply chain is interrupted, in which case the supplier is going to focus on fixing the interruption and may well give a discount or rebate to waiting customers in order to keep them as customers and prevent them from choosing a competitor's product rather than trying to exploit the interruption to temporarily hike prices. That will make your students a lot more informed about price determination. You could use text books in which firms are asked about their own price setting behavior, giving answers that would infuriate any micro econ professor.

The second problem is one of intellectual honesty -- are you teaching students how they should act or how they do act? An electron doesn't need to struggle to find its position, even though the math is complex. That's a law -- you follow it even if you have a low IQ. It requires no effort to follow the law, you can't help but follow it. No such laws exist in describing human choices. When students are given baskets of different goods and asked to rank these baskets in order to find their indifference curves, they find the process extremely frustrating and difficult. They tend not to be able to repeat their ranking, and are so discouraged by the exercise that they give up and start picking random baskets just to get out of the study. Many people find shopping extremely stressful. Utility optimization based on indifference curves is _not_ a natural mode of subconscious behavior. It requires enormous focused effort. But we need to rapidly select thousands of products and do so with minimal effort because we need some focus left for enjoying ourselves and working. If the point of your class is to teach them how best to optimize, then that is very different from a description of human behavior, and you may want to highlight that your class is prescriptive rather than descriptive.

[rsj]

Another question is how often do the rules of thumbs need to be wrong before you stop teaching them? As basic micro gives the wrong answers on immigration, minimum wage, and welfare, then you will be creating voters who will be on the wrong side of this issue, voting ideologically when they believe they are voting based on "economics". To me, even a 10% failure rate is reason enough not to give people false information, or to at least temper the class discussion with sufficient warnings to the effect "hey this stuff is often the wrong way to look at policy decisions, here are some examples when these rules don't work". It would be better to not teach anything than to teach something that is wrong.

[rsj]

Tel,

Anyone can take two variables and plot them, getting some functional form, but unless the form is predictive on out of sample behavior, and here we want to be very broadly predictive, then this isn't a "law" of behavior. And of course when the topic is *decision making* for an animal that has billions of neurons, there aren't going to be _any_ known laws of decision making. There will only be phenomenological heuristics whose boundary of applicability is unknown, precisely because we don't know how to model decision making for anything more than a couple of neurons. Even the 100 neuron nematode that we can kinda model is too complex to reason about theoretically, and can only be modeled numerically with state of the art supercomputers. A human has a 100 billion neurons, with about 20 billion in the cerebral cortex and a chimpanzee has about 1/3 of that. Given that we can only reason about a handful of neurons, and that these are not the be all and end all of human decision making, there is _no_ known law of consumer choice for either chimpanzees or humans, and certainly not a convex function into which we can dump a few numbers in order to get the "objective" that the human or animal brain is trying to maximize. That is complete bullshit.

[Bob Murphy]

Nick,

Sorry to go off-topic, but I am going to kinda sorta defend Dean Baker from you and Sumner. But first, I need you to answer this question (which I posted at Scott's blog probably after you stopped looking at that post):

What do you think house prices should do, if nobody made any obvious mistakes? Just rise with CPI so that the expected appreciation in real housing prices is 0% for any time horizon, at the moment of purchase?

[Tel]

Anyone can take two variables and plot them, getting some functional form...

They tried that with the Phillip's Curve and discovered that no matter how you jigger it, still you get a cluster of dots all over the place.

If you want to demand predictions from economists you are going to have to be quite open and explicit about that demand, you cannot simply presume that's what they are talking about, especially when the words themselves said it was descriptive.

And of course when the topic is *decision making* for an animal that has billions of neurons, there aren't going to be _any_ known laws of decision making.

Well I just gave you a known law: all decisions made are the outcomes of decision making. What's that? Don't like it? Sheesh, tough crowd.

Look, while it's fun yanking your crank here, please don't blame me for the situation, I'm pointing my finger at the moon. Whatever you do to me, the moon is still going to be there like it was before I pointed, but to actually climb up and knock down that moon is a really big job.

In case you have the slightest doubt about how big that moon is (looks small from far away, don't be fooled) let's talk about the Theory of Evolution: "Survival of the Fittest," is what they tell me, but who are the fittest, exactly? Why the fittest are the ones who survive. Wait! Haven't we just gone around to where we started from? Hmmm, yeah looks like biologists have a theory with excellent descriptive power, but not a whole lot of predictive power. Tautology rules, OK? No wonder there's such a huge argument about whether we can teach this self affirming theory in schools.

Let's try that "Efficient Market Hypothesis" where you can't beat the market, but if you ever do happen to beat the market, that's taken as an example of why no one else can beat the market. At least in this instance they explicitly disclaim any capability of prediction... because all prediction is done by the market already. Pretty good huh?

A recent classic that you might have heard of is the "Dunning–Kruger effect" which states that unskilled individuals believe they are right when they are really wrong. So how do we know who are the "skilled individuals" and who are the "unskilled individuals"? Easy, first figure out which one is right and which is wrong, and whoever got it right must have been skilled... pretty obvious really. Excellent descriptive theory... ahhhh, not so good for prediction, when you think about it.

Let's try a thought experiment: I get into an argument over some Austrian vs Keynesian thing and the other guy says, "Stupid Austrian! You guys get everything wrong, it's the Dunning–Kruger effect in action" to which I would no doubt reply, "Stupid Keynesian! You guys get everything wrong, it's the Dunning–Kruger effect in action." Eventually one side or the other will be proven correct, and guess what? Either way it will be the Dunning–Kruger effect in action. Brilliant!

[Avon Barksdale]

nivedita,

You have a misunderstanding of rational choice and what it means for economics. You say that, "I drink more coffee and eat more food than is healthy for me, and I know that", that statement makes you rational. You know that your behaviour is unhealthy, but you do it anyway because the pleasure you receive from doing it is high enough. That means I could probably predict your behaviour given a set of choices presented to you. It allows me to construct a model of how you discount future payoffs, etc. I would almost certainly be able to build a rational choice model around your consumption of coffee and food. What would be irrational is if you hated coffee, disliked its taste and all of its effects and when presented with alternatives, you kept picking coffee by mistake, showing no ability to learn that you were making such a poor choice. I bet you're rational.

The two laws that I pointed out are deeply useful, and almost universally applicable.

And as far as stock markets a derivative markets are concerned - there is nothing forced there.

[Avon Barksdale]

rsj,

Indifference curves describe human behaviour very well. It matters not that people don't work them out - they behave as though they do.

Cheetahs know nothing of the theory of differential games, or vector calculus. When they chase gazelles on the savannah, they behave as they they do. I can predict the types of paths that a cheetah will use based on vector calculus and differential game theory even though cheetahs have no idea what that stuff is. Humans are just animals like cheetahs.

[Avon Barksdale]

rsj,

Indifference curves describe human behaviour very well. It matters not that people don't work them out - they behave as though they do.

Cheetahs know nothing of the theory of differential games, or vector calculus. When they chase gazelles on the savannah, they behave as they they do. I can predict the types of paths that a cheetah will use based on vector calculus and differential game theory even though cheetahs have no idea what that stuff is. Humans are just animals like cheetahs.

[Nick Rowe]

One of the things that ECON 1000 students learn is that some things cause both demand and supply curves to shift. Immigration increases the supply of hairdressers, but also increases the demand for haircuts. A late Spring frost reduces the supply of Fall apples, but probably doesn't affect the demand for Fall apples. Immigration of skilled hairdressers only, who supply more haircuts than they demand, is more like the late Spring frost.

[Tom Brown]

rsj, it's true that a human brain is beyond our current capabilities to model with high fidelity, but consider that this may not be necessary to model some aspects of a huge group of humans. As an analogy, it's not necessary to accurately model the internal structure of each sub-atomic particle making up the constituent parts of a particular type of gas molecule to derive the ideal gas law. Very few micro properties of individual gas molecules carry over into the aggregate. Likewise the aggregate has emergent properties that the individual gas molecules don't have: for example pressure.

Even if we had perfect models of individual human brains, perhaps very few of the properties of those individual agents would carry over to the aggregation, and likewise, perhaps the aggregation has properties not present in individual agents.

This physics paper discusses a means of analyzing some types of aggregate properties w/o necessarily having to have models of the individual constituent parts.

[rsj]

Tom.

Exactly, which is why it's wrong to insist that macro economics be micro-founded in human behavior. For some reason, the economics profession thinks it's OK to have fables as the micro foundation rather than not having any micro foundations at all, or using simple rules of thumb that may admittedly need to be constantly re-written.

That makes no sense to me, given that the current micro stuff is just wrong in so many cases -- e.g. supply curves slope down, not up, for almost all products (it's cheaper to buy more than to buy less, and it's cheaper to produce more, per unit, than to produce less), the quantity transacted is never going to satisfy a tangency condition, optimization is extremely costly and difficult to do; design, marketing, logistics, and warehousing costs account for more than 80% of the total cost of most tradeables, etc.

If you can find general rules that work for the whole economy, then they wont be micro-founded, if you have something that is micro-founded, it wont be a good basis for a macro model, since the micro stuff is too hard, and just BS-ing your way through it with some Grand Contemplative Theories doesn't cut it in a modern era where we are starting to learn something about our brain as well as about what firms and consumers really do.

[Tom Brown]

rsj, I think we're mostly on the same page, although I'd argue we might be able to skip the neuroscience and examine aggregates directly. That reference I posted has some concrete examples from physics: showing how the law of gravitation and different forms of the ideal gas law can be obtained principles the lay out (i.e. as entropic forces). Here's a follow up paper inspired by the above paper which uses the same basic framework to develop supply and demand curves (1st 8 pages) and a general equilibrium solution (with no utility maximization, rational expectations, representative agents or brain or behavioral science required!). Basic calculus is all you need to understand it.

Some related concepts in econ have been examined in the past, notably by economist Gary Becker in his 1962 paper “Irrational Behavior and Economic Theory" published in JPE and reprinted in Becker’s The Economic Approach to Human Behavior. Quoting economist David Glaser on Becker's paper "...he showed that budget constraints were sufficient to imply negatively sloped demand curves and other standard microeconomic results. He credited Alchian’s 1950 paper in JPE “Uncertainty, Evolution, and Economic Theory” for anticipating his argument. Becker speaks of "irrational agents" but the same results from simply taking agents' behavior as so complex that it appears to be random. Effectively, we can think of there being an emergent representative agent who's behavior is not determined by micro foundations.

[Avon Barksdale]

rsj,

You have absolutely no idea what you're talking about and your comments make little sense. Nick, start policing this blog better by blocking inane comments or you'll end up with an ocean of nut jobs.

[Oliver]

...Much of the critical discussion of methodological individualism in the philosophy of social science concerns the relationship between what Watkins called “rock-bottom” explanations and “half-way” ones – or those that do and those that do not specify an action theoretic mechanism. In general, there is no question that, given any particular half-way explanation of a social phenomenon, it would always be nice to know what agents are thinking, when they perform the actions that are involved in the production of that phenomenon. The question is whether the explanation is somehow deficient, or unscientific, in the absence of this information. The answer to that question will depend upon one's broader commitments concerning the status and role of the social sciences. Nevertheless, it is worth noting two very common types of social-scientific inquiry that fall short of providing the sort of rock-bottom explanations that methodological individualism demands:

6.1 Statistical analysis

Consider the following example of a social-scientific debate: During the 1990s, there was a precipitous decline in violent crime in the United States. Many social scientists naturally began to apply themselves to the question of why this had occurred, i.e., they set out to explain the phenomenon. A number of different hypotheses were advanced: the hiring of more police, changes in community policing practices, more severe sentencing guidelines for offenders, decreased tolerance for minor infractions, an increase in religiosity, a decline in the popularity of crack, changes in the demographic profile of the population, etc. Since the decline in crime occurred in many different jurisdictions, each using some different combination of strategies under different circumstances, it is possible to build support for different hypotheses through purely statistical analysis. For example, the idea that policing strategies play an important role is contradicted by the fact that New York City and San Francisco adopted very different approaches to policing, and yet experienced a similar decline in the crime rate. Thus a very sophisticated debate broke out, with different social scientists producing different data sets, and crunching the numbers in different ways, in support of their rival hypotheses.

This debate, like almost every debate in criminology, lacks microfoundations. It would certainly be nice to know what is going through people's mind when they commit crimes, and thus how likely various measures are to change their behavior, but the fact is we do not know. Indeed, there is considerable skepticism among criminologists that a “general theory” of crime is possible. Nevertheless, we can easily imagine criminologists deciding that one particular factor, such as a demographic shift in the population (i.e., fewer young men), is the explanation for the late-20th century decline in violent crime in the United States, and ruling out the other hypotheses. And even though this may be a “half-way” explanation, there is no question that it would represent a genuine discovery, one that we could learn something important from.

Furthermore, it is not obvious that the “rock-bottom” explanation – the one that satisfies the precepts of methodological individualism – is going to add anything very interesting to the “half-way” explanation provided by the statistical analysis. In many cases it will even be derived from it. Suppose that we discovered, through statistical analysis, that the crime rate varied as a function of the severity of punishment multiplied by the probability of apprehension. We would then infer from this that criminals were rational utility-maximizers. On the other hand, if studies showed that crime rates were completely unaffected by changes in the severity of punishments or the probability of apprehension, we would infer that something else must be going on at the action-theoretic level.

Results at the action-theoretic level might also prove to be random or uninteresting, from the standpoint of the explanatory variables. Suppose it turns out that the decline in crime can be explained entirely by demographic change. Then it doesn't really matter what the criminals were thinking – what matters is simply that a certain percentage of any given demographic group has the thoughts that lead to criminal behavior, so fewer of those people translates into less crime. The motives remain inside the “black box” – and while it might to nice to know what those motives are, they may not contribute anything to this particular explanation. In the end, it may turn out that each crime is as unique as the criminal. So while there is a concrete explanation in terms of actual people's intentional states, there is nothing that can be said at the level of a general “model” of rational action. (In this context, it is important to remember that methodological individualism in the Weberian sense explains actions in terms of a model of the agent, not the actual motivations of the real people.)
...

[Oliver]

Copied from:

http://plato.stanford.edu/entries/methodological-individualism

[Oliver]

6.4 Fallacies

...it is worth noting that too much emphasis on the action-theoretic perspective can generate its own fallacies...

Avon and others seem oblivious to this.

[Nick Rowe]

rsj: Can you answer this ECON 1000 question, from my post: "What is the relation between the slope of the PPF and the relative price of apples and bananas? Why is the PPF curved, and what does it mean? (And unless you can answer that question, you do not understand why supply curves slope up.)"

Hint: suppose the country stretches North-South, for one simple model.

For a slightly more complicated model, assume growing apples or bananas requires both land and labour. All land and all labour is the same (to keep it simple), but apples and bananas are not equally labour-intensive.

By the way, and this question is for all of you commenting here: have you ever taken Intro Economics, or read an Intro economics textbook?

[Nick Rowe]

Here's an example. All you commenters here have both a comparative and absolute advantage over me in doing math, so you can solve it for me:

The production function for apples is: A = L^a.N^(1-a) where A is quantity of apples, L is labour and N is land, and 0 < a < 1.

Same for bananas, just replace A with B and a with b.

Solve for the PPF, for a given total L and N, then solve for the slope of the PPF as a function of A and B. That's your supply curve. Under what conditions will it slope up?

[Nick Rowe]

And nobody gets to make any more comments here until you have solved that simple problem (unless it's to ask for help in solving it).

Because Avon does have a bit of a point (though rsj is smart, and on the whole does make some good comments). And I am getting really pissed off at arrogant bloody engineers (or whatever the hell you are) shooting off about economics when you don't understand the basic stuff.

[Nick Rowe]

And a second question, for rsj: Assume increasing returns to scale. For example assume Q = L - F where L is labour input and F is some fixed cost. Assume the firm maximises profit. What is the slope of the supply curve? (Yes, it's a trick question, and the real question is to tell me why it's a trick question).

[Oliver]

have you ever taken Intro Economics

To the extent that that was directed at me (among others): I have, but I admit it's been a long time and I never went on to use it in any meaningful sense. Dangerous half-knowledge, as the Germans say, which I try to amend on occasion by visiting here and other places in Webistan.

My quote rant was in reaction to Mr. Barksdale's comment on rationality, his comparison between humans and cheetahs and his subsequent off-the-mark comment vs. rsj, not to your post. Feel free to delete all of them, though.

And to attempt to answer your question Why is the PPF curved, and what does it mean?: a straight line reflects a case in which two goods can be substituted for one another at no extra cost. A convex curve reflects increasing opportunity costs whereas a concave curve could be used to show economies of scale (which rsj seems to be denying?). And as for the follow up question: given a fixed supply of land, an increase in the labour supply would tend to shift production towards the more labour intense fruit production process?

Does that get me banned from comments?

Architect, not engineer, btw. (as in engineer minus the math). And I have an in-house migration specialist to pull me back to the left, should I stray.

[Nick Rowe]

Oliver: "A convex curve reflects increasing [marginal] opportunity costs" [You missed a word, for clarity]

But WHY (under what conditions) would it be convex, as opposed to a straight line?

"And as for the follow up question: given a fixed supply of land, an increase in the labour supply would tend to shift production towards the more labour intense fruit production process?"

That depends on preferences. But that was not the question. Hold total labour and land fixed. Is the PPF convex, concave, or straight? (how does the slope vary as we move along the PPF?

[Avon Barksdale]

Oliver,

Please read economics before commenting about how it works.

Crime and enforcement has a feedback problem – increased crime causes more enforcement but more enforcement causes less crime. Because criminals are RATIONAL, they change their behaviour in the face of more or better enforcement, you can't simply regress one variable on the other. That's the micro foundation of the problem. This is a classic problem in econometrics – simultaneous equation bias. There are ways to handle this problem. The very fact that you have this problem is because both sides of the equation act rationally. You need to understand the micro foundations of economics to make sense of this data.

If you honestly care about understanding phenomena, economic or otherwise, get busy reading. You certainly cannot understand how quantum mechanics works through some journey of introspection and you can't understand economics that way either. Read. Calculate. Then read more.

[Nick Rowe]

Avon: sorry, but you gotta solve my problem first. Should take you only a couple of minutes.

And BTW, have you taken Intro? (I suspect yes, but confirm for me.)

[Oliver]

At the risk of making an even greater fool of myself than I already have:

The PPF is convex if the supply curve of the products slopes upward / the marginal cost of each product increases.

That would generally be true when looking at manual ouput of say individuals.

That might not be true when looking at say firms that can put economies of scale to use.

So, taking apples and bananas, if both are equal, the PPF will be concave and the efficient output will be 1/2 each. If apples are more labour intensive to produce and capital (land) is given, then the PPF will also be concave but shaped such that maximum efficiency is biased towards the production of bananas?

[Nick Rowe]

Oliver: totally wrong. Horribly wrong.

Your second sentence should be reversed: It is the curvature of the PPF that determines the slope of the supply curves of apples (and bananas), not vice versa. You can derive the PPF from technology and resources alone. Supply curves (if they exist) come later, when we add profit-maximising producers to the model.

"So, taking apples and bananas, if both are equal, the PPF will be concave and the efficient output will be 1/2 each."

God no.

If A=B, that defines a 45 degree line ray from the origin, which picks a point on the PPF. It tells us nothing about the slope of the PPF at that point, and even less about how that slope changes as we move along the PPF (so A =/= B).

And there is nothing "efficient" about producing where A=B. Suppose nobody likes eating bananas?

[Nick Rowe]

Let me make it really easy for you:

Country 1 is an island, with 100 acres of identical land. Each acre can produce either a apples or b bananas. No labour needed. What is the slope and shape of the PPF?

Country 2 is a different island, that is very long and thin on a North/South axis, so land is not identical. And (a/b) steadily increases from South to North. What is the shape of the PPF?

[Oliver]

@ Nick

:-) Fair enough. I'll give it more thought.

@ Avon
Criminals are RATIONAL? You sound like an architect lecturing the hoi polloi about how modern architecture is right. I know what you mean, the apparent contradictions can be addressed within the framework, but you still have a whole body of philosophy and social science telling you that it's only one way to look at things and that it has well known short comings. There are reasons for the existence of the other social sciences and philosophy. In reading your comments, one wouldn't know. And all rsj was saying to my mind, is that first year students should know.

[Nick Rowe]

Oliver: stop arguing with Avon, until you have answered my very simple 11.14 questions.

[Oliver]

ok, sorry only just saw your last question. the ppf is more convex on the long island because the more apples are substituted for banans, the more banans will have to be grown on land actually more suitable for growing apples.

[Nick Rowe]

Oliver: correct as far as it goes, but not very precise. What is the shape and slope of the PPF on the first island?

[Tom Brown]

"Country 1 is an island, with 100 acres of identical land. Each acre can produce either a apples or b bananas. No labour needed. What is the slope and shape of the PPF?"

Are you saying that if N = acres of land used for apples

0 <= N <= 100
A = N*a
B = (100-N)*b
?

Then the slope of the PPF is -b/a (if A is on the x-axis and B is on the y-axis). On those same axes, the PFF is a straight line: B = (100 - A/a)*b

How'd I do?

[Nick Rowe]

Tom: full marks on that (easy) question.

I prefer to write it as: 100 = A/a + B/b, but same thing.

Now, assume each acre is owned by one profit-maximising farmer (100 farmers in all). Let P be the price of apples in terms of bananas (so P=2 means 2 bananas trade for 1 apple). With P on the vertical axis, and A on the horizontal, what will the supply curve of apples look like?

[Oliver]

The shape for the short island is a straight 45° line.

[Tom Brown]

OK, thanks... but before I get to that, for Country 2 it appears that
dB/dA ~ A^(-1/2)
The PPF is a concave curve: B ~ b*(100 - sqrt(2*A/a_max))
You can replace "~" with "=" when a=0 in the extreme South, and a=a_max in the extreme north, linearly varying inbetween and b constant

[Nick Rowe]

Oliver: nope.

Tom: for country 2 (the long thin North-South island) you can't give me a mathematical answer, because I only told you that a/b is an increasing function of latitude. I didn't tell you what that function was. But you are right that the PPF is bowed out. Because the slope (with B on the vertical axis) is -b/a, which gets flatter (in absolute value) as you grow more and more bananas, pushing the margin of cultivation between apples and bananas further and further North.

[Tom Brown]

Nick, you're right: I can't give you a formula. I made assumptions consistent w/ the problem, and did that problem instead. I also assumed a single contiguous apple growing area in the North, with the remainder to the South used for bananas.

[Tom Brown]

Nick, what's your favorite 1st year syllabus look like? (books and/or papers?)

[Tom Brown]

Nick, would these be at all appropriate as part of a syllabus (say for any of the 1st four years)?:
Gary Becker's 1962 paper "Irrational Behavior and Economic Theory" and Irving Fisher's 1892 thesis "Mathematical Investigations in the Theory of Value and Prices?"

[Nick Rowe]

Tom: "I also assumed a single contiguous apple growing area in the North, with the remainder to the South used for bananas."

That is not an *assumption*; it is a *conclusion*. It follows from the premise that a/b is an increasing function of latitude. If you grow some apples south of some bananas, you are *inside* the PPF, not *on* it. You could grow more of both fruits by rearranging what grows where.

No. Just read a first year text. Mankiw, Krugman, whatever.

[Tom Brown]

"That is not an *assumption*; it is a *conclusion*."... Good point! I knew there must have been a good reason I concluded that. ;^)

[Nick Rowe]

Tom: you are getting it.

Now solve for the apple (or banana) supply curve for country 1. (It might help you think about it more clearly if you put P on the horizontal axis and A on the vertical, because economists draw it the wrong way round, for obscure historical reasons). We want A as a function of P.

Hint: for each value of P, figure out what fruit would be most profitable for a farmer to grow.

[Avon Barksdale]

Nick,

To start we should not write thing like 100 = A/a + B/b. Your apples and bananas form a tuple, you can't add them.

Here's the answer. Find the marginal rates associated with each input. The negative of the ratios give the marginal rate of technical substitution and RTS_A = RTS_B. (You can derive that condition from the full Lagrange multiplier set up or argue from market clearing). To simplify the notation here's the problem:

u= x^ay^(1-a); v = (A-x)^b(B-y)^(1-b)

u are the apples, v are the bananas, A is the total amount of the first input and B is the total of the second. RTS_u = RTS_v:

u_x/u_y = v_x/v_y

where the subscript denotes partial differentiation. A bit of high school algebra yields:

y = Bb(1-a) x/[Aa(1-b) – (a-b) x]

which gives the PPF:

(u,v) = ( [Bb(1-a)/(Aa(1-b) – (a-b)x)]^(1-a) x, (A-x)^b[B(1- b(1-a)x/(Aa(1-b) -(a-b)x))]^(1-b) )

Now that you have u and v as functions of x, simple application of the chain rule gives du/dv. I'll leave that one for you!

To answer your second question – I have never studied economics at school. But I have carefully worked through Mas-Colell Whinston and Green's Micro Theory, Ljungqvist and Sargent's Recursive Macro, Hamilton's Time Series Analysis, Greene's Econometric Analysis Cochrane's Asset Pricing, Romer's Advanced Macro, and Osborne and Rubinstein's Game Theory text. And that's just the my core econ – I've read far more math finance, which is my actual speciality. I have so much fun with this stuff, I take it on vacation with me, seriously. My wife has to pry it out of my hands.

[Oliver]

I see Tom has answered correctly. I'm afraid I can't do the math. But to see whether I at least understood correctly in hindsight: The slope is determined by the relative price of one good in terms of the other. It is straight on the short island because marginal cost of each good in terms of the other remains equal, irrespective of the mix of goods? But it cannot be said to be 45° because you didn't specify how apples exchange for bananas and thus can only be presented as a relation?

In any case, thanks for the free lesson. I doubt I'll ever reach any level of proficiency in anything involving even the slightest bit of math. At the very least I'll have to save that task for when the kids grow older and I get to sleep & read again. And god knows what I learnt back in business school. I only vaguely remember the macro intro class (Volkswirtschaft) and slighty more clearly law and accounting. But none of the above rung the faintest bell. So in case it wasn't apparent, I just threw everything together ad hoc by skimming through the internet while pretending to work. I hope that my public display of ignorance is repentance enough for what I regret came across as arrogance. That was not intended.

But I still remain curious whether you, Nick or Frances, consider Avon's statement 'criminals are rational' to be A)universally true or, if not, at least B) indispensable for practicing economics? Which I realise is quite separate from asking whether such an assumption can be useful or not - something I am certainly not qualified to opine about.

[Nick Rowe]

Full marks to Avon (I think).

But what I wanted to prove is that if a=b the PPF is a straight line, and if a=/=b the PPF is bowed out. I should be able to prove that by differentiating the PPF twice. (Note the "should".) I was hoping this would be simpler, so I could do a very simple intuitive post on this.

And that is important because a=b is very unlikely to be true, and if a =/= b, so the PPF is bowed out, the supply curve for apples (and bananas) will slope up. Which is something I wanted to explain to rsj.

"To start we should not write thing like 100 = A/a + B/b. Your apples and bananas form a tuple, you can't add them."

I don't understand you there. A/a is amount of land needed to grow apples, and B/b is amount of land needed to grow bananas. We can add them. It's just the resource constraint.

Well done on the reading. And I know that intro texts are written in babytalk. But I still think you would gain by holding your nose and reading/skimming one. There's a breadth/depth trade off. I would be very surprised if you found nothing interesting.

[Nick Rowe]

Oliver: I give you credit for trying. Let me make some small changes to correct what you said:

" But to see whether I at least understood correctly in hindsight: The slope [of the PPF] is determined by the [marginal opportunity cost] of one good in terms of the other. It is straight on the short island because marginal [opportunity] cost of each good in terms of the other remains [constant], irrespective of the mix of goods? But it cannot be said to be 45° because [it would only be -45 if a=b]"

Yep.

The slope is -a/b (or -b/a, depending what's on which axis). Which is a constant in country 1.

And profit maximising farmers will choose to grow apples or bananas according to whether Pa/Pb (the relative market price of apples in terms of bananas is > or < (b/a). So the supply curve of apples will be perfectly elastic (horizontal, with Pa/Pb on the vertical axis and A on the horizontal) at a height of b/a. Ditto for the supply curve of bananas, except its at a/b. And this means that in competitive equilibrium (as long as people consume some of both goods) Pa/Pb = b/a.

For country 2, the slope of the PPF is still -a/b, but a/b is not a constant. As you grow more and more bananas, pushing the margin of cultivation north, a/b gets bigger, so the marginal opportunity cost of a banana increases. (Ditto for apples.) Which is why the PPF is bowed out, and why supply curves slope up, except in the very special case where all land is identical.

"Criminals are rational" is useful. We can always posit a utility function that makes it universally true, but then it may no longer be useful. Why did he do X? Because he likes doing X. Where X is anything whatsoever. Why did he do X yesterday, but Y today? Because he liked doing X yesterday, and likes doing Y today. We need some sort of stability of preferences across people and across time, if it's going to be useful. And that's even before we start talking about beliefs.

Rationality is not strictly essential for practicing economics. But economics would be very different without it, and we would need to replace it with something else.

[Avon Barksdale]

Nick,

For an intro text, I read Hal Varian's Intermediate Micro. That's a nice book, I like it very much, but it shows the difference between the expectations of physics students and economics students at the undergrad level. Baby Varian is often used as a second year text but most of Mas-Colell Whinston and Green is at the level of upper second year physics. What I would like to see is baby Varian used in first year (but calculus based) and Mas-Colell Whinston and Green started in second year. That is how a physics curriculum would treat micro theory. In fairness, the top of the econ field catches up completely by the end of grad school and the top of econ is as every bit as technical as physics. I have deep respect for it. The problem is that so many people want to take econ at the undergrad level (the joys of subsidized post secondary education), there is no way the discipline can be treated seriously until grad school. The graduating class in a typical physics department from 4th year is less than 10 at Carleton (maybe less than 5). Most go to grad school. I bet econ majors out numbered the physicists by 20 to 1 at least, and almost none go to grad school. They wouldn't have even gotten through second year if held to the same standard as physics. Take a walk over to the physics department, Nick, you'll see what I'm talking about.

You need to be careful what adding variables with different units. What's 3 kilogram + 2 metres? Apples and bananas carry different units. When you see budget constraints on a graph with quantities, they have implicitly converted units, usually to prices or made the system unitless by some other conversion. Treat the problem as a tuple and then solve.

[Hugo André]

I've been trying to keep away from this comment section because I have a lot of work to do but I just can't resist.

@Oliver

A big part of the problem is that being 'rational' means something very different in economics jargon. Perhaps you know that already but if so it doesn't seem like the point has sunk in.

Take an example: A man goes to a club with some acquaintances. These are not close friends of his so he may never meet them again and they have no influence over his career. He drinks "far too much" which makes him throw up on the way home. The next day he is stuck with a pounding headache which makes it hard for him to focus at work and he makes several mistakes. Was this person being rational when he drank too much?

Regular person: No of course not! Drinking gave him only a little pleasure while throwing up and the headache harmed him greatly. His overall well-being was clearly reduced as a result and therefore it was not rational of him to drink as much as he did.

Economist: Yes. His actions that evening show that he did not at the time care much about how he would feel the next day or even later that evening (he heavily discounted the future). The man may also have believed that the cost (having to think about something he doesn't want to) of considering the consequences of his actions was larger than the benefit. It would then be a case of imperfect information. In sum: given his constraints and the information he had available the man was rational in his drinking behaviour.

With this kind of reasoning criminals are always "rational". I honestly think it would be better if we made up some jargon word and used that instead since there are times when even economists mix up economics rationality with common sense rationality (and plenty of times when they give the wrong impression to journalists and the general public because they use this word).

[rsj]

Nick,

Are you doing macro or micro? Because I thought this was micro. It costs $100 million to develop a drug and then 1 penny to make each pill. The drug company is willing to sell 1 pill for $100 million + 1 penny. 2 pills for $50 million plus 2 pennies. Etc. As the quantity increases, the price per pill that the company is willing to supply those pills at decreases. What is the shape of the supply curve?

If many people travel between two points, an airline can use a bigger plane, which means that it is cheaper, per passenger, to transport those passengers. What would the supply curve be for the airline?

A word processing program costs $100 million to make. What is the shape of the supply curve for the company that sells this program?

Reading over Adam Smith's remarks about division of labor, what do you think the supply curve is of his Pin factory?

The fact that this doesn't aggregate nicely into a "you have one island that you spend half the territory growing apples and half bananas" is not my problem. It just doesn't aggregate nicely and firms need to find some way to survive, or in the case of airlines, of going bankrupt gracefully.

But given that this is the most common state of affairs, shouldn't we be teaching it?

[Nick Rowe]

rsj: "Are you doing macro or micro?"

Good question. In this case I am doing General Equilibrium theory, which is usually said to be part of micro, but is about the economy as a whole. Yep, micro supply curves are a GE phenomenon.

BTW, let's switch this PPF/supply curve stuff to my new post. See the end bit of my new post, on Increasing Returns to Scale. (Though I'm not happy with that bit yet.)

[notsneaky]

" I honestly think it would be better if we made up some jargon word and used that instead since there are times when even economists mix up economics rationality with common sense rationality"

Consistency (since there is that implicit assumption that preferences don't change from one situation to the next buried in there). Or (I think it was David Colander who suggested it) "purposefulness"

[Tom Brown]

For the record: I should have written "convex" and my math as a bit off, even for the problem I did solve. For that problem:
B = b*10*sqrt(100 - 2*A/a_max), and dB/dA = (-b/a_max)/sqrt(1 - 2*A/(a_max*100)). I think. I'm sure nobody cares.

[Nick Rowe]

I care Tom.

[Avon Barksdale]

rsj,

Everything you've talked about is covered in a basic introduction to micro. Seriously, read something like Hal Varian's Intermediate Micro. Work out the details for yourself. Your basic misunderstandings is like saying physics has all wrong - heavy things fall faster than light things in gravity fields. It's clear, just look at it - a feather falls slower than a bowling ball so all this physics nonsense that say everything falls with the same acceleration is just silly! Seriously, you look that foolish.

[rsj]

Nick,

Good question. In this case I am doing General Equilibrium theory, which is usually said to be part of micro, but is about the economy as a whole. Yep, micro supply curves are a GE phenomenon.

OK, but you are not going to get a Law of Supply from GE. You wont even get a law of Demand. GE gives you very little beyond existence. The fact that (I hope) we all agree that, at a minimum, MC firms are the norm + IRS and large fixed costs are also the norm means that supply curves, as commonly introduced to first year students, should be shown to go down rather than up. You can say, "Well these firms -- i.e. pretty much all firms -- don't actually have supply curves", in which case why teach the scissors as the basis of what a first year will remember? They are not going to remember the SMD, they will remember the scissors, and they will vote accordingly, creating a lot of misinformed, yet smug voters who think they know what's going on because they read a first year text.

I'll take a look at your next post, at which point I will needle you about revolving capital.

[Too Much Fed]

“By the way, and this question is for all of you commenting here: have you ever taken Intro Economics, or read an Intro economics textbook?”

Nick, do you think I took an Intro Economics course?

What grades do you think I get/got in school?

“And a second question, for rsj: Assume increasing returns to scale. For example assume Q = L - F where L is labour input and F is some fixed cost. Assume the firm maximises profit. What is the slope of the supply curve? (Yes, it's a trick question, and the real question is to tell me why it's a trick question).”

I am going to try this one. I am probably missing the point on this one. I might be making a different point. With fixed costs, is it possible that the supply curve slopes both up and down?

[john]

hello Nick, I have a question (off-topic)

I've read you before referring to Canada as a 'small' open economy and I was wondering what is the metric you use?

I (think I) know from my international macro that to be 'big' in that context, you need to be able to influence the international interest rate or, what seems to be equivalent, to be able to exert an influence on global savings/investment decisions

Is Canada 'small' in that sense?

thank you

[Nick Rowe]

john: well, we normally reckon that Canada is about one tenth the size of the US, which is about one quarter the world economy. So if canadian saving or investment increased by 100% (which is very big), that would increase world saving or investment by around 2.5%. Yep, small enough to ignore, given all the other things we ignore when we build a model.