What's Wrong With How We Teach the Supply and Demand Model?

Question

The standard undergraduate narrative argues that prices fall in the face of excess supply and rise in the face of excess demand. This argument can be traced back at least to Smith. I do not have problems with this intuition but rather with the claim that this a property of the supply and demand model that we teach to undergraduates.

The supply and demand model is a static model; it is always in equilibrium, because it is closed with an equilibrium condition. Further, the model is supposed to represent a perfectly competitive market and so price adjustment by firms and households is precluded by assumption.

Should we not inform our students, at least at the intermediate level and especially economics majors,  that we have to assume that prices adjust as hypothesized? (This is similar to the irritating fact that we can prove that demand curves do not necessarily have negative slopes and so we are forced to assume that they do.)

There are many other problems with the supply and demand model that seem to go unremarked in our teaching. Is this an intentional lapse or an inadvertent one?

Comments

I see this is an old thread, but I'm hoping to revive it a bit. First, thank you Allan for this post -- I found it by Googling something along the lines of "supply and demand is a lie," which really speaks to my inner turmoil here. I'm a beginning econ student (returning to school in mid-life as a stay-at-home mom, actually), and when I'm up late at night with the baby, I can't stop thinking about what's fishy about supply and demand. Normal, right? :) Here goes...

When we graph a supply curve, we start with the assumption that the quantity supplied is a response to a change in price; a given quantity is what a supplier will provide when they can fetch a given price for it. But then something strange happens when we start talking about surpluses and shortages that lead to convergence on the equilibrium price: all of a sudden, the suppliers are changing their prices as a response to quantity. We've reversed the independent and dependent variables, and we've violated the underlying premise of the supply curve, which is that the price on the supply curve is a reflection of how much is supplied as long as the suppliers get that price. When we talk about surpluses and unsold product, the supplier is no longer fetching the price on the y-axis, and that very same supply curve is no longer valid because we've changed the conditions of the universe that the curve was created in. Same for demand: we're initially graphing how quantity demanded depends on price assuming that any given quantity is available to meet that demand, but then we change the rules by superimposing a supply curve that creates a shortage, which invalidates the conditions under which the demand curve was created. It doesn't make sense to chart supply and demand curves on top of one another; they were drawn in different universes.

Supply alone and demand alone make sense, but using them together to find an equilibrium price--or to describe what happens out of equilibrium--doesn't make sense to me at all. While the S/D graph is a useful visualization tool to demonstrate S/D relationships and surpluses and shortages conceptually, it seems to me that it's mathematically completely wrong. It seems that there's no magical 'equilibrium' price where the S/D curves cross because that model depends on switching causality of the variables mid-game, and you can't have it both ways; the 'equilibrium' price has to be any point on the supply curve or demand curve because those are the assumptions of the creation of the curves.

I don't have a problem with the intuition behind the model, but I do take issue with what we're being told the model can be used for. I agree with Eric that a better way to present the information is like a "story" between buyers and sellers. But that story is getting turned into bad math ("bad" in my current understanding, anyway), and then all sorts of other insights are gleaned from that questionable foundation. So in moving forward with my indoctrination into the world of economic thought, I want to know exactly what's in the kool-aid I'm swallowing.

There's the nature of my inner turmoil in a nutshell.

I don't have a feel for how economists use the supply/demand model in the real world. How literal is this model? Again, I'd have no issue if it were just a conceptual diagram, but it is presented in texts and classes as a graph, as an actual mathematical phenomenon. Do economists actually use this as a mathematical model to analyze prices and markets and predict 'equilibrium' price?

And if S/D curves are used in the real world, how? Are S/D curves ever actually observable? Since supply and demand are constantly interacting to create market prices, one wouldn't be able to just look at price and quantity because you wouldn't be able to disentangle the effects from the supply side vs demand side, right? You'd just have a curve of what the 'equilibrium' prices ended up being from the combined effects of S&D?

And the shape of those curves... Couldn't a supply or demand curve take on many different shapes depending on the nature of the market? It seems to me that for a supplier, there's a window of profitability within which supply should match the demand curve as closely as possible (and outside of which, supply is zero). Why aren't the supply and demand curves more closely matched? Isn't it also important to consider where prices are coming from? Depending on how prices are set (whether the supplier is setting the price or responding to it), wouldn't the curves look very different?

If anyone can point me in the right direction for more resources on this, I'd be very grateful.  I'm sure that if I'm asking these questions, they've been asked before, but I can't find much about it except this thread, and my attempt to discuss these issues with my professor failed. What am I missing here? Is there some fundamental issue that I'm just not understanding properly? If you've read this far in my diatribe, thank you. And thanks in advance to any respondents!

---

Hi Erica,

I can understand much of your confusion. If you will send me an email address I will send you some papers that deal with the various issues you raise and connect you with my Sleemanomics site, which contains the materials I use to teach S&D.

I hope that you continue to be interested in economics. David Hilbert once said that physics was too important to be left to physicists; a similar argument could be made about economics.

Thank you for your interest

Allan

allan.sleeman@icloud.com

---

Hi again Erica,

I have looked at your comment again and I think that you are confused about some aspects of S&D although your instincts are excellent.

The d and s curves are graphs of the d and s functions. Those functions map prices to quantities. For each price there is a corresponding q ded or sed. But the model has three components. The two functions and an equilibrium condition that says q ded is equal to q sed. The model is a <<static>> model and is therefore <<always >> in equilibrium. This is what my most important S&D paper emphasizes, the one that no-one wants to read, although they read my much less interesting paper on S&D and general equilibrium. Early drafts can be found on my <<ResearchGate>> page but I’ll send you the latest one if you contact me.

So all textbooks that I have consulted, including intermediate and advanced ones, make an error when they discuss disequilibrium, because the model has nothing to say about disequilibrium. Where are the equations that explain how constrained satisfaction consumers react to disequilibrium prices? Where are the equations that explain how constrained profit maximizing firms react to disequilibrium prices?

The standard narrative says that if at a price there is excess supply then firms will lower prices until equilibrium is achieved, and if there is excess demand at a price then households will bid up prices to equilibrium. But the S&S model is a model of a perfectly competitive market where, by assumption, consumers and firms are too small to effect prices by changing the qs they will offer to buy and sell. Therefore all the economist can legitimately say is that we <<assume>> that prices fall if there is excess supply and rise if there is excess demand.

One very important thing for you to understand, that receives almost no emphasis in standard texts and courses is the the demand curve shows for each price how much the consumer would be willing to consume given that the consumer has the necessary ability to pay (ATP) - has sufficient income, wealth, or access to credit. This means that I have no demand for $12.6m Bugatti sports cars (although I really want one!), the homeless have no demand for accommodation because they lack the necessary ATP, children living in poverty (about 25% of US children) have no demand for dental care and many of the things that your child will have easy access to.

I have a rather rough paper on whether S&D models have any empirical validity, also on <<ResearchGate>>  but I will send it to you if you like.

I hope that you continue with your economics. Economics seems to be unfriendly to women although we are trying to make amends. If you go on to do what is laughingly referred to as “serious” economics you will need to be happy with basic mathematics.

Enjoy your baby.

Allan Sleeman

---

Hi Allan,
Thanks so much for your response -- you've given me quite a bit to mull over here. Much of my confusion does stem from trying to use the model to explain disequilibrium. I'd love to read your paper on this topic (which means you can no longer claim that no one wants to read it!).

This comment in particular helps to clarify something for me:
"the S&S model is a model of a perfectly competitive market where, by assumption, consumers and firms are too small to effect prices by changing the qs they will offer to buy and sell. Therefore all the economist can legitimately say is that we <<assume>> that prices fall if there is excess supply and rise if there is excess demand."
Thinking of this as an assumption we make rather than an insight we are able to glean from the model itself is a very useful mental shift.

Also, you're right to assume that I haven't encountered the concept of ATP -- this gives me a new topic to stew over in the middle of the night.  

Thanks again,
Erica

Answer 1

One of the big problems with supply curves is the assumption that they exist -- that is, the assumptions surrounding perfectly competitive markets on which they rely. In the age of amazon, google, uber, etc. it seems crucially important to give more weight to the discussion of monpoly, oligopolies, price discrimination, information economics etc.

Comments

Perfect competition is the only market structure with a well defined supply curve. However, if we teach S&D then we are assuming a PC market. My issue is with the disequilibrium argument - among others which I won't go into here.  Aggregation is a major problem at all levels in economic theory. The way that we derive the market demand curve ignores  income and wealth distribution.

---

If we switch our emphasis to non-competitive markets then do we cut out the S&D chapter or move it to later in the course/text? But in either case we would have to modify other sections of the course (text chapters) that also use the S&D framework.

Answer 2

You're thinking about the model backwards. I tell my students to think of supply and demand as telling a story, not as two curves or two equations.  There are buyers and sellers, trying to maximize their payoffs. An equilibrium is NOT just where two curves cross, or the solution to two equations. It is a price and quantity such that nobody wants to deviate from their actions. In other words: a Nash equilibrium.

Comments

I often get things backward. And I agree that  students should be discouraged from confusing geometry, or even mathematics. with economics.

However, My problem is not with the equilibrium concept but with the disequilibrium adjustment story we tell. Since the model is always in equilibrium then there can be no disequilibrium adjustment, and it could not be via consumers and firms adjusting prices because the S&D model is a model of a perfectly competitive market where prices are given.

---

I see. Yes, it's good to tell a story of why a non-equilibrium price can't be an equilibrium, but bad to tell a story of gradual adjustment, like a hog cycle story.

---

I think that when going through the excess demand/supply stories we should tell our students, especially majors, that the S&D model is a static model and so our stories are <<assumptions>> and cannot be <<derived>> from the <<theoretical model>>, which is closed by an equilibrium condition and has nothing to say about what happens out of equilibrium.  I have a paper that I can send to you if you are interested.

---

Don't be shy! How about posting the paper's web address or cite here? Then anyone who wants can take a look.

---

Hi Eric, My apologies for not responding earlier - at 80 my health is sometimes a problem (not now thank goodness) and I have been busy redoing my microeconomics course, which you can look at by typing sleemanomics into your browser. My S&D papers are on my <<ResearchGate>> page, although everyone seems to be obsessed with S&D and general equilibrium. I would be happy to send you the latest version of the most important S&D paper if you are interested.

---

It's great that you're still thinking economics. You can send me the paper at erasmuse@indiana.edu, but, to be honest, I probably won't get round to reading it.... Information overload.

Answer 3

You have identified a big weakness of microeconomic theory. It is very good at rigorous analyses of equilibrium; it is much worse when it comes to disequilibrium. The closest theory I know is the following: “Take any large group of people. Let them communicate and strike deals. Protect their property rights and stand ready to enforce any and all voluntary agreements. These agreements are tentative; they can go on looking for better deals until they can’t find any. Nothing is finalized until everything is finalized. (This by the way is how many negotiations in politics and international trade work.) The outcome of this process will be close to the allocation determined by the equality of supply and demand, or its multi-commodity version, general competitive equilibrium.” This is a very very rough statement of the Debreu-Scarf core convergence and equivalence theorems. A rigorous statement needs several underlying conditions, and a rigorous proof needs a lot of math. For that I must refer you to Werner Hildenbrand and Alan Kirman, “Introduction to Equilibrium Analysis”, chs. 1 and 5, and Jerry Green, Andreu MasColell and Michael Whinston, “Microeconomic Theory”, Ch. 18.

As a practical matter, my advice would be: Take supply-demand equilibrium as the starting point for your thinking. Don’t be surprised if the outcome you observe in a specific context differs substantially from what equilibrium theory would predict, but then look for reasons for that departure, for example monopoly, asymmetric information, or institutional failure. You will find many such departures and many such reasons, but still the starting point will give you a better foothold on the problem than any other approach that I know of.

Comments

Much of Vernon Smith's work (and Nobel prize) was attributed to his experiments measuring how (and how fast) supply and demand converged towards equilibrium. Experimentalists have gutted rational choice theory (I exaggerate) but also contributed to evidence of market efficiency...

Answer 4

One more point is worth making in this context. It is to do with the possibility of private information. The well-known idea is "the winner's curse". If one object is being auctioned and there are n bidders, the potential winner (who gets a high private signal about the value of the object) should ask him/herself: "Others got lower value signals; what should I learn from that?" A mirror image is the loser's curse. If (n-1) objects are being auctioned and there are n bidders, the only potential loser-out (someone who gets a very low value signal) should think: "All others are getting higher signals; what should I learn from that?" There are indeed theoretical models that show how bidders can strategically adjust their bids to differ from their signals to take into account these curses. But suppose k objects are being auctioned with n bidders, and the ratio k/n is neither close to 0 nor close to 1. Then neither the winner's curse nor the loser's curse has significant bite. Each bidder can bid based on his/her own signal, ignoring the fact that others have their private signals, and not worrying about what information they may have. This is the clearest and most rigorous statement I know of the Hayekian idea that markets allow decentralization of information. Again this theorem needs a lot of underlying conditions and a lot of math. The paper that does it (Wolfgang Pesendorfer and Jeroen Swinkels, "Efficiency and information aggregation in auctions," American Economic Review, 90(3), June 2000, pp. 499-525) is not as well known as Debreu-Scarf etc., but in my judgement it deserves a place right up there.

Comments

Excellent — this gives me quite a bit to chew on, and I very much appreciate the additional resources you’ve mentioned. I find particularly helpful your advice to use the standard supply and demand model only as as starting point and from there, to look for reasons behind any departure from its predictions. Thank you!

---

Funny you should bring this up. Just yesterday I was at a conference-- the BLS conference on employment and price stability, and in it, it was made clear that industrial firms do not raise prices in response to input price rises.

Where you don't have a Walrasian Crier, supply and demand completely falls apart. The law of diminishing marginal utility, the law of diminishing marginal returns are inaccurate and/or are antiquated, leaving microeconomics a largely false field. Some of their stuff is an interesting thought experiment that may be helpful in an early analysis, but to be taught as representative of economic behavior is asinine. And this is for getting their absurd efficiency claims. Definitely microeconomics needs to be rethought and retaught. For example the industrial goods industry Market supply curve, as discussed, is perfectly elastic. Economists at the conference were even joking that no firm (in the industrial sector) raises its prices in response to input price increases. Microeconomics is like Geology teaching that there are no tectonic plates that shifted.