Why is supply and demand so confusing?
Maybe you don’t find it confusing, but I do. It all started a few years ago when we noticed that we had a “trick question” on our placement exam at Bentley. The question asked what would happen to the demand for tea if there was a health scare regarding coffee. Obviously coffee and tea are substitutes. So if the health scare reduces the price of coffee then the demand for tea will . . . well, let’s think about it a bit more.
Then a few years later I noticed the following statement in chapter 4 of Mankiw’s intro text:
“Suppose that the price of frozen yogurt falls. The law of demand says that you will buy more frozen yogurt. At the same time you will probably buy less ice cream.”
Well at least that clears up the coffee/tea example. A health scare will reduce the price of coffee. Since coffee and tea are substitutes, if the price of coffee falls you will buy less tea.
Or will you? Doesn’t Mankiw’s yogurt example also imply that “the law of supply says you will sell less yogurt” when the price falls. And since the “law of logic” says that the number of units sold equal the number of units bought, we’ve got a problem here somewhere.
Things got more complicated in chapter 4’s “Problems and Applications.” Question 12 asked:
An article in The New York Times described a successful marketing campaign by the French champagne industry. The article noted that “many executives felt giddy about the stratospheric champagne prices. But they also feared that sharp price increases would cause demand to decline, which would then cause prices to plunge.” What mistake are the executives making in their analysis of the situation?”
I thought the answer was that the marketing campaign had caused the demand to increase, and that caused both price and quantity to increase. But that can’t be right because Mankiw already told us that when the price goes up people will buy less yogurt—surely the same must be true of champagne? So there seems to be a conflict.
That was the 3rd edition. Then the 4th edition came out, and question 12 mysteriously disappeared, like a missing face from a Stalin-era group portrait. All the other questions were still there. I emailed Mankiw to ask him why my favorite question was removed, and he indicated that with each new edition changes were made based on feedback from professors who used the book.
The yogurt example is still there in the 4th edition. Perhaps most professors agreed that a higher price of yogurt does reduce consumption, but thought that an accurate treatment of supply and demand such as question 12 would be “too confusing.”
That’s when I started notice that I was in the minority. Many economists seem to like to do what I call “reasoning from a price change.” It goes like this:
1. If interest rates fall sharply we can expect investment to . . .
2. If the price of oil falls then consumers will . . .
3. If the dollar depreciates then our imports/exports/trade balance will . . .
Let’s take these one at a time. This year interest rates have fallen to low levels, and investment has plummeted. So lower interest rates must reduce investment. The price of oil has plunged, and oil consumption is also down. So low oil prices would seem to reduce oil consumption. Now at this point you might be getting fidgety. “Wait, you’re not holding other things equal. Something else changed.”
But that’s not really the problem. Of course something else changed, how else could the price/interest rate/exchange rate have changed in the first place! There is nothing “tricky” about the examples I just provided, they are straight-forward applications of supply and demand. Draw a supply and demand curve. Then shift demand to the left. You will clearly see that at a lower price consumption will drop.
So what do we know about prices? We know that if the price falls because supply increases, then consumption will increase, and if the price fell because demand fell, then consumption will decrease. In other words we know that if the price (or interest rate or exchange rate) changes, we can predict with 50% confidence that quantity will increase, and 50% confidence that quantity will decrease. So that’s progress, I guess.
Recently I read that the current low energy prices are actually a “bad thing,” because they will discourage consumers form conserving. I found this interesting, as I think the current low prices are a bad thing because they reflect consumers conserving energy. Why are they “conserving” energy? Because they are out of work.
Often I read that the Fed made a huge mistake cutting interest rates to 1% in 2003, because it blew up a housing bubble. I thought interest rates fell in 2003 because investment fell.
I notice that people really like to pontificate about exchange rates. I really don’t know much about open economy macro, but I always wonder what people mean when they talk about the implications of a falling dollar. Why would the dollar fall in the first place? More sophisticated pundits will talk about a scenario where foreigners suddenly don’t want to hold American dollars anymore. So now we finally have a root cause, and we can see what happens next. But why do foreigners not want to hold dollars anymore? Does it mean that they no longer want to run trade surpluses with the US? And if so, why not? Do they want to save less, or invest more? And how will their preference for balanced trade affect us? All we really know is that our trade deficit will disappear. But what does that mean for the US economy? More exports? Or less imports? Suppose the Fed is targeting NGDP when this collapse in confidence in the dollar occurs? What then? Here is the Sumner Rule, derived from 54 years of watching predictions fail to come true:
If over several business cycles people keep saying a trend is unsustainable, it is sustainable. Or at least you will not live long enough to see it reversed. It may be reversed some day, but since you will not live long enough to see it reversed, it’s not worth thinking about.
What sort of unsustainable trends am I thinking about? How about Asians lending money to Aussies and Americans? Or how about health care costs rising as a share of GDP? By the way, on this point it might be easier to think in terms of “non-health care costs.” Suppose the non-health share of GDP falls in half every 500 years. Then total spending on non-health would still be able to rise as long as we had any kind of decent RGDP growth rate. Can’t happen forever? Don’t bet on it.
Lessons for teaching economics
Is it possible that economics students don’t learn anything when we teach them supply and demand? It seems to me that there are basically two things we’d like our students to know about S&D (before we get into applications like price controls and taxes):
1. The impact of a supply or demand shift.
2. How to draw inferences from changes in prices and quantity.
I teach at an institution that is well above average, and here is what I have found. Almost every single student comes into EC101 knowing the impact of supply and demand shocks. Tell them a frost hits the Florida orange crop, and they can explain what happens to the price of oranges. Tell them millions of Chinese start buying cars and they can tell you what happens to the price of oil.
I also find that almost no student comes into my class knowing how to interpret price and quantity data. And what is worse, they leave the course equally ignorant. I often ask the following question to upper level econ or MBA students who have already taken principles:
Question: A survey shows that on average 100 people go to the movies when the price is $6 and 300 people go when the price is $9. Does this violate the laws of supply and demand?
Very, very few can answer this question, especially if you ask for an explanation. Even worse, I think there is a perception that there is something ‘tricky’ about this question, something unfair. In fact, it is as easy a question as you could imagine. It’s basic S&D. It’s merely asking students what happens when the demand for movies shifts. I cannot imagine a less tricky question, or a more straightforward application of the laws of supply and demand. In the evening hours the demand for movies shifts right. Price rises. Quantity supplied responds. What’s so hard about that? And yet almost no student can get it right. Our students enter EC101 knowing one of the two things they need to know about S&D, and they leave knowing one of the two things they need to know about S&D. Maybe instead of having them memorize mind-numbing lists of “5 factors that shift supply,” and “5 factors that shift demand,” we should just tell them to read something that will explain what economics is all about, something that portrays economists as detectives trying to solve the identification problem, something like Freakonomics.
If there are any other economics instructors out there I’d like to know what you think. I really don’t think we need to teach students what happens when frost hits the Florida orange crop. Perhaps we should just put supply and demand into an appendix and tell them to study it if they need to. Instead devote 100% of chapter 4 to the identification problem. Leave all the technical stuff for students majoring in economics to take in their intermediate level courses. Or maybe the identification problem is too hard, and we should simply forget about teaching supply and demand. Devote the whole course to opportunity costs, incentives, marginal analysis, etc.
I’m open to suggestions. But when I read the newspaper, even the elite newspapers like the NYT, FT, WSJ, etc, I feel like something is wrong. What they are doing is about as closely related to economics as astrology is to astronomy. Talking about the implications of price/interest rate/exchange rate changes is about as useful as talking about the implications of Saturn being in Aquarius.
PS. In case Greg Mankiw is reading: The fact that I use your book shows that I think it is the best one out there.
PPS. In case Tyler Cowen is reading: Ignore the previous PS, I haven’t had a chance to check out your (and Alex’s) book yet.
Tags:
28. July 2009 at 12:34
Scott, I tried to make this point once in relation to another post on causal statements, but I think the problem has to do with the difficulty of translating mathematical concepts involving recursive, simultaneous relations into verbal terms; that plus the fact that many people find it hard to use the English language with precision. You were guilty of one such slip in this paragraph:
“So what do we know about prices? We know that if the price falls because supply increases, then consumption will increase, and if the price fell because demand fell, then consumption will decrease.”
The statement “then consumption will decrease” seems incorrect in its use of the future tense. If demand fell, isn’t that the same as saying that consumption already has fallen which is why the price fell? (If I misunderstood what you meant, that also proves my point. I can’t lose this argument!)
I never took an economics course but from my reading I always thought supply and demand was simple enough to understand in terms of a sequential process in which various adjustments take place over time. (I’m thinking of something like the Marshall period analysis, you know first price adjusts, then supply, etc.). Unfortunately people end up trying to describe something in 25 words or less than really needs ten times as much. And sometimes they don’t have the space, like in newspaper articles, so they end up with a crude, misleading shorthand.
28. July 2009 at 12:36
Casey Mulligan also thinks many economists have a hard time grasping supply and demand.
28. July 2009 at 12:45
Scott,
I couldn’t agree with you more. I took AP Econ in high school, Micro and Macro in undergrad, and Micro in grad school, along with several other complimentary economics courses i.e. Money & Banking, Health Econ etc.
I think that I have a pretty good intuition for economics but the standard way of teaching it really frustrated me and it wasn’t until several years after undergrad that I really started to understand and love economics. Mark Skousen’s text book “The Making of Modern Economics: The Lives and Ideas of the Great Thinkers” really helped me understand the intellectual evolution of economics and thus provided a solid starting point.
The biggest “aha” moment for me was when I finally grasped the idea that incentives matter and everything else flows from there. Public choice theory was also a critical moment in my intellectual journey. I am still trying get my head around Monetary Econ but that is why I read your blog.
I just thought that I would give a “student’s” perspective and keep up the good work.
28. July 2009 at 13:00
The fundamental challenge (and it’s not just students who suffer) is understanding causality. Economics introduces the notion of causality as a “shift in the curve” vs. “movement along the curve”. Properly, it should be treated as exogeneity vs. endogeneity, but students don’t get this until they get to econometrics (e.g. stats). And if you think students find economics unintuitive, it has nothing on stats…
And yes, that question on movies IS horribly worded:
“A survey shows that on average 100 people go to the movies when the price is $6 and 300 people go when the price is $9. Does this violate the laws of supply and demand?”
Does that mean: On average (and holding all else equal), if movies were to be priced at $6, then 100 people would attend. And if priced at $9, then 300 would attend.
Or does that mean: On average, at times when movies were priced at $6 there were 100 people in attendance. And at times when movies were priced at $9, there were 300 people in attendance.
Also, putting the $ before the people implies a temporal sequence which implies (usually) causality – which in this case is deliberately deceptive.
28. July 2009 at 13:08
[…] Scott Sumner does a small discussion on prices. The exchange rate is a price, as he notes the important thing is “why the price has […]
28. July 2009 at 14:29
Well if pundits do, they aren’t that sophisticated. Foreigners–such as Asians–hold American assets not currency or demand deposits. Assets tend to be a hedge against inflation of the currency of their market. So their exposure to monetary policy (isn’t this what’s really cited as the cause?) is lessened. Indeed, they are more exposed to exchange-rates drifting inconsistent with inflation. Certainly this happens in the short-term: real rates are equalized by the exchange-rate in the short-term but in the long-term, exchange-rates compensate for relative-inflation.
28. July 2009 at 14:57
I see what you mean in our other discussion Scott.
As the others mention I think the problem is explanation in time.
There is a confusing sf story related to this by Matt Hughes. In the story a set of physics laws of time and space are set out, then some odd things happen. The solution to the mystery is that a particular species of beings live in one location in space, but across all of time. The example question you give does the same thing in economics, I’m not surprised it’s confusing.
28. July 2009 at 15:03
Back in my physics days, one of the things I liked about thermodynamics was the convention, when writing a partial derivative, of including a subscript to denote what was being held constant. If PV=nRT, and free energy is a function of P,V, and T, then the partial derivative of free energy with respect to pressure is different if you hold volume constant than if you hold temperature constant.
Ceteris paribus is great, but logically inconsistent in a literal sense; if “every thing else is held constant”, why would supply change? You need to specify which ceteris are paribus.
28. July 2009 at 15:35
If you engage your students on successful (e.g. student aid, progressive income tax rates) and unsuccessful (The early 90’s attempt at progressive sales taxes, i.e. the luxury tax) attempts at price discrimination in various settings, then one would think that they ought to “get” the movie theater example wicked* fast. I could be wrong, noting I don’t teach econ.
Anyway, as a non-teacher, I’d try to teach first year econ students about 1)supply, 2)demand and 3)price by holding 1 constant and showing them what happens to 2 when 3 changes, and what happens to 3 when 2 changes. Then I’d hold 2 constant and have them show me what happens to 1 and 3 when they vary. Finally, repeat the exercise with 3 held as a constant. Does anyone teach it that way? The last exercise ought to be the most fun (when/if they get it), and most useful to society at large when they go on to the real world as non-economists.
*I grew up in suburban Boston. I don’t get to use “wicked” too much anymore now that I am in a land far far away.
28. July 2009 at 15:47
I’m not an economist, but I have tried on occasion to teach some elementary stuff to law students. For me, the key thing was distinguishing movement along the demand curve from a change in the location of the curve. Just saying “the price has changed” doesn’t tell you which of these has happened, so you can’t infer anything from a price change alone; you have to know why it changed. (Sorry if this observation is too banal for this site, but it seemed to me that it’s this chat’s the source of the trouble.)
One of the econ texts, I forget which, has a nice example of a change in the location of the curve. In the early 1960s, the Catholic church allowed Catholics to eat meat on Fridays, except during lent. In Massachusetts, a heavily Catholic state, both the price of fish and the amount of fish sold fell sharply.
28. July 2009 at 16:00
As a non-econ, I’m never sure if people are using the correct terminology when they say demand. For instance, when people conserve energy because the price is high, it is often called demand-destruction. But that isn’t really demand-destruction, is it? Because the demand curve for oil hasn’t shifted. Or can movement along the curve be correctly called demand-destruction? It’s the language I’m unclear on.
28. July 2009 at 16:13
[…] 2008 (3)October 2008 (1)September 2008 (3) For those who teach principles of economics, this post by Scott Sumner is worth contemplating.Scott does a good job explaining why the model of supply and demand is not always as […]
28. July 2009 at 16:25
I also think the question is a bit misleading, particularly given the way a principles course is usually taught.
Such courses often focus on the mechanics of a market in which all goods are homogeneous. In that sense, it seems your question refers to two different markets; the market for movies during the day, and the market for movies during the night. While someone who’s completed a principles course should be able to understand cyclical fluctuations of demand over time, such a course doesn’t specifically train them to do that, and by asking them about the law of demand it’s likely they’ll be focusing on everything they know about that law, rather than thinking about how and why there might be such cyclical fluctuations.
I suppose you could argue that they’re part of the same market in which price discrimination is being practiced.
28. July 2009 at 16:33
Economics is just a big made up game, what happens is whatever you decide happens in the model.
28. July 2009 at 16:43
I use a multi-stage queue model for the distribution of a good. The queue has a forward direction in flow, the goal is to match inventory level variances at one level of the queue across good types, including money. (All goods at a grocery store will have the same inventory variance at equilibrium)
Since we know, ex-post, that queues do not generally go to zero. we can make a broad assumption because of the zero bound. One statement is that supply and demand have met.
28. July 2009 at 16:59
No, the basic concept of supply and demand is not confusing. It’s simple. The champagne price goes up because the marketing campaign shifted the demand curve to the right, meaning that at any given price, consumers would now buy more, and even though the equilibrium price is now higher, they are buying more – not as much as they would if there were no price-elasticity (a vertical demand curve), but more than they were buying before the marketing campaign and shift in demand curve. So the effect of higher price is not lost, nor does it offset the demand-increasing effect of the marketing campaign that increased the equilibrium price in the first place. Rather, it is reflected by the fact that equilibrium quantity is less than it would be if demand were perfectly inelastic, with a vertical demand curve shifting to the right. It ain’t complicated.
And yes, obviously the cause of a change in equilibrium price, as represented by a shift in the supply or demand curve, affects whether equilibrium quantity will increase or decrease. That’s why those curves — i.e., the concept of supply and demand — are taught. I really don’t see what the conceptual problem is or why it would be difficult to communicate or to learn.
As a side note, I’m sure some people cringe at the redundancy of “French champagne”.
28. July 2009 at 17:45
“We know that if the price falls because supply increases, then consumption will increase, and if the price fell because demand fell, then consumption will decrease. In other words we know that if the price (or interest rate or exchange rate) changes, we can predict with 50% confidence that quantity will increase, and 50% confidence that quantity will decrease.” What if the price fell in part because of an increase in supply (the supply _schedule_ or _curve_, that is) and in part because of a fall in demand (the demand schedule)? Then the quantity bought/sold (though not necessarily “consumed”) might stay constant. So maybe our degrees of confidence should be: quantity increase–33%, quantity decrease–33%, quantity remain the same–33%. (I have left over 1% for _metaphysical doubt_; also to avoid fractions.)
28. July 2009 at 18:17
“Suppose that the price of frozen yogurt falls. The law of demand says that you will buy more frozen yogurt. At the same time you will probably buy less ice cream.”
This is entirely correct.
If a health scare reduces the demand for frozen yogurt,
then you will probably buy less as a result of the health
scare. If, as a result of the health scare, the price falls, then you will probably more as a result of the price
drop. At the same time, you’ll buy more ice cream as a
result of the yogurt health scare, and less ice cream as a
result of the yogurt price drop. By shifting supply and
demand curves, you can easily see which effects are dominant.
Of course you knew all that already. But that’s how I
explain it to my students (and essentially how I explain it
in my textbook) and it seems to me to be the best way to
make the issues clear.
28. July 2009 at 18:21
Scott,
Great post. I hated the movie example though. My first guess, the $9 movie was a better movie. We didn’t have enough information to answer the question. Same theater? same movie? same time? weather? I hated it when my professor’s gave these types of questions, especially when they were multiple choice and the prof. didn’t realize there were two correct answers in the multiple choices. As long as you use it as a verbal question to get everyone thinking about possible cause and effect, I’ve got no problem.
28. July 2009 at 19:41
To be honest, I haven’t had this same problem with my students. Granted I spend a lot of time on supply and demand, cutting out time for other things, but I think once students get 1. and 2. (and for my students I do have to teach both) the intuition is very valuable to them.
I think the key is to explain not just what supply and demand are mathematically, but what they are all about. I tell my students that demand is the *relationship* between price and the quantity buyers are willing and able to purchase, given various factors that affect this relationship called determinants. If a determinant changes, it changes the whole *relationship* between P and Qd, ie a shift in the curve. I use a similar definition for supply. Once my students have down that supply and demand describe relationships, I show them equilibrium and how these two relationships determine P and Q.
The key for my students at this point, particularly when talking about what causes price and quantity changes, is knowing what kind of things will change each relationship and, most importantly, what kinds of things will *never* change the relationship. I find that once students get this idea, they can graph and explain the effects of shocks to supply and demand as in 1. I also find that they can look at a real world situation in which P and Q change and come up with plausible stories about what could have caused those changes. I have my principles students write a paper in order to demonstrate this fact, and I’ve had great success with it so far.
28. July 2009 at 19:47
As a semi-educated Econ guy, I find the craziest thing about S&D is that it seems intuitive to the layperson. Of course, they understand the implications of a spring frost in florida on the price of oranges! It’s logical!
Only later does one realize that the more you learn, the less you intuitively “get”. Explaining elasticity and tax incidence and dead-weight loss? Things start getting sticky- quickly.
Econ is one of the few areas that the less you know about it, the easier it seems. Your Econ confidence is inversely proportional to your Econ knowledge.
28. July 2009 at 19:50
[…] — Scott Sumner […]
28. July 2009 at 19:51
I have trouble believing anyone would have trouble with any of questions you posed in you article, let alone anyone with a passing knowledge of economics. Hopefully you helped some very confused people out, but I bet this post was just a waste of your ingenuity.
28. July 2009 at 19:57
I like the way you presented the case. You made a very nice point that that faculties are not able to teach the subject properly to the student. In his book The Economic Naturalist: “In Search of Explanations for Everyday Enigmas” Robert H. Frank has given an example on knowledge level of students and professor on the simple concept of OPPORTUNITY COST and how maximum of them failed miserably. It is the reason why one of USA President Truman, Harry S said “Give me a one-handed economist! All my economists say, On the one hand on the other.”
I personally believe that we must have a valid reason and data to prove our thoughts, expectations and views. The main reason why all this is not present in todays world is just because people lack common sense. The main reason for this lack is availability of too much information many of them is not required.
It seems to me after all this year of studies is that I wasted a considerable amount of time in studies.
28. July 2009 at 20:02
Harry Truman said one more good thing “If you can’t convince them, confuse them”
28. July 2009 at 20:16
Here’s my understanding:
Coffee/Tea
At the current price fewer people want to buy coffee. The price will fall and quantity supplied will fall until equilibrium. This happens because sellers compete by bidding lower prices to keep more buyers than their competitors.
Tea is the opposite, at that price more people want it, so the buyers will bid higher until the equilibrium is met.
Bidding doesn’t happen explicitly in real life, but the effect is the same.
28. July 2009 at 20:19
I teach Micro to MBA students in Pune, India. I’ve nearly finished the course, and I have not taught them Supply and Demand yet – not because of the problems you mentioned, but because they’d already done the ‘framework’ in an earlier course. But I’ll ask them the movie question the next time around – I can’t wait.
And my own answer? Nothing intrinsically wrong with a higher price for a higher quantity, but more information needed (same movie? different time? different movie hall?)
Great point, though – food for thought for sure!
28. July 2009 at 22:31
[…] brings me to the real reason I wrote this post. Sumner had a post on why he is confused by rudimentary supply and demand models. In it, he refers to a question that […]
28. July 2009 at 22:42
I simply loved this piece (HT: Mankiw)
29. July 2009 at 01:08
If coffee and tea are direct substitutes then if there is a decrease in the demand for coffee there will be an increase in the demand for tea. This is too simple! As direct substitutes the total consumption for the two would remain the same. Therefore if demand for one goes down the demand for the other goes up. The same with frozen yogurt and ice cream. It may be a big if on the substitution of either, but assuming it is a given….
The movie example does violate the laws of supply and demand, period. UNLESS there is another variable that changed and it is implied in the question that no other variable changed. If the question were how the higher price and higher demand could have occurred then it would make sense.
If you teach at such a highly regarded school I’ll be applying for your job. Should be easy to get if you can’t see this. Oh and I don’t even have a PHD just a BS. (and several decades of life experience) Here is where you get offended, but you couldn’t explain it so……. No offense intended, but I couldn’t resist. This really does seem simple to me.
Supply and demand is easy to understand. The more something cost the more people want to sell and the less people want to buy and vice versa. See N. Greg Mankiw’s reminder that supply and demand curves are always drawn with “all things being equal.” That means no other variables like time of day for the movie have changed.
To Philp P above
These aren’t mathematical concepts they are human behavior observations. Big difference, Big, Huge.
Not so humbly, but with a very dry sense of humor. I don’t really know it all, but I play one online.
Fletch
P.S. Supply and demand curves don’t always go left and right. sometimes they go up and down. Think about it. I doubt they would necessarily remain parallel either.
29. July 2009 at 01:31
[…] curso de Economia seriamente, certamente terão dificuldades para fazer abstrações. Scott Sumner apontou o problema e Mankiw reforçou. Mas nunca se esqueça que o mais difÃcil conceito, mesmo para economistas mais […]
29. July 2009 at 01:36
1. I find that it helps to always distinguish verbally between “demand” and “quantity demanded,” with the former referring to the whole demand curve and the latter to the number of units purchased at a given price.
2. Word problems, especially with multiple choice, do require careful pinning down of ceteris paribus assumptions. As students get more experience they will become more sophisticated in understanding from context what is probably meant, but in a principles course you really can’t assume that.
3. I’ve found that for managerial-oriented micro a big problem for a significant fraction of the students is the concept of marginal revenue. The idea of having to lower the price on all the inframarginal units you would have sold if you had produced a little less seems to baffle a subset of the class. This difficulty makes it very hard for them to understand optimal pricing, etc.
29. July 2009 at 02:54
[…] The full article can be found here. […]
29. July 2009 at 03:03
Thanks Scott for this illuminating article. I really feel we should clarify these general terms because economic concepts do not have the concreteness that the subject matter of natural sciences have. You’ve articulated some of my personal concerns in studying economics, and I’ve turned to the philosophy and history of economics just to clarify what we mean by value, utility, price, etc.
A word on libfree’s post: I also had the same immediate impression. I was thinking that the higher-priced movie was valued more — if a kind of prestige pricing could have stimulated demand.
29. July 2009 at 03:13
I think several previous commenters identified the problem correctly when they noted it boils down to the students understanding when to “move the curve” vs. when to “move along the curve.”
It took me a while to understand it, but when I did I had my “aha” moment.
29. July 2009 at 03:14
Scott:
My students have trouble with changes in supply and demand.
In assignments, I require that they write out the answer. So, the problem might be–the economy goes into recession. Income falls. Gasoline is a normal good. They draw the diagram, and then write out–
The equilibrium price of gasoline falls. The equilibrium quantity of gasoline falls.
I explain that the reason they must do this is to emphasize that prices and quantities are social phenomenon in the real world. That is what we want to explain. What is happening to them. And, of course, recessions and incomes are also things that happen too. The supply and demand apparatus are tools that we use to explain these things.
I must certainly do discuss (and have a web discussion topic) of “the demand for gasoline rises, which causes the price of gasoline to rise. A higher price of gasoline is associated with a higher demand. So the law of demand doesn’t apply to gasoline.”
I also sometimes use the “everlasting cycle” story. Demand rises, price rises, demand falls, price falls, demand rises, price rises… it is an ever lasting cycle.” That is in the context of explaining why I make such a big deal about changes in demand and changes in quantity demand. In reality, both involve changes in how much people are buying, but we use those terms to avoid that confusion. Remember, of course, this business about movements of the curve and shifts in the curve are all about tools we economists use to explain reality. They aren’t the reality. The reality is people exchange money for goods and producing, selling, buying, and consuming goods.
I also tell them about how one purpose of the tool is to be able to figure out what sort of thing must have caused price and quantity to change. If price and quantity both rise, look for something that would impact demand.
But, I can’t claim that my students get it.
29. July 2009 at 03:36
Scott,
This post would sparkle if Andy Rooney were to read it on “60 Minutes”.
That’s what I was imagining when I read it.
Especially when I got to the part:
“I notice that people really like to pontificate about exchange rates. I really don’t know much about open economy macro, but …”.
29. July 2009 at 03:36
I agree with what several other commenters wrote; in most of these cases, the problem is indeterminate because there is insufficient information about other factors. In teaching, it is better to avoid setting questions to which the “right” answer depends on assuming some information that you think is obvious but which might not be so clear to other reasonable people. Such questions can handicap the more thoughtful student. This is one reason why I believe that academics should be expected to spend a substantial period – say two years – early in their career working in a practical environment in which they encounter wider considerations and have to explain their work to non-deferential colleagues. I suspect that the arrogance of some academics reflects success in their stylised, theoretical world that they have not tested in the real world.
29. July 2009 at 03:39
@Phil P, @Gordon:
Excellent replies, couldn’t agree with you more. Causality in supply and demand theory is not that difficult. Of course, reality is more complicated and dynamic, but that does not invalidate these very helpful analytical tools!
29. July 2009 at 03:51
Minkiw approves:
http://gregmankiw.blogspot.com/2009/07/why-is-supply-and-demand-so-confusing.html
29. July 2009 at 04:27
great read,
im going into EC101 with a passion for economics prior to picking it, solid references are used and im glad i have more than just the basic knowledge going into school, its good to know reasoning apart from plain textbook definitions and explanations. thanks!
adam
29. July 2009 at 04:29
My favourite part:
“So what do we know about prices? We know that if the price falls because supply increases, then consumption will increase, and if the price fell because demand fell, then consumption will decrease. In other words we know that if the price (or interest rate or exchange rate) changes, we can predict with 50% confidence that quantity will increase, and 50% confidence that quantity will decrease.”
Corollaries:
Price changes require a shift in at least one curve.
Shifts in one curve are prerequisite to movement along another.
Shifts dominate movements along.
Two existing curves are not long for this world.
Shifts in curves can result in any combination of higher or lower prices and higher or lower quantities. There are 4 possible directional combinations.
Don’t rule any of them out absent other information (e.g. the movie question).
29. July 2009 at 04:46
Phil P, Yes, you are right, consumption should have already fallen.
TGGP, Yes, I agree that many economists slip up on S&D. The real point of my post was to criticize economists who “draw inferences from a price change.” I’ve done that a few times as well, but it is bad economics.
Thanks Scott.
Stasguy, Sorry, but I don’t think is it horribly worded, but even so it doesn’t explain the results I get. When I ask the question in class I describe an eager young economist who goes out and conducts a survey of how many people go to movies at different price points. The term ‘survey’ suggests that we are looking at real world data, not “other things equal.”
Let me just clarify one point, because I imagine people are going to nitpick this post. There is no question that my basic point is right–that most people don’t understand supply and demand. And it isn’t just terminology, saying “demand” when they should say “quantity demanded.” They don’t understand supply and demand, period. I can’t tell you how often I have picked up a newspaper and read something like, “In violation of the laws of supply and demand, consumers continue to buy more gasoline despite much higher prices.” This isn’t just bad wording, it shows a complete lack of understanding of the fundamentals of S&D. And you see it in elite newspapers with reporters who studied econ at Ivy League schools. When you get to changes in interest rates and exchange rates (which are also prices) it is ten times worse.
Jon, You are right, and that annoys me too. I was giving them a break on calling dollar assets “dollars”, but you are right.
Current, I’m not sure I follow, but I’ll read the others and come back.
dWj, I agree that partial derivatives make things clearer. Of course I am not attacking the underlying theory but rather how people misuse it.
happyjugger0, I should mention that the first things they think of are “price discrimination” and popular vs unpopular movies. But then I tell them that both prices refer to the same movie, and refer to the price of adult tickets (not seniors and children.) Then they are stumped. So price discrimination is not (necessarily) involved.
Alan, That’s right, and I teach that too. But students have trouble applying it to the real world, especially when then start with a price change. They need to start with a shift in a curve, but we don’t directly observe curves shift in the real world.
Rob, You’re right that people misuse the term “demand,” but I am after bigger fish. They don’t even understand the concept. I don’t quibble with people misusing the term, as long as they get the basic idea.
Mike, All shifts in demand occur over time. It’s like saying there is no seasonal demand shift for gasoline, because gasoline in the summer is different than gasoline in the winter.
More importantly, trust me, I can do this with any product, and 95% of students don’t know the answer. If I say people tend to buy more houses when interest rates are high, or they tend to buy less gasoline when gas prices are low, students will scratch their heads just as mush as in the movie example.
Mattyoung, I’m not sure I follow, but I steer clear of queuing problems with basic S&D.
Gordon, It may seem very easy to you, but obviously most people (including many economists) don’t find it so easy. Especially applied to the real world. Obviously all economists understand the abstract model.
Philo, Good point.
Steven, Thanks for responding. You may not remember me but I was in many of the same classes as you at Chicago. Since you are a better writer than me, I hesitate to disagree, but Mankiw’s statement still seems wrong as written, although he clearly understood the concept in the same way that you and I do. You seem to argue that there are partial derivatives implied in a simple sentence sequence like “if A happens then B will happen.” I took it as “if you observe A, then you can expect to observe B as well.”
And of course even if you are right it doesn’t change my basic point, that when people in the real world draw inferences from anticipated prices changes they are often making unconditional predictions. As I said in response to earlier commenters, it is especially bad when interest rates or exchange rates change.
By the way, a note to my readers, Steven Landsburg’s “The Armchair Economist” is highly recommended to anyone who likes to see very bright economists take a counterintuitive look at the world. I believe he also has some other books, but that’s the only one I have read.
I’ll take more comments later . . .
29. July 2009 at 04:49
Bill:
“I also sometimes use the “everlasting cycle” story.”
Just like my big “aha” moment in understanding supply/demand came from the discussion of exogeneity/endogeneity in dynamic systems (e.g. instrumental variables), the “aha” moment in understanding the business cycle came in differential equations. Specifically, the predator/prey cycle.
http://mathworld.wolfram.com/Lotka-VolterraEquations.html
The insight: In a time-dependent closed loop system, the existence of a capacitance in components of the system will create cycles.
29. July 2009 at 04:54
Champagne is not like the other products considered, because it is sold as a luxury good — part of its appeal is that it is expensive (“conspicuous consumption”). At one time the Champagne industry found itself in the odd position of selling a luxury good that was not particularly expensive compared to many other wines; that’s why it introduced super-premium lines (e.g., Roederer Cristal). Somewhat similarly, independent colleges have also been known to to reposition themselves in the marketplace by raising tuition charges to match an aspirational peer group.
29. July 2009 at 05:17
Scott,
Great post — I know this is one thing that often tripped me up in intro econ. I just wanted to second your recommendation of Landsburg’s “The Armchair Economist” — it’s one of the best introductions to economics I’ve ever read.
29. July 2009 at 05:26
I’ll admit that I was fooled initially by the movie question. On further reflection its easy to explain if it is rephrased as “The market clearing price for 100 units is $6 per and the clearing price for 300 units is $9 per”. Given fairly common assumptions about supply curves the response is obvious and accurate.
29. July 2009 at 05:30
libfree, That’s exactly the problem. You assume we need more information, when we don’t. Nothing in the question violates the laws of S&D, period. In fact no set of price and quantity information for any good of any kind could possibly violate the laws of S&D. These laws allow for anything to happen in P-Q space.
Norman, It sounds like you are teaching it the right way. But I try to do the same, and they still get the movie question wrong (I ask it at the end, or at the beginning of an upper level course. And of course many economists make the mistake of drawing inferences from price changes.
Gray, That is a good point. People often ask me for my view on the crisis. When I explain it they often say “that makes sense.” That’s how I know they haven’t understood a word I have been saying. Otherwise they would have said “Are you crazy? You think tight money caused the crash, and not the financial crisis.”
Greyswan, Pick up any newspaper and you’ll see reporters for top newspapers misusing S&D. You’ll see economists misusing S&D. I could ask this question to undergrad econ students at Harvard and most would get it wrong.
manish, I think common sense makes it harder to learn econ, as the field is very counterintuitive.
MattW, Yes, the demand for tea probably rises, but you can’t be sure from the information. Suppose most tea drinkers were poor people who didn’t are about health. Maybe they’d switch over to coffee because it was cheaper.
Ashish, Thanks. Actually you don’t need more info–see answer to libfree above.
Thanks Sriram.
Fletch, You are wrong because if the price changes we can infer one of the curves shifted. So it is consistent with S&D.
srp, See answer to libfree above.
Stan, See answer to libfree. I am skeptical of prestige pricing for movies, but even if it occurred, it wouldn’t change the answer, which is “no.”
Thanks Tim.
Bill, I do a lot of that stuff too, but I think S&D is harder than economists realize. How often do even grown-up economists say “if the dollar rises US exports will fall.” That’s exactly the same mistake. Suppose the dollar rises because a tech boom here results in foreigners wanting to buy more of our nifty new computers?
I make this mistake sometimes.
So no matter how hard we try, students will slip up in the real world.
JKH, Thanks. Yes, that line does sound like Andy Rooney. I’m getting to be a grouchy old reactionary! 🙂
Rebeleconomists, The question I posed about movies does not depend on any real world information. There is no set of assumptions that would change the answer. The answer is “No.”
anon, It’s good Mankiw likes it, but now I need to spend hours responding to comments. Be careful what you wish for . . .
JKH, Exactly right.
29. July 2009 at 05:34
Scott: “I’m not sure I follow, but I’ll read the others and come back.”
My point is that the question doesn’t say anything about time. The person who reads the question must impute that it relates to time by Sherlock Holmes’ method “when you have eliminated the impossible, whatever remains, however improbable, must be the truth”.
29. July 2009 at 06:02
Scott, what you are doing here violates the guild rules of the economics profession. You’re stopping to think carefully about the assumptions behind what you are doing and what you are teaching. That’s not allowed in economics! That’s “philosophy”! It’s forbidden!
More seriously, what you have discovered is that our background understanding places an irreplaceable role in economic explanation — and “blackboard” economics falsifies the explanatory context because it is impossible to capture and model the background understanding that is part of the explanatory framework.
And I would respectfully suggest that you are mistaken if you think that econometrics adequately makes up for this deficiency — just as economists falsely mistake the “given” assumptions of their “blackboard” models for the real world and all the understanding found in their background knowledge, they do the same with with their econometric constructions.
Finally, I’d say your movie problem is a trick. A 3 o’clock movie that same good as a 9 o’clock movie? Are these even near substitutes? Does one have to be a “philosopher” to think carefully about this stuff?
29. July 2009 at 06:14
I disagree Scott. Unless you say that variables other than price and attendance are allowed to vary, and the law of demand is that, other things equal, demand decreases as price increases, then the answer should be “Potentially, yes. In the absence of changes of other variables, this example does breach the laws of supply and demand”.
29. July 2009 at 06:22
“Pick up any newspaper and you’ll see reporters for top newspapers misusing S&D.”
Maybe that’s why they’re reporters.
Really, this kind of thing is nothing new. Ferraro and Taylor showed several years ago that most economists don’t even understand opportunity costs. Understanding economics requires the ability to think clearly. Most people can’t.
29. July 2009 at 06:22
Any reasonable person would think that the expected costs/utility of getting sick from consuming coffee function as a tax – $3/lb coffee might cost me $10000/lb if I end up in the hospital.
The substitution of coffee and tea would take place based on the total cost.
Why is any of this confusing?
29. July 2009 at 06:23
I agree with Scott’s reason why we don’t need more info in the Movie question (contrary to my previous opinion), I just think the sentence construction is misleading in denoting causal relation.
29. July 2009 at 06:28
Rebel,
Here’s yet another take on the movie question:
the answer yes or no depends on whether:
not necessarily allows the answer no,
or possibly requires the answer yes.
I’d say yes to the first and no the second.
So the answer is no.
🙂
29. July 2009 at 06:29
Rob:
Just to add a layer of complexity, it depends on how long low supply has led to high prices. If people expect the price increase is temporary, they may drive a little less this week with the expectation of resuming normal behavior shortly. This would be a shift along the demand curve. If they expect the price increase to be permanent, they are more likely to invest in a more fuel-efficient car and shift the demand curve (not to mention changing the elasticity of demand in the process). Perhaps this is another reason to get NGDP expectations fixed?
29. July 2009 at 06:35
Nota bene:
http://blogs.wsj.com/economics/2009/07/29/secondary-sources-fed-politics-excess-reserves-economists-and-the-queen/
(The one by Todd Keister and James McAndrews (in which Scott is footnoted)
The article linked here shows that the Fed is entirely conscious about what they are doing (no surprise – these are smart people) – and that they are getting exactly the results they (and taylor, volcker, and the WSJ) wish – deflation and the reserves necessary to save Goldman and screw around in this or that approved of credit market.
Japan – here we come. Unemployment – deflation and stagnation.
No wonder Congress wants to nationalize this institution. I don’t recall voting for these policies.
Am I being conspiratorial here? Do they really wish what we are now getting? Or do they just think there is no better alternative?
Again I say, Obama has forgotten that Roosevelt, when forced, chose to represent the interests of debtors, not creditors. The WSJ does not think like that. But we have always been a country of cheap money. Europe is the land of gold and hard money. Lets go back to where we come from. Debase the currency. Get on with it.
29. July 2009 at 06:41
[…] ~ July 29th, 2009 in Teaching In a though provoking post Scott Sumner goes into detail about his difficulties getting students to really get Supply and Demand. He asks […]
29. July 2009 at 06:42
Scott & All,
Seems that this all comes down to either
(1) the difference between a shift OF the curves (i.e., quantity that will be demanded or supplied at a given price, and resulting equilibrium) vs. movement ALONG the curves and shape of the curves (elasticity due to price-sensitivity), which I would think is something that would be taught and fairly easily understood early in the first class of Econ 101, or
(2) first-order effects followed by opposite second-order (or third, etc.) effects (e.g., marketing campaign shifts demand curve to the right, increasing equilibrium price, which then draws in new suppliers (or other expansion of capacity or capacity utilization), shifting supply curve to the right and lowering equilibrium price from that peak (probably not all the way down to the original price).
Either way, I really don’t see what would cause conceptual confusion. Yes, in the real world it is often difficult to adequately account for all these effects and produce a prediction that is precise or even sometimes directionally correct, and it is also difficult/impossible to fully account for all other factors (even outside the changed variable and subsequent chain of effects) that could affect a future state (e.g., equilibrium price), but those obvious qualifiers are categorically different from unnecessary conceptual confusion of the sort represented by the examples above (e.g., champagne).
29. July 2009 at 06:48
Reading these comments, I am amazed that people really do seem to be confused about moving along a curve vs. shifting the curve. I guess the problem people have is that people mix this up:
Supply and Demand —–> (imply a) Price.
However, when price changes, it can be because of either a change in supply, demand, or both. So we can not deduce anything from a change in price except by additional clues in a question. So its foolish to ever see a change in price and come to any conclusion without provided information (e.g. “the movie was at night” which would imply an increase in demand.)
29. July 2009 at 06:56
Scott,
A graph to bring to your attention on your favorite topic: http://dmarron.com/2009/07/29/why-are-banks-holding-so-many-excess-reserves/#more-1144. It shows excess bank reserves and they are monstrously huge compared to required reserves.
29. July 2009 at 07:07
Scott,
Although not much new, the New York Fed paper referenced above by JimP is worth looking at:
http://www.newyorkfed.org/research/staff_reports/sr380.pdf
It does emphasize the intended purpose of excess reserves as a Fed liability management tool more than a commercial bank asset motivator, something I’ve believed from the beginning. Hence the payment of interest on reserves, etc.
Obviously, the fact that the Fed is paying interest on reserves means they aren’t too warm to the idea of charging interest on reserves, at least not right now. Still, as I’ve said before, I’d be very interested to see something written from the Fed as to why they don’t want to go that route, and if it’s something they’ve considered already and/or are prepared to consider in future, depending on their assessment of deflation risk at the time. They haven’t really publicly addressed the idea at all yet.
29. July 2009 at 07:25
Great article! I love your movie question. I think the bottom line here is that looking at a supply-demand chart doesn’t really tell you anything about the relationship betwen P and Q. It gives you one point only – which is not enough to infer anything about future movements. Everything besides that one point is unobservable in the market – the Supply and Demand curves as pictured are purely hypothetical, because the observable variables of P & Q can’t move along those curves – at least one of the curves must change to have any movement. So perhaps a supply-demand chart is just a confusing way to think about the relationship between P and Q.
29. July 2009 at 09:19
Scott,
Great post. It is amazing how badly people mangle supply/demand but as someone pointed out above, it does often require more thought than one might think.
Anyway, here’s a supply/demand question. Why does most of the healthcare reform debate concentrate on the demand for healthcare services? Why isn’t there more discussion about supply (of healthcare not health insurance)? Shouldn’t we be looking at ways to reform the supply side of healthcare services as well as managing the demand?
29. July 2009 at 09:53
“I’m not an economist, but I have tried on occasion to teach some elementary stuff to law students. For me, the key thing was distinguishing movement along the demand curve from a change in the location of the curve. Just saying “the price has changed” doesn’t tell you which of these has happened, so you can’t infer anything from a price change alone; you have to know why it changed.”
The answer to this question in my principles level class is the most useful thing that I took away from the course. I actually didn’t understand the distinction (and it was not taught) until I asked the difference, and then forced my Professor to explain it more completely because I didn’t understand. That was nearly 3 years ago and I am almost finished with my degree in Economics now.
The morons who write newspaper articles throw around “supply” and “demand” like a parrot might if you said these words too much. But the reality is that if people don’t understand that curve shifts are different from a shift along the curve, your movie tickets example serves no purpose.
I love Economics and I wish we had more of it available to undergraduates. Unfortunately, the push seems to be away from research methods and writing (even at the 300 level) and more towards spending the time teaching the exceptions to mainstream economics’ policy prescriptions (market failure, collusion, monopolies) instead of teaching people how markets actually function.
Young people often think that the world is as polite and nice as they are, and then they wonder why the market tends to produce results that are efficient, even if it means that people must work longer hours, overtime, etc. An explanation of the rule (that getting more for less is efficient, that growth is objectively desirable, etc.) rather than dwelling on the exceptions is the difference between students who actually understand Economics and those who parrot vocabulary.
Great post, and thanks. Thank Gary Mankiw for linking me to you. 😀 And tell your students not to give up on learning these issues! Even coming from a younger guy like me, learning about how markets function has totally changed the way I’ve looked at the world–and I never knew that would happen when I was sitting in my principles classes 3 years ago.
29. July 2009 at 10:43
Mr Punch, You are refering to a “Giffen good,” a product that becomes more desirable at a higher price. It is unlikely that there are any Giffen goods in the real world, although it is possible. In any case, it would not change the answer to question #12, their fears were groundless as the price rise reflected higher demand.
Johnleemk, Thanks.
Dan, Don’t feel bad, economics is very counterintuitive.
Current. I thought time was obvious in the question. I used the term “when.” But even if it isn’t as written, it is obvious when I presented the problem in class. I described the person going around and surveying theatres, one at a time. Obviously one cannot do all this simultaneously.
29. July 2009 at 14:03
Greg, I agree with much of what you say, except the last part. You are free to substitute any other product, and change the prices in any way you want, and change the quantities any way you want, and the answer will still be the same: NO!!!
There is nothing tricky about it at all. Any price and quantity combination is completely compatible with the laws of supply and demand. If you are told that gasoline prices rise 300%, and yet people buy twice as much gas as before, that is still completely consistent with S&D.
Rebeleconomist: You said:
“I disagree Scott. Unless you say that variables other than price and attendance are allowed to vary, and the law of demand is that, other things equal, demand decreases as price increases, then the answer should be “Potentially, yes. In the absence of changes of other variables, this example does breach the laws of supply and demand”.”
You are wrong. No other variables change, all that happens is demand shifts. Obviously if the price changes either supply or demand has shifted–that is a given for any question about supply and demand. If one hadn’t shifted, then there could not have been a price change in the first place. The laws of supply and demand do allow for prices to change over time, and they also allow for quantities to change. Do you deny this?
Jeff, I loved that study. That’s exactly my point (which most people aren’t getting.) The problem isn’t students, it’s reporters and economists. I had that study tapped to my door for months–I seem to recall that economists did worse on the (multiple choice) op. cost question than you’d expect based on chance.
Why bring it up now? Have you noticed how many people believe low interest rates are “easy money?” Maybe not understanding S&D caused the crash of 2008.
29. July 2009 at 14:11
I agree with StatGuy, that was the same thought I had. While Sumner is technically correct in that he did not say there were no other variables, the understanding by most students of any major is that questions are to be taken with all inputs seeminly implied as such, and to an extent without regard to possible external variables unless the question suggests otherwise. It’s simply a matter of perspective, but Sumner should be aware before marking the majority of his students’ answers wrong that they are conditioned to look at a simple academic question in the manner I just described. A matinee and a prime time showing are not valued the same, and nothing in his question implies there is any difference between the value of the two situations. Professor Sumner, I would ask you to consider wording the problem differently. You may find that it is not necessarily the case that students don’t grasp supply and demand, so much as within their experiences most professors asking a question worded as you did would expect a different answer from you. Even if you simply changed the wording of the final sentence to something like “Can you think of any situation in the real world where this would not violate the laws of supply and demand?”. You may get answers comparing matinee prices to prime time prices, you may get answers comparing theatres in different states, or even in different decades. If students took the liberties when answering this question the way you see as correct, and applied that same logic to other questions in other classes, they would much more often than not find themselves marked incorrect. I could give examples, but I think it’s easy to imagine such scenarios in any field of study. Now having said that, if the same students were faced not with a simple academic question but a real world problem outside the academic environment I further hypothesize they would be more likely to search for answers incorporating other factors. And now having said that, if more professors thought like you (and were both more careful with their words and more accepting of qualifiers within students’ answers) you would probably have found more students answering the question in the way you consider correct.
29. July 2009 at 15:07
Joe, That’s almost certainly right. I suppose one could argue that tea drinkers don’t care about getting sick and would switch over to coffee because it’s now cheaper, but that’s not very likely.
Stan, JKH, D. Watson, I agree.
JimP, Thanks, It is good that they noticed me, but I wish they had a good explanation for the policy of interest on reserves. I’ll take a closer look later. It might be worth a post.
Gordon, I agree.
Greyswan, Very well put.
D. Watson, Yes, those graphs are amazing.
JKH, I do agree with the Fed’s current explanation. They keep emphasizing that it is a contractionary policy that will prevent inflation from rising despite the huge monetary base. I’m completely with the Fed on that point. I agree that inflation won’t rise very much for the exact reason they say–the policy is contractionary. My only question is why was a policy so contractionary that it was capable of completely negating the effect of a doubled monetary base, adopted last October?!?!?
Nathan, Yes, the weird thing about supply and demand is that literally any P&Q combination is possible. Now if they tell you that something occurred that clearly affected one or the other (a frost hits Florida) then you can start making predictions, but not until then.
Joe, Great point. There are lots of regulations that drive up costs on the supply side. Excessive restrictions on who can do certain medical tasks, barriers on interstate competition in health insurance which drive up insurance rates, mandates that health insurance even cover things that many people don’t want to buy insurance for, etc. The point isn’t whether a certain medical issue is legit, it’s whether people want to buy insurance for it. You don’t insure every valid expense for your car.
Tim, Thanks for your kind comments.
29. July 2009 at 15:22
A lawyer with zero economic training would answer your movie question by writing, “not necessarily” and the lawyer would be correct.
If you requested more meat on the answer, the lawyer would send young associates on a long discovery process to develop the facts: are the theaters showing the movies in different geographic regions? Do the movies play in different local theaters, some more attractive than others (sound, chairs, parking, clientele, other amenities?) Do all the movies themselves have similar appeal to local cultural tastes, or are some disfavored? Do the movies play during different seasons? Different hours of the day? And through that investigative process, the legal method is to discover the proximate cause or causes for the discrepancy in price.
Then the lawyer would send you a bill for $180,000. Then, they would use all the developed facts for a series of strike suits based on OSHA, labor, employment, business practice and other violations by the theatre owners. And settle for another $180,000.
The economic method seems to this neophyte less profitable, it seems to jam the question into a model and accept or reject (hold constant) facts based on how they serve the model. Clearly that’s a wrong assumption, but I don’t understand your ontology. That’s why I visit, thanks for putting in the time on this blog!
29. July 2009 at 16:36
I prefer to teach S&D as having direct or inverse relationships with their determinants and prices. I’ve found it is a lot easier for my students to grasp the effects this way.
Question: A survey shows that on average 100 people go to the movies when the price is $6 and 300 people go when the price is $9. Does this violate the laws of supply and demand?
Answer: No, because the period of time is not stated and ceteris paribus is not assumed. If both the period of time was stated and ceteris paribus was assumed, this could still hold true if movies are a Giffen good meaning price and quantity demanded have a direct relationship.
29. July 2009 at 19:30
[…] I knew Mankiw was going to link to my last post, I wouldn’t have made it so meandering and confusing. So here is the mop-up, what I really wanted […]
29. July 2009 at 23:01
It intrigues me when such briliant economic-related article is on the internet. I love the example frommy economic lecturor, which is about a simulated trade zone, and we studnets were trading apples, only a few factors are involved: free to trade with eachother, the seller with the most profit and buyer with the most money unspended win.
My question is what can we get if we increase one or two factors or rules into the trade environment each time
There probably will never be a chance in my life for me to run this kind of economic experiment, So hope that you can keep my contribution in mind, and find the answer later on.
29. July 2009 at 23:04
Scott
Thank you for replying to the many posts here. I am pleasantly surprised.
You say I was wrong, but we agreed? Perhaps I didn’t word my answer well.
You also said the question wasn’t worded poorly and S&D is hard for students and even professional economists to understand. Seems that changing something is in order. If you have asked this question many times to students and they don’t get your point then what could you change to help them understand better?
Oh, and apologies for my poor attempt at humor in my comments above.
Fletch
29. July 2009 at 23:12
Good points all around. I wish I had a better econ professor in the supposedly prestigious university I studied at – mine didn’t even require that we buy a textbook (yup).
Oh, and it’s PPS, not PSS. 😉 Just a matter of whether you meant post-postscript or postscript-script.
30. July 2009 at 01:42
Perhaps you need to define the law of (supply and) demand here Scott. My understanding is that it includes a “ceteris paribus” proviso. If you invoke a “shift” in demand – I assume you mean a shift of the demand schedule – then why does this happen if it is not because a variable other than price and quantity changes? You mentioned that the time of day might differ between the two observations; is that not a change of another variable?
Essentially, I agree with Ian at 16.36 above, except that I would say “potentially yes”, because my default position would be to assume ceteris paribus unless told otherwise, but note that I explained this in my answer. That is why your students are confused; a reasonable person can give a no or yes (although a good student should explain their reasoning). Frankly, a remark like “you are wrong” in such a context is unscientific.
30. July 2009 at 03:00
> Often I read that the Fed made a huge mistake cutting
> interest rates to 1% in 2003, because it blew up a
> housing bubble. I thought interest rates fell in
> 2003 because investment fell.
The problem with understanding supply and demand is that it demands a market that mostly does not exist. That is why the vaguely worded questions test acceptance of fiction, rather than the derivations of the tautologies of the model.
The fed may have made a make a mistake cutting interest rates in 2003. There was a housing bubble.
Investment may have fallen in 2003. Interest rates fell in 2003 because the fed wanted them cut. The fall had little to do with supply and demand because, they are not set by supply and demand.
Did oil demand drive oil prices to $150 prior to the GFC and then suddenly disappear to make them fall to vastly lower level? Short term oil prices do not seem to be driven by normal supply and demand curves either, how about market manipulation and speculation?
30. July 2009 at 03:47
Fiveweight, Even if they assumed the products were different in some way, that would not change the answer at all. You might want to see my new post, which explains this a bit more.
Good luck, You might want to see my new post on S&D.
Ian, I don’t like the “cereris paribus’ approach here, because if other things were equal the price could not have changed in the first place. So the implication is that ceteris is not paribus.
Fletch, I have changed the way I teach in response to this question.
Catherine, Thanks, I knew it was PPS, I just forgot to proofread.
Rebeleconomist, You are probably confusing the law of demand with S&D. The law of demand says that ceteris paribus, a high price will lead to less consumption. There is nothing similar in of S&D. If there were, prices could never change for any reason. I have a follow up post that tries to explain this.
30. July 2009 at 04:49
Thanks for the response!
One way to find an answer when you are only given the outcome (changes in price and quantity) is to draw out S&D and see what has to happen to get that outcome.
Assuming we start out at P=6 with Q=100 as our original equilibrium point and P=9 with Q=300 as our new equilibrium point, demand must shift to the right (increase). Now if the original Supply curve intersected the point where P=9 with Q=300 then it will not shift, but if it did not intersect the new equilibrium point Supply must also shift to make this new point the equilibrium.
We can now explain these shifts in demand and supply(or movement along the supply curve which would be explained by the law of supply) with their respective determinants.
30. July 2009 at 05:17
I see; that is why I said that you should define what you mean by “the laws of supply and demand”. If it is what I think you mean, then either can shift and any (P,Q) combination does not necessarily violate the laws of supply and demand. For the same reason, I still think that “potentially, yes” is a valid answer, but “yes” would not be.
30. July 2009 at 11:02
Great post, it kept me up most of last night (that and the 100+ heat wave in Seattle :).
1) Even though it seems you were tongue-in-cheek when you wrote “Or maybe the identification problem is too hard, and we should simply forget about teaching supply and demand [and] devote the whole course to [other topics]”, I actually think a somewhat-less-exaggerated version of this would be swell. In my intro class I spend only the last 1/3 of the class on S&D. (The first 1/3 is individual optimization, the second 1/3 is game theory.)
2) My take on it is that S&D says one and only one thing: that market prices and quantities are determined by the intersection of two curves. I emphasize to my students (guilty caveat below) that market prices and quantities are always at the end of the story, and that the beginning of the story is S&D curves that are built up from preferences based on hypothetical questions like “If prices were xyz, how much chicken would you buy?” So reasoning from price changes is bad bad bad: a student who tells me that demand for chicken changes when the price of chicken changes does not understand what a demand curve is.
3) My guilty caveat is that I do talk about price changes for one good affecting the market for another good, e.g., defining two goods as substitutes when an increase in the price of one increases demand for the other. (I could say that two goods are substitutes “if an increase in the price of one caused by supply-side factors increases demand for the other” or “if an increase in the price of one increases demand for the other, assuming consumer preferences don’t change”, but if I used these phrases then it seems that I’d have to put similar phrases into discussions about, e.g., what happens to demand for chicken if the price of chicken increases… and I dunno, maybe you think that I should!) I suppose what I’m saying is that I’d like to know how you define substitutes and (if different than the usual) how you talk about what happens to demand for chicken if the price of chicken changes.
31. July 2009 at 12:53
artsrc, The oil market is one of the best examples of supply and demand in operation. Both the supply and demand for oil are very inelastic in the short run. Thus when there was a big increase in demand for oil from developing countries the prices soared to $147, and then when the world economy turned sharply lower the price also fell sharply.
Regarding 2003, I don’t agree that rates fell because the Fed cut them, rates used to fall when the economy was weak even before the Fed existed. The cause was a big drop in the business sector’s demand for loanable funds after the tech bubble burst.
Ian, That’s exactly right. When you draw the graph it is just so obvious that it doesn’t violate the laws of S&D.
Rebeleconomist, The reason the only answer is no is because of the way the question was worded. Does the information provided violate the laws of S&D? No it does not. There might be other factors that do, but not the evidence given.
Yoram, I put price in the middle. One curve shifts, then the price changes, than the quantity along the other curve adjusts because of the price change.
31. July 2009 at 16:26
I think an easy solution to the movies problem is simply to draw supply-demand curves and see what has to change in order that more tickets are sold at a higher price. This makes it obvious that the demand curve has to shift. There would seem to be no need of assuming anything at all about the real world (except that the signs of the slopes are the usual).
1. August 2009 at 05:50
Smoking Frog, Yes, you are right.
1. August 2009 at 10:35
The way I see things:
Teach incentives first in your course. Economics is simply the social science that treats humans as rational agents, and seeks to understand the incentives at play (the detective idea that was mentioned earlier.)
Then teach demand. (Just demand.) Teach what affects demand. (this will be very easy.) Then teach how price is a very powerful incentive. Draw one curve (the demand curve) on the board, and make price exogenous. Force a particular price with no explanation, and ask the students to consider themselves as the actor. I.E. if you went to the store, and suddenly XBOX’s were $2, would you buy one? Would you buy more than one? Explain that their demand hasn’t shifted, but that the price has. Then explain what happens when demand shifts and ‘everyone wants one.’ (an effective way to do this is to estimate a simple demand curve in class: take a survey, and ask how many people would buy an XBOX at $250, $240 $230 etc., down to zero. (not precise, but it’ll show the point.) Then ask what would make them willing to pay more for one?) Notice that this
is just a simply way of holding “all things equal.”
Then, explain supply, and the analogous principles (i.e. that suppliers produce more at a higher price, if price is exogenous.) Have a test, or a big quiz, that finishes the section. THEN teach the combination of the two–including equilibrium–and ask the question, ‘what causes price?’
At this point, explain that price is not exogenous, but is explained by either 1) a shift in demand or 2) a shift in supply. (or market inefficiencies, like government intervention–but save that for later.) Show what would happen if demand shifts in (i.e. the XBOX’s are poisoned. Would you still buy them at $2? What about -$50?) What if supply shifts out?
Last request: please write good tests. Don’t ask “when supply decreases, what will happen to price?” Your tests should test economic principles clearly, not the ability of students to understand what you wanted to say. As someone mentioned earlier, I would not have been able to answer the movie question either, because it is poorly framed. Ask clear questions.
1. August 2009 at 14:05
I hesitate to add to the mass of comments but wanted to say a) I really like the way you express the problem of analyzing price changes without additional information – it’s given me some ideas about how to make this clearer to my students so thank you! And b) I think your point about how often you see these mistakes made in the media is exactly why we need to teach S&D better to our students. I kind of love when I find quotes that say the laws of supply and demand have been violated (and they haven’t, but the writer is confused), because I think those are great examples to give students – I point out to them that if they can understand why the writer is wrong, they are on their way to being much more informed consumers of media “news” outlets. But I also agree with some earlier comments that these make frustrating multiple-choice questions for students – I prefer to use them as discussion starters or essay questions.
1. August 2009 at 14:07
Phrase your question differently:
Question: A survey shows that on average the price is $6 when 100 people go to the movies and the price is $9 when 300 people go. Does this violate the laws of supply and demand?
I think you’ll find that more people get the answer right.
1. August 2009 at 20:48
ssumner said, “Smoking Frog, Yes, you are right.”
I knew I shoulda been an economist! 🙂
Bill
1. August 2009 at 21:39
I apologize for not having the ambition to read all of the messages in this thread before I say what I’m about to say, but it seems to me that, for the movies question, some people may be missing the fact that “demand curve shifts to the right” is the only possible explanation (and since there is an explanation, the laws of supply and demand are not violated). Just draw the curves and try everything, and you’ll find that’s the only thing that works. So the question is analogous to a very simple algebra question or anything like that. It may appear not to be elementary, but if you just stick to the very basics, you see that it is.
2. August 2009 at 01:07
You seem to be confused indeed. 🙂 Please have a look at the related reddit discussion:
http://www.reddit.com/r/Economics/comments/96o5q/why_is_supply_and_demand_so_confusing/
2. August 2009 at 09:04
[…] Source […]
2. August 2009 at 11:30
To:
StatsGuy
28. July 2009 at 13:00
I completely disagree with your criticism of the movie question. I understood the import of it instantly and would have answered it correctly if I were one of Scott Summer’s MBA students.
The thing is, my advanced degree is the Humanities and I would never be one of Scott’s MBA students.
Because I wouldn’t be in an advanced econometric class, I simply accessed my personal experience of “100 people in a theater at $6 each” versus “300 people in a theater at $9 each” and the solution is obvious. It’s the difference between a matinee and evening showing of the same film.
2. August 2009 at 11:49
Benjamin, If you would not have been able to answer the movie question then you don’t understand supply and demand. There is literally no set of price and quantity data in any market of any kind that would refute the theory of supply and demand. If you think there is, then you don’t understand S&D. Block out movies and just think “product X” And pick any pair of prices and any pair of quantities, and the answer is always the same: NO. It doesn’t refute S&D. Not “it depends.”
Jennifer. Thanks, I used the movie question in class, not on an exam. I don’t want students to hate me!
Nathan, Thanks, I’ll try that. You might be interested in the first part of my next post, as I argue that a use of ‘consumption’ and ‘production’ can also confuse students.
Smoking frog, Yes, a few others noticed that too.
Hermel, Yes, I understood all of what you said about coffee, but I was really after a different issue, so I just left it for others to think about.
I hope professors aren’t avoiding products because they are giffen goods, as giffen goods probably don’t exist. In any case, the answer to that question would be the same even if it was a giffen good, they have nothing to worry about.
2. August 2009 at 12:10
Re: “Question: A survey shows that on average 100 people go to the movies when the price is $6 and 300 people go when the price is $9. Does this violate the laws of supply and demand?”
Your question implies that demand increased after prices went up. All else being equal this apparently violates S&D. However, if you worded the question differently and said something like: “the price is $9 when 300 go to the movies and is $6 when 100 want to”, all your students would get the correct answer.
2. August 2009 at 12:50
Scott B:
No it does not. Stand outside a given screen in a given cineplex and count the people coming OUT of the matineee and then those going IN to the evening showing. It’ll take you 45 minutes total and you won’t have to move.
That’s the “survey.”
It’s simply two data points. The causal factor is what you IMAGINED the question implied (“that demand increased BECAUSE the price went up” when in reality the question simply stated a straightforward empirical observation. The price of a movie ticket DOES go up between 5:00 pm and 7:30 and the demand for tickets also goes up.
The fact that you mistook what you imposed as causal for what the question actually asked is, um, pretty much the perfect example of what the article above is about.
2. August 2009 at 12:51
I think there are two things going on here.
(1) Distinguishing between movements in the S&D curves and movements up/down the S&D curves.
(2) The reason many students miss the question after learning #1 above has to do with which neural decision system they use to arrive at the answer. If they use the fast, intuitive system that depends on learned associations, they get it wrong. If they use the slow, hard-work system that depends on rule application, they should get it right.
DeMartino et al. (Science, Vol. 313, p. 684-7) had a very interesting paper on this in the context of Kahnemann and Tversky’s Asian disease problem. Subjects who were resistant to the preference reversal framing effect showed activation in precortical areas that subjects susceptible to the framing effect lacked.
2. August 2009 at 14:10
Scott –
I think that all you have demonstrated is that simple S&D problems (I am thinking of the movie question in particular) can be presented in a confusing way. I think we all knew that already.
Supply and demand analysis is only good for what it is good for: It is a method for analyzing the direction of change in equilibrium P & Q of a good after an event occurs. Logically, that causative event can not be a change in P or Q of the good in question.
We economists often teach in a way designed to intimidate or confuse rather than instruct. A good example is the practice of saying “it’s a change in demand, not a change in quantity demanded.” Ugh. Another example is this posting.
Mankiw’s practice of breaking the process down into three questions: “1) Which curve shifts? 2) Which way? 3) What’s the effect on P&Q?” makes it very clear what we are trying to accomplish and how it works. It also gives students practice in orderly thinking. It is eminently worth while to teach, and it is valuable and instructive about how professional economists do economics. Much more so than a vague and smug statement such as “incentives matter.” Vague and smug statements are how bad economists do economics.
I do agree with you, though, that it would be great to add a emphasis in chapter 4 on some detective exercises giving P&Q changes, and asking “which curve shifted?” I also always liked the champagne question and had not noticed its absence.
2. August 2009 at 14:12
I think part of the problem here is that economics has inappropriately tryed to model itself after physics. In physics you have absolute laws such as f = ma that are precise and hold for all conditions, and can therefore be used to make perfect predictions
Economic laws, on the other hands, are conditional truths. An example of a conditional truth would be the statement if you turn the ignition key in an automobile, the engine will start up. This is usually true, but there are innumerable conditions, such as adequate gasoline in the tank, a charge in the battery, clean spark plugs, and so on, that must be met if the phenomenon of interest is to occur. This is because the starting of an engine is the product of very complex causalities.
Economics phenomena such as prices are likewise the product of highly complex causalities, including many psychological, sociological, technological, and legal factors. And as a result expected phenomena may not occur, as in the various types of market failure, and so economic truths are conditional. As a consequence understanding of real economic phenomena requires analysing the various underlying causalities.
Alas, economists, in part for ideological reasons, like to pretend that economic laws are simple and absolute. And this of course leads to innumerable bad policy decisions.
2. August 2009 at 18:36
jfxgillis:
I knew the correct answer, but as a B-school professor myself, I would not have worded the question the same way, as the wording SEEMS TO imply a certain causal sequence that is not actually present.
Worded the way our host worded it, I would expect only the better students to get it right. Worded differently, most students would. (If the purpose of the question was to make that distinction among students, then it is fine, if not, you’ve entered non-random error into the question results).
Finally, I want to say that this was a great article and my picking nits by no means should be taken as critical od the overall analysis.
2. August 2009 at 19:18
Dear Scott:
My first reaction was that either: a) you are a really bad teacher or b) you don’t really understand supply and demand. However, you have been teaching this stuff for 27 years, and have a PhD in economics, so I find both of those to be implausible.
I don’t know what else to tell you; the question as stated (re: movies) is badly done. Fletch, StatsGuy, Tim, and a few others have it right: you have to differentiate between a shift in the demand curve and movement along the demand curve.
Look: It doesn’t matter how long you have been teaching, this *is* a trick question. It’s a trick question because one would naturally assume that “all other things are equal” holds. The “right answer” is reachable only if this doesn’t hold; “different times of day” means that things are not equal. If all other things are equal, then this result DOES violate the law of supply and demand … i.e., people value the higher priced movie more than the lower priced movie for no reason other than price.
Imagine the ticket taker, who says to each customer, “That ticket will be either $6 or $9, depending on which you would like to pay.” Clearly, it is *possible* that 300 will choose $9 and 100 will choose $6. But this is a non-standard, non-rational response; hence, violation of the law. If there is ANY other reason for choosing one price over the other – time of day, quality of movie, lack of information, whatever – then all other things are not equal, and the law of SD is not violated.
These concepts are easy if you explain them correctly. When you try to trick your students, and you don’t give them all the information, then you will not get good answers. When you don’t even realize you are trying to trick your students, then you won’t understand why the answers are bad.
However, it seems that you are trying to prompt the students to consider reasons why the demand curve would shift, changing the equilibrium price from $6 to $9. Perhaps you want to get them out of the habit of thinking “all other things are equal.” I don’t know why you want to do that, but as long as you understand that that’s the goal, then I’d guess that’s okay for an advanced econ class.
Cheers.
2. August 2009 at 22:37
In my opinion this clearly is a “trick question”.
In Switzerland (at least at the University of Berne) Macro is taught with the following basic assumptions or laws:
1. space: there is no distance
2. time: time doesn’t matter
3. goods: the considered goods are always homogeneous
4. persons: there are no preferences or dislikes
5. information: the actors know all relevant information
Unless stated otherwise by the interrogator the laws mentioned above are always true. If one of the laws is violated we don’t look at a perfect market. Your answer is obviously only true for a non-perfect market.
3. August 2009 at 02:02
“Your Econ confidence is inversely proportional to your Econ knowledge.”
I can relate to this. I’m getting my Ph.D. in economics I feel that with every passing year I am even less confident about answering seemingly simple questions.
I also find that this specific topic makes it incredibly frustrating as a TA to grade work for intro undergrad courses. Some professors expect it to be “obvious” that if the price is lower, quantity has gone up due to the law of demand, and this is the answer they judge to be correct, whereas others are trying to weed out the students who understand that you cannot make a statement about quantity solely based on an observed change in the market price. Quantity could’ve gone up, down, or stayed the same.
I find myself explaining to students, “well, the answer for THIS professor is …” Students do not find professor-specific explanations to be very satisfactory.
3. August 2009 at 02:17
[…] * Scott Sumner: Why supply and demand is so confusing. […]
3. August 2009 at 03:15
Scott:
I would say “imaginative” or “flexible thinking” rather than “better,” but either way, I’d say making that distinction among students is the purpose.
The larger issue of public discourse and/or general public understanding arises though. If most people, indeed most public commentators, if even most people supposedly “expert” in business by virtue of an MBA, can’t answer that question, that eventually filters up to public policy. That is because dolts expressing the S & D understanding of a typical high school junior dominate the media devoted to business and economics (CNBC, Fox Business, WSJ op-ed page, etc.)
When, as occasionally happens, some glibertarian accuses me of needing to take Econ 101, I usually reply that unlike them, I read chapter 2 and beyond in the text book.
3. August 2009 at 03:27
Ted:
WHY imagine that? That seems like a ridiculous scenario borne from a line of training and a way of thinking that encourages students to rigidly apply overly simplified rules to ridiculous scenarios. Why would you imagine such a silly, silly event when the survey described is surely applicable to the ordinary experience of everyone being asked the question?
Unless, that is, advanced economics and MBA students don’t go to movies.
3. August 2009 at 05:09
As a former econ instructor,
and one who skimmed this blog in response to being re-directed from mankiw’s blog, I can say that what my foggy memory recalls is that:
1. mankiw’s yogurt/ice cream example is an example of a “movement” along a given demand curve. Fragment aside, which amounts to a movement along a given demand curve. Its purpose is to illustrate reservation wages.
2. the health scare for coffee is to illustrate a “shift” of a demand curve or if you may, a change in demand.
two totally different concepts.
3. Whereas, the marketing campaign has a few effects. First, the campaign shifts the demand curve outward, as the perceived value of the product has increased demand. Whole concept of building brand image to seek supranormal rents…real world examples are plentiful.
No confusion in the logic of the Man’s text.
I stopped reading at this point though…
3. August 2009 at 09:29
Scott,
After answering these mind numbing econ questions for four years as an undergrad, I can tell you that the future lawyers in your econ class should have insisted that that question be thrown out upon further review. As your question is worded, the student must assume that the movie theaters are showing the same movie, at the same time, in similar quality facilities, on opposing sides of the street in the same town, with equal access to parking and identically priced popcorn. If you introduce variables (like the timing of the movies), you open the floodgates to any number of rational explanations. Oh let’s see…the $9 theater is showing Harry Potter in IMAX 3D and the $6 theater displays in regular def. The $9 theater is in a nicer part of town…or gives free parking…or free popcorn…or waiter service at your seat…or shows the newest movies…OR the $6 theater doubles as an adult movie theater with sticky floors. Whatever.
It’s a bs question, because you don’t explicitly control for the myriad economic and environment factors that govern behavior. No wonder most of your students get it wrong. You are arbitrarily choosing time as the differentiator when any number of answers are reasonable. Further, econ students are EXPLICITLY taught to control for these variables when answering exam questions. Given the wording of your question, a student would have to spin some yarn about giffen goods or other non-sense.
3. August 2009 at 10:31
StatsGuy wins the thread.
Causality. Causality, causality, causality. Yes, this is movement along the curve versus moving the curve, but the basic issue is that these questions are all phrased to obscure the causality involved. Two things happen. Are they correlated or coincidental? Are they causally correlated? If so, which way? The examples are all operating in a universe where some direction of causality is assumed but concealed.
It costs more to house a prisoner than to educate a student. Since cost reflects demand, which ultimately reflects quality, it therefore must be the case that prisons, being more expensive than schools, deliver a better product.
3. August 2009 at 10:34
Demand can go up because the supply curve moves (or changes shape) and demand adjusts. Or it can go up because the demand curve moves. Which is it? If you just declare two different demand quantities under two different conditions, which one of those actions caused the demand to go up?
When physicists, chemists, statisticians, or any other scientist or mathematician draws a curve, they make it clear which axis is a function of the other axis — x “causes” f(x).
Economists don’t do that, and thus they are constantly confusing correlation with causation.
3. August 2009 at 10:44
“what would happen to the demand for tea if there was a health scare regarding coffee. Obviously coffee and tea are substitutes. So if the health scare reduces the price of coffee then the demand for tea will…”
A coffee health scare *causes* a change in the coffee demand curve. A change in the coffee demand curve will, immediately, cause a change in the tea demand curve, because coffee and tea are substitutes. This will raise the demand for tea.
Yes, the quantity of coffee supplied will drop. But there has not been any information indicating that a change in the quantity of coffee supplied changes the *supply curve* of tea. Thus, the tea demand curve will move, but supply curve will not, and both the quantity demanded and price of tea will go up.
The thing being missed in any insistence on a differnet conclusion is that there is a demand curve for (coffee+tea) (since they are complimentary items). Anytime less coffee is consumed, more tea will be consumed, unless the supply curve of one of the two commmodities changes.
“Suppose that the price of frozen yogurt falls. The law of demand says that you will buy more frozen yogurt. At the same time you will probably buy less ice cream.”
Depends on whether the price of frozen yogurt fell as the result of the demand curve moving. If the frozen yogurt demand falls as a result of the frozen yogurt *supply curve* moving, then more yogurt (and less ice cream) will be consumed.
If it falls as a result of the *demand curve* moving, less frozen yogurt (and thus more ice cream) will be consumed.
The “trick” is not just counterintuitive, it is wrong.
3. August 2009 at 12:13
Scott B, I don’t want them to get the right answer, I want them to understand it. But I see your point.
jfxgillis, Thanks
Mersault, That sounds interesting. I will try to read it sometime. My next post addresses the framing effects in a slightly different way.
Strat, You said;
“We economists often teach in a way designed to intimidate or confuse rather than instruct. A good example is the practice of saying “it’s a change in demand, not a change in quantity demanded.” Ugh. Another example is this posting.”
That is exactly my point. I think we do students a disservice trying to get them to memorize quantity demanded and demand. They’ll never remember that. But they might remember Freakanomics.
What you described Mankiw doing is what I also try to do.
Eduardo, You said,
“I think part of the problem here is that economics has inappropriately tried to model itself after physics. In physics you have absolute laws such as f = ma that are precise and hold for all conditions, and can therefore be used to make perfect predictions”
One of my recent posts contests this. Physics cannot make perfect predictions.
Ted, All you are saying is that one could bring in additional information that violates the laws of S&D, but my question asked whether the info given violates those laws, and it clearly does not.
Swiss student, It is the correct answer for both perfect and non-perfect markets.
Michael, That is very funny, it might be worth a post.
Jay, If it is a movement along a demand curve you don’t say that people buy more yogurt when the price is low, you say that other things equal people want to buy more yogurt when the price is low. Big difference. otherwise the laws of S&D would imply that Americans would be expected to cut back on oil consumption when oil hit $147 last year.
brendan, None of the assumptions you made, nor any you could make, would change the answer, which is unambiguous.
What you say about economics is true, but it doesn’t affect anything I said in this post. I was discussing how students misinterpret the S&D model
eyelessgame, You said:
“Depends on whether the price of frozen yogurt fell as the result of the demand curve moving. If the frozen yogurt demand falls as a result of the frozen yogurt *supply curve* moving, then more yogurt (and less ice cream) will be consumed.”
I agree, this was exactly my point.
3. August 2009 at 13:53
Scott – I do see your point, but I’m afraid I’ll have to agree with StatsGuy that the wording implies causality. It’s as if I said, “Darn! I don’t like going to the movies when it’s only $6. I prefer to go when it’s $9.” This is an obscure way of saying I prefer going at night, and would probably confuse people.
However, I was very happy to see your follow-up comments on teaching strategy. It seems you try to teach concepts rather than rules; examples rather than abstractions. This is exactly the right way. Richard Feynman once gave a great talk about his experience in Brazil, where he found the students excelled at memorizing physics formulas, but had no idea what the formulas meant or how to apply them.
eyelessgame: I agree … StatsGuy wins the thread. And it allows me to post one of my favorite cartoons on causality: http://xkcd.com/552/
jfxgillis: yes, it is a ridiculous example. It is hyperbole, one of the best ways to focus on a key concept. Here, the key concept is “all other things being equal.” I couldn’t think of a better way to express this idea with the movie example, in a way that preserves – as much as possible – perfect information. Note that the only ways for someone to choose $9 over $6 are: a) they don’t have perfect information; i.e., they think the $9 movie is somehow “better;” b) they get more pleasure out of paying more for the exact same good; or c) they violate the law of S&D.
I will, however, concede the point that the law of S&D cannot be violated, if once considers the following: Any reason, no matter how ridiculous or irrational, for choosing to pay $9 instead of $6, means that paying more fulfills some need for the person paying. Either they prefer paying more (fills a need) or they think they will prefer it (imperfect information). [Or, I suppose, they don’t understand the question, which is also imperfect information.]
They may even be trying to prove a point, that the law of S&D can be violated. In which case, it is “worth” $3 extra dollars to prove this point, and the law is thus preserved. So it’s kind of a catch-22 situation.
But I don’t think there is a better way to logically equate the two choices as being completely equal in every way. Let me know if you think of one.
3. August 2009 at 16:29
This from an Econ 101 drop out:
The struggle I always had with Economics, best captured in the S&D question you ultimately pose, is that Econ is taught very well when taught in narrative form but is absolutely inscrutable as a “language”. Perhaps I’m a nincompoop, which is likely, but consider Exhibit A: Your opening example. In it, you mention neither supply nor demand in clear terms; demand is inferred relatively easily, but less so is supply.
In study, I often found that Economists often relayed information by referring not to other things which may mean something to the uninitiated. Rather, they referred back to the principles themselves. It’s linguistically an endless recursion, thus not immediately intelligible.
Yeah, so I’m the dumb guy, the rube. So be it. Bottom line — there’s a reason why so many get through Economics without actually learning it. And I suspect it’s the same reason that Freakonomics was / is so successful where 101 fails.
4. August 2009 at 05:34
The post illustrates that the way economics is taught tends to ignore CAUSALITY. Things happen for reasons, this sets the chain of events in motion.
In other words, students (and journalists, and sometimes economists) confuse correlation and causality. If A changes then B happens. Yes, but that’s just a correlation. B could have equally be the cause of change A.
4. August 2009 at 07:05
Ted, You said;
“Scott – I do see your point, but I’m afraid I’ll have to agree with StatsGuy that the wording implies causality. It’s as if I said, “Darn! I don’t like going to the movies when it’s only $6. I prefer to go when it’s $9.” This is an obscure way of saying I prefer going at night, and would probably confuse people.”
I still say that the problem is not the question, it is the readers. Otherwise why would people always say things like a rise in the dollar will hurt exports, or a rise in interest rates will hurt investment. I still believe that the problem is not the wording of the question, but a misunderstanding of supply and demand. There is no assumption one could make that makes the answer any different from “no.” The theory of S&D says that price changes mean absolutely NOTHING, and I still think many people don’t see that.
Ted, You said;
“jfxgillis: yes, it is a ridiculous example. It is hyperbole, one of the best ways to focus on a key concept. Here, the key concept is “all other things being equal.” I couldn’t think of a better way to express this idea with the movie example, in a way that preserves – as much as possible – perfect information. Note that the only ways for someone to choose $9 over $6 are: a) they don’t have perfect information; i.e., they think the $9 movie is somehow “better;” b) they get more pleasure out of paying more for the exact same good; or c) they violate the law of S&D.”
Now I am completely sure that you are confused about S&D. This statement is simply false. You forget the most obvious explanation—the demand curve shifted! According to the theory of S&D, every change in price reflects a shift in S or D. Why in the world couldn’t D have changed? If price changed, we know there is a 50-50 chance it was S and a 5o-50 chance it was D.
Michael, I think you are agreeing with me, but I can’t quite tell. My points were:
1. Students shouldn’t feel dumb if they don’t understand S&D, because many experts also do not.
2. We teach it the wrong way.
3. Maybe we should just use Freakonomics
I think those are your points, but I’m not sure.
4. August 2009 at 19:22
Hi, Scott. You said:
“Now I am completely sure that you are confused about S&D.”
Sorry for the confusion … I was talking about my reformulated example at that point, not your original one. I don’t think a shift in the demand curve can explain people choosing to pay $9 instead of $6 for the exact same movie under the exact same conditions. jfxgillis thought it was a silly example, and I was merely pointing out it was silly for a reason.
You also said:
“I still say that the problem is not the question, it is the readers. Otherwise why would people always say things like a rise in the dollar will hurt exports, or a rise in interest rates will hurt investment.”
I don’t have a problem with the statement that a rise in the dollar will hurt exports. Obviously, fewer goods will be sold at a higher price point than at a lower one. And if all American goods are more expensive, due to a stronger dollar, then replacement goods from other countries will displace some of our exports.
The key, again, is “all other things being equal.” Sure, exports MAY happen to go up as the dollar goes up – it may even be the REASON the dollar goes up. Nevertheless, a stronger dollar exerts negative pressure on the quantity exported, simply because our goods are more expensive for the rest of the world than they otherwise would be.
You may say: perhaps the dollar is going up because American exports are becoming more popular … a shift in the demand curve. But I’m talking about a situation (“all other things being equal”) where there is NO shift in demand. In this case, more expensive goods = fewer goods sold. That’s what a demand curve represents.
And frankly, if the dollar gets stronger, then that pretty much means the supply curve has shifted. At least, that is the most reasonable assumption. Other things can certainly mitigate the downward pressure on exports that a shifting supply curve represents; but such a shift will ALWAYS put negative pressure on the quantity of goods sold. Honestly, I don’t see the problem here.
Rising interest rates make it more expensive for businesses to borrow money. That *does* put downward pressure on investment. It is a clear first-order effect. There may be other things going on as well, and those other things may mitigate or even outweigh the first-order effect. But investments have become more expensive, and you shouldn’t ignore that in your calculations.
Still, I am enjoying the exchange, so thanks for that.
5. August 2009 at 10:57
Ted, You said:
“Sorry for the confusion … I was talking about my reformulated example at that point, not your original one. I don’t think a shift in the demand curve can explain people choosing to pay $9 instead of $6 for the exact same movie under the exact same conditions. jfxgillis thought it was a silly example, and I was merely pointing out it was silly for a reason.”
When the Chinese bought more cars there was an increase in demand for gasoline. That increase in demand could easily cause consumers to buy more of the exact same 87 octane gas at a higher price. The same is true for movies. If they is a shift toward more demand for movies (say due to higher incomes), then people may want to see more of exactly the same type of movie at a higher price.
You said:
“I don’t have a problem with the statement that a rise in the dollar will hurt exports. Obviously, fewer goods will be sold at a higher price point than at a lower one. And if all American goods are more expensive, due to a stronger dollar, then replacement goods from other countries will displace some of our exports.”
I should have worded it more clearly. I meant that often people are predicting that if the dollar goes up there will actually be fewer exports, and they wrongly think this prediction is based on S&D theory. You are right that ceteris paribus a higher dollar will cause firms to want to export less. But many of the errors you see are clearly not making a ceteris paribus claim, they are making a prediction.
You said:
“And frankly, if the dollar gets stronger, then that pretty much means the supply curve has shifted.”
This is wrong. There might be more demand for US goods. Perhaps a high tech boom. Or consider when a country discovers oil, both its currency and its exports rise.
On your final point, I don’t see a change in interest rates as a first order effect. The first order effect is the thing that caused interest rates to change. And that is usually a shift in the demand for loanable funds. That is why when interest rates rise investment usually rises, rather than falls.
Suppose you could look into a crystal ball and just see interest rates. I give you two options. Interest rates in three years are 1%, or 5%. If that’s all you knew after looking at the crystal ball, in which case would you forecast higher investment?
5. August 2009 at 14:59
The typical S&D plot has Price on the vertical and Quantity on the horizontal. This is practically mathematical shorthand for dependence and independence. If you want your students to intuititively consider the inverse, switch axes every other week. At first they’ll be nauseated, but they’ll get the point after a while.
5. August 2009 at 19:46
Hi, Scott. I think we agree on the dynamics of supply and demand, and disagree on the semantics of econ test questions. Which is fine, now that I know where you are coming from … we seem to be focusing on different things: I point out “if ceteris paribus holds” and you point out that we shouldn’t assume it holds.
I agree with both. I would add that my feeling is, ceteris paribus should be assumed – particularly in econ test questions – as a first step in analysis. After the direct effects are calculated (or taught), then other factors can be layered in to get the whole picture.
You said: “The same is true for movies. If they is a shift toward more demand for movies (say due to higher incomes), then people may want to see more of exactly the same type of movie at a higher price.”
Yes, but that violates the “same conditions” statement I qualified my example with. We’re just talking about two different things, I think.
You said: “I meant that often people are predicting that if the dollar goes up there will actually be fewer exports, and they wrongly think this prediction is based on S&D theory. You are right that ceteris paribus a higher dollar will cause firms to want to export less. But many of the errors you see are clearly not making a ceteris paribus claim, they are making a prediction.”
I agree completely. Reporters, etc., should emphasize this situation is a concern – downward pressure on exports – not a prediction. The smart ones do; the lazy or confused ones simply say exports will go down when the dollar goes up, which is just bad reporting.
You said (first quoting me):
“”And frankly, if the dollar gets stronger, then that pretty much means the supply curve has shifted.”
This is wrong. There might be more demand for US goods. Perhaps a high tech boom. Or consider when a country discovers oil, both its currency and its exports rise.”
I may be confusing myself here. I understand that an increase (shift) in demand moves us along the supply curve. However, I don’t see how a change in the value of a dollar can be anything other than a change in the supply curve, regardless of whether or not demand also shifts.
Here’s an example: Say there are 20 people in the US willing to export X at $20 per unit, and there are 15 people willing to export X at $18 per, and no one else makes X, and the value of the dollar suddenly drops by 10%. Then – to everyone outside of the US – the supply curve for X just shifted.
I guess you could say: relative to the US folks, the demand curve is the one that shifted. That’s true. But my feeling is, if you’re talking about a fall in the dollar, that’s the wrong way to look at it. Otherwise, you have to have a S&D graph with a variable y-axis.
You said: “On your final point, I don’t see a change in interest rates as a first order effect. The first order effect is the thing that caused interest rates to change. And that is usually a shift in the demand for loanable funds. That is why when interest rates rise investment usually rises, rather than falls.”
Again, the difference seems to be semantics, or at least emphasis. I didn’t say a change in interest rates was a first-order effect; I said that downward pressure on investment is a first-order effect of the interest rate changing. There can certainly be things that happen that *cause* interest rates to rise, but a decrease in the demand for loanable funds isn’t one of them.
So I would say that your third sentence gets the causality backwards … an increase in investment puts upward pressure on interest rates, but not the other way around. An increase in interest rates puts downward pressure on investment, but not the other way around. If interest rates and investment both go up, something else is going on.
You asked: “Suppose you could look into a crystal ball and just see interest rates. I give you two options. Interest rates in three years are 1%, or 5%. If that’s all you knew after looking at the crystal ball, in which case would you forecast higher investment?”
I would say, all other things being equal, 1% interest rates means higher investments than 5% interest rates. But since we don’t know whether or not ceteris paribus holds, we can’t make any prediction as to what is going on here.
However, we can probably assume ceteris paribus doesn’t hold between these two scenarios. Since the Fed will often lower interest rates in order to spur investment, and because 5% represents a moderate-to-high federal discount rate (moderate historically; high in the current environment) and 1% represents a low discount rate, I would guess that a 5% rate is intended to taper down investment and mitigate inflation in an overheating economy.
Therefore, I would *guess* that the level of investment is higher in the 5% scenario. However, even if we can CONCLUDE that investment is higher under 5% interest rates, we should never say that investment is higher BECAUSE of the high interest rates. That would imply a causal relationship that doesn’t exist.
6. August 2009 at 04:08
[…] post on supply and demand produced a lot of good comments, but none better than this one from Michael: I can relate to this. […]
6. August 2009 at 07:07
tg, Yes, that might be a good idea.
Ted, You said;
“Yes, but that violates the “same conditions” statement I qualified my example with. We’re just talking about two different things, I think.”
You can’t assume ceteris paribus with a price change if you are trying to explain how quantity changes. If ceteris paribus held then the price could not have changed. Ceteris paribus is used to explain why demand curves slope downward and supply curves slope upward.
You said;
“Here’s an example: Say there are 20 people in the US willing to export X at $20 per unit, and there are 15 people willing to export X at $18 per, and no one else makes X, and the value of the dollar suddenly drops by 10%. Then – to everyone outside of the US – the supply curve for X just shifted.”
Never start an analysis with a price change. The dollar never just drops. It drops because of less demand or more supply.
You said:
“Again, the difference seems to be semantics, or at least emphasis. I didn’t say a change in interest rates was a first-order effect; I said that downward pressure on investment is a first-order effect of the interest rate changing. There can certainly be things that happen that *cause* interest rates to rise, but a decrease in the demand for loanable funds isn’t one of them.”
I never said a decrease in the demand for loanable funds would cause interest rates to rise. It would be an increase in demand for loanable funds.
7. August 2009 at 19:50
Scott,
Isn’t the answer to Question 12 more simple than that? The answer is that the executives are confusing demand and quantity demanded. If they think the higher price of champagne will decrease demand for champagne, they are confusing the fact that price changes don’t affect demand, only the quantity demanded.
8. August 2009 at 05:50
Enrique. That’s right.
9. August 2009 at 11:09
I think there is something tricky, or at least unclear, about the way your movie question is worded. If it wasn’t a test question, I would have read it as saying that someone took a survey asking if people would go to the movies at particular prices and more would go at a higher price. Now, when it’s a “does this violate some principle” question, you obviously can’t only use your first interpretation. But my reaction as a test taker would be “I don’t really know what he’s getting at, so I’m just going to list a bunch of true things.” You might be asking if I know what demand shocks do to P and Q, or you might be asking about whether Giffen good violate the law of supply and demand.
10. August 2009 at 06:00
Kyle. No the question described a survey of real world movie theaters, it was not a hypothetical question. Nor was it in the slightest way “tricky.” The only possible answer is no, under any set of assumptions.
12. August 2009 at 07:06
[…] iktisat öğreteneler ve iktisat öğrenmeye çalışanlar! Scott Summer’ın yazısını okuyun. Greg Mankiw’in yazıyla ilgili notuna da bakın. Comments [0]Digg […]
13. August 2009 at 03:52
Mankiw’s two questions “specify which ceteris are paribus” (dWj, above) by considering a price shift from the perspective of (respectively) consumer and producer.
In the frozen yogurt question, the “you” is the average consumer. The price falls (not because of a yogurt bacteria scare that would decrease his individual demand, but for reasons unrelated and unintelligible to him) and he buys more. Conversely, the champagne question asks one to view the situation from the perspective of the industry executives. The reason for the price increase that makes them “giddy” is obviously not a large-scale failure of champagne makers, but an increase in consumer demand.
Your three example questions mislead as to the cause of the price shifts by implying that consumers are reacting to a shift in supply. The answers that they suggest are correct in supply shift situations – which, as you point out, the current moment is not. The applicable questions for the current situation would be, “If interest rates fall sharply we can expect lending to … ” (decrease), and so on.
A scenario that specifies a perspective of consumer or producer implies that the cause of the price shift is exogeneous (to use StatsGuy’s terminology) to that party. If perspective is removed, the direction of the consumption shift in response to a price change is indeterminate, as you say. Alternatively, if considered from both perspectives simultaneously – if the price change is exogeneous to both producer and consumer – disequilibrium results, with divergent quantity supplied and quantity demanded.
In sum, the questions are not incorrect, but they do assume that the market responses – supply or demand curve – of the party in consideration are paribus. Generalizing the questions to situations in which that is not the case causes the errors you discuss.
13. August 2009 at 07:52
Dan, You said;
“In the frozen yogurt question, the “you” is the average consumer. The price falls (not because of a yogurt bacteria scare that would decrease his individual demand, but for reasons unrelated and unintelligible to him) and he buys more.”
The distinction between the average consumer and total consumption is not relevant here. Obviously if the average consumer consumes more then total consumption will also rise. And if total consumption rises then the total quantity supplied will also rise. Does a lower price cause an increase in the quantity supplied? I don’t think so. this confusion explain why newspapers often say silly things like “consumers are buying ever more oil, despite record high prices, in apparent violation of the law demand.”
I don’t agree with your comments on the three questions I posed. There are completely neutral. The reason why people get so confused is that they think the terms are not neutral, Thus if they hear “oil consumption increased” they mentally form an image of demand rising, and if they hear “oil production increased” they mentally form an image of supply increasing. Neither assumption is valid. But people make those assumptions all the time, and hence get confused by supply and demand.
If interest rates fell sharply I’d expect investment to fall, and I’d be right far more often than those who would expect investment to rise.
18. August 2009 at 13:08
[…] Scott Sumner’s post in which he tries to answer this question is well worth contemplating. Greg Mankiw also weighs in. […]
15. March 2010 at 08:12
[…] know that Chris knows this, but I’ll make the point anyway, because Scott Sumner taught me to: interest rates reflect supply AND demand for investment funds. So, of course, a […]
3. April 2010 at 20:38
[…] the equilibrium of supply and demand is a very confusing concept. Indeed, Scott Sumner has mused on this subject. However, while Scott focuses upon the methods by which we teach supply and demand, I want to take […]
22. April 2010 at 10:10
Scott,
were you aware of the textbook “The Economic Way of Thinking” by Heyne et al. ? It is the most non-technical intuition that you can get, I guess. They do argue from price changes of inputs and subsitutes, but not the good itself, and always make the distinction clear between supply (the curve) and quantity supplied (the actual amount) etc.
Maybe you find that a better book for your courses.
Steve
29. April 2010 at 05:19
Steve, Yes, I have heard of it, but haven’t read it. I will try to check it out sometime.
19. May 2010 at 10:24
[…] a comment, steve (not co-blogger Steve) reminded me of a very good post by Scott Sumner that illustrates the endogeneity of prices with respect to quantity. It turns out I had read it, […]
14. December 2010 at 15:37
[…] in den letzten Jahren so viel gelernt habe, wie Scott Sumner, den Autor von TheMoneyIllusion. Hier ein sehr guter Eintrag zum scheinbar trivialen Zusammenspiel von Angebot und Nachfrage, den man gleichzeitig als […]
28. November 2018 at 08:09
Demand can go up because the supply curve moves (or changes shape) and demand adjusts. Or it can go up because the demand curve moves. Which is it? If you just declare two different demand quantities under two different conditions, which one of those actions caused the demand to go up?
When physicists, chemists, statisticians, or any other scientist or mathematician draws a curve, they make it clear which axis is a function of the other axis — x “causes” f(x).
Economists don’t do that, and thus they are constantly confusing correlation with causation.
13. June 2021 at 05:42
[…] Sumner makes the same point in one of his […]