A great example of "deadweight loss" are restrictions on surge pricing for ride share companies. No surge pricing to get drivers on the road=no ride home from the ballgame for me.
Great explanation of deadweight loss. I’d add the concepts of fixed costs vs. variable costs as a contributing factor to why suppliers decrease quantity with a price cap.
* Fixed costs: Costs a supplier pays regardless of quantity sold
* Variable costs: Cost per-a-unit of production
In the Popeyes example
* Fixed: Lease, maintenance, wages of a minimal crew
* Variable: Ingredients, wages of additional workers at busy times
With the price ceiling, Popeyes could find that sales after 10 PM doesn’t provide enough revenue to offset the costs of the minimum crew needed to operate a store. They could also find that some locations should be closed since they don’t bring in enough revenue to cover their fixed costs regardless of hours of operation.
This is a technical quibble but in producer theory we learned that in the short run a business should decide whether to operate or not based on whether revenue covers variable costs.
Getting off topic here, but actually the point it APPLES ARE NOT CANDY. I kind of doubt that candy apples were ever a popular thing to give trick or treaters, since making them and wrapping them involves a lot of time and effort and expense—the issue was just plain apples, which some adults (you know the kind) saw as a “healthy” alternative to candy. The story about the razorblades arose from word of mouth, started to show up in local media and warnings in classrooms, and was probably an urban legend invented by children to ensure that trick or treaters got proper, manufactured, sugar-loaded CANDY, not fruit.
The urban legend only makes sense if it’s a candy apple, though. If you stick a razor blade in a plain apple, there will be a big slit in the side of the fruit, surrounded by a bruise.
Moreover, after you jam the razor in, a little corner will still remain sticking out, and give the game away. You need to cover the whole thing in candy if the scheme is going to work.
Scary stories don’t need to be that accurate. Most urban myths don’t stand up to the slightest bit of scrutiny, but that doesn’t keep them from spreading.
While perhaps not as venerable as some of the commentariat, I was around for Halloween before that story became part of the cultural milleu, or at least before it was widely heard, and have never, ever seen a candy apple on Halloween. And my buddies and I were the ones who used to blitz like 5 different subdivisions with bikes and a wagon.
I have absolutely seen people giving out fruit, vegetables, water bottles, and pretzels. The last two were often a welcome break from binging on sugar while foraging for more sugar, though.
I think that the issue is that it’s not obviously a situation with “transactions” per se. But beyond that, the change described is more of a shift in preferences than the imposition of a price or quantity control. It’s more than a little unusual to use DWL to describe a change in quantity because of a voluntary shift in demand or supply. I can see how you get there by squinting and saying that the shift wasn’t voluntary because it was premised on a moral panic or bad information, but that’s really a stretch. Generally, consumer theory doesn’t have a carve out to specify which preferences are legitimate.
So, I get the joke, but I would expect to only give it partial credit on a test question asking for an example of deadweight loss. I know some of my own professors would have just marked it wrong.
No, we shouldn’t over analyze jokes, but once you write an explainer based on it…
The deadweight loss is the loss to all those kids who like candy apples *even with* a chance of razor blades in them! The regulators (parents) who insist that you should only take candies with professional wrapping are causing loss to both the kids and treat-givers who would voluntarily exchange at the risky margin.
Had to laugh at this. Read the whole 2,000 words and had no idea what 'deadweight loss' is. Some people like some things and others like other things. Is this what the economics profession has been scratching their heads over? Econ speak for gibberish? Pretty sure your deadweight loss is my gain if what you dislike is what I like, and your calculation of your disutils versus my utils is proof economists are full of it and have no idea what 'deadweight loss' is or why we should care about it. My advice is stop reading Greg Mankiw and get on with your life.
I teach this stuff to college freshman, and I think he did a very good job explaining it. But, like, I probably could have done it better, but my job isn't writing blog posts.
DWL is a valuable concept. But it's worth being skeptical of the way it's quantified on a supply and demand plot, especially when (though the post does not do this) economists start referring to it as a measure of "welfare."
Willingness-to-pay is a very convenient proxy for welfare, but I think most people would acknowledge that its a pretty bad one. The fact that I was willing to pay up to $100 for an uber ride to the airport yesterday (I actually paid $80) while the homeless guy I passed probably wasn't willing to pay anything for it says basically nothing about how much welfare that uber ride produced. But on the supply and demand plot, my willingness-to-pay gets represented by a point along the demand curve at P = 100, while the homeless man's willingness to pay gets represented as a point a P = 0.
This isn't a profound or novel insight; plenty of economists acknowledge the issues with market valuations of welfare. But I think it's worth having in a comment here given all the people likely to see this post.
It occurs to me that a lot of ways of cutting inflation would be really destructive. Price controls on basic consumer items that are produced in competitive markets would certainly cut inflation, but would also result in constant shortages. A rationing system would also cut inflation but only by forcing people to reduce the amount of stuff they normally buy.
There are lots of ways charlatans can screw things up for everybody with daft policies. Price controls and rationing do have a use in certain circumstances, but often when there has been a policy failure elsewhere.
Whenever someone says "this is my number one priority", then I always think "so, if you could achieve this by destroying the entire human species in a nuclear war, then you would?"
It's surprising how many problems you can solve by destroying humanity: crime, for instance.
The very general rule should be to always try to increase supply before resorting to reducing demand. There will of course come times where the former isn't practical in a reasonable timeline.
During the world war II, in the US, there were very restrictive price controls on everything because the government was attempting to redirect all output to war production. It worked - the US was the arsenal of democracy, outproducing pretty much the entire world in war production** - but the black market thrived. Post-war, while price controls remained in force, the price of new cars stayed the same, so dealers came up with assorted hacks to charge extra for various essential car pars, like the steering wheel. If you bought everything you needed the combined price would wind up being equivalent to the market price. Economy-wide price controls suck, but they work (which is why communism survived in the USSR as long as it did). Price controls on a very limited set of goods or a market segment work, but they can be quite costly. They can also be used to neutralize the effects of monopolistic pricing - but that's fixing the effects of an existing monopoly - which is the first-order problem.
elm
it depends
** The US essentially outproduced the entire rest of the world combined (ex the USSR) in military goods
This was also when the original sin of the US healthcare system was created. Because of wage controls, employers paid for health insurance as a work-around and the IRS declared those benefits to be tax-free to the employee. The resulting 80 years of a lack of a functioning market for healthcare has led to price spirals, opaque pricing, backdoor rebates and limited supply.
Any sort of initiative to push people to take a cash subsidy from their employer and go buy healthcare coverage from the Marketplace would be an immediate step up. It would start untangling the ratfuck of mis-aligned incentives that the healthcare sector outside of actual providers is shot through with, albeit not yet the wildly mis-aligned incentives among providers themselves.
Preferential tax treatment needs to end, preferably by ending deductions for employer-sponsored coverage, but more likely by creating one for personally-purchased coverage, lol.
As I said, likely easier to extend to all personal health insurance expenditures and allow companies to directly fund an FSA-equivalent for purchase of insurance than to wind it up.
We can always make up the revenue elsewhere, it's the unlevel playing field that's killing us.
"Economy-wide price controls suck, but they work (which is why communism survived in the USSR as long as it did). "
In the case of WW2 there was a huge positive externality to the rationing/controls - "win the war" - that could not easily be done on any individual level, so the deadweight loss incurred could be beaten out by the rewards from winning the war. But that's pretty atypical.
I don't see that they helped communism in the long run.
>>Economy-wide price controls suck, but they work<<
If combined with rationing, yes, otherwise the shortages will be intolerable. Also, it helps if the cause is popular, so that the policy is enforced in part by societal disapproval of cheaters.
Those regulators, insisting that *every* car have a good steering wheel, are causing deadweight loss for those producers and consumers who would voluntarily transact to exchange a car with a bad steering wheel for a low price!
"And to understand those concepts, you need to understand a really basic supply and demand chart with the price of widgets on the vertical axis and the quantity of widgets sold on the horizontal axis. The line for demand slopes down, and the line for supply slopes up "
Ok, as a scientist and math minor, this kind of graph abuse is what makes economics basically incomprehensible to me and everyone else. You're not using "chart" to mean the same thing we do - you can't, because a graph with "price on the vertical axis" and "quantity on the horizontal axis" can't have two lines. There's only one line: price, as a function of quantity. That's what it means for the chart to have axes - they define your dependent and independent variables, and the line you graph in the coordinate space is the relationship (the function) between the axes.
You're doing something else with your chart, clearly, but since it's so radically different from how data is typically charted and you don't explain how to interpret it, it's incomprehensible and that's where I fall off the rest of your post.
True. But you need to write to your audience. It seems these graphs weren’t sufficiently explained and many people here unfamiliar with the concepts didn’t understand them.
I don’t know how many highly-technical classes you’ve taken Milan, but it might be worth mentioning here that the standards economists use to draw these graphs are much less rigorous than you would ever find in, say, mathematics or engineering. The people most confused seemed to have more technical backgrounds where you wouldn’t see a graph like that.
Milan, in time you’ll learn that one is never the best judge of the legibility of their own work. A professional scientist just told you your presentation of data is incomprehensible to them. Your post is supposed to be geared to a general audience, not to those already in the know. Perhaps it is the wiser course to consider that there might be something to the critique ? And at any rate not to be so quick to dismiss it ?
As a counterpoint - I'm a professional scientist, and while I think that sentence could have been edited for clarity, the idea of "two functions, f(x) and g(x) that have the same units and can be plotted on the same graph" really isn't that exotic or alien to scientific fields. Off the top of my head, "mixed potential theory" uses the exact same kind of graph.
That’s kind of besides the point I’m trying to make here. It’s not about the standards of this or that field. It’s about writing to a general audience and learning to listen to feedback, especially with regards to issues of clarity and what conventions can or cannot be taken for granted when addressing such audience. I mentioned that op was a professional scientist as a “a fortiori” move. Frankly Milan’s answer would have been equally unhelpful had op just been Joe shmo subscriber- it’s the latter that the post is meant for after all.
I mean, you can criticize the tone of the response but there is no other way I can think of off the top of my head to relay the information.
About the only criticism I have often heard repeated about the Econ 101 supply-demand charts is that they reverse the standard for dependent and independent variables on the X- and Y-axes. Which is... debatable. There is nothing to prove that price is the independent variable except blind intuition, and not even always that.
Certainly, the two lines criticism makes *no sense at all* within any discipline's framework; they're two different functions plotted on the same graph. Full stop. This is completely and utterly ordinary for anyone whose work involves math and I cannot imagine why someone regarded it as unreasonable.
OP says they don’t understand the chart. There are multiple ways to respond but I would think trying to explain the chart to them in other words might be a good thing to include in the response, and being dismissive might not. The discussion of other commentators here, some likewise admitting their confusion, others patiently trying to elucidate this for them, offers a good model, imho.
The initial line of criticism "what's the other line?" seems just completely ridiculous to me, having dealt with graphical representations of functions in a professional capacity, yes. Like, seeing two functions on the same plot is not the least bit unusual.
Counterpoint: If someone has trouble following supply and demand curves, then the concept of deadweight loss will be beyond their comprehension. And I shudder to imagine a professional scientist who is ignorant of the concept presented in Milan's chart.
I have to say, I had a hard time making sense of the chart. Usually I’m someone who prefers visual information, but I found the written explanation much easier to understand, and in fact had to entirely depend on that explanation to make head or tail of the chart, which makes the chart kind of superfluous.
Supply and demand charts are deceptively complicated things. I TAed for for undergads this past spring who still made basic mistakes with them 4 months later.
But they're also incredibly powerful tools once you get the hang of them. If you're interested, you might try googling "marginal benefit interpretation of demand curve" or "marginal cost interpretation of supply curve" - that'll get you into the reasoning underlying these charts, which I think is the most intuitive starting point.
Ok, maybe you can explain it. There's two axes, one dependent and one independent. One of the lines is the relationship between price and quantity. What's the other one?
One, demand, shows the price required to induce consumers to buy a given quantity. The other, supply, shows the price required to induce producers to provide a given quantity.
The real issue to complain about is that it’s much more intuitive to think about price (y-axis) being the independent variable. It can make sense both ways, but many students initially have trouble seeing how price can be a function of quantity.
The short answer is it depends on the market. In general, price is closer to being the independent variable, and you assume a latent utility function of the marginal consumer and producer yielding the equilibrium price, amount produced / consumed, and the surplus amount, *but* this isn’t always actually true because there are natural upper bounds to supply - not only in land, but you’ll notice no one is selling Passenger Pigeons these days. (Technically this is still a function of price, though, it's just a function of slope zero).
That said, yes, the common practice of using price as the Y-axis in Econ is AIUI to be convenient for various reasons of use in econ but it’s contrary to typical usage in every other field.
EDIT: To respond more directly to your specific question -- "supply" and "demand" usually aren't functions of quantity, they're *measures* of quantity - they're *functions* of price, at least in the common graphical representation Milan is using.
The only way I can make it work is: (1) flip the whole thing so Price is the X and Quantity is the Y (2) Change the "S" label to "Supplier" and "D" to "Consumer" (3) Go back to reading each line where Quantity (willingness to sell or buy an amount of something) is a function of Price (4) Find the intersection which is where both parties want to exchange the same amount for the same price. And it does somewhat make sense that the area under the curve (in this case, properly "under" because you flipped the axes- now just a good ol' integral) represents a total amount of money that could be exchanged, though it's a little fudgy because it's all hypothetical.
Where they lose me is when they start moving lines around willy-nilly ;) There seem to be a ton of assumptions and intuitions about what would cause the slopes/intercepts of the lines to shift for the supplier or customer.
For the supply curve, basically the lines shift based on cost of production. If production costs are low, then you'll be willing/able to produce more at every price point. If production costs are high, then it's the opposite.
For the demand curve, it has to do with how badly people need a given product, or how popular the product is at a given time, etc.
Think of it as if you were standing on a hill and threw a ball, and you wanted to ask where the ball hits the ground. You would graph two functions, each altitude as a function of distance. One would be the path of the ball, the other would be the terrain. Simple physics problem.
Traditionally you take quantity and price is determined. That is how a lot of markets historically worked. Eg a bunch of people show up at the market with produce.
To draw the graph in two dimensions, you need to hold a lot of things fixed, such as the quality of products. To stay in two dimensions and think about those things, you represent the lines themselves shifting.
You can certainly do things that are more complex, but then you are beyond a graph and you need to write down the equations. It's not really any different than saying the simple physics problem I described above doesn't take into account air resistance, spin of the ball, heterogeneity in the gravitational field, etc. If you want to do those things, you need to write it out. But like those things in my physics example, relaxing a lot of these assumptions doesn't make a huge difference in many cases.
For example, in oligopoly the supply problem would be much more complex, but in general none of the things that get more complex interact with the kind of examples that Matt gave (market-wide changes impacting all competitors), so it doesn't impact the intuition very much.
Supply and demand aren’t points, they’re functions. So if supply or demand change for any reason, you don’t represent it as moving a point along a fixed line, you represent it as the entire line moving.
In the image where he shifts the supply line, it’s because there was a new tax levied. This shifted the line because it now cost more to produce the same quantity.
(Disclaimer: I didn’t actually read this article.)
They are both independent functions of quantity. The supply curve gives the price at which the suppliers would supply that quantity. The demand curve gives the price at which the customers would purchase that quantity. The place where they intersect is the market price, the price where the amount people want to produce and the amount customers want to buy is the same.
The thing to realize is that the “demand” and “supply” at a single point in time is not a value, it’s a function. That means that changing the supply/demand of a product isn’t represented as moving along the line, it’s represented as shifting the entire function.
The demand function would tell you that at this moment, if you sold your product for $X, then you would sell a total of Y products; this function slopes downward.
Similarly, the supply function says “If we charge $X, then we can afford to produce Y products,” and it slopes upwards. The market equilibrium occurs at the point where the two lines intercept. Deadweight loss is when the price of the product is forced downward, creating a wasteful gap between the two lines.
You can’t necessarily create these charts with real data, but they’re still conceptually very meaningful.
Sorry if i wasn’t clear enough. What I meant to say is that the entire function shifts, as Milan illustrated in the second graph.
If the supply/demand were a straight line (which they almost never are) then this would be the same thing as changing the b value in the standard line formulation of y = mx + b.
And remember, you can think of the demand at a given time as the entire line, not just an individual point on it.
This is literally an Econ 101 graph. Anyone who has ever taken an economics class or read a Paul Krugman column should be familiar with them. I honestly can't overemphasize how fundamental and basic Supply/Demand graphs are to Economics as a discipline. It's like saying that using letters in Algebra makes mathematics incomprehensible.
The idea that Milan invented Price/Quantity graphs is maybe the funniest thing I've ever heard on this site. I'm not even here to say that it's a great way to express these things! It's just the way basic Economics works. If you have a better way to graph Economics concepts you should go monetize it right now because it will be revolutionary.
A better way being argued elsewhere is using price as the X-axis because "The X-axis represents the value of the independent variable in a Cartesian Graph of a function f:R1 -> R1" is the convention every other field uses.
Right but price isn't always the independent variable in a Supply/Demand function. I assume (but can't confirm) that economists feel like it is more straightforward to have the chart use the same axes rather than switching depending on the context of the analysis. There are lots of economists with substantial math backgrounds, I assume they're not unfamiliar with Cartesian coordinate conventions.
I think big picture these graphs are supposed to illustrate concepts rather than be used for mathematical analysis. Bigger picture these are taught to 15 year olds. The number of people claiming they are scientists or mathematicians in this thread but are apparently unable to grasp the impenetrable secrets of the Supply/Demand graph is absolutely wild.
Traditionally quantity was the independent variable. Many markets historically worked that way. You showed up at the market with what you grew, price changed until the market cleared.
You can definitely model it the other way, and that's what you spend your time doing in higher level courses. For markets with a lot of firms or just one firm, it doesn't really make a difference.
The issue is that price really is a function of quantity. I think it's less helpful to think of the y-axis here as "price" - maybe instead try thinking about it as "value." The demand curve represents marginal value (or "marginal benefit") to consumers, the supply curve represents value *lost* (or "marginal cost") to producers.
Relatedly, you need to use the area under the curve to show market value lost/gained at different arbitrarily set prices or quantities. That interpretation becomes less obvious with the axes reversed.
Can you not graph two functions on the same graph in your field? Quantity supplied and quantity demanded are both quantities with a relationship to price.
There are two different sets of observations on the same graph. If you wanted to graph the relationship between age and average height for men and also graph it for women, can you imagine doing that on the same graph? Or would you do it on two different graphs because “average height of women” is a different y variable than “average height of men”?
This would be super straightforward: https://www.researchgate.net/figure/Median-height-by-age-of-male-and-female-adolescents-in-Ibadan-in-comparison-with-median_fig1_44655202 And easy to read: Given a person who's 10 years old, a female is 143 cms tall on average and a male is 137 cms. Implied here is that one's height is function of one's age. We don't typically also say one's age is a function of one's height, and we wouldn't bother plotting one's age as a function of how many years ago they were born. And we wouldn't start sliding around the lines and say "people are getting taller, so now you can see a deadweight loss in the difference in their ages"...?? Again, I trust the S&D curves do make sense, but there's something about them that's just fundamentally different from the way the rest of the scientific world uses XY plots.
Maybe if I imagined values on the axes, and read it in the same way: "Given a $3 price for a chicken sandwich, Popeyes is only willing to sell 1,000 of them but people would be willing to buy 4,000 of them. At $6 per sandwich, Popeyes would be willing to sell 4,000 but people are only willing to buy 1,000." I assume this is what's meant by the figures. It still feels a bit weird to me. I think of price as more of the independent variable- as a consumer, it feels like the supplier gets to pick the price and I get to pick whether I buy at that price or not as a function of the price- quantity sold is then a function of price. But maybe from the supplier perspective, they set price as a function of the quantity they're selling? Still feels weird/backwards. (It's also not intuitively obvious to me that a supplier wants to sell less of something when it's a lower price. Certainly that would depend on the relative variable and fixed costs to manufacture the item? Maybe this is where the different slopes come in? I promise I took Econ 101 20 years ago, it just has always felt a little hand-wavey to me)
Stop thinking of *a* seller and think of a market of sellers. One thing that can occur when market price drops is that marginal sellers get out of the market, which decreases supply.
Yeah it would be more intuitive if price were the independent variable, but from the supplier perspective it is in fact the opposite. Think of a fisherman. You bring in your boat in with your catch along with all the other fishermen, and together you caught X fish that day. The price is the dependent variable that's based on whatever that day's X is.
I still think the axes should be flipped, but this is how Alfred Marshall did it the first time, and everyone has followed suit since then.
How does this differ at all from the supply-demand plot?
In both cases, there are multiple different relationships/functions for which the unit and scale are shared and they can be plotted in a single chart.
The big conceptual difference is that the integral of a height plot has no meaning and we do not care about it, which is not the case for a supply-demand plot.
The quantity axis is actually two different variables; "quantity demanded" and "quantity supplied." But they share a unit and a scale so can be displayed like this, and the interaction between the two functions graphed is of such interest that they basically *must* be displaced like this to have any utility at all.
I can envision other circumstances in which it makes sense to plot things like this. Predator-prey population curves, which have "number" as the dependent variable when they actually mean "number of predators" and "number of prey", for instance.
The supply curve represents the quantity that would be supplied at that price, the demand curve represents the quantity that would be demanded at that price
It's quantity as a function of price, and there are two lines because one is quantity as a function of price for sellers and the other is quantity as a function of price for buyers. Obviously, more producers will sell the item if they can get more for it, and more buyers will buy the item if they can pay less for it. Where those two lines intersect is the quantity that will result.
I think the non econ people must have woken up without coffee or something. I guess you are right that you could make two graphs with the producer and consumer responses to changes in price. But the point of this graph is to show you: 1) where the two lines cross is where we discover the quantity of the good or service actually transacted; 2) how ‘shocks’ to demand or supply will impact the number of transactions at different price levels.
Unfortunately you have (at least) two parties involved and their respective responses to the price signal will in fact change consumption and production….and do so differently, dependent on the specifics of each and every market. Without the two curves you can’t really figure out the deadweight loss….
I think I'm just used to XY-plots where the X is clearly an independent variable and Y is clearly a function of X. That's not the case here and it always takes me like 2x longer to wrap my head around the message of the diagram than I'd expect, given how simple the diagram looks. Also there are all kinds of rules about "when ABC happens, this line moves left or right or up or down" that you're just supposed to know, or trust that the writer did correctly. I'm fairly confident I'm not a dumb person, but these plots always make me feel dumber than usual (like when someone is watching you while you work on an Excel spreadsheet...)
"...this kind of graph abuse is what makes economics basically incomprehensible..."
Exactly my reaction. Feeding this to a mathematician or scientist is like feeding a piece of bad syntax into a compiler.
The econ bros downthread all claim that it's an innocent combination of several graphs, each of which is well-formed.
Okay, show me.
Give me the decomposition into multiple graphs, each of which has a proper independent variable on the horizontal axis and dependent variable on the vertical axis, and shows a genuine function.
Thanks for calling me an Econ bro. You’re being rude and dismissive of something you’d don’t (want to?) understand. What’s wrong with Y1=10-X and Y2=X? These are two functions with two lines. If we call Y1 the price at which billy buyer wants to transact for X units and Y2 the price at which sally seller wishes to transact for X units, then solving for their intersection turns out to be economically meaningful.
I’m happy to explain econ and the use of graphs to someone who’s generally curious instead of “science” peacocking. But no, that won’t extend to attempting to post graphs into the substack comment system. I can get you some khan academy econ links if you’d like. I’d imagine Sal Khan is more mathematically fluent/acceptable than I am given the (impressive!) breadth of his work.
This idea that the supply-demand sketch used in Econ is offensive to professional scientists/engineers/mathematicians is strange to me (retired Chemical Engineer). Not every x-y graph needs to show a single y axis variable’s dependence on an x axis independent variable.
Sure, by convention, many graphs do show exactly that. But there are also important cases in science/engineering where it’s helpful to graph the relationship between mutually dependent variables. Chemical equilibrium comes to mind as an example-- for example the concentrations of ethanol and water in a gasohol distillation column. Neither the ethanol nor the water concentration is independent but generations of chem eng students learned to design distillation columns with x-y plots of these.  Probably graphical design has been replaced with computing but I suspect that the chart is still shown in class and in textbooks as an aid to understanding.
I think the problem is that the people who show economics graphs are very familiar with the genre conventions of these graphs, and forget how to explain how those conventions differ from the conventions of many other graphs.
I think this would be very fine with a little bit more verbal explanation, that wasn't given in the original post, and was given only sketchily in the comments.
The explanation should say something like this. "The supply line shows how many suppliers would be willing to sell this good if they could get a given price for it. The demand line shows how many consumers would be willing to buy this good if they could get it for a given price. Both of these are simple functions of price, when it is considered as an independent variable. Once we have this set of counterfactuals, we then see that any point other than the intersection of the two lines will be an unstable situation for a market - either some supplier will ask for a higher price or some consumer will ask for a lower price or some other sort of change will happen."
We then need a bunch of words to explain how a change in tax rate, or a regulatory cap, or whatever, can be interpreted as moving one or another line, if we are being careful about how we interpret the variable.
An Econ 101 professor can often browbeat the students into understanding what is meant, without knowing how to put it precisely into words, and once you understand it, the chart without many words is quite helpful. But if you don't have a full semester with your audience, then getting the words right is very helpful.
This is a fundamental problem you discover as a teacher. Everyone knows* Socrates adage about the challenge of knowing that you don’t know (something you truly learn to appreciate as a researcher) but as a teacher you learn to really know and appreciate what you *do* know and not to take it for granted , but rather consciously attempt calibrate yourself to your audience, based on what *they* presumably know. It’s a huge challenge.
(*“Everyone” = presumably you, Kenny, and a fair number of SBers who’ve gone deep enough down the rabbit hole to read this…)
The ideas are well-formed and useful even if they aren’t rigorous. Here’s a formulation if you like:
Define two functions S and D which take in a price and some vector X. Here, X represents the entire state of the world. S and D represent the amount of product supplied or demanded when the world is in state X and the product is sold at price p.
Then, for some range of prices P, the graphs above plots two sets of points: {(p, S(p, X))} and {(p, D(p, X))} for some fixed X.
Now, we can’t actually compute these graphs in the real world, but they’re still useful. X is complex and unwritable, but if we change it certain controlled ways we may still be able to reasonably discuss how it would modify the two sets of points.
>>"There's only one line: price, as a function of quantity. That's what it means for the chart to have axes - they define your dependent and independent variables, and the line you graph in the coordinate space is the relationship (the function) between the axes." << THANK YOU- as an engineer, the supply/demand diagram drives me insane (and always makes me feel dumb), because it doesn't work the way I'd intuitively expect it to- the way graphs are supposed to work, in math, in the rest of the world. (Not Milan's fault, though!! He's just using the graph the way everyone else does.)
Hopefully, Milan is taking the wheel today so Matt can research an epic post about Congressional politics.
It looks like the Republicans will have 220 or 221 House seats in the next Congress. They’ll have alot of fun with oversight and may even embarrass the Biden administration. Progressive legislation will be DOA. However, a dozen of the Republican members will be from suburban NYC or LA, so I’m not really sure there will be government shutdowns. The really intriguing question is whether a vital center can develop. Can the NY Republicans get together with Spanberger and Murkowski and Manchin and actually fix some problems? Or will McCarthy use the gavel to keep centrist legislation from getting to the floor? (I expect very little legislation to pass, but my confidence isn’t high). Do the moderates have an actual policy agenda beyond not being crazy? The answers to these questions will all be personality driven, and I don’t pretend to understand the House Republican zeitgeist.
A great example of "deadweight loss" are restrictions on surge pricing for ride share companies. No surge pricing to get drivers on the road=no ride home from the ballgame for me.
Great explanation of deadweight loss. I’d add the concepts of fixed costs vs. variable costs as a contributing factor to why suppliers decrease quantity with a price cap.
* Fixed costs: Costs a supplier pays regardless of quantity sold
* Variable costs: Cost per-a-unit of production
In the Popeyes example
* Fixed: Lease, maintenance, wages of a minimal crew
* Variable: Ingredients, wages of additional workers at busy times
With the price ceiling, Popeyes could find that sales after 10 PM doesn’t provide enough revenue to offset the costs of the minimum crew needed to operate a store. They could also find that some locations should be closed since they don’t bring in enough revenue to cover their fixed costs regardless of hours of operation.
This is a technical quibble but in producer theory we learned that in the short run a business should decide whether to operate or not based on whether revenue covers variable costs.
Minimal crew wages are pseudo-variable - as you note, they can always close earlier - but the lease probably won't get any cheaper if they do.
Maintenance too - put it this way, an annual service on a car is a fixed cost, whereas a service every 10,000 miles is a variable cost. I think?
There is also marginal cost - the cost to make one more.
1) Thought I knew what deadweight loss was.
2) Read the joke. Didn’t get it.
3) Wondered if if I didn’t know what deadweight loss was.
4) Read the article to confirm that I knew what deadweight loss was.
5) Still didn’t get the joke. 😔
Kids miss out on getting candy, adults miss out on the joy of handing out candy
Getting off topic here, but actually the point it APPLES ARE NOT CANDY. I kind of doubt that candy apples were ever a popular thing to give trick or treaters, since making them and wrapping them involves a lot of time and effort and expense—the issue was just plain apples, which some adults (you know the kind) saw as a “healthy” alternative to candy. The story about the razorblades arose from word of mouth, started to show up in local media and warnings in classrooms, and was probably an urban legend invented by children to ensure that trick or treaters got proper, manufactured, sugar-loaded CANDY, not fruit.
The urban legend only makes sense if it’s a candy apple, though. If you stick a razor blade in a plain apple, there will be a big slit in the side of the fruit, surrounded by a bruise.
Moreover, after you jam the razor in, a little corner will still remain sticking out, and give the game away. You need to cover the whole thing in candy if the scheme is going to work.
Scary stories don’t need to be that accurate. Most urban myths don’t stand up to the slightest bit of scrutiny, but that doesn’t keep them from spreading.
While perhaps not as venerable as some of the commentariat, I was around for Halloween before that story became part of the cultural milleu, or at least before it was widely heard, and have never, ever seen a candy apple on Halloween. And my buddies and I were the ones who used to blitz like 5 different subdivisions with bikes and a wagon.
I have absolutely seen people giving out fruit, vegetables, water bottles, and pretzels. The last two were often a welcome break from binging on sugar while foraging for more sugar, though.
In central Canada people used to say “Halloween apples” instead of Trick or Treat!
Marge is one of those adults.
https://www.youtube.com/watch?v=ZFLE2KApo4k
I think that the issue is that it’s not obviously a situation with “transactions” per se. But beyond that, the change described is more of a shift in preferences than the imposition of a price or quantity control. It’s more than a little unusual to use DWL to describe a change in quantity because of a voluntary shift in demand or supply. I can see how you get there by squinting and saying that the shift wasn’t voluntary because it was premised on a moral panic or bad information, but that’s really a stretch. Generally, consumer theory doesn’t have a carve out to specify which preferences are legitimate.
So, I get the joke, but I would expect to only give it partial credit on a test question asking for an example of deadweight loss. I know some of my own professors would have just marked it wrong.
No, we shouldn’t over analyze jokes, but once you write an explainer based on it…
The deadweight loss is the loss to all those kids who like candy apples *even with* a chance of razor blades in them! The regulators (parents) who insist that you should only take candies with professional wrapping are causing loss to both the kids and treat-givers who would voluntarily exchange at the risky margin.
Just because you know what deadweight loss is doesn’t mean you’re a total econ nerd, which is required to find that joke funny.
Had to laugh at this. Read the whole 2,000 words and had no idea what 'deadweight loss' is. Some people like some things and others like other things. Is this what the economics profession has been scratching their heads over? Econ speak for gibberish? Pretty sure your deadweight loss is my gain if what you dislike is what I like, and your calculation of your disutils versus my utils is proof economists are full of it and have no idea what 'deadweight loss' is or why we should care about it. My advice is stop reading Greg Mankiw and get on with your life.
Nothing quite like having a college freshman explain something they just learned in a college freshman class to you.
Considering that I've often forgotten some details over the years, the fresh perspective has some value.
Imagine what we could accomplish if even half our politicians already understood what this freshman just learned.
I teach this stuff to college freshman, and I think he did a very good job explaining it. But, like, I probably could have done it better, but my job isn't writing blog posts.
Great post!
DWL is a valuable concept. But it's worth being skeptical of the way it's quantified on a supply and demand plot, especially when (though the post does not do this) economists start referring to it as a measure of "welfare."
Willingness-to-pay is a very convenient proxy for welfare, but I think most people would acknowledge that its a pretty bad one. The fact that I was willing to pay up to $100 for an uber ride to the airport yesterday (I actually paid $80) while the homeless guy I passed probably wasn't willing to pay anything for it says basically nothing about how much welfare that uber ride produced. But on the supply and demand plot, my willingness-to-pay gets represented by a point along the demand curve at P = 100, while the homeless man's willingness to pay gets represented as a point a P = 0.
This isn't a profound or novel insight; plenty of economists acknowledge the issues with market valuations of welfare. But I think it's worth having in a comment here given all the people likely to see this post.
It occurs to me that a lot of ways of cutting inflation would be really destructive. Price controls on basic consumer items that are produced in competitive markets would certainly cut inflation, but would also result in constant shortages. A rationing system would also cut inflation but only by forcing people to reduce the amount of stuff they normally buy.
There are lots of ways charlatans can screw things up for everybody with daft policies. Price controls and rationing do have a use in certain circumstances, but often when there has been a policy failure elsewhere.
There are much worse options than that!
Whenever someone says "this is my number one priority", then I always think "so, if you could achieve this by destroying the entire human species in a nuclear war, then you would?"
It's surprising how many problems you can solve by destroying humanity: crime, for instance.
It's why technically enough bleach would have worked as a Covid cure AND a Covid vaccine!
The very general rule should be to always try to increase supply before resorting to reducing demand. There will of course come times where the former isn't practical in a reasonable timeline.
It's interesting that government pretty much always does the unreasonable alternatives.
A gas tax holiday, for instance, is the epitome of this. It's basically: "Increase supply? Nope, best I can do is stimulate demand!"
Matt tweeted out this image a while back, and I facepalmed multiple times as I could tell people had no clue on how to actually reduce inflation.
https://i.ibb.co/FbC2JT4/image.png
During the world war II, in the US, there were very restrictive price controls on everything because the government was attempting to redirect all output to war production. It worked - the US was the arsenal of democracy, outproducing pretty much the entire world in war production** - but the black market thrived. Post-war, while price controls remained in force, the price of new cars stayed the same, so dealers came up with assorted hacks to charge extra for various essential car pars, like the steering wheel. If you bought everything you needed the combined price would wind up being equivalent to the market price. Economy-wide price controls suck, but they work (which is why communism survived in the USSR as long as it did). Price controls on a very limited set of goods or a market segment work, but they can be quite costly. They can also be used to neutralize the effects of monopolistic pricing - but that's fixing the effects of an existing monopoly - which is the first-order problem.
elm
it depends
** The US essentially outproduced the entire rest of the world combined (ex the USSR) in military goods
This was also when the original sin of the US healthcare system was created. Because of wage controls, employers paid for health insurance as a work-around and the IRS declared those benefits to be tax-free to the employee. The resulting 80 years of a lack of a functioning market for healthcare has led to price spirals, opaque pricing, backdoor rebates and limited supply.
Any sort of initiative to push people to take a cash subsidy from their employer and go buy healthcare coverage from the Marketplace would be an immediate step up. It would start untangling the ratfuck of mis-aligned incentives that the healthcare sector outside of actual providers is shot through with, albeit not yet the wildly mis-aligned incentives among providers themselves.
Preferential tax treatment needs to end, preferably by ending deductions for employer-sponsored coverage, but more likely by creating one for personally-purchased coverage, lol.
As I said, likely easier to extend to all personal health insurance expenditures and allow companies to directly fund an FSA-equivalent for purchase of insurance than to wind it up.
We can always make up the revenue elsewhere, it's the unlevel playing field that's killing us.
"Economy-wide price controls suck, but they work (which is why communism survived in the USSR as long as it did). "
In the case of WW2 there was a huge positive externality to the rationing/controls - "win the war" - that could not easily be done on any individual level, so the deadweight loss incurred could be beaten out by the rewards from winning the war. But that's pretty atypical.
I don't see that they helped communism in the long run.
>>Economy-wide price controls suck, but they work<<
If combined with rationing, yes, otherwise the shortages will be intolerable. Also, it helps if the cause is popular, so that the policy is enforced in part by societal disapproval of cheaters.
Price controls aren't successful long-term in a democracy. Ask the UK Labour Party circa 1953.
"...to charge extra for various essential car pars, like the steering wheel."
You see, I would pay extra for a *good* steering wheel, that doesn't fly off while you driving.
Those regulators, insisting that *every* car have a good steering wheel, are causing deadweight loss for those producers and consumers who would voluntarily transact to exchange a car with a bad steering wheel for a low price!
Consumer choice is good, but when you get in there, you're like, "if the steering wheel fly off, I'm toast."
"And to understand those concepts, you need to understand a really basic supply and demand chart with the price of widgets on the vertical axis and the quantity of widgets sold on the horizontal axis. The line for demand slopes down, and the line for supply slopes up "
Ok, as a scientist and math minor, this kind of graph abuse is what makes economics basically incomprehensible to me and everyone else. You're not using "chart" to mean the same thing we do - you can't, because a graph with "price on the vertical axis" and "quantity on the horizontal axis" can't have two lines. There's only one line: price, as a function of quantity. That's what it means for the chart to have axes - they define your dependent and independent variables, and the line you graph in the coordinate space is the relationship (the function) between the axes.
You're doing something else with your chart, clearly, but since it's so radically different from how data is typically charted and you don't explain how to interpret it, it's incomprehensible and that's where I fall off the rest of your post.
Pretty easy to interpret to me
These are pretty standard Econ 101 textbook graphs, they’re acting like you invented them
True. But you need to write to your audience. It seems these graphs weren’t sufficiently explained and many people here unfamiliar with the concepts didn’t understand them.
I don’t know how many highly-technical classes you’ve taken Milan, but it might be worth mentioning here that the standards economists use to draw these graphs are much less rigorous than you would ever find in, say, mathematics or engineering. The people most confused seemed to have more technical backgrounds where you wouldn’t see a graph like that.
Milan, in time you’ll learn that one is never the best judge of the legibility of their own work. A professional scientist just told you your presentation of data is incomprehensible to them. Your post is supposed to be geared to a general audience, not to those already in the know. Perhaps it is the wiser course to consider that there might be something to the critique ? And at any rate not to be so quick to dismiss it ?
As a counterpoint - I'm a professional scientist, and while I think that sentence could have been edited for clarity, the idea of "two functions, f(x) and g(x) that have the same units and can be plotted on the same graph" really isn't that exotic or alien to scientific fields. Off the top of my head, "mixed potential theory" uses the exact same kind of graph.
That’s kind of besides the point I’m trying to make here. It’s not about the standards of this or that field. It’s about writing to a general audience and learning to listen to feedback, especially with regards to issues of clarity and what conventions can or cannot be taken for granted when addressing such audience. I mentioned that op was a professional scientist as a “a fortiori” move. Frankly Milan’s answer would have been equally unhelpful had op just been Joe shmo subscriber- it’s the latter that the post is meant for after all.
I mean, you can criticize the tone of the response but there is no other way I can think of off the top of my head to relay the information.
About the only criticism I have often heard repeated about the Econ 101 supply-demand charts is that they reverse the standard for dependent and independent variables on the X- and Y-axes. Which is... debatable. There is nothing to prove that price is the independent variable except blind intuition, and not even always that.
Certainly, the two lines criticism makes *no sense at all* within any discipline's framework; they're two different functions plotted on the same graph. Full stop. This is completely and utterly ordinary for anyone whose work involves math and I cannot imagine why someone regarded it as unreasonable.
OP says they don’t understand the chart. There are multiple ways to respond but I would think trying to explain the chart to them in other words might be a good thing to include in the response, and being dismissive might not. The discussion of other commentators here, some likewise admitting their confusion, others patiently trying to elucidate this for them, offers a good model, imho.
The initial line of criticism "what's the other line?" seems just completely ridiculous to me, having dealt with graphical representations of functions in a professional capacity, yes. Like, seeing two functions on the same plot is not the least bit unusual.
Counterpoint: If someone has trouble following supply and demand curves, then the concept of deadweight loss will be beyond their comprehension. And I shudder to imagine a professional scientist who is ignorant of the concept presented in Milan's chart.
I have to say, I had a hard time making sense of the chart. Usually I’m someone who prefers visual information, but I found the written explanation much easier to understand, and in fact had to entirely depend on that explanation to make head or tail of the chart, which makes the chart kind of superfluous.
Supply and demand charts are deceptively complicated things. I TAed for for undergads this past spring who still made basic mistakes with them 4 months later.
But they're also incredibly powerful tools once you get the hang of them. If you're interested, you might try googling "marginal benefit interpretation of demand curve" or "marginal cost interpretation of supply curve" - that'll get you into the reasoning underlying these charts, which I think is the most intuitive starting point.
Ok, maybe you can explain it. There's two axes, one dependent and one independent. One of the lines is the relationship between price and quantity. What's the other one?
One, demand, shows the price required to induce consumers to buy a given quantity. The other, supply, shows the price required to induce producers to provide a given quantity.
The real issue to complain about is that it’s much more intuitive to think about price (y-axis) being the independent variable. It can make sense both ways, but many students initially have trouble seeing how price can be a function of quantity.
Ok, but which of "supply" and "demand" is the function of quantity? And what's the other one?
The short answer is it depends on the market. In general, price is closer to being the independent variable, and you assume a latent utility function of the marginal consumer and producer yielding the equilibrium price, amount produced / consumed, and the surplus amount, *but* this isn’t always actually true because there are natural upper bounds to supply - not only in land, but you’ll notice no one is selling Passenger Pigeons these days. (Technically this is still a function of price, though, it's just a function of slope zero).
That said, yes, the common practice of using price as the Y-axis in Econ is AIUI to be convenient for various reasons of use in econ but it’s contrary to typical usage in every other field.
EDIT: To respond more directly to your specific question -- "supply" and "demand" usually aren't functions of quantity, they're *measures* of quantity - they're *functions* of price, at least in the common graphical representation Milan is using.
Neither is a function of the other, it’s parametric.
What is the unshown independent parameter, then? (It makes more sense to me if I read it as Q(P), honestly.)
The only way I can make it work is: (1) flip the whole thing so Price is the X and Quantity is the Y (2) Change the "S" label to "Supplier" and "D" to "Consumer" (3) Go back to reading each line where Quantity (willingness to sell or buy an amount of something) is a function of Price (4) Find the intersection which is where both parties want to exchange the same amount for the same price. And it does somewhat make sense that the area under the curve (in this case, properly "under" because you flipped the axes- now just a good ol' integral) represents a total amount of money that could be exchanged, though it's a little fudgy because it's all hypothetical.
Where they lose me is when they start moving lines around willy-nilly ;) There seem to be a ton of assumptions and intuitions about what would cause the slopes/intercepts of the lines to shift for the supplier or customer.
For the supply curve, basically the lines shift based on cost of production. If production costs are low, then you'll be willing/able to produce more at every price point. If production costs are high, then it's the opposite.
For the demand curve, it has to do with how badly people need a given product, or how popular the product is at a given time, etc.
Think of it as if you were standing on a hill and threw a ball, and you wanted to ask where the ball hits the ground. You would graph two functions, each altitude as a function of distance. One would be the path of the ball, the other would be the terrain. Simple physics problem.
Traditionally you take quantity and price is determined. That is how a lot of markets historically worked. Eg a bunch of people show up at the market with produce.
To draw the graph in two dimensions, you need to hold a lot of things fixed, such as the quality of products. To stay in two dimensions and think about those things, you represent the lines themselves shifting.
You can certainly do things that are more complex, but then you are beyond a graph and you need to write down the equations. It's not really any different than saying the simple physics problem I described above doesn't take into account air resistance, spin of the ball, heterogeneity in the gravitational field, etc. If you want to do those things, you need to write it out. But like those things in my physics example, relaxing a lot of these assumptions doesn't make a huge difference in many cases.
For example, in oligopoly the supply problem would be much more complex, but in general none of the things that get more complex interact with the kind of examples that Matt gave (market-wide changes impacting all competitors), so it doesn't impact the intuition very much.
Supply and demand aren’t points, they’re functions. So if supply or demand change for any reason, you don’t represent it as moving a point along a fixed line, you represent it as the entire line moving.
In the image where he shifts the supply line, it’s because there was a new tax levied. This shifted the line because it now cost more to produce the same quantity.
(Disclaimer: I didn’t actually read this article.)
They are both independent functions of quantity. The supply curve gives the price at which the suppliers would supply that quantity. The demand curve gives the price at which the customers would purchase that quantity. The place where they intersect is the market price, the price where the amount people want to produce and the amount customers want to buy is the same.
The thing to realize is that the “demand” and “supply” at a single point in time is not a value, it’s a function. That means that changing the supply/demand of a product isn’t represented as moving along the line, it’s represented as shifting the entire function.
The demand function would tell you that at this moment, if you sold your product for $X, then you would sell a total of Y products; this function slopes downward.
Similarly, the supply function says “If we charge $X, then we can afford to produce Y products,” and it slopes upwards. The market equilibrium occurs at the point where the two lines intercept. Deadweight loss is when the price of the product is forced downward, creating a wasteful gap between the two lines.
You can’t necessarily create these charts with real data, but they’re still conceptually very meaningful.
Moving the vertical line I assume? When you say moving the line, I picture the slope changing but I'm pretty sure that's wrong.
Sorry if i wasn’t clear enough. What I meant to say is that the entire function shifts, as Milan illustrated in the second graph.
If the supply/demand were a straight line (which they almost never are) then this would be the same thing as changing the b value in the standard line formulation of y = mx + b.
And remember, you can think of the demand at a given time as the entire line, not just an individual point on it.
Oh, I see it now. S1 -> S2. Thanks!
Don’t blame Milan for this - he’s presented the standard DWL chart that every economist uses. He didn’t reinvent the wheel here.
This is literally an Econ 101 graph. Anyone who has ever taken an economics class or read a Paul Krugman column should be familiar with them. I honestly can't overemphasize how fundamental and basic Supply/Demand graphs are to Economics as a discipline. It's like saying that using letters in Algebra makes mathematics incomprehensible.
The idea that Milan invented Price/Quantity graphs is maybe the funniest thing I've ever heard on this site. I'm not even here to say that it's a great way to express these things! It's just the way basic Economics works. If you have a better way to graph Economics concepts you should go monetize it right now because it will be revolutionary.
A better way being argued elsewhere is using price as the X-axis because "The X-axis represents the value of the independent variable in a Cartesian Graph of a function f:R1 -> R1" is the convention every other field uses.
Right but price isn't always the independent variable in a Supply/Demand function. I assume (but can't confirm) that economists feel like it is more straightforward to have the chart use the same axes rather than switching depending on the context of the analysis. There are lots of economists with substantial math backgrounds, I assume they're not unfamiliar with Cartesian coordinate conventions.
I think big picture these graphs are supposed to illustrate concepts rather than be used for mathematical analysis. Bigger picture these are taught to 15 year olds. The number of people claiming they are scientists or mathematicians in this thread but are apparently unable to grasp the impenetrable secrets of the Supply/Demand graph is absolutely wild.
Traditionally quantity was the independent variable. Many markets historically worked that way. You showed up at the market with what you grew, price changed until the market cleared.
You can definitely model it the other way, and that's what you spend your time doing in higher level courses. For markets with a lot of firms or just one firm, it doesn't really make a difference.
The issue is that price really is a function of quantity. I think it's less helpful to think of the y-axis here as "price" - maybe instead try thinking about it as "value." The demand curve represents marginal value (or "marginal benefit") to consumers, the supply curve represents value *lost* (or "marginal cost") to producers.
Relatedly, you need to use the area under the curve to show market value lost/gained at different arbitrarily set prices or quantities. That interpretation becomes less obvious with the axes reversed.
Can you not graph two functions on the same graph in your field? Quantity supplied and quantity demanded are both quantities with a relationship to price.
You can graph two functions but how can two variables have two relationships?
There are two different sets of observations on the same graph. If you wanted to graph the relationship between age and average height for men and also graph it for women, can you imagine doing that on the same graph? Or would you do it on two different graphs because “average height of women” is a different y variable than “average height of men”?
This would be super straightforward: https://www.researchgate.net/figure/Median-height-by-age-of-male-and-female-adolescents-in-Ibadan-in-comparison-with-median_fig1_44655202 And easy to read: Given a person who's 10 years old, a female is 143 cms tall on average and a male is 137 cms. Implied here is that one's height is function of one's age. We don't typically also say one's age is a function of one's height, and we wouldn't bother plotting one's age as a function of how many years ago they were born. And we wouldn't start sliding around the lines and say "people are getting taller, so now you can see a deadweight loss in the difference in their ages"...?? Again, I trust the S&D curves do make sense, but there's something about them that's just fundamentally different from the way the rest of the scientific world uses XY plots.
Maybe if I imagined values on the axes, and read it in the same way: "Given a $3 price for a chicken sandwich, Popeyes is only willing to sell 1,000 of them but people would be willing to buy 4,000 of them. At $6 per sandwich, Popeyes would be willing to sell 4,000 but people are only willing to buy 1,000." I assume this is what's meant by the figures. It still feels a bit weird to me. I think of price as more of the independent variable- as a consumer, it feels like the supplier gets to pick the price and I get to pick whether I buy at that price or not as a function of the price- quantity sold is then a function of price. But maybe from the supplier perspective, they set price as a function of the quantity they're selling? Still feels weird/backwards. (It's also not intuitively obvious to me that a supplier wants to sell less of something when it's a lower price. Certainly that would depend on the relative variable and fixed costs to manufacture the item? Maybe this is where the different slopes come in? I promise I took Econ 101 20 years ago, it just has always felt a little hand-wavey to me)
Stop thinking of *a* seller and think of a market of sellers. One thing that can occur when market price drops is that marginal sellers get out of the market, which decreases supply.
Yeah it would be more intuitive if price were the independent variable, but from the supplier perspective it is in fact the opposite. Think of a fisherman. You bring in your boat in with your catch along with all the other fishermen, and together you caught X fish that day. The price is the dependent variable that's based on whatever that day's X is.
I still think the axes should be flipped, but this is how Alfred Marshall did it the first time, and everyone has followed suit since then.
How does this differ at all from the supply-demand plot?
In both cases, there are multiple different relationships/functions for which the unit and scale are shared and they can be plotted in a single chart.
The big conceptual difference is that the integral of a height plot has no meaning and we do not care about it, which is not the case for a supply-demand plot.
The quantity axis is actually two different variables; "quantity demanded" and "quantity supplied." But they share a unit and a scale so can be displayed like this, and the interaction between the two functions graphed is of such interest that they basically *must* be displaced like this to have any utility at all.
I can envision other circumstances in which it makes sense to plot things like this. Predator-prey population curves, which have "number" as the dependent variable when they actually mean "number of predators" and "number of prey", for instance.
The supply curve represents the quantity that would be supplied at that price, the demand curve represents the quantity that would be demanded at that price
Boy, this subthread is way over my head as I futilely try to discern the very strong opinions people have on this subject.
It's quantity as a function of price, and there are two lines because one is quantity as a function of price for sellers and the other is quantity as a function of price for buyers. Obviously, more producers will sell the item if they can get more for it, and more buyers will buy the item if they can pay less for it. Where those two lines intersect is the quantity that will result.
I think the non econ people must have woken up without coffee or something. I guess you are right that you could make two graphs with the producer and consumer responses to changes in price. But the point of this graph is to show you: 1) where the two lines cross is where we discover the quantity of the good or service actually transacted; 2) how ‘shocks’ to demand or supply will impact the number of transactions at different price levels.
Unfortunately you have (at least) two parties involved and their respective responses to the price signal will in fact change consumption and production….and do so differently, dependent on the specifics of each and every market. Without the two curves you can’t really figure out the deadweight loss….
I think I'm just used to XY-plots where the X is clearly an independent variable and Y is clearly a function of X. That's not the case here and it always takes me like 2x longer to wrap my head around the message of the diagram than I'd expect, given how simple the diagram looks. Also there are all kinds of rules about "when ABC happens, this line moves left or right or up or down" that you're just supposed to know, or trust that the writer did correctly. I'm fairly confident I'm not a dumb person, but these plots always make me feel dumber than usual (like when someone is watching you while you work on an Excel spreadsheet...)
"...this kind of graph abuse is what makes economics basically incomprehensible..."
Exactly my reaction. Feeding this to a mathematician or scientist is like feeding a piece of bad syntax into a compiler.
The econ bros downthread all claim that it's an innocent combination of several graphs, each of which is well-formed.
Okay, show me.
Give me the decomposition into multiple graphs, each of which has a proper independent variable on the horizontal axis and dependent variable on the vertical axis, and shows a genuine function.
Thanks for calling me an Econ bro. You’re being rude and dismissive of something you’d don’t (want to?) understand. What’s wrong with Y1=10-X and Y2=X? These are two functions with two lines. If we call Y1 the price at which billy buyer wants to transact for X units and Y2 the price at which sally seller wishes to transact for X units, then solving for their intersection turns out to be economically meaningful.
I’m happy to explain econ and the use of graphs to someone who’s generally curious instead of “science” peacocking. But no, that won’t extend to attempting to post graphs into the substack comment system. I can get you some khan academy econ links if you’d like. I’d imagine Sal Khan is more mathematically fluent/acceptable than I am given the (impressive!) breadth of his work.
This idea that the supply-demand sketch used in Econ is offensive to professional scientists/engineers/mathematicians is strange to me (retired Chemical Engineer). Not every x-y graph needs to show a single y axis variable’s dependence on an x axis independent variable.
Sure, by convention, many graphs do show exactly that. But there are also important cases in science/engineering where it’s helpful to graph the relationship between mutually dependent variables. Chemical equilibrium comes to mind as an example-- for example the concentrations of ethanol and water in a gasohol distillation column. Neither the ethanol nor the water concentration is independent but generations of chem eng students learned to design distillation columns with x-y plots of these.  Probably graphical design has been replaced with computing but I suspect that the chart is still shown in class and in textbooks as an aid to understanding.
I think the problem is that the people who show economics graphs are very familiar with the genre conventions of these graphs, and forget how to explain how those conventions differ from the conventions of many other graphs.
I think this would be very fine with a little bit more verbal explanation, that wasn't given in the original post, and was given only sketchily in the comments.
The explanation should say something like this. "The supply line shows how many suppliers would be willing to sell this good if they could get a given price for it. The demand line shows how many consumers would be willing to buy this good if they could get it for a given price. Both of these are simple functions of price, when it is considered as an independent variable. Once we have this set of counterfactuals, we then see that any point other than the intersection of the two lines will be an unstable situation for a market - either some supplier will ask for a higher price or some consumer will ask for a lower price or some other sort of change will happen."
We then need a bunch of words to explain how a change in tax rate, or a regulatory cap, or whatever, can be interpreted as moving one or another line, if we are being careful about how we interpret the variable.
An Econ 101 professor can often browbeat the students into understanding what is meant, without knowing how to put it precisely into words, and once you understand it, the chart without many words is quite helpful. But if you don't have a full semester with your audience, then getting the words right is very helpful.
This is a fundamental problem you discover as a teacher. Everyone knows* Socrates adage about the challenge of knowing that you don’t know (something you truly learn to appreciate as a researcher) but as a teacher you learn to really know and appreciate what you *do* know and not to take it for granted , but rather consciously attempt calibrate yourself to your audience, based on what *they* presumably know. It’s a huge challenge.
(*“Everyone” = presumably you, Kenny, and a fair number of SBers who’ve gone deep enough down the rabbit hole to read this…)
The ideas are well-formed and useful even if they aren’t rigorous. Here’s a formulation if you like:
Define two functions S and D which take in a price and some vector X. Here, X represents the entire state of the world. S and D represent the amount of product supplied or demanded when the world is in state X and the product is sold at price p.
Then, for some range of prices P, the graphs above plots two sets of points: {(p, S(p, X))} and {(p, D(p, X))} for some fixed X.
Now, we can’t actually compute these graphs in the real world, but they’re still useful. X is complex and unwritable, but if we change it certain controlled ways we may still be able to reasonably discuss how it would modify the two sets of points.
>>"There's only one line: price, as a function of quantity. That's what it means for the chart to have axes - they define your dependent and independent variables, and the line you graph in the coordinate space is the relationship (the function) between the axes." << THANK YOU- as an engineer, the supply/demand diagram drives me insane (and always makes me feel dumb), because it doesn't work the way I'd intuitively expect it to- the way graphs are supposed to work, in math, in the rest of the world. (Not Milan's fault, though!! He's just using the graph the way everyone else does.)
Thank you -- Justin
Agree. This is why I fully understood the concepts in econ classes, but really struggled with the unique chart/graph style.
Hopefully, Milan is taking the wheel today so Matt can research an epic post about Congressional politics.
It looks like the Republicans will have 220 or 221 House seats in the next Congress. They’ll have alot of fun with oversight and may even embarrass the Biden administration. Progressive legislation will be DOA. However, a dozen of the Republican members will be from suburban NYC or LA, so I’m not really sure there will be government shutdowns. The really intriguing question is whether a vital center can develop. Can the NY Republicans get together with Spanberger and Murkowski and Manchin and actually fix some problems? Or will McCarthy use the gavel to keep centrist legislation from getting to the floor? (I expect very little legislation to pass, but my confidence isn’t high). Do the moderates have an actual policy agenda beyond not being crazy? The answers to these questions will all be personality driven, and I don’t pretend to understand the House Republican zeitgeist.