YT YouTube2Text

YouTube2Text → Transcripts → Price Theory Summer Camp, Day 1: Lecture by Kevin Murphy

Price Theory Summer Camp, Day 1: Lecture by Kevin Murphy — Transcript

by Becker Friedman Institute University of Chicago · 10,434 words · 1,666 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:01All right, let me Somebody just asked me
0:04a question about the monopoly case. It's
0:07important to realize that when you talk
0:10about cost increases or taxes on a
0:13monopoly, it's actually quite different
0:16than in a competitive firm space because
0:19in in the case of a competitive
0:23industry, when we put a tax on the
0:25competitive industry, I'm going to come
0:27back to this in a moment anyway. So I
0:28draw the picture case of a competitive
0:31industry. We have
0:33demand, we have
0:35supply, we put our tax on our tax pushes
0:40the buyer's price up to here, pushes the
0:42seller's price to
0:45here. In general, the sellers are able
0:48to pass some of the tax increase on to
0:50the buyers. And typically, unless it's a
0:54very large tax, the loss to the
0:58sellers, which is this area here, is
1:03actually less than the amount of tax
1:06collected. And in that sense, a
1:08competitive industry is able to pass
1:11some of the tax on to the buyers because
1:15this box is which is part of the tax
1:17paid by the buyers is usually bigger
1:19than that triangle. Okay?
1:22In the monopoly
1:24case, there's no pass through in that
1:27case. In that
1:29sense, in the sense that the monopouist
1:33is able to pass the price increase on to
1:36buyers. In fact, in many cases, the
1:38monopouist will raise the price to the
1:40buyers by more than the amount of the
1:42tax. So you might put a dollar per unit
1:44tax on monopoulos might raise the price
1:46a
1:47$150 but still in
1:51equilibrium the amount of loss in profit
1:54to the
1:56monopouist. I'll ask you how does it
1:59compare to the taxes that he
2:01pays? If I see a
2:04monopouist paying
2:07$100,000 in
2:10taxes, how does that amount paid to the
2:12government compared to his loss in
2:23profits from the
2:26tax? Equal, bigger, smaller.
2:34know it's actually always smaller than
2:37his loss in
2:39profits. The amount of tax he pays in
2:42government to the government understates
2:44his loss in profits. How do I know
2:48that? It's easy. Start from the post tax
2:53equilibrium. Let's just say you will
2:56remove the tax and the monopouist
2:59doesn't change a thing.
3:01How much would his profits go up?
3:08100
3:11$100,000. But we know in fact his
3:13profits are going to go up more than
3:15that because he's going to move his
3:18output. So a monopouist never has this
3:21ability to pass
3:23through in a profit sense the tax to
3:26consumers. Doesn't mean he doesn't hurt
3:28consumers.
3:30In fact, he may hurt consumers more than
3:32dollar for dollar. The consumers may
3:34lose. He may lose 100,000 and they may
3:36lose
3:38150,000. Government raises 100,000 in
3:40taxes, he loses 110 and the consumers
3:43lose 150. No problem with
3:47that. Why is it so
3:49different? You have a question? Yeah. I
3:52was wondering how do we know for certain
3:53that the monopouist changes prices when
3:56the tax is instituted? because his
3:58marginal cost goes up. And when his
4:01marginal cost goes up, the place where
4:03marginal revenue equals marginal cost is
4:06going to occur to the left of where it
4:07did before. And therefore, he'll raise
4:10price in equilibrium.
4:13Perhaps by more than the tax price by
4:15more than the tax. Does anybody see why
4:17he might be raising price by more than
4:19the tax?
4:27dependent very much on that. Well, what
4:31what determines the amount the
4:33monopoulos raises his
4:35price? Anybody
4:38know? We'll get I'll get back to this
4:40when we talk about monopolies. But
4:43anyway, why is the monopoly case so
4:46different?
4:48Well, the way I like to think about it
4:50and and this is this is sort of a
4:52mathematical side to this, which is from
4:54the monopouist problem, we know that the
4:58monopouist already had the ability to
5:00raise
5:01price. And therefore, from the
5:03monopoly's point of
5:05view, when you raise the
5:09tax, any gain he gets from raising the
5:13price is really just second order,
5:14right? He there was a first order gain
5:16to raising price. he had done it
5:18already.
5:20Okay? Whereas in the competitive case,
5:22typically because they're producing
5:24where price equals marginal cost,
5:27there's no first order loss from
5:28reducing
5:30output. Price equals marginal cost. A
5:32little reduction in output doesn't hurt
5:33them at all. So a first order reduction
5:37in output is profit increasing for the
5:39competitive industry, but profit neutral
5:42for the monopoly.
5:44And therefore that ability to pass on
5:46the tax increase and shield yourself
5:48from profit loss is present in the
5:51competitive case but not in a monopoly
5:53case. So competitive industries are
5:56actually in some sense easier to tax
5:58than monopoly industries from a
6:00political point. Yeah. Could profits
6:03suppose you put a tax on like a unit
6:06tax. Could profits remember we just went
6:08through the case where if you increase
6:11the wage rate or other factor input you
6:13might get an increase in profits to a
6:14competitive industry not to an output
6:17could an output tax could an output tax
6:20raise the profits of the competitive
6:21industry? Could they actually like to
6:23have the output
6:28tax? Well, look at the picture.
6:32Could the sellers could the seller's
6:34price go
6:37up the net price received by the
6:43seller? Right? In no case would a output
6:47tax be profit increasing for the
6:53sellers. What determines the
6:56split? In fact, I have
6:58the formula right here on the board.
7:01Right? The split
7:03between the increase in the price
7:07depends on the elasticity of supply and
7:10the elasticity of demand and the change
7:13in marginal cost. So in the case of a
7:15tax, think of delta MC as the tax. The
7:19increase in the price is increased by
7:21this factor which is always less than or
7:23equal to one because in my notation
7:26elasticity of demand is a negative
7:28number. So for example elasticity of
7:31supply = 4 elasticity of demand=
7:35minus1 this number would be 4 fths. So
7:39the price would rise by 80% of the
7:42tax. So unless demand is perfectly
7:46inelastic, which means this is zero,
7:49sellers will not be able to pass the
7:51full tax on to
7:53buyers. In that case, you would just
7:55hold their price profits the same level,
7:57right? Profits wouldn't rise, they would
7:59stay hot. See, but we saw in the other
8:01case with the cost increase with
8:02inelastic demand, uh that's a great
8:06case. Price would rise by the amount of
8:07the increase in marginal cost. So as
8:09long as the increase in marginal cost
8:11was greater than the increase in average
8:13cost in the fully in elastication, no
8:15profits will arise. Doesn't hold for a
8:17tax. All right. So now let's let's put
8:19this let's put this to use. Let's say I
8:22have a competitive industry out there
8:25and the government wants to put a tax on
8:27this competitive industry for some
8:29reason. Maybe they create pollution and
8:31they want to reduce pollution. Maybe
8:34they want to tax this industry to raise
8:36revenues. whatever they have a desire to
8:38put a tax on this
8:40industry. Now, as we spoke about
8:42earlier, one of the most organized
8:44groups of opposition to that tax is
8:47likely to be these
8:49producers, right? Consumers likely to be
8:52less organized because they're they're
8:54this smaller fraction of their
8:57livelihood. So, is there let's assume
8:59these guys are very powerful in
9:02Washington. They have veto power. Let's
9:04say they can veto a tax if they want
9:09to. What is the government to do in that
9:13case? What is the government to do? They
9:16want to put
9:17on dollar per unit
9:20tax and they know they can't get this
9:24through without the consent of the
9:28industry. Is there something the
9:30government could do? Yeah. Yeah, I mean
9:32you could give back part of the tax
9:33revenue as a lump sum if you're just
9:35concerned about the level of output in
9:37the industry to compensate the
9:40producers. Now, do you have to give it
9:42all back? No, just enough to compensate
9:45them for the loss in profit. So my
9:47example here, you have to give them back
9:5020% of the subsidy.
9:53What's important is that I mean of the
9:55tax you give back
9:5720%. But you can't you can't lower
10:00marginal cost, right? If you gave them
10:02back 20% as a per unit subsidy, it would
10:06defeat the whole purpose. You have to
10:08give it back in some other
10:10way. Any examples you see of that? Any
10:15practical examples you see when the
10:17government does
10:27Think about I come to an industry and
10:29think about not don't think about a tax.
10:31Think about like I'm going to put some
10:33pollution requirements on this industry.
10:36I'm going to require that for every unit
10:38you put on some pollution cost abatement
10:41that raises your cost by the same as the
10:43tax.
10:46What do they do? Yeah. Do we give away
10:48some of the sulfur dioxide permits?
10:50Okay, so you might give away some of the
10:52permits. How many of them do you have to
10:54give away here? 20%, right? I mean, you
10:58have to give away 20%. Or maybe you
11:00grandfather the existing plants. You say
11:03some of those existing plants you don't
11:06have to bring up to the current
11:09requirements. So you bribe it. You
11:12wouldn't want to do it on the new
11:13plants, right?
11:16grandfathering the new plant. Even if
11:17the old plants are more polluting than
11:19the new plants. So you really don't want
11:22to grandfather the old C plants from the
11:24point of view of reducing pollution.
11:26You'd really love to have them upgrade
11:28those plants. You can't do it on the new
11:30plants.
11:33Why? What happens if you grandfather?
11:36What if you you say I'm going to tax all
11:38the old plants and the new plants?
11:41they're
11:43exempt entry. That's the worst of all
11:46worlds from their point of view. That's
11:47like this whole thing turned on its
11:49head. That's lowering that's not raising
11:52marginal cost and raising average cost.
11:54It's not going to be it's not going to
11:56fly. Okay? So, you have to do it through
12:00grandfathering the old plants. Or maybe
12:03you just make them upgrade all their
12:05plants, but you give them some kind of
12:07quote adjustment assistance. You give
12:10them some tax credits to help them
12:12adjust to all the new things, right? We
12:14hear about those things all the time.
12:15We're going to have a special tax credit
12:17for this industry to compensate them for
12:20all the costs that we're imposing on it
12:22through this regulation or something
12:23else. Again, it can't be tied to output.
12:27Can't be tied in a way that's going to
12:29lower marginal cost. a battle going on
12:32now in the energy area uh whether the
12:37administration originally said they
12:40wanted to impose a carbon tax. Unions
12:42been fighting that as proposes said they
12:44cap the trade. Um maybe you'll get into
12:48that at some point. Yeah, think about
12:50that because that difference between
12:52those two approaches has an enormous
12:54different effect on the firms who are
12:57producing the carbon. one case they
12:59might increase their profits. The other
13:02case for sure they'll reduce their
13:03profits. Yeah. If you give them if you
13:05give away all the permits very likely
13:08you're going to over you're going to
13:10overcompensate the firms, right? Because
13:12remember you only have to compensate for
13:15the firms for the part that they can't
13:17pass through. If they can pass through
13:1980% of the of the cost, you even giving
13:23them 20% of the permits in some sense is
13:25going to likely almost fully compensate
13:27them. You got to worry a little about
13:29this triangle. That's one. You got to
13:30worry about that triangle a little bit.
13:32But in some sense, if they can pass
13:35through 80% of the cost, you really only
13:36want to give them about 20% of the
13:38permits to make them whole. It's not
13:40clear you want to make them whole, but
13:41if you're forced to give them 100% of
13:43the permits, you've like benefited
13:46twice, right? They've been able to pass
13:48it through to consumers and at the same
13:50time you've compensated them for the
13:52full amount. who's getting hit
13:54by ei in either case by what's going on
13:58the consumer will end up paying a higher
13:59price in either case but how that
14:01affects the profits of the firm be very
14:03different in both these cases okay any
14:07questions that that people
14:09have give you another
14:11example how many people have heard of
14:13the master settlement
14:15agreement there's an agreement between
14:18tobacco
14:19manufacturers and states attorney
14:22general from I think 46 of the
14:26states that was
14:28intended to have and it had the effect
14:31of having tobacco manufacturers pay to
14:35the attorney generals
14:37uh compensation for past harm
14:41uh and cost incurred as a result of
14:43tobacco smoke. How was that
14:47structured? They owe all this to they
14:50according they to get themselves out
14:51from under
14:53lawsuits. The tobacco companies agreed
14:57to pay the attorney generals a very
14:59large sum of money for these past
15:04liabilities. You think they just wrote
15:06him a
15:08check? You think they agreed, okay, you
15:10got us. Let's write you a
15:15check. Is that was that the best deal
15:19that the attorney generals and the
15:21tobacco companies could
15:26strike? I think a lumpsum check from the
15:29tobacco companies to the attorney
15:32generals would be a a jointly maximizing
15:36deal from the point of view of the two
15:37players at the
15:40table. What would the tobacco
15:42manufacturers like to do?
15:48If they read a dollar's worth of checks,
15:50it cost them a
15:54dollar. How do you think they set it
15:58up? How do you think they set up those
16:02payments? What did they condition your
16:04payments on? Yeah, I might be wrong, but
16:06wasn't it a match between uh the state
16:08funding and the No, no, never mind. They
16:13set it up. Yeah. uh on the number of the
16:16packs of cigarettes that you sold in
16:17that state was in the past um in the
16:21future. Yeah. Going forward, they said
16:24we will agree to pay to the attorney
16:27general a certain amount which is
16:29currently about 38 cents I think per
16:34pack going
16:37forward. Why was that better than so
16:41what they did and how did they determine
16:42that 38? Well, they forecast how many
16:44they were going to sell. They took the
16:46amount they agreed to pay them and
16:49divided that by the output and got 38
16:51cents. Why was that a better way for
16:54them to do that
16:58than like doing it on the past than just
17:00writing them a check which would the
17:02past let's say they did on the past you
17:04say well they compensated all the harm
17:07they caused in the past or looks
17:09attractive to the tax on the past.
17:12Yeah. But
17:14what what did what did this tax going
17:18forward do
17:20different? Yeah. Reduce the likelihood
17:23of more hassle.
17:26No, that was a
17:28claim. What's the most important effect?
17:30Changes marginal cost rather than Yeah,
17:33absolutely. And so what it meant is the
17:35tobacco
17:36companies didn't bear most of the cost.
17:40Most of the cost was actually passed
17:42through to
17:43smokers because the elasticity of supply
17:46is much greater than the elasticity of
17:48demand. Most of the cost was passed on
17:51to the smokers and came off of the
17:53tobacco companies. So the tobacco
17:56companies viewed it as a much less
17:58costly alternative to writing a check to
18:01the government. The attorney generals
18:03got their money either way. So they
18:05didn't really care. In fact, they could
18:08go back to the state and say, "Look, we
18:09got all this and then by top of that's
18:11going to reduce smoking in the future."
18:12So, it's double benefit for us. But I
18:16guarantee you the biggest reason they
18:18did it is cuz they wanted these guys to
18:22pay. That was the big deal by the two
18:24parties at the table having a system set
18:28up that was based on future output with
18:32marginal
18:33costs. I mean, I don't understand. They
18:35didn't write them a check. They could
18:36have set it up in the same way that they
18:38would be passed on afterwards to
18:39consumers. How do they agree? How do
18:40they enforce that where they have a
18:43collusive agreement among them like
18:45among all uh manufacturing firms? I'm
18:48guessing we have a thing called antirust
18:50laws which would make that illegal,
18:53right? You cannot as a group of
18:54producers get together and say we're
18:57going to impose a tax on each other and
18:59split the revenues among ourselves. No,
19:01just the cost increase because you're
19:02writing a check to the government. Yeah.
19:04What kind of cost? Well, why would that
19:06raise prices? Why would it cost increase
19:08increasing? It's increasing their cost
19:11and right. What's they have to finance
19:13that cost somehow? No, they're
19:15increasing their marginal cost.
19:18Does it increase their marginal
19:21cost? If you just write a check to the
19:23government for all your past bills, is
19:26that effective? It doesn't increase your
19:27marginal cost. I agree. But it could be
19:29that if you just finance it in one
19:31period. So maybe one period would be
19:32extremely high prices. Can you raise
19:34prices if it doesn't raise your marginal
19:36cost? How's that cont consistent with
19:42equilibrium? If they have the ability to
19:44raise prices, why would they do it at
19:46this cost,
19:47right? You have to have some me that's
19:50the whole point of markets. You have to
19:52have some mechanism that pushes the
19:54market price up. I mean, but here you're
19:56describing a case where the car that
19:58every every firm has to write a check to
20:00the government, right? So it's not like
20:01only independent of the amount of output
20:03they produce. If the check they have to
20:06write is independent of the amount of
20:08output they have to produce, why is that
20:11going to affect market prices?
20:18Market prices are determined and I don't
20:20care if it's monopoly, igopoly, whatever
20:22it is, right? Market prices are
20:25determined by the interaction of
20:27marginal cost on one side of the market
20:29and demand or willingness to pay on the
20:31other side of the market. That's even
20:32true with igopoly or whatever structure
20:35you want. Doesn't even matter that it's
20:36perfect competition or
20:38competition. Point is if you want to get
20:41past it, if you want to pass it through
20:43to consumers, you have to raise marginal
20:45cost. And by structuring it in such a
20:48way that the payments that individual
20:50firms made were conditional upon the
20:54amount of output sold, that's the only
20:56way they could get past it. On the other
21:00hand, if there had been monopoly, if
21:02there had been if there was a single
21:03cigarette producer who was a monopouist,
21:05he wouldn't care. Make it conditional on
21:08output or make it the other way. In
21:10fact, you wouldn't make it conditional
21:11on output because that would distort his
21:13incentives and actually reduce the size
21:15of the
21:16pot. In fact, it's only because it was
21:19comp has some element at least of
21:21competition that it was optimal to
21:23structure it the way we
21:25did. So the key here is you cannot you
21:29can't affect prices without affecting
21:31marginal cost.
21:35might affect the pro the pricing of the
21:38monopas, but it won't affect his profits
21:39depending upon which way you do it.
21:41Yeah, it'll be worse if you make it
21:43contingent because you're going to
21:44distort his incentives and the total pie
21:47will be smaller. So actually in a
21:50monopoly case, you'd want the lumpsum
21:52payment. Monopolist would like the lump
21:54sum and regard I mean the parties
21:56together would prefer the lumpsum
21:57payment in the monopoly case but in the
21:59competitive case they like it because
22:02it's like an agreement to reduce alpha
22:05which is profit maximum. Yeah. So does
22:07it violate antitrust law for the tobacco
22:10companies to sign an agreement that 38
22:12cents of your charities? That is a
22:14question about whether that agreement
22:16violates the antitrust laws. Actually is
22:19I think still being litigated, but that
22:22that is a that is a question because it
22:24wouldn't violate the state antitrust law
22:27because they're all doing it and maybe
22:29the federal government. We have a
22:31federal system. So in general, states
22:33were doing it uh as a revenue raising
22:36method. Whether it violated federal
22:38antirust law is a separate question. The
22:42fact that this is a settlement with the
22:43attorney generals helps them a lot. This
22:46was a settlement with a private party.
22:47They definitely could get a trust in
22:49this because you you know you can
22:52settlements can violate antitrust laws.
22:55There's a potential that this
22:57could but it certainly has the effect
22:59it's no doubt that it's a collusive
23:02agreement to reduce output and raise
23:04prices. There's no question that that's
23:05what it is and it was in the interest of
23:08the parties to do that because the
23:09demand for cigarette the aggregate
23:11elasticity is less than once and kind of
23:14more like a half. So they're going to be
23:17able to pass a lot of that on. Yeah. I'm
23:21finding this easy to understand by
23:22thinking of the flip side of it. So like
23:25if we don't want people to adjust
23:27quantities then we don't tax the margin
23:31you know and that's really like you know
23:32when you take public finance that like
23:34they pound that into you they pound that
23:37into you but remember that's the theory
23:39of optimal
23:40taxation. We're talking here about the
23:43theory of kind of political t you
23:45political theory of taxation which the
23:48two don't always agree even the no
23:50externalities from smoking. No, there
23:53was no externalities from smoking. They
23:54I'd still reach this agreement, right?
23:57There's still uh raise revenue for the
24:00states and they and the firms might be
24:02happy with it as well, but there's no
24:06external effects that you should have an
24:08exposure. There was no external effect.
24:09So there' be no case for any from an
24:11optimal tax point of view. Yeah. Well,
24:14yeah, that's true. That's true. But I'm
24:15thinking so we're thinking of this as
24:17what's optimal from the perspective of
24:19the firm you know and maybe from the
24:21state government and the government I'm
24:23thinking of the firm and the government
24:25at the table deciding what's from their
24:28joint interest whereas the public
24:29finance course you're talking about has
24:32the government the firm and the consumer
24:35all sitting at the table and asking from
24:37our joint point of view exactly what's
24:39maximizing exactly you know it's key in
24:42this case that the political process is
24:44kind of cut one of those parties out
24:46from the table so that the optimal or
24:49equilibrium tax structure doesn't look
24:51like it would come out of the simple PF
24:54model. In fact, it's coming out of a
24:56model of interest group politics that in
24:59fact says from the interest group point
25:01of view, which are these two parties at
25:02the table, this tax makes a lot of
25:05sense. And the consumers often get hurt
25:08because they're much less organized
25:09politically than many producer groups.
25:14Yeah. Any questions that people have?
25:17But again, this gets back to where we
25:18started today that when you think about
25:21markets, you have to think about
25:23marginal cost and you have to think
25:25about the impact of marginal cost
25:27affecting price. And when you think
25:29about profits, it depends on how price
25:31is affected relative to average cost.
25:35And many changes in markets whether they
25:37be induced by changes in the prices of
25:39inputs, taxes, regulation and the like
25:42are very profitably viewed through that
25:44kind of
25:45lens. If you just focus on the firm then
25:48you get a not a very rich model because
25:50on the firm's point of view anything
25:52that increases cost is bad. Anything
25:53that lowers cost is good and that's the
25:56end of the story. What gets more
25:58interesting is when you combine the firm
26:00and the market. When you combine the
26:02firm and the market, you get our
26:04interesting example with regard to the
26:05ethanol subsidy. You get this
26:08interesting example of the MSA
26:10grandfathering in the case of pollution
26:12regulation. All those kinds of things
26:15kind of come together out of that same
26:16kind of basic framework. There's a lot
26:18of
26:19literature, contemporary literature that
26:22puts a lot of emphasis on the individual
26:24consumer and the individual firm and
26:26never really goes back to the market.
26:29You can see how that works out in the
26:31market. There'll be lots of examples
26:33where what the intuition you get out of
26:35thinking about the firm uh doesn't hold
26:38the market level like in the profit
26:40case. At the firm level, raising cost
26:43clearly lowers profits absolutely and
26:46necessarily whether it's competitive or
26:48not. At the market level, as we saw,
26:50there are many cases in a in a
26:53competitive case where at the market
26:54level, what holds for the firm doesn't
26:56hold for the market. Profits, for
26:58example, could go up. And we're going to
26:59see many other applications of that
27:02idea. Yeah. And that's because prices
27:04again in the market are not determined
27:08by average cost. Just because your costs
27:10go up doesn't mean you're going to be
27:12able to pass that on. The degree to pass
27:15to which it's passed on depends on how
27:16marginal cost is changed. That insight
27:19is very very important because that's
27:22what makes it that's the whole reason
27:24why the MSA was structured the way it
27:27was while there's so much farmer support
27:29for ethanol substitutes because you get
27:32all those units priced according to the
27:34market. Okay, any questions? Yeah. Yeah.
27:38So I I'm a little confused about the
27:40difference between market power and
27:42rent. I've always thought of market
27:44power as equivalent to being able to
27:46extract rent. But now from this
27:47discussion I can see that maybe that's
27:50not quite well. The market power can be
27:52what? Some source of rent on time
27:55because rent fundamentally is about
27:59receiving a return bigger than your
28:01opportunity cost.
28:05Absent market
28:07power that always owes itself to a
28:11difference between your opportunity cost
28:13and somebody else's opportunity cost.
28:15Because in a competitive environment,
28:18price is going to be equal to
28:19opportunity cost for the marginal
28:22producer. And therefore any rents earned
28:25by other producers
28:27owes to the fact that they have lower
28:29opportunity cost than the others than
28:32the margin as shown in the elasticity of
28:35supply. Right? Their opportunity costs
28:38are all the same that you'd have a
28:41infinitely elasticity elasticity of
28:42supply. So it's only when they differ
28:45that you get the upward supply. Now in
28:48the market power
28:50case, so rent, think of
28:53[Music]
28:54rent, we're going to that it happens
28:58when price is bigger than average cost,
29:00right? That's kind of what it
29:03is. And under
29:06competition, we have price equals
29:08marginal cost, which is bigger than
29:10average cost. And that's where the
29:11redstone comes. Under market
29:17power you have price big I mean price
29:21bigger than marginal
29:23cost and therefore even if you have
29:25marginal cost equal to average cost you
29:27can have rents in the market power case.
29:30You see in the market power case you get
29:33the gap between price and average cost.
29:36Not necessarily by there being a gap
29:38between marginal cost and average cost
29:40but there being a gap between price and
29:43marginal
29:44cost. So there's two ways to get a gap
29:47between price and average cost. One is
29:49to have a gap between price and marginal
29:51cost. Have marginal cost equal to
29:53average cost or have price equals
29:56marginal cost but have marginal cost
29:57bigger than average cost. So rents can
30:00come from either of those two channels.
30:02So then in the basketball
30:05example, would would the player have
30:08market power? Because I mean I'm having
30:10a hard time. It depends upon whether by
30:13changing his supply, he can affect the
30:16market price of basketball talent. If
30:20DBon James earns a high wage because he
30:22just has lots of talent, but the price
30:25per unit talent is independent of how
30:28much he supplies to the market, then
30:30we'd say there's no market power there.
30:32LeBron James has no ability to raise the
30:36market price of basketball talent. He
30:39just takes he's like the price taker who
30:41takes the price of talent as what it is
30:44and he produces where price equals
30:46marginal cost.
30:48So there's
30:50clearly a lot of competitive rents for
30:53LeBron James. That is most of his rent
30:57is due to the fact that the market price
31:00for basketball talent times how much
31:03basketball talent he has greatly exceeds
31:05his opportunity cost off the basketball
31:08court. So there's no question that's the
31:09big source of rent. He may be a unique
31:12enough asset that he has some small
31:15amount of market power in itself that
31:17he's not only a great basketball player,
31:19but he's not that substitutable with
31:21other basketball players, probably less
31:23so on the court than he is as a
31:25spokesman and other things like that. So
31:27there may be some monop market power
31:30there. So, I'm not saying LeBron James
31:33has no market power, but the fact that
31:36he's earning this high wage relative to
31:38his opportunity cost probably doesn't
31:39have that much to do with market power.
31:42So, for a lot of talents, an actor, an
31:45actor has, so to speak, a lot of acting
31:47units. So, in that sense, there's no
31:50market power. On the other hand, one of
31:51the top actors may also have, you know,
31:54some monopoly power that they're not
31:56they're not substitutable by uh 15 other
31:59actors, so to speak. They're more than
32:02just the equivalent of 15 actors. So
32:04they're said, yeah, exactly. They have
32:07some ability to affect by adjusting
32:09their supply the price that they receive
32:12per unit talent. Yeah. So, like if
32:14LeBron is gaining a lot of extra
32:16attention for the NBA and getting other
32:18players noticed, is he having market
32:19power then? Because if he wasn't playing
32:21in the NBA,
32:23no, but presum What do you mean by
32:25market power? I mean, that's just part
32:27of his marginal
32:28product. Part of his marginal product is
32:31that he raises demand for the you know,
32:34you know, he in he's gets compensated in
32:37some sense for that. Now it depends
32:39depends on how the contracting works
32:42whether in fact that's internalized
32:44through the NBA that the NBA
32:46internalizes that that benefit and that
32:51raises his salary at the Cleveland
32:54Cavaliers. That's a question. But I
32:56don't mean think that means he has
32:57market power. That just means there's
32:59another source to his Marshall product.
33:02Market power has to me market power is
33:05the
33:06ability through my choices to affect the
33:10market price for services. That's really
33:12what market power is about. Yeah. It's
33:14just the demand curve for for your
33:16services, right? Yeah. I mean I mean I'm
33:20sketchy on that because there are a lot
33:21of people who have price bigger than
33:24marginal cost who I don't like to think
33:27of in any meaningful sense of having
33:30substantial market power.
33:34Maybe you and I disagree.
33:36I mean they can have a completely
33:40elastic demand with respect to the
33:42changes in their supply. They could
33:43still be earning a lot of rent. I would
33:45say they have no monop market power in
33:47that case. Yeah. I mean they're not
33:50affecting the PR. Let's say they they're
33:52equivalent to 30 other units. Yes. As
33:54they change their amount of units,
33:56that's small in terms of total number of
33:58units in the market. They have no that's
34:00clearly what's going on in LeBron James.
34:01I call that I mean I don't think the
34:03definition any different in this case
34:06than other case of monopoly or monopoly
34:08yet on the elasticity of the demand for
34:12your your services. more it's more more
34:15when we go in the other direction. So I
34:16think it's pretty clear that you would
34:18want to think of LeBron as largely
34:20competitive. Absolutely. That there's a
34:21price for talent and he earns a lot
34:24because he had lots of units of talent
34:27and the price per unit talent wouldn't
34:28be much different if LeBron retired
34:30Mark. There'd be some reduction in
34:32talent in the NBA, but the price per
34:35unit talent would be wages in the NBA
34:37wouldn't change a lot if LeBron left the
34:40market. It's not like Alcoa. If they
34:42went out of the aluminum market when
34:44they're in their heyday, world would
34:46have been really different. Prices would
34:47have adjusted. It's more on the other
34:50side. You got the guy running the hot
34:52dog stand on the corner and he's pricing
34:55his hot dogs above marginal cost. Do you
34:58want to call him as having market power?
35:01I've always found it hard to think of
35:02him as it's it's the reverse case for
35:05the guy who has maybe a positive price
35:08cost margin. Do you really want to call
35:10that substantial market? Certainly in an
35:12antirust sense, you probably don't. I
35:15would say the hot dog guy has no
35:16monopoly power because if he tries to do
35:19too much, he'll probably be replaced by
35:21another hot even though he has price
35:23bigger than marginal cost. Yeah, even
35:24though he has price bigger than marginal
35:26cost. Yeah, you have to you have and he
35:27has no rent.
35:30It's like the reverse of LeBron James.
35:32He's got price bigger than marginal cost
35:33and no rents. LeBron James has price
35:36equal to marginal cost and lots of
35:37rents. Lots of rents. That's the
35:39difference. All right. Any questions
35:41before we go on? All right. Next next
35:45one I want to go back to. So we did the
35:47markets. I want to go back to marginal
35:50value equals marginal cost because that
35:53really is
35:56important. Why? What makes it important?
35:59Where where do we use
36:02that? Where do we use marginal value
36:05equals marginal cost?
36:08Turns out that's probably our more one
36:11empirical link between
36:15theory and data. And I let me try to
36:18explain why. For example, let's think
36:21about welfare economics. How do we do
36:27welfare? So
36:29remember remember our where we start. So
36:34you think about some consumer. I'm going
36:35to
36:36maximize U of X1 up to XN. I got a bunch
36:42of goods and goods. Prices are
36:45P1 up to
36:47PN. And my first order conditions are
36:50partial u partial x1 = lambda
36:54p1 down to partial u partial xn equals
36:59lambda pn where
37:02lambda is the marginal utility of
37:05income. All right. What are those? How
37:07do you interpret those
37:11equations? What makes those equations so
37:13important?
37:17makes those so
37:22important. What makes them so important
37:25is these are fundamentally things we're
37:27interested in which is how much do
37:30people value various units of the
37:33goods. These are things that we can
37:35observe namely the prices. So these
37:38would imply partial u partial x1 /
37:44partial u partial x2 = p1
37:52p2.
37:54So what do you call
37:57this? What's p1 over p2?
38:01What price? Relative
38:04price. Is there another term? What does
38:06it express? What does it tell me? Give
38:09me a verbal description of what it
38:13says. What's the units is it
38:17in what units is that number in?
38:26What units like inches, pounds,
38:30dollars, P1 P2?
38:33What? times how how many how many times
38:36I value?
38:37No, it's in units of what? What is P1
38:41over P2?
38:42P1. What units is P1?
38:45No. Dollar for X1 for X1. Dollar for
38:50X2 equals X1 for
38:53X2.
38:55Right. Sorry. X2 for X1.
38:59It tells me how many units of good two I
39:04have to give up to get a unit of good
39:07one, which is what we would call the
39:10cost of good one. You would say cost of
39:16good one is three units of good two. So
39:18if this is six and this is two, then a
39:21cost of a unit of good one is equal to
39:24two units of good two. So this is the
39:28cost of good
39:31one in terms of good
39:38two. What's
39:47this? What's this? This is the cost of
39:51being of good one in terms of good two.
39:54What must this be?
39:57Value. This must be the value of good
39:59one in terms of
40:09good terms of
40:14good.
40:16So, and here I'm thinking of value as a
40:19measure of what the person is willing to
40:21pay. That's what value is. Value is a
40:24statement about what you're willing to
40:26pay. Cost is a statement about what you
40:29have to pay. And in equilibrium, those
40:32two things are going to be equal. A
40:34maximizing agent is going to equate the
40:37marginal value of the good to its
40:39marginal
40:41cost. That's what we have. Okay. So
40:45what's nice about this
40:48framework is it allows us to take an
40:51observable thing like
40:53this which we think of as measurable in
40:57the market the cost of good one in terms
40:59of good two and infer from that the
41:04value that people place on good one
41:07relative to good two. We learned
41:08something about their
41:10preferences from the data.
41:14This marginal value equals marginal cost
41:17framework allows us to translate
41:19observable quantities and prices into
41:23the underlying concepts of individual
41:26value. And once I know value, I can
41:29start thinking about things like welfare
41:33economics.
41:35So let's say I had two two points in
41:39time or two alternative states of the
41:42world.
41:46Matt guy had income
41:49M1 prices
41:52P1 I'm sorry M1 P11 up to
41:57PN1 and he bought
42:00X11 up to
42:02XN1 and then I have another state M2
42:07P12 up to PN2. So I change these income
42:10and
42:11prices and I get
42:15X12 up to X
42:18N2. And I want to know and let's assume
42:22these are pretty close. I made small
42:23changes in the prices and income. And I
42:26want to know is he better off afterward
42:29than he was
42:30before. How does this framework allow me
42:33to do that?
42:38How does just knowing this, knowing
42:40nothing other than what we have here,
42:43how does this allow me to ask or answer
42:45that
42:49question? Well, let's think about
42:54u depends on x1 up to xn.
42:58Therefore,
43:00du equals the sum over i
43:05ui dx
43:08i i i = 1 to n, right? Or let me write
43:12it as partials. partial u partial x i dx
43:17i which is equal
43:20to the sum from i = 1 to n of lambda
43:27pi dx
43:29i equals lambda the sum from i = 1 to n
43:34pi
43:37dxi. So what does that allow me to
43:41do? Just knowing that
43:45formula, what does it allow me to
43:48say? Can I measure whether this guy is
43:51better or worse off?
43:55Yeah, if I know the prices which I can
43:59observe and I know the changes in
44:01quantities, I can measure whether
44:04utility for this person is going up or
44:05down based on that formula.
44:09So knowing nothing more than marginal
44:13value equals marginal cost is important
44:16because the guy got more haircuts and
44:19fewer hot dogs and and more
44:21entertainment and less sleep and all
44:24those things. I know how much he values
44:26each of those things cuz I see their
44:29price. I know how much he values each
44:32one. I can therefore know how he trades
44:35those off and I can figure out whether
44:38he's better or worse
44:39off. That's the basic idea behind
44:43welfare
44:44analysis is that we use this equality
44:47between prices and marginal
44:50values to tell you whether he's better
44:53or worse off. I can divide this formula
44:57through. du over lambda equals the sum
45:01from i = 1 to n of pi
45:06dxi. What does that
45:15measure? What is
45:18that? Is
45:21that what is this?
45:29So now I'm not just telling you whether
45:31he's better or worse off. I'm telling
45:33you how much better or worse off he is
45:37based on
45:38what? What units is this
45:41in? Dollars. This is his dollar gain or
45:45loss from the change in consumption that
45:48we just
45:49saw. So we're getting the dollar gain or
45:52loss quantified that way.
45:57Maybe I could do it this way. du over
45:59lambda divided by his
46:02income equals sum of i = 1 to n p
46:08i x i / m dx i / x i which is equal to
46:17sum from i = 1 to n. Let's call it si.
46:21That's the share of good
46:23I times delta x I. This is the
46:29percentage
46:31change in x
46:35i. So now I'm going to say he got a 2%
46:40gain in effective
46:42income measured as a share weighted
46:45average of the changes in consumption.
46:49So just that formula
46:52alone allows me to do a lot in terms of
46:56going from the
46:57data to thinking about changes in
47:00welfare and measuring this person's
47:03values. Okay, I haven't done anything
47:06yet with demand theory, substitution,
47:08and all those things. Just the fact that
47:11people are optimizing gives me a window
47:15into thinking about how changes in the
47:17world affect them. It's a very effective
47:21way for looking at the world. And what's
47:23nice about these
47:25equations is you don't need
47:29econometrics because these things only
47:32depend upon shares, things that are
47:34directly measurable. I don't need
47:35exogenous variation. I didn't assume
47:38that anything here was
47:40exogenous. There's nothing nothing's
47:42exogenous in this world. I just have to
47:45measure the shares. They only depend on
47:47kind of first order. Yeah, that equation
47:50are using one or two. Well, that's a
47:53good question. For a small change like I
47:55have here, it doesn't matter. In
47:57practice, probably we'd want to use some
48:00kind of combined average between the two
48:02of them. This is what we would call a
48:05quantity index.
48:07Okay, think about how we do this with
48:08discrete data,
48:10right? Little digression here about
48:13price and
48:15quantity. Okay, so price and quantity
48:18indices, how do they
48:19work?
48:21So, think of it this way. We got the sum
48:24from i = 1 to n of x
48:29i 2 pi 2 / the sum from i = 1 to n of x
48:37i1
48:39pi. So that's like the ratio of kind of
48:44nominal expenditures in the two
48:47periods. Well, you can decompose this in
48:50lots of ways. One way to do it is the
48:53sum x i2 p i2 / the
48:58sum of x i 2
49:04pi1
49:05[Music]
49:06times the
49:08sum of x
49:10i 2
49:13pi1 divided by the sum of x i1 pi1
49:20one. That's one decomposition that we
49:23often use. What does that look? What is
49:26that? Well, we're one of these is what
49:28we call a price index and one of them
49:30was what we call a quantity index. So,
49:34which is the which is the price index?
49:40This is a price index. What makes it a
49:42price index?
49:45the same quantities are the same and the
49:47price is different. So I've decomposed
49:50this into a price index. This is what
49:53kind of price index the usual
49:56terminology. This be a Laspar's price
49:59index. No, this would be a posh price
50:01index. I'm sorry because we're using
50:02period
50:032 posh price index. This is what we call
50:07a quantity
50:08index which is analogous to what we just
50:11did up there. We held the prices fixed
50:14and looked at the change in
50:16quantities.
50:18Okay. So you were asking about period 1
50:21versus period 2. This is one method.
50:23What's the other decomposition? Well,
50:25the other decomposition is x i1 p i 2
50:30sum of x
50:32i x i1
50:36pi1 times sum of x
50:40i1 p
50:42i 2 i sorry x i2 p i2 sum of x i 2
50:53pi1 I do that
50:55That should be the second one should be
50:57one and two. Oh, I'm sorry because I'm
51:00doing uh this is my
51:03one and
51:06two. There we go. In this case, this is
51:10still our price index. In which case,
51:12this is period one price index and this
51:15is period 2 quantity index. So whenever
51:19you do one of these decompositions, you
51:21always end up with one of them being in
51:24period one and the other one being in
51:26period two. Yeah. Just make sure to
51:28answer. So you want to understand
51:29whether du is positive or negative for
51:31instance and you're assuming that in
51:32situation one is maximizing also
51:35situation two. So even if the price
51:38difference is very small and if you
51:40insert there you want the law is going
51:42to be negative. Yes. Should be either
51:45way. Correct. should be should be you
51:49know these can
51:50disagree but they don't you know usually
51:52they don't you know usually when they
51:54disagree if the bundles are pretty close
51:56these are going to give you pretty much
51:57the same answer. Okay.
52:01Now
52:03the question is is well how do you
52:06choose between these two measures? The
52:08answer is neither one is right. They're
52:11kind of biased in particular ways right?
52:13Everybody's studied price indices
52:15before. So, let's just review this one
52:18here. This price index is biased which
52:20way
52:22what? To make it look like prices went
52:24up or prices went
52:26down. It's biased when making it look
52:28like prices went up because you're
52:31evaluating it. You're ignoring the
52:33ability to substitute when prices
52:36changed. So, price index here is biased
52:39up. This price index is biased in the
52:42other direction to make it look like
52:44prices went down. The quantity indices
52:47of course are biased in the other
52:49direction. In fact, quantity indices are
52:52easy to think
52:54about. What do the quantity indices look
52:57like? What are the amounts x1 and
53:03x2? When I look at a quantity index,
53:06what am I looking at?
53:09What am I doing in quantity
53:12index? Anybody
53:15know? Remember, we're approximating the
53:17guy's indifference curve. The quantity
53:20index does
53:23what? It approximates his indifference
53:26curve by the budget line. It uses the
53:29budget line which is tangent to the
53:30indifference curve as an
53:32approximation to the indifference curve.
53:35So the quantity index any of these
53:38quantity index always amounts to using
53:39the budget line as an approximation to
53:42the indifference curve. And you're
53:44saying he's better off if he ends up
53:47above the budget line and worse off if
53:49he ends up below the budget line. Now of
53:52course the quantity index is kind of
53:54always
53:55biased toward making it look like he's
53:58better off, right? because there's
54:01points like here and here that he's
54:05above the budget line but below his
54:07indifference
54:08curve. Okay, so that quantity index is
54:11always kind of biased just like a price
54:13index is always biased in the other
54:15direction. So quantity and price
54:17industries each have their corresponding
54:19bias. One of the interesting things
54:20about the quantity index
54:23is it's a first order approximation
54:25regardless of what the utility function.
54:27I see that in fact have any utility
54:30function that in terms of its properties
54:33out outside of the tangency and it'll
54:37still be a first order approximation
54:39regardless of the utility. What Gary's
54:41saying
54:42is this term sum of
54:46pi
54:48dxi is a first order approximation to a
54:51utility function.
54:54first
54:56order
54:58approximation to the
55:05utility. Similarly, the sum of xi
55:10dpi which is the corresponding thing
55:14which is a price index is a first order
55:18approximation to what?
55:26What's that? A first order
55:28approximation.
55:32No, remember think of this as like a
55:34first order approximation to the utility
55:36function because the PS are proportional
55:39to the derivatives of the utility
55:42function. What is
55:46this? Where does that come from as a
55:48first order approximation?
55:52What is a price index intended to
55:55measure? Cost of
55:57living. What is the perfect measure of
56:00the cost of living? The cost
56:03function. What's a first order
56:05approximation to the what are the
56:07derivatives of the cost
56:08function? The derivative of the cost
56:11function with respect to the price of
56:12good I is equal to
56:15what? Quantity of good I. So this is
56:19just equal to the sum of partial C
56:22partial PI DPI which of course is a
56:26first order approximation to the change
56:28in the cost
56:29function. So one of the things that
56:31makes price and quantity indices so
56:34useful
56:36empirically is that they provide each
56:39provide first order approximations to
56:41the things we most care about. the cost
56:44function in the case of a price index
56:47and the utility function in the case of
56:50a quantity
56:51index. So much empirical work that we do
56:54when we talk in you know you want to go
56:57to the data you go to the data you want
56:59to talk about consumption of
57:03cars. Well, how do you measure the
57:05consumption of cars and how do you
57:07measure the price of cars? Just add up
57:10cars or television sets. How do you
57:13imagine the quantity of television sets
57:15and the price of television
57:17sets? Do we want to just add up the
57:20number of television
57:22sets? 65 in TV with a 9 in TV and say
57:26that's two
57:28TVs. No, that doesn't make any sense.
57:31You have to
57:33aggregate. And this theory gives us a
57:36very natural way to aggregate.
57:39It says if I want to measure the change
57:41in the consumption of TVs, I probably
57:43ought to wait them by their
57:46prices. If I want to measure the
57:48consumption of basketball players, this
57:50gets back to what Gary said. LeBron is
57:52more of a basketball player than the
57:5412th man on the
57:55bench. It counts as more units. So, I
58:00want to wait them by their prices when I
58:02go to add them up.
58:05So this idea of a quantity index is a
58:08very natural thing. We go out and we
58:11measure TV consumption by the total
58:14value of the different TVs. You measure
58:17you measure car consumption by the value
58:19of the cars consumed things like that.
58:21Yeah. So if you have a bunch of people
58:24in the same market buying a good
58:27regardless of what their utility
58:29functions are or their cost functions
58:31are to a first order approximation this
58:34holds every single one of them. I mean
58:36that's how we use it when we go into a
58:37market. Exactly. I think that's a very
58:39important drop. Exactly. So the reason
58:42and the way I like to think about we
58:44have these empirical
58:45constructs which are price and quantity
58:48indices. They have a very strong
58:51theoretical foundation because in fact
58:54they're providing us with first order
58:56approximations to the things we care
58:58about which is the cost function and the
59:02utility functions of
59:04individuals. All right. Any questions
59:06that people have? Now in
59:09practice not everything is derivatives.
59:11So you have to worry about this. So one
59:13of the things is there there's always
59:14two
59:16decompositions and one
59:20decomposition is in terms of initial
59:23change in prices at period 2 quantities
59:26and change in quantities at period one
59:29prices or the change in prices at period
59:321 quantities and then the change in
59:34quantities at period 2
59:36prices. We can also combine them.
59:40If you take this to the 1/2 times this
59:43to the
59:441/2, you get a, you know, you get a
59:48Fisher type ideal index that combines
59:50the two. And you can do the same with
59:53Y. And so you ask me what I would do
59:55before. Probably I would do something
59:57like that. I would calculate this
59:59measure and this measure, take some
1:00:01geometric average, calculate these two,
1:00:03and take some geometric. That's probably
1:00:05what I would do. Okay. Any questions?
1:00:09that people
1:00:10have. Now, how do we think about this in
1:00:14terms of our simple model we had on
1:00:15board? You don't really need utility to
1:00:18do this,
1:00:27right? So, here's a demand
1:00:30curve. Here's an initial
1:00:33price P 0 quantity
1:00:37Q0.
1:00:39Here's the new price P1 quantity
1:00:47Q1. How
1:00:50does So what's the loss in consumer
1:00:55surplus from the rise in price? How do
1:01:00we usually measure
1:01:01that? How would we measure that?
1:01:08Well, one thing we might look at is we
1:01:11might look at this area
1:01:13here and say that's a measure of about
1:01:15how much consumer
1:01:18lost. What does the Laspair's price
1:01:20index amount
1:01:25to? What's the Laspair's price index?
1:01:29How does it measure the
1:01:31loss from the rise in price?
1:01:38It uses this
1:01:40box, right? So that's why the laspair's
1:01:44number is too big. It it it it adds that
1:01:46extra
1:01:47triangle. What does the posh measure do?
1:01:52It uses this
1:01:54box, right? Which is too
1:01:57small. What does Fiser do? He uses kind
1:02:01of a triangle a kind of a kind of half
1:02:04that box which is pretty which in this
1:02:06case looks like it's pretty good.
1:02:08Okay. So that would be an example of how
1:02:13to do that. Now what do we do when we
1:02:16have multiple periods? What do we do? So
1:02:19like for example let's say this was
1:02:22let's say this was
1:02:2419 99 and this was 2000.
1:02:28Well, then chances are these two
1:02:31decompositions are going to be pretty
1:02:33similar because the numbers just aren't
1:02:35that different. Which base you use isn't
1:02:37going to matter very much. But let's say
1:02:41I wanted to do the sum from y = 1 to n
1:02:45of x i
1:02:482009 pi
1:02:512009 / the sum from i = 1 to n x i
1:02:581959 pi
1:03:001959. So I want to decompose the growth
1:03:04in in nominal income say between 1959
1:03:08and
1:03:092009
1:03:12into prices and quantities. How much did
1:03:14real prices go up? How much did prices
1:03:17go up? And how much did real consumption
1:03:19go
1:03:20up? Now what do I
1:03:24do? What do I do?
1:03:29Do you have any years in between?
1:03:31Okay. If I Yeah, but let's let's assume
1:03:33I didn't. If I didn't have years in
1:03:35between, I'm left with a pretty bad
1:03:37choice, right? I could use the 1950
1:03:42quantities, in which case I'd ignore a
1:03:45whole bunch of goods that I have today
1:03:47because they didn't even exist in
1:03:491950. Or if I tried to use the 2009
1:03:52quantities, I'd be in real trouble
1:03:54because I'd have a bunch of goods. I'm
1:03:56stuck. So your answer is, let's use
1:03:57years in between. So if I have years in
1:03:59between, what am I going to do? You um
1:04:02you calculate price growth from year to
1:04:04year. So you chain the the Fisher index.
1:04:07So I'm going to look at price growth
1:04:10from 59 to 60 using like the 59 or 60
1:04:14bundle. Then I'm going to do it from 60
1:04:16to 61 using the combination of 60 and 61
1:04:20bundles. And I'm going to chain them all
1:04:23together. What does that correspond to
1:04:25in terms of my graph?
1:04:30the third dimension? No, not necessarily
1:04:33a third dimension. Think about something
1:04:34like where the price of this good has
1:04:36been going up and up and up over time or
1:04:38down and down and down over time. What
1:04:40does it correspond to?
1:04:44So I
1:04:47got
1:04:48P
1:04:511959, Q
1:04:571959, P
1:04:592009,
1:05:02Q
1:05:062009 and
1:05:12got quantities moving along like
1:05:18that. So what am I doing when I use that
1:05:21chain
1:05:24index? Kind of like integrating, right?
1:05:26I'm kind of like cuz this from 59 to 60,
1:05:29I'm going to be using the 59 to 60
1:05:32category. So my Fisher index is going to
1:05:34be like some little approximation to
1:05:37that trapezoid. And then my next Fisher
1:05:40index is going to be some approximation
1:05:42to that trapezoid. So when things are
1:05:44kind of moving monotonically like this,
1:05:47the Fisher index is a pretty good idea
1:05:49because what it's doing is it's always
1:05:51kind of calculating your day on the
1:05:53margin given what you're consuming at
1:05:55that point. It's kind of like
1:05:57integrating rather than taking either
1:05:59this huge box. If I do it in one swoop,
1:06:02I either use this huge box or this
1:06:04little box. Here I'm kind of
1:06:06interpolating and I'm using kind of the
1:06:09quantities as they go through time to
1:06:11infer how much he values them. Now, when
1:06:15prices are moving around, the Fisher
1:06:18index can be kind of a bad idea because
1:06:20you can have prices go down and then
1:06:23prices come back up and they keep moving
1:06:25around and you end up back where you
1:06:27started even though your Fisher index is
1:06:29going to say you have this growth in
1:06:31value because in some sense all these
1:06:34appro you're adding approximation errors
1:06:36as you go. Okay? Whereas in this case,
1:06:40you're kind of getting a better and
1:06:41better approximation.
1:06:43So for long periods of time, these
1:06:45Fisher indices have these chained
1:06:47indices have something going for them.
1:06:49They amount to kind of always adjusting
1:06:51the marginal value. In fact, they're
1:06:53easier to think about in in the quantity
1:06:56index case because in the quantity index
1:06:58case, what you're always doing is you're
1:07:00looking at the quantity growth from year
1:07:021 to year two using the marginal value
1:07:05as of that point in time. And then you
1:07:06go from two to three using the marginal
1:07:08value at that time. So, you're just
1:07:10accumulating all this growth and
1:07:12marginal
1:07:13value. All right, any
1:07:17questions? All right, another question.
1:07:21Let's see how we're going to use
1:07:22marginal value one more time. I got a
1:07:24couple other examples, but we're not
1:07:25going to get to them all.
1:07:32So let's
1:07:34say let's go back to
1:07:37cigarettes
1:07:39and let's just say
1:07:42people
1:07:45learn that smoking is
1:07:53bad in terms of the
1:07:58might be good in other
1:08:00videos. And I see
1:08:05consumption
1:08:07falls
1:08:1110%. How can I use that
1:08:14fact to help me
1:08:17measure how
1:08:20much they value that change in perceived
1:08:23value?
1:08:26Any ideas?
1:08:36Yeah, we're going to use the same idea.
1:08:38So, let's go back the implication of
1:08:41these
1:08:43equations when it comes to a demand
1:08:49curve. Here's the price of
1:08:51cigarettes. Here's the quantity of
1:08:55cigarettes.
1:08:58Okay. The implication of these equations
1:09:01is that the marginal value of
1:09:04consumption is equal
1:09:07to the price that people are paying.
1:09:09Right? That's what it says. Just divide
1:09:11that equation by lambda. It says P
1:09:14equals the marginal dollar value with
1:09:17consumption. So when people learn that
1:09:19smoking is bad for them, what happens to
1:09:21demand?
1:09:26shifts
1:09:33in. So, what did I just tell you
1:09:36happened when they learned that smoking
1:09:37was
1:09:41bad? Consumption
1:09:44fell 10%.
1:09:51How much do we
1:09:53estimate they thought the marginal cost
1:09:57of smoking went
1:10:01up? How do I measure? What do I need to
1:10:06know to
1:10:11measure that distance?
1:10:24the elasticity of demand. That's all you
1:10:26need.
1:10:28Right? That
1:10:31is all I need to know is the elasticity
1:10:34of demand. If I know the elasticity of
1:10:37demand, let's say I figure out that
1:10:41cutting that raising the price of
1:10:44cigarettes 50 cents would have the same
1:10:48effect as this of a 10% reduction in
1:10:51consumption. I can infer that their
1:10:53marginal value they place on their
1:10:55health is about 50 cents a
1:10:58pack. Okay. To do that we use nothing
1:11:02more than this first order condition
1:11:05here. I can infer the change in their
1:11:08marginal value of
1:11:10consumption from the change in their
1:11:13quantity. I can infer because their
1:11:15consumption fell 10% that their demand
1:11:18curve must have shifted down by about
1:11:22the amount equal to 10% divided by the
1:11:25elasticity of demand. The less elastic
1:11:28is demand, the more we must infer that
1:11:32the value of smoking is
1:11:34declining. So if cigarettes are selling
1:11:37for $5 a pack, I get a 10% reduction in
1:11:42smoking. If the elasticity of demand is
1:11:45minus a half, that means I must have had
1:11:48the equivalent of a 20% decline in
1:11:51value. So that would be a dollar a pack
1:11:55in perceived health cost.
1:11:57See how I did that? I inferred how much
1:12:00the demand curve shifted down off of how
1:12:03much the quantity changed at a given
1:12:07price. Now, I assumed in this example
1:12:10that the price of cigarettes was
1:12:13unchanged because I assumed I had a
1:12:15perfectly elastic supply. What if supply
1:12:17was upward slope?
1:12:22Then what happens? Then what's delta Q?
1:12:27What's delta Q in that case? Is it
1:12:30bigger or
1:12:32smaller? What would what would I have if
1:12:36supply were upward sloping to get a 10%
1:12:39reduction in quantity? Now, what would I
1:12:41need? Would it would the value of
1:12:43consumption have to be bigger or
1:12:45lower? Would the health costs have to be
1:12:48bigger or smaller? It'd have to be
1:12:51bigger because not only would it have to
1:12:53be pushing me up the demand curve, it
1:12:55have to be pushing me down the supply
1:12:57curve. So I'd have to divide by the
1:13:00elasticity of demand plus the elasticity
1:13:04of supply and absolute value, right? I'd
1:13:06have to be dividing by the two. I'd have
1:13:09to be dividing by actually I be careful
1:13:12divide by
1:13:17this stuff like that. That's what I
1:13:21have. We can come back to those formulas
1:13:24tomorrow. Any questions that people
1:13:27have? All right. Right? So, I'd have a
1:13:29more I'd have to infer that they had
1:13:31even a greater value on health because
1:13:34what's going to happen in that case is
1:13:35the price of cigarettes going to fall
1:13:37and that's going to keep them from
1:13:39reducing consumption as much when when
1:13:42uh when health costs
1:13:45rise. All right. So I think we'll do a
1:13:48little more
1:13:49tomorrow. But it turns out that just
1:13:51that basic first order
1:13:54condition is really fundamental for so
1:13:58many problems that we deal with in
1:13:59economics. It's fundamental for a theory
1:14:01of of measurement where we use price and
1:14:04quantity indices. It's fundamental for
1:14:07aggregation of goods.
1:14:10It's fundamental for inferring value for
1:14:12things like health consequences, value
1:14:14of
1:14:15life, cost of transportation, all kinds
1:14:18of things. We'll come back to more
1:14:20examples tomorrow. There's a lot of
1:14:21things that come right out of there. And
1:14:24the great thing about everything that
1:14:26comes out of those equations is
1:14:28impirically it's easily measured because
1:14:30they only involve things that are
1:14:33directly observed. prices as proxies for
1:14:37for
1:14:38uh for marginal values. You don't have
1:14:41to have something where you estimate
1:14:43elasticities. Elasticities are going to
1:14:45be important. We're going to come up
1:14:47with elasticities later. And for this
1:14:49example to get a magnitude often you'll
1:14:52need an elasticity for as in this
1:14:54example, but for
1:14:56valuation we only need to know the
1:14:59shares. Okay. All right. Why don't we
1:15:02stop here? We'll pick up from here
1:15:04tomorrow.

About this transcript

This page contains the full transcript of Price Theory Summer Camp, Day 1: Lecture by Kevin Murphy by Becker Friedman Institute University of Chicago, generated from the public captions YouTube serves with the video. The transcript has 10,434 words across 1,666 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.