YT YouTube2Text

YouTube2Text → Transcripts → ECON20110 Fall 2013 - Lecture 16: Public Goods

ECON20110 Fall 2013 - Lecture 16: Public Goods — Transcript

by Glen Weyl · 10,091 words · 1,814 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:01okay everybody um
0:03so
0:04uh competition policy which we talked
0:07about in the last lecture
0:09basically tries to stop firms from
0:12colluding with each other tries to stop
0:14them from internalizing the
0:16externalities across one another but
0:18just as often we want to promote this
0:20sort of collaboration between different
0:24individuals organizations to try to
0:26solve externality problems
0:30and in particular
0:32the most extreme form of this is when we
0:34have public projects that bring diffuse
0:37costs or benefit that are spread across
0:39societies
0:40in which individual voluntary initiative
0:43is insufficient
0:44so i'm going to start just to link it to
0:46our last lecture on competition policy
0:49by talking about cornea's collaboration
0:52problem which is sort of a duel to
0:55cornos competition problem but has the
0:56opposite implications
0:59and
1:00where in the cornea uh
1:03competition problem we found that as the
1:04number of people got large things
1:06started working very well in cornea's
1:08collaboration problems the number of
1:10people gets large things will work very
1:12poorly
1:14and we'll talk about the relationship of
1:16this to public goods problems
1:20and
1:21we'll talk about the definition of
1:23public goods and bads about private
1:25solutions
1:27to
1:28public goods problems and and the
1:29difficulties that they face
1:33and therefore why it's necessary to use
1:35coercion in public goods and bads
1:37problems
1:39and as an example of that we'll talk
1:40about eminent domain and the assembly of
1:42complementary goods
1:44and then we'll talk about given that
1:46coercion is necessary how we can go
1:48about deciding
1:50on uh whether to apply coercion in a
1:53given case or not
1:54and that will lead us into questions of
1:56voting and the flaws of standard voting
1:59procedures
2:01and
2:03this will be sort of a downer ending
2:05because we'll say that you know
2:07none of these voting procedures are a
2:08particularly good way to solve these
2:10problems and when we come back from
2:11break we'll talk about an alternative uh
2:14procedure that that might work better
2:16okay
2:17so corno uh was is one of my sort of
2:21heroes he came up both with the first
2:25model of monopoly and the first model of
2:28competition
2:29and he also came up with
2:31uh
2:32this model of collaboration
2:34between firms so um he imagined a case
2:38where there were n firms rather than
2:40producing products that were perfect
2:42substitutes for one another they were
2:43ones that were perfect complements we
2:45needed all of the things in order for
2:47the firm to in order for the product
2:50uh to be valuable
2:52uh a classic example of this is i don't
2:55know if you guys remember when the
2:56challenger crashed it was probably
2:57before most of you were born it was
2:59before most of you were born uh and
3:02um they found that the reason it had uh
3:07blown up was because there was a single
3:09o-ring that had failed o-ring is like
3:12something about this size
3:13and so because one o-ring had failed the
3:16whole thing blew apart right
3:18so that's what uh
3:21as a classic case of complementary goods
3:23things are only as good as the weakest
3:25link right
3:26and so um in this case
3:30uh the
3:32value of the good as a whole is really
3:35only the value if you get all of the
3:36pieces together
3:38which has a price which is the sum of
3:40the prices of the individual components
3:43and demand depends on this sum of prices
3:48now again for simplicity let's
3:50assume that each of these goods has
3:52constant marginal cost
3:55then
3:57the
3:58profits that each firm makes
4:00is going to be the price
4:03minus their cost times the quantity
4:05which depends on the sum of all the
4:06prices
4:07another way of representing that is that
4:09it's their markup price minus cost
4:13times quantity which depends on the sum
4:15of all the markups plus the sum of all
4:17the costs
4:20okay
4:21so um
4:22sue well
4:24what would be the optimal
4:27price for an individual firm
4:30to set
4:31uh if they were choosing their price in
4:32this market okay so you take it through
4:35zero yep
4:40to maybe do this one that one yeah for
4:44example um
4:45what's the optimal price
4:48you'd want markups well
4:50maybe you want to do it on the board
4:52yeah
4:53yeah
5:03okay so prophet is
5:10sorry
5:14quantity of
5:17n plus c so i'm listening some of the
5:19markers of c's that
5:21cost so
5:23i guess
5:24c is
5:26um
5:27c is a constant right okay and
5:31m depends on price so
5:33um
5:34well does that okay so how does that
5:36depend on price so m is the sum of
5:37everybody's minus c buys summed up so
5:40that's yeah um
5:43and let's say every firm takes all the
5:45other firms prices and therefore markups
5:47is given okay sure in which case it's um
5:50i guess it's like your price p i plus
5:52the sum of
5:54i don't know like j
5:57yeah
6:00u i minus
6:03c of n
6:14you're maximizing right okay so
6:16i mean because choosing your markups
6:18equivalent to choosing your price right
6:20right because
6:21right cost is
6:25so let's
6:35i'm just going to call this
6:36stuff yeah yeah
6:39that's great um so take the derivative
6:41of that it's right to p i it's n equals
6:43zero you have
6:44um
6:46one
6:50so you have a pi times this term yes i'm
6:53going to use a product
7:06is that that's i'm just like i do need
7:09to separate those things no no no can i
7:11just because you can just treat that
7:12whole first term
7:14as a term and you can use the product
7:16rule applied to that
7:18right so what's the derivative of first
7:19term with respect to pi um just
7:22just
7:24um
7:25just one yeah one times
7:46yeah
7:47times the derivative of the inside which
7:49is just one yeah
7:51times this term here yeah which is just
7:53the markup question
7:55so
7:56okay so now let's solve for the markup
7:58of i have to okay right so
8:01markup of i is equal to um
8:05so negative
8:08q
8:09di plus
8:17and i guess this is just this can just
8:19be written as m plus pi or m plus c
8:22would that just make more sense
8:33exactly
8:35nice work
8:37um
8:42okay and let's
8:45and remember that we can get capital m
8:47by just summing all these up right
8:49so capital m is then going to be equal
8:51to negative n times q over q prime
8:55right
8:59so note that this is exactly the
9:01opposite of the formula we got under
9:02competition
9:03under competition mi was going to be
9:06equal to
9:07capital q over n divided by q prime
9:10because
9:11because uh each q i was only one over n
9:14of the capital q
9:16here it's multiplied by n so rather than
9:19the markup diminishing as n increases
9:21the markup is going to increases and
9:23increases right
9:24so this is called the double
9:26marginalization or multi-marginalization
9:28problem and the more firms you have the
9:31more
9:33markup there is so the basic intuition
9:36here
9:37is that um the firms are all gonna
9:41take a cut right all of them are
9:43complementary products right each one is
9:46going to demand a markup on their chunk
9:49of the product and all those markups are
9:51going to
9:52stack up and multiply with the number of
9:54firms
9:55so let's consider a simple example
9:58imagine that capital q is equal to 1
10:00minus p this is our like linear demand
10:02example from before and ci was equal to
10:05zero
10:06okay tanner what is going to be q over q
10:09prime in this example well q is just one
10:12minus t yes
10:15which is negative one so times negative
10:18yeah so you guys have to make it long so
10:20so it's just
10:22one minus p negative one yeah yeah
10:25right
10:26um okay
10:28so
10:29uh so well what's then going to be the
10:30equilibrium price
10:32uh using your formula
10:35okay so
10:37p i minus
10:40so remember we said that if capital m is
10:42equal to n times this
10:44okay right and there's no cost
10:48okay so if capital m is equal to n
10:52maybe you can do it on the board again
10:53yeah it's probably gonna be the easiest
11:04um so capital m okay
11:07that's a or capital p which is equal to
11:09capital n because there's no cost right
11:11okay so capital p
11:20was it just n times one minus p yeah
11:22okay exactly so we're gonna
11:24solve out for p okay
11:28so um
11:37yep
11:53what that gives us
11:55is that the price is equal to n over n
11:57plus one
11:59so uh
12:01is dake here
12:04okay
12:05uh
12:06does somebody else want to tell me what
12:07the quantity is then
12:09yeah mike galvin um you substitute the
12:12thing that's well just derived into the
12:13equation above and you get one minus n
12:15over n plus one which and what's that uh
12:21just make a common denominator okay uh
12:25n plus one minus and so one over
12:30one over n plus one right yeah
12:32uh
12:34so what happens to this is the number of
12:36people gets larger so well
12:38it just goes to zero yeah exactly
12:42um
12:43and so eventually
12:45all these people
12:46taking their little cut
12:48kills the market by the death of a
12:50thousand cuts right uh
12:53the mark it's exactly the opposite um
12:57of the
12:59uh cournot
13:01compliments problem i mean the corner
13:03competition problem where the price went
13:04to the efficient level here price is
13:06going to the point where it destroys all
13:07demand
13:09right in fact the formulas are exactly
13:12the opposite right because
13:19remember that in corneau
13:21um
13:22competition i don't know if you guys
13:23remember but the price was the markup
13:26was equal to one over n plus one and the
13:28price and the quantity was equal to n
13:31over n plus one and how they're exactly
13:33the reverse uh for this demand at least
13:36in fact these two problems are exactly
13:37dual to one another and um i just want
13:40to try comparing that
13:42so um in the cornea quantity competition
13:45problem everyone takes each other's
13:47quantities as given and the price
13:49depended on the sum of all the
13:50quantities and profits were the product
13:53of markup and quantity
13:55under cornell's collaboration problem
13:57everyone takes each other's prices as
13:59given the quantity depends on the sum of
14:01the prices and the profits are the
14:04product of markup and quantity right so
14:05the problems are exactly identical
14:07except reversed
14:09um
14:10and uh therefore any result that's true
14:13in the cornea competition model is true
14:15here but reversed
14:17firms hurt each other by raising their
14:19price rather than bending rather than
14:21hurting each other by raising their
14:23quantity and therefore prices are going
14:25to be too high even from the perspective
14:27of the firms so if they could coordinate
14:29if they could
14:30collude if they could merge they would
14:32reduce the price and make everyone
14:34better off
14:36okay
14:37so
14:41mergers between firms producing
14:43complementary products which i think
14:45someone brought up in an earlier class
14:47is called
14:49a vertical merger
14:52and vertical mergers the reason they're
14:54called vertical is often
14:56one way that firms end up being
14:57complementary is one firm produces
14:59something that they sell to another firm
15:01they then do something with that and
15:03they sell it to another firm so they if
15:04you draw a diagram of it they're
15:06vertically related rather than
15:08horizontally competing with one another
15:10um and vertical mergers are usually not
15:13uh considered to be anti-competitive
15:16um in fact
15:17there's a
15:18reverse version of corneau's theorem
15:21that applies in this context that we've
15:23just sort of been going through and
15:24philip can you guess what that is um
15:27let's see as the number of
15:29uh
15:30we're producing perfect compliments
15:32tends to infinity uh quantity tends to
15:35zero yeah and the price tends to
15:36whatever price will get no one to buy
15:38the good absolutely
15:40yeah austin what's this like
15:44derivation
15:46so robert this was originally derived by
15:49corneau in 1838 but did robert burke use
15:52this to argue against competition policy
15:54at some point probably because every
15:56argument that there exists against
15:58competition policy robert burke has made
16:00at some point
16:01so
16:04driving forces against getting rid of
16:06like vertical integration wise
16:08uh yes i mean this is a classic chicago
16:11school argument and he was part of that
16:13broad camp so i would not be surprised
16:16but i don't know in particular
16:18um
16:21so markups become individually small but
16:24they collectively kill all demand
16:26because there's all these layers of
16:27monopoly um you can also think of this
16:30as being like a tragedy of the commons
16:32problem everyone wants their markup but
16:34they don't realize that in the process
16:36of getting their markup they're killing
16:37the market that's
16:39you know the goose that lays the golden
16:40egg right
16:42so
16:43let me relate this to an even more
16:45famous problem
16:46imagine that everyone can make costly
16:48contributions to a charity ai
16:51and the value that each person gets
16:53depends on the total amount of
16:54contributions that are made which is
16:56capital a which is the sum of ai and
16:58imagine each person's valuation
17:01vi prime
17:02uh
17:04is greater than zero
17:06so then each individual
17:10gets a value of this but the value of
17:12society is the sum of that
17:15and um imagine that everyone's the same
17:18the contribution
17:20that is going to bring
17:22total value of this only gives a value
17:25of this to each individual
17:26so if they're symmetric
17:28the individuals are going to solve v
17:31prime is equal to n whereas society
17:34would set v prime equal to 1.
17:36and v prime equal to n if v
17:38is concave is going to lead to much much
17:41much lower
17:43contributions towards the public good
17:45then would be socially optimal
17:48so this gives you basically the same
17:49result as corneau
17:51as n gets large you're going to be
17:53driven down to basically no provision of
17:55the good
17:56because everyone will free ride on other
17:58people
18:02their hope that other people will
18:03provide the good
18:05and so this is called the collective
18:07action problem in a large population
18:09with voluntary contributions
18:11public goods become arbitrarily
18:13underprovided
18:15so
18:15what exactly uh do we
18:18mean
18:19uh by a public good
18:21this was sort of an example of it but
18:24more broadly
18:26christian what's the definition of a
18:28public good more broadly
18:30um
18:33um so i probably won't be able to get a
18:35very precise yeah sorry but um
18:38isn't something that would be
18:40more like socially optimal for like a
18:42central government to produce because
18:44it's
18:45um generally a
18:48um non-rival
18:49what does that mean yeah um it means
18:52that you can't really exclude
18:54anyone from using it no that's not
18:56excludable wait but that's the other
18:58part yeah yeah
19:00and not right yes so yes that is the
19:02definition and non-rival means um that
19:05like my use of the product doesn't
19:06affect you exactly
19:08exactly
19:09so non-rival which i think is the most
19:11important part is that one person
19:12benefiting it from it does not preclude
19:14others from benefiting from it
19:16there's zero marginal cost of supplying
19:17it to other people
19:19non-excludable which is secondary but
19:21still important is that everyone
19:23receives the benefit regardless of
19:24whether they're willing to pay or not
19:26there's no way to exclude people or
19:27charge them for participation
19:30so examples failing the first condition
19:32are called congestible
19:34public goods
19:36and examples failing the second
19:37condition are called excludable public
19:39goods
19:40so
19:41intellectual property is actually a good
19:44example of an excludable public good
19:46because like if you create music for
19:48example there's really no marginal cost
19:50of other people listening to that music
19:52but it is often possible using copyright
19:55protections to exclude people from using
19:56it right
19:58now in reality exclusion can be more or
20:00less costly so
20:02in the internet age it's become more and
20:05more difficult to actually exclude
20:07people from using
20:09uh
20:11music right and eventually it may become
20:14a non-uh excludable good because it may
20:16become just too difficult to stop file
20:18sharing so um that's why i think
20:21non-rivalry is sort of the more
20:22fundamental condition
20:24however you know we had a whole lecture
20:26on ip right
20:27and we didn't call that stuff public
20:29goods and i think that the real reason
20:33the real thing that differentiates
20:34public goods from ip is ip is usually an
20:38example where we think that there is
20:41like a genius type person or some
20:43entrepreneur who has a vision about what
20:47needs to happen in the marketplace like
20:48you know uh bill gates or steve jobs
20:51with the iphone or whatever or there's
20:53some small group of people with this
20:55vision for the broader public whereas a
20:57public good is something where we think
20:58the information about whether it should
21:00be created or not or the most important
21:01information resides not with one
21:04individual but with the broader public
21:07and therefore it's important that we
21:08elicit that information from the public
21:11this resembles to some extent
21:13the distinction we were making between
21:15things that were driven by expertise and
21:17externalities and things that were
21:19driven more by diffuse information among
21:21the public and as we discussed almost
21:23everything is some mixture of these two
21:26and so these are really two categories
21:28into which we put things rather than a
21:30completely sharp dividing line
21:32okay but using the more traditional
21:34definition of public goods philip could
21:36you provide some examples of pure public
21:38goods
21:49yeah yeah those are good examples
21:51um environmental services like the air
21:54water
21:55weather
21:56you know climate change could affect the
21:58weather and that's sort of a pure public
22:00good the existence value of some goods
22:02charity
22:03um
22:05ideas often are very hard to exclude
22:07anyone from they just spread and uh
22:10everyone benefits from them
22:12um yeah austin so is ideas like basic
22:15like research yeah i think to a large
22:18extent basic research is basically a
22:19public good
22:21um
22:24uh
22:25i mean could you give an example of a
22:26congestable public good
22:29congestable is if it's non-violence it's
22:32non-exclusive you can't exclude people
22:35but more people using it does diminish
22:37its value
22:38um
22:41what bridge no
22:43is that big one
22:44yeah that i mean it depends whether
22:46you're able to like put guard posts on
22:48the bridge and charge people for getting
22:49onto it
22:51then you know
22:52yeah yeah
22:55so that's a good example
22:56amenities
23:00amenities like uh
23:02lakes parks
23:05things that are hard to keep people out
23:06of
23:07but
23:08which do get congested
23:11uh kim could you give an example of an
23:12excludable public good
23:16um
23:19uh yeah that's that's a good example
23:21another example
23:23would be um land reform so imagine i
23:27don't know if you guys know but you know
23:28in many developing countries
23:30at least during some periods there was a
23:33group of very wealthy people who owned a
23:35lot of land and they broke up the land
23:38and they gave it to the people who were
23:39tenants on that land
23:40now you could say that that's sort of
23:42like a public good because once you've
23:44broken up the land you might as well
23:45give it to all the people who are living
23:46on it because they know how to farm it
23:47better but of course you could exclude
23:49one of those people from that and like
23:51kick them off their land and so that
23:53that's sort of a excludable public good
23:56okay
23:57now um
23:59many uh yeah mike why are those lands
24:01not congestible
24:04well i mean the idea is that each late
24:06piece of land sort of belongs to someone
24:08who's been farming it for a long time so
24:10the value that it would give to anyone
24:12other than that person is very small
24:13okay
24:14that's sort of the idea okay that makes
24:16sense
24:16yeah christian i just i don't know
24:20an idea about the ideas does that yeah
24:22specify like specifically non-rival
24:25ideas or like the feel good ideas
24:26because i mean wouldn't there be a lot
24:28of ideas like that might give you an
24:29advantage in a market or some new method
24:31or something that's not covered by ip
24:34but
24:34yeah i mean but there are certainly
24:35ideas that go the opposite way where
24:37they're actually anti-rival
24:40right if everyone knows it then you yeah
24:42i mean like for example like most like a
24:45lot of pop culture is that way it
24:46depends whether you're a hipster or a
24:48preppy but if you're a preppy then then
24:50the more people who know the i you know
24:52the cultural idea the better and if
24:54you're a hipster it is rivalrous right
24:59uh you know the the famous uh
25:02do you guys know the famous joke about
25:04um
25:05why did the hipster burn his uh mouth
25:11because he tasted the soup before it was
25:12cool
25:18or how many hipsters does it take to
25:20screw in a light bulb
25:23anybody know
25:26it's a number you wouldn't have heard of
25:31um
25:34so
25:39many things
25:42many things rather than uh bringing
25:45diffuse harms
25:49rather than bringing different huge
25:50benefits people bring diffuse harms so a
25:53classic example of that is land takings
25:55so this is when the government comes in
25:57and they're going to build some like you
25:58know new uh
26:00shopping mall or bridge or public work
26:04and they need to take a bunch of land
26:06from people
26:07right in order to build that
26:09and
26:10it's only worth taking it if they can
26:12get the whole lot of land right because
26:14if they only get some subpart they're
26:16not really able to build the project or
26:17not one that says beneficial
26:20and so that's a public bad because
26:23either you take it from everyone or you
26:24take it from no one
26:26right so it either hurts all the people
26:28or none of them
26:30another example
26:32is
26:33corporate acquisitions if someone's
26:35coming to buy a corporation either
26:36everyone who currently owns the shares
26:38is going to lose their shares or nobody
26:40will because the person wants to have
26:41control over the whole company
26:43similarly if you make a settlement on a
26:45debt so
26:47where like you decide not to take the
26:49company into bankruptcy because
26:50someone's going to come and buy it out
26:52of bankruptcy well then everyone has to
26:54give up their claims on the company or
26:56nobody does right another example of
26:58this is that the federal communications
27:00commission has been trying to reassemble
27:02parts of the spectrum so that they can
27:04be repurposed to do ymax do you guys
27:06know what wimax is it's a system where
27:09you can blanket a whole city and wi-fi
27:11basically it's so like normal wi-fi is
27:13like a range of 30 feet this would have
27:15a range of like miles
27:16um and in order to do that you need to
27:19like clear out chunks of the
27:20electromagnetic spectrum which has been
27:21auctioned off to all these different
27:23people but you need like a coherent
27:24block and so you have to either get all
27:26of that or none of it right
27:28um yeah just a quick question fcc
27:31spectrum so when they auction off
27:32originally is that like a set period of
27:34time no it's it was an ad infinitum oh
27:36so that really sucks yeah
27:38or we're always like a 99 year lease or
27:40something like that you know which is
27:42very long relative to the length of the
27:44telecoms industry i mean after 99 years
27:46we might not even be using the
27:48electromagnetic spectrum anymore who
27:49knows right so
27:51are there any restrictions on water
27:56no i don't think there were many
27:57restrictions actually
27:59um
28:00so i i have a paper about this
28:03application in in particular
28:05um
28:07patent pools are another example of this
28:09so remember we talked about the
28:11multi-marginalization problem
28:13one case where i think that arises a lot
28:15in practice is when there's like some
28:17technology and there's a bunch of
28:19different patents that are necessary to
28:20make that technology work they're all
28:22complementary and people are really
28:24scared that like if they're all just
28:26licensing on their own it will like kill
28:28anyone's ability to use the technology
28:30because they're going to have to be
28:32negotiating with all these different
28:33people they're all going to take a cut
28:34the price is going to be way too hot so
28:36sometimes what they do is they build
28:37what's called a patent pool
28:39where they put all these things together
28:41and they label them all as what's called
28:44standard essential patents
28:46and that forces people to license them
28:49at what are called fair reasonable and
28:51non-discriminatory rates which makes it
28:54much easier to license all the things
28:56together to avoid the
28:57multi-marginalization problem but in
28:59order to do that you need to get all the
29:00technologies together which again
29:03is going to be one of these things
29:04that's going to take the tech the patent
29:06rights away from all the people
29:09polluting projects are another version
29:10of this where it'll hurt a few set of
29:13people right
29:15now fundamentally these aren't that
29:17different from externalities problems
29:18but there are cases where we think that
29:20the victim's knowledge
29:22about how costly it is to them might be
29:25really important so think about the
29:26example of land takings someone might
29:28know that this land is really valuable
29:29to them that this is where their
29:30grandparents grew up other
29:33people might just be you know a yuppie
29:35who's happy to take it as long as they
29:36get another uh spot right
29:39and so this again creates a collective
29:42action problem
29:43but now in this reverse case it's
29:45usually called the holdout problem
29:48and the idea is that everyone demands a
29:49large payment in order to accept the
29:52harm in the same way the free riders
29:54were making small payments because they
29:56didn't want to uh
29:58make a large contribution to the benefit
30:00right
30:00okay
30:02so
30:02um
30:03these problems are
30:05pervasive they show up in lots of areas
30:09of the economy
30:12and
30:13i'm going to argue that we need coercion
30:15in order to solve these problems in most
30:17cases
30:20but before we turn to that i want to
30:22consider some potential private
30:23solutions the most famous one is
30:25probably altruism people often just out
30:28of the goodness of their heart do public
30:30spirited things like giving to charity
30:32voting
30:33volunteering for military service
30:37and while people choose to do these
30:38things it's not motivated by classic
30:40self-interest
30:43culture plays an important role in
30:45encouraging these things
30:46people may want to do it in order to uh
30:49have the prestige like i think that
30:51plays an important role in many of the
30:52charitable foundations why rockefeller
30:54and vanderbilt gave stuff it also make
30:56can make you look good to you to a
30:58potential mate so if you're uh someone
31:02like bill gates who maybe uh had a hard
31:04time in his romantic relationships
31:07uh it could make you much more
31:09attractive if you're seen as being
31:11public spirited
31:15warm glow can be another important part
31:17of this so this is when people feel good
31:19about themselves for doing something
31:21right regardless of the actual effects
31:23that it has on society
31:25another thing is you might just be
31:27signaling you might be trying to show a
31:29potential employer that you're an honest
31:31person who cares about society and not
31:32someone who's just going to steal from
31:34the company right and while i think
31:36there's no import question that these
31:38things are important and help
31:41i also think that they often require a
31:43slow development of culture in order to
31:46embed these values in people and they're
31:48not don't do a very good job responding
31:51to exactly in a particular circumstance
31:54how much people
31:56value something because if they're all
31:58relying on inside of an individual's
32:01head how much do they think uh other
32:03people will respond to this yeah rishi
32:05generally when you
32:06try to work altruism into into a model
32:12well i think that those are usually the
32:14least compelling ways of doing it i
32:15think the more compelling ways of doing
32:17it are thinking about things like these
32:19where like you might be altruistic but
32:22it's sort of motivated by something else
32:24or those tend to be the most interesting
32:26models i find but yeah sometimes you do
32:28put it straight into the utility
32:29function
32:30uh yeah bart um
32:32so for a lot of these those examples
32:34like are essentially self-interest
32:37so like the slow development
32:39the culture towards like putting these
32:42sort of behaviors yeah isn't that also
32:44like somewhat perverse
32:46like you're encouraging attitudes that
32:49might not
32:50be great
32:52well i mean
32:53culture
32:55can be useful to overcome some of these
32:57problems
33:00it also can be perverse i mean so for
33:02example like around environmental issues
33:05there has developed this view that you
33:08really should by all this green stuff
33:10but if you look at all the green stuff
33:12almost all of it like nobody would be
33:14buying it under the optimal carbon tax
33:16because it's actually like way too
33:18expensive and like most of the stuff
33:19that people would actually be doing
33:20under the optimal carbon tax would be
33:23like very minor changes that are
33:25extremely not costly to people and that
33:28would account for most of the carbon
33:29mitigation so this stuff can easily go
33:31overboard right because in order to like
33:33get the public spirit going so it can so
33:36i actually think that you know the
33:38problem with a lot of these cultural
33:40ways of solving these problems is
33:42they're just not like very
33:45accurate at conveying the relevant
33:46information like
33:48to address huge widespread problems that
33:52have relatively simple solutions they
33:54can work okay
33:55but they're not all that good like
33:57staying up to date
33:59with
34:00like what's the way that a price system
34:02would be with like what exactly is the
34:04informational conditions and how much
34:06should i be doing of this thing exactly
34:08i think it's a little bit like prior to
34:11um like market exchange with money
34:14people used to have barter systems or
34:16even before barter they would like do
34:17something good for someone else and
34:18someone else would owe something back to
34:20them and so forth but it was very hard
34:22in that system to communicate like
34:23exactly how much i owed to you and like
34:26to trade with someone who we don't know
34:28and all this sort of stuff right and the
34:31market system by formalizing those
34:33systems of gift exchange made things
34:35work a lot better and so i think that's
34:37the problem with trying to solve these
34:40by these sort of voluntary cooperation
34:42the reason why states have been so
34:44important to the development of
34:45civilizations is it just gets really
34:47complicated trying to keep track of all
34:48that stuff without some centralized
34:50course of authority to to make all the
34:52stuff work yeah
34:54austin i remember reading like following
34:56september 11th
34:58afghanistan refused to give up the sun
35:00but one due to the fact that there's
35:01like a cultural one there yeah it says
35:03like you'll never give up a guest
35:05basically yeah
35:06yeah and that seems like it's very
35:08similar to kind of this like strong
35:10cultural norms yeah i mean of course
35:12laws
35:13can also get outdated right but um i
35:17think formal systems generally have
35:20proved themselves to allow interactions
35:22over broader ranges in more flexible
35:24ways than do cultural institutions
35:27though cultural institutions have been
35:29crucial
35:30in areas where it would be very hard for
35:32the law to observe things
35:35so they they are important as well but i
35:38think that the more that we're able to
35:40take something out of the realm of the
35:41cultural and into the realm of the
35:43formal
35:44the the
35:45broader is the possibilities for social
35:48development so i'm just wondering
35:49because like obviously that turned out
35:51badly for afghanistan yeah
35:53if there's any kind of anybody's ever
35:55estimated
35:59uh
36:01yeah i mean i think you have to figure
36:02out exactly which cultural norms you
36:04wanna but yeah i mean there's been some
36:06work on stuff like that like avner grave
36:08uh has worked on that and in fact we're
36:10having him come give a talk in february
36:13so
36:13um yeah
36:15i'm not sure if you mentioned it later
36:17but how exactly would you define foreign
36:22that's a very good question that's an
36:23excellent question um
36:27so the coercion is an extremely
36:29problematic term the reason why it's a
36:31very problematic term is that it's
36:34always relative to some status quo that
36:37we assume yeah uh all what coercion
36:40means is we can force people to
36:42participate
36:43but
36:45we could
36:46the only thing
36:47that was making them not participate is
36:49that we were protecting their property
36:50rights to begin with but of course it's
36:53not necessarily coercion to protect
36:55to to take away their property rights
36:57because
36:59it could be coercion just for them to
37:00stop us from taking away their property
37:02right so it's all defined relative to
37:04some baseline so i think really um if
37:08you look at it from a
37:10even higher level of abstraction there's
37:12like an implicit assumption being made
37:15that we're sort of in like some property
37:17rights regime where people make sort of
37:20like linear pricing transactions with
37:22one another and it's coercion relative
37:25to that baseline but i think that like
37:28i mean one thing that we're going to
37:29sort of come to is that that like linear
37:32pricing with property rights baseline is
37:34sort of like not the right way to be
37:36thinking about these situations
37:38so in that sense we need coercion we
37:40really need some fundamental deviation
37:43from linear pricing in some sense
37:46so i mean that's getting to a very very
37:48high level of abstraction but
37:51does that make sense
37:52yeah
37:54okay
37:55so um
37:56if
37:57uh
37:59public goods and bads are local enough
38:02and it's easy for people to move
38:05then people can just move to the place
38:06where a lot of public goods are provided
38:09well and this is called tibu competition
38:14where local areas compete to attract
38:17uh people
38:18and i think this is quite important for
38:20many local public services such as
38:22street cleaning
38:23schools police fire etc but it depends
38:27crucially on people being able to move
38:29uh to the best area cheaply and in
38:32particular on the value of the good to
38:34them being independent of their location
38:36so you have to be willing to like move
38:38from where you've grown up where all
38:39your friends are to wherever happens to
38:41provide the best public goods which most
38:43people aren't willing to do
38:45and effectively this makes things
38:47both excludable and rival because in any
38:50particular area those things will tend
38:52to get congested and you can keep
38:53someone from moving to the area so what
38:56this says is that you know some things
38:57that might appear to be public goods
38:59maybe aren't that public goods at least
39:01if people aren't too committed to the
39:03particular location where they're living
39:05and that what makes something really
39:07subject to being public goods is someone
39:08who's sort of stuck there or they've got
39:10some value of being there in particular
39:12right
39:13um so i don't think that this sort of
39:15competition is a real solution to the
39:17public goods problem but it does put
39:18some bounds on how bad it can get before
39:21people will up and leave
39:24okay now the most famous solution to the
39:27public goods problem private solution
39:30came from two swedes uh
39:32uh eric lindahl and knute vixel
39:35and they uh imagined that we were trying
39:37to build some big project imagine it's
39:40just a binary decision yes or no are we
39:42going to build this project
39:46and i'm going to focus mostly on this
39:48case for public goods from now on and
39:50imagine each citizen gets some value vi
39:53from the project being built and the
39:56cost of building it is capital c
39:59so the solution they proposed was to ask
40:01each person
40:03to pay a share s i
40:05uh which is equal to their share of the
40:08total value
40:10and
40:12if you do that
40:13then
40:15note that
40:17if
40:18the total value of the project is
40:20greater than its cost that is it's
40:22efficient
40:23then everyone will find the value
40:27their own value to be greater than their
40:29share of the cost
40:31so they proposed a mechanism
40:33in which everyone if everyone agrees and
40:36the project gets built oh yeah mike um
40:39sorry what does capital b
40:41oh sorry capital v is the sum of all the
40:43values okay
40:45um if everyone agrees then the project
40:48is built
40:49and you pay your share of the cost if
40:51anyone disagrees the project is not
40:53built and nobody pays anything
40:55and this
40:57should succeed exactly when the thing is
40:59efficient right because everyone is
41:01pivotal to the decision
41:03and therefore they're all willing to
41:04make the contribution because they know
41:06that it won't happen
41:07otherwise this is called lindal pricing
41:09or equilibrium or vexillion unanimity
41:13um and in fact this was the mechanism
41:16that was originally the idea that turned
41:19into groupon
41:21so they were trying to fund public goods
41:23this way where ever where there would be
41:25some provision point you would put money
41:27in and only once it got to that point
41:29would your money actually be spent right
41:31and otherwise it would just keep
41:33accumulating right
41:34now
41:35um
41:37in practice uh i don't think that this
41:40would actually work
41:41and uh yeah tanner how is this group
41:45okay so this was this turned into
41:47groupon so initially groupon's idea was
41:50this
41:51uh for like some public goods but then
41:53they said oh well one public good could
41:55be negotiating a coupon with the things
41:58but then of course the like actual group
42:00aspect became irrelevant because once it
42:02was just coupons like the companies just
42:04didn't care whether they were giving it
42:05to a large group or not right
42:08so
42:08christian do you have a question um this
42:10might be a dangerous question yeah but
42:12has anyone tried to fix this like
42:14problem with like a vcg-esque mechanism
42:18well we're gonna get we're gonna get
42:19into that
42:20we'll get into that very soon
42:24um
42:27uh okay
42:28so
42:29um
42:30this is probably not practical for a few
42:34uh reasons christian wha wha why why do
42:36you think this look up
42:37why do you think that this wouldn't work
42:39in practice
42:41um well one thing like you could have to
42:43hold out like you said like the one guy
42:46that doesn't want to pay a share second
42:48one
42:48but but if everyone if we if we knew
42:51exactly what share of the value everyone
42:53had then that wouldn't be a problem
42:55right because people would do it if and
42:57only if it was efficient
42:59you're assuming that you know the value
43:00of the share exactly
43:03uh that that's exactly the biggest
43:05problem and furthermore this isn't very
43:07credible right imagine one person
43:08doesn't do it everybody else does
43:10wouldn't everyone just then like chip in
43:12and say okay i'll pay a little bit extra
43:14just to pay that guy's share right
43:16uh
43:17more more importantly with the issue
43:18christian raises that you don't know
43:20everyone's value exactly right
43:23imagine you only know people's shares
43:25approximately
43:26so that there's always some chance that
43:30even if it's efficient their share of
43:33the cost is greater than their value
43:36right yeah
43:37if v is greater c than what happens like
43:40for excess
43:42um
43:43like money raised well every because
43:46everyone will get their share of that
43:48excess right because they only pay their
43:49share of the cost not their share of the
43:51value
43:54um
43:56yeah dave
43:57wait so if their share is much is
44:00significant you're basically saying that
44:01there's a probability that the um
44:04their valuation is lower than their
44:06share
44:08yeah there's some probability just
44:10because we don't know perfectly what
44:11everyone's share of the total value is
44:14um
44:16and i think that this is the reason why
44:18it's so hard to imagine
44:20using this is that there's always going
44:22to be some chance that you've missed
44:24estimated somebody and that that person
44:27says no genuinely
44:28right
44:29um
44:30and we can compute what a problem this
44:33would cause by crunching some numbers so
44:35imagine that we have
44:37capital f being the cumulative
44:39distribution function of these values
44:40imagine everyone has equal shares so
44:43everyone we think is pretty much one
44:44over n but there's some uncertainty
44:47about that so then the probability that
44:49any individual rejects will be the cdf
44:52of c over n
44:53and the probability that any individual
44:55consents will be 1 minus f of c over n
44:58so the probability that everyone
45:00consents will be 1 minus f of c over n
45:03to the nth power
45:05now imagine that we have such an
45:07accurate estimate and this seems pretty
45:09unrealistic but we have such an accurate
45:10estimate of people's valuations that we
45:12can make sure there's only one chance in
45:14a thousand that they'll reject
45:16right
45:18so
45:19um even then once we have a thousand
45:21people there's only a 35 percent chance
45:23that the thing succeeds
45:25and with a small town
45:27there will be less than one in 20 000
45:29chance that it
45:30succeeds um
45:33so
45:33uh this shows what happens as the number
45:36of citizens grows large just the logic
45:39of exponentials means that it's very
45:41unlikely that this thing is going to
45:42succeed
45:45okay
45:46and that's not simply true
45:49of this mechanism where we make
45:52people take it or leave it offer like
45:53you have to have
45:55uh this fraction of the cost and you
45:58know you can accept or reject
46:00um if you start saying okay well maybe
46:03some people
46:04don't want to do exactly this much maybe
46:07we can allow them to say well look i'm
46:09willing to pay a little bit less but if
46:11you start doing that then people have an
46:13incentive to free ride right because
46:14everyone's gonna try to
46:17be the one who uh is the holdout right
46:21um so either way you go you're sort of
46:23damned right
46:25and this is a fundamental theorem a lot
46:27like corneau's theorems it's called the
46:29male of poster weight theorem it says if
46:31everyone must consent to the funding of
46:32a project
46:34and no external funds are available the
46:36probability of the project being
46:38undertaken goes to zero as n gets large
46:42uh what that says is that no
46:44private that is voluntary in self
46:46financing system works and again to
46:49florence's question that's voluntary and
46:51self financing relative to the set of
46:53things that we've sort of endowed to
46:55people before we
46:56uh do this project right
46:59so that's we always run in
47:01yeah uh
47:03um
47:04how was that related like
47:07i believe
47:10we're going to talk a lot about voting
47:12in just a second so let me just get
47:13there
47:14okay so um
47:17thus we always run into a collective
47:19action problem public projects
47:22good or bad always require collusion
47:25um and we did it for a binary case but
47:27it's more true more broadly
47:29and coercion can operate really in two
47:32ways that are quite closely related so
47:33well what are those two ways that it can
47:35operate
47:36so we we defined private as voluntary
47:39and self-financing so what are the two
47:41ways we could break that
47:42um
47:48yeah
47:49and what's how about
47:51how what form of coercion could break
47:53self-financing
47:54like government
47:57if you had some money from the outside
47:59right which would have to come from
48:00somewhere presumably from making
48:02somebody else pay for it right
48:04um
48:06yeah so you can make the members of the
48:09yeah what do you mean by making members
48:11of the community
48:13well you can force them to pay right you
48:16can use the coercive power of the
48:17government
48:21i mean or implicitly taxing them like
48:23put forcing them off their land
48:26right
48:27uh or you can raise some external funds
48:30uh
48:32but of course unless that's voluntary it
48:34has to come from someone so in some
48:36sense they're both coercive strategies
48:39right
48:41so in eminent domain um
48:44uh
48:47um
48:50in eminent domain how would this work in
48:52practice evan
48:54domain well
48:56you have the government that wants to do
48:58some kind of project and so it goes it
49:00takes evaluation of whatever it is they
49:02want to take using the land yeah and
49:04they go and make a price
49:05to or they offer the price of whoever it
49:07is they can either choose to accept it
49:09or a lot of times people will actually
49:11go to court to try and extract a higher
49:12price
49:14yeah so the government can but the
49:16government has the right to undermine
49:17the domain to force people off their
49:18land right yes but it also has to be
49:21considered like by law a fair price or
49:23else it's cruel unusual punishment in
49:24the united states well it's not cruel
49:26and unusual punishment but it's not just
49:28it's not just compensation well i think
49:30that i think there is a law that has to
49:33be considered a fair price yeah yeah
49:35it's called just compensation yeah so
49:37yeah that's the that's the
49:38constitutional provision if it's if it's
49:40ruled a fair price you have to accept it
49:42yeah exactly so that's one thing is that
49:46they can
49:47tax some of the beneficiaries of the
49:49public uh
49:51in the eminent domain situation
49:53and then they can use that to subsidize
49:55this fair price
49:57or they can just force people off of
49:59their land and the supreme court has
50:02really expanded this so that you can do
50:04this in a very wide range of cases
50:07similarly in debt and corporate
50:10restructuring there's usually some sort
50:12of voting where you can force people who
50:14don't want to sell their things to to
50:16sell them
50:19okay
50:20so
50:21um
50:25what what this has taught us is that we
50:27need to use some form of coercion
50:30in order to address
50:32these problems
50:35but of course when we use coercion you
50:37know we don't any longer have the
50:39market's
50:40magic decision making mechanism guiding
50:44the decision that we make because
50:45there's going to have to be some central
50:47coercive authority that makes the
50:49decision
50:51on
50:52for example when is it worth uh building
50:55the project or kicking people off their
50:56land etc
50:58um and this is very similar problem to
51:01the externalities discussion we had
51:03before
51:04and what ways did we talk about then uh
51:08making a decision
51:09about uh whether to impose
51:12um
51:13a particular uh externalities tax uh
51:16evan what were some of the
51:17uh procedures that we talked about so we
51:19have things like contingent evaluation
51:21yeah
51:22gross mechanism yeah
51:24um
51:29yeah that's good expertise using having
51:32experts assess
51:34yeah
51:35so
51:36uh exactly
51:38and uh
51:40these are used with sort of declining
51:41frequency often it's accessed by experts
51:44sometimes they use these crazy surveys
51:46and then the mechanisms that christian
51:48was referring to are almost never used
51:50right
51:51now um
51:53the
51:54i think the key difference
51:56uh here if there is any is that we are
51:58focusing on a case where there's like
51:59some binary decision to be made whereas
52:02there we were more focusing on like how
52:04high the tax should be and we think it's
52:06really important to get input from
52:07individuals so we don't think the
52:10expertise is going to be sufficient in
52:11most cases and there is a very large
52:14number of individuals involved now in
52:16these settings in practice we often use
52:17some form of voting to make the decision
52:20in particular everyone gets some weight
52:22and the thing is approved if
52:24people
52:25uh people's support is above some
52:27threshold
52:28okay
52:29so um
52:31let's think about how voting is as a way
52:33to make decisions on these issues so
52:35imagine that everybody
52:37has even shares again and has those iid
52:39values we talked about now we want the
52:41thing to pass if the sum of all the
52:43values is greater than the cost
52:45which is to say if the average value is
52:47greater than the
52:48per person cost right
52:50but when is it going to pass under
52:52voting can somebody
52:54tell me
52:55uh yes you are a majority of people have
52:57their value over the average value yeah
53:00so
53:01can you express that in a
53:03like some statistic being greater than
53:05the uh per person cost so like n over
53:08two times
53:12two
53:13okay so if the average if the
53:16majority has their values be above that
53:18what will that say about how the some
53:21statistic compares to the per
53:24person
53:25uh so like
53:26would that include the um cdf
53:29yeah it's related to that
53:31i guess
53:32um
53:34yeah i'm sorry
53:36anybody else uh
53:38uh
53:39yeah adam
53:40um
53:42exactly
53:44exactly so
53:45if
53:46the majority have vi greater than c over
53:48n this is the same as the median person
53:51having
53:52something that's above that right
53:55so this is going to be efficient if and
53:57only if the median person has the same
53:59preferences as the mean person
54:01and the sufficient condition for this is
54:03that there's a symmetric distribution of
54:04valuations but even in a fi when there's
54:07a large population in a finite
54:09population that's not even enough
54:11so
54:12uh ted bergstrom in 1979 used this to
54:15argue for an interesting tax principle
54:17he said that it's probably not likely
54:19that there that the median person
54:22uh is has the same value as the mean
54:24person because rich people tend to value
54:27public goods more you know they value
54:29having nice streets and dollar terms
54:31more
54:32and the we know that the distribution of
54:34income is not symmetric it's skewed to
54:36the right but imagine that we have
54:38people pay taxes that are proportional
54:40to their income then it might be that
54:42their valuation as a fraction of their
54:44income might be closer to being
54:46symmetric and so democracy might work
54:48better if things are funded up out of
54:50taxation that's proportional to income
54:52rather than out of a flat
54:54tax that everybody pays which i think is
54:56an interesting argument yeah more um why
54:59wouldn't it be efficient if the median
55:01were greater than the mean
55:03uh it would
55:05um it wouldn't be efficient because
55:07imagine that the mean were less than c
55:09over n
55:11and then the median were greater than c
55:13over n then you would build it even
55:14though it's not efficient
55:17so just like to generally be efficient
55:18it requires that
55:21um okay
55:24so um
55:26a sort of more broad principle of
55:28taxation is that if you can tax people
55:30in a way that sort of organizes it so
55:33that we know
55:34what part of the distribution
55:37corresponds to
55:39the mean conditional on the taxation
55:42then things are going to work better
55:43because you could always set the voting
55:44threshold so that it's exactly equal to
55:48whatever part corresponds to the mean so
55:50for example if you know the distribution
55:52is right skewed you could have a
55:55sub-majority rule
55:57for approving the thing so that you
55:59don't need a
56:00full majority if it's left skewed you
56:02could have a super majority rule uh in
56:04order to
56:06um
56:07in order to
56:08get something that's more efficient but
56:09again you need to know what quantile of
56:11the distribution corresponds to the mean
56:13in order to do that
56:16yeah
56:16everybody relates to filibusters yeah
56:20so would that mean that
56:22filibusters be like more efficient if
56:24you have a more extreme wing of one
56:26party
56:27yeah so if it filibusters
56:30are a more efficient mechanism
56:33then uh
56:36well
56:37and sometimes submajority rule right so
56:40you you want to
56:41have on positive votes
56:44uh filibuster if the people who feel
56:47intensely are opposed yeah
56:50uh and you want to have sub-majority
56:52rule if they're in favor
56:54okay you see what i mean
56:56so like like here's a here's a simple
56:58example of it like gay marriage right
57:01so gay marriage is an issue where like
57:03it's very clear there's a very small
57:05part of the population that's
57:06tremendously affected by this issue and
57:08that's in favor of gay marriage and
57:10another part of the population that
57:11doesn't care as much but might be in
57:13favor of or against so it seems very
57:15clear to me on like gay marriage that we
57:17should have not majority rule but like
57:19sub-majority rule to pass it or super
57:21majority rule to stop it you see what i
57:24mean because only if the rest of the
57:26population is overwhelmingly against it
57:28should we overrule the preferences of
57:29that minority who we know feels very
57:31strongly about it
57:34um
57:35okay
57:36so
57:39in addition
57:40to this sort of general problem that
57:42there can be the tyranny of the majority
57:44uh
57:46under majority
57:47rule
57:48uh there can also
57:50this can also create particular
57:52perversities when people have the
57:53ability to manipulate
57:56how much value people get out of the
57:58issue
57:59so let me
58:02give you
58:03an example
58:04imagine that
58:07we are the government or maybe a private
58:10person who's in league with the
58:12government and we want to buy out some
58:14land
58:15and imagine that we're going to make an
58:17offer for this land
58:19and then we're going to have a majority
58:20vote among the people who whose land is
58:22being taken about whether the offer they
58:24want to accept the offer or not and then
58:25everyone's going to be forced to go
58:26along with that
58:27if you were the developer adam what
58:29would what strategy would you follow in
58:31order to get the land for the lowest
58:32possible price well you
58:34if it's like uh like a 51 yeah authority
58:38rule then you
58:39just
58:40offered just enough to get to that like
58:43that 51 of people would support it and
58:46then like
58:47in particular you would target all of
58:49your offer at those like you'd identify
58:5151 of like poor people who'd be willing
58:53to sell for
58:54a little bit right you would give them
58:56all just enough to get them to sell and
58:58then you'd offer nothing to everybody
59:00else
59:01and so then you would have stolen
59:03at least 49 of the land probably more
59:06right
59:10yeah so that can lead to inefficient
59:13expropriation and under investment in
59:16fact
59:17like in general and this is the example
59:19of christians giving uh
59:22like
59:23democratic governments are known to like
59:26be prone to populism which is basically
59:30when you find something that like
59:31benefits
59:32like a majority of people in the
59:34relevant place
59:35and then you screw everybody else right
59:38even though they're benefiting less than
59:40everybody else is
59:41losing right
59:42um
59:43and
59:45there are tons of other examples of this
59:47tyranny of the majority problem
59:49so that are more severe than the iowa
59:51caucuses so one of them is you can get
59:55just 51 of a company
59:57and then you could like just sell the
59:59whole instruct the company to be sold to
1:00:02a shell company that you control
1:00:04right we talked about that in the theory
1:00:06of the firm lecture
1:00:08populist governments in latin america
1:00:11do all sorts of things that maybe
1:00:12benefit the majority but screw over the
1:00:14rest of the population and ruin the
1:00:15economy
1:00:17majoritarian governments in the middle
1:00:19east have recently as people have
1:00:21probably noticed been doing things that
1:00:24like really hurt minorities in those
1:00:26areas so like the
1:00:28um muslim brotherhood government in
1:00:30egypt
1:00:31was you know doing all sorts of things
1:00:32to oppress religious minorities which
1:00:35led to all sorts of instability
1:00:37uh a simple example
1:00:40is that in doodle i don't know if
1:00:41anyone's ever used doodle for scheduling
1:00:44but like it's very easy for people to
1:00:46like you know only check the one date
1:00:48that they most want to do and then maybe
1:00:50a bare majority does that and then
1:00:51there's all sorts of people who can't
1:00:53make that date at all who don't get
1:00:54included right gay marriage referenda
1:00:57etc right
1:00:58so to fix many of these problems in
1:01:00practice we don't actually use majority
1:01:03rule right
1:01:04what we use is some combination majority
1:01:07rule with all sorts of informal
1:01:08institutions like log rolling
1:01:10like i scratch your back you scratch
1:01:12mine uh you know lobbying the supreme
1:01:16court tries to protect minorities
1:01:18there's like checks and balances that
1:01:20create super majority rules etc so we
1:01:22have all these like we've got democracy
1:01:24and we're like oh that's a great idea
1:01:25well actually maybe not such a great
1:01:27idea let's put all these bandages around
1:01:29it in various ways right
1:01:31um and the problem with all of these
1:01:34fixes is even though maybe they're
1:01:35better than just simple majority rule
1:01:37because it would be overwhelmed by the
1:01:38tyranny of the majority they lead to all
1:01:40sorts of inefficiencies in corruption
1:01:42and paralysis as we've noticed uh in
1:01:45congress recently yeah philip do you
1:01:47have any evidence that like log ruling
1:01:49and lobbying actually fix more problems
1:01:51than they create no
1:01:54but i think they do probably
1:01:56uh because i think that if you
1:01:59like if you look at like the movie
1:02:00lincoln
1:02:01you know the 13th amendment i don't
1:02:03think would have passed without that
1:02:04sort of corruption and i think that for
1:02:06every example that you have of like some
1:02:08bad corruption you see actually i do
1:02:10have one sort of example they got rid of
1:02:12earmarks in congress a few years ago
1:02:15and i think congress has turned into
1:02:16like much worse of a disaster since they
1:02:18got rid of them
1:02:19so i think that's like one example
1:02:22uh yeah dick does this also tie back to
1:02:24our earlier discussion about when you
1:02:26have um
1:02:28because uh entities that are lobbying
1:02:30tend to have
1:02:32more income i suppose you could say um
1:02:35they get higher valuation out of
1:02:37decisions that they're more impacted by
1:02:39decisions yeah
1:02:41but on the other hand like that that
1:02:43leads them to like get all the influence
1:02:45like if you just turn it over to like a
1:02:46lobbying system where it just can be
1:02:48bought out well then that's not very
1:02:50efficient either but it's just then like
1:02:52the person who values it the most wins
1:02:55it so
1:02:56that's not a very good system either so
1:02:58kind of like a filler
1:02:59yeah exactly yeah right um how is it
1:03:02related to like
1:03:04the most familiar
1:03:07like uh there's a probability that it's
1:03:10higher observable
1:03:12no um
1:03:14voting is pretty invariant to that
1:03:17uh because the thing is male 1 really
1:03:19comes out of the absence of coercion
1:03:22not out of the inefficiencies associated
1:03:24with voting so voting is one potential
1:03:26solution to that but it has its own
1:03:27inefficiencies
1:03:30good question now
1:03:31yeah jake um with loading isn't there
1:03:34also the problem that there's a
1:03:35transaction cost to the actual accuracy
1:03:38itself and so that might filter out
1:03:40people who don't care that much about
1:03:41the details but might feel strong
1:03:45so that can actually improve things but
1:03:47like there's a whole culture around
1:03:49voting and democracy which i think
1:03:51encourages people not to take those
1:03:53votes like those costs into account and
1:03:55to like vote just because it's their
1:03:58obligation to vote because like
1:04:01the mechanism is not really vote if and
1:04:03only if it's like you know above the
1:04:05transactions cost to you the real
1:04:07mechanism is everyone should vote and
1:04:09even though you're not literally
1:04:10required to in transaction costs do
1:04:12filter out some people and maybe that
1:04:13can make it a little bit better actually
1:04:15um
1:04:16that's not sort of what
1:04:18people
1:04:19uh want you know and i guess because
1:04:21it's better because that means that the
1:04:24media person is closer
1:04:26to the mean yeah probably at least in
1:04:28some cases yeah christian isn't there
1:04:30actually like a paradox or some old dead
1:04:32guy like the voting paradox that like in
1:04:34a large enough society the actual
1:04:36benefit of your vote is basically zero
1:04:39yeah and so since voting does have some
1:04:41transactional costs i go no one should
1:04:43be voting yeah that's true there's a
1:04:45whole lot of literature about that
1:04:47question who's who's what's that
1:04:50well i mean it's a very old old idea i
1:04:52think condorcet first pointed that out
1:04:54but um
1:04:55but there's a lot of recent literature
1:04:57there's a book by mark bleis that's
1:04:59pretty good on this
1:05:01if you're interested in this

About this transcript

This page contains the full transcript of ECON20110 Fall 2013 - Lecture 16: Public Goods by Glen Weyl, generated from the public captions YouTube serves with the video. The transcript has 10,091 words across 1,814 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.