YT YouTube2Text

YouTube2Text → Transcripts → ECON20110 Fall 2013 - Lecture 8: Assessing Externalities

ECON20110 Fall 2013 - Lecture 8: Assessing Externalities — Transcript

by Glen Weyl · 11,555 words · 1,696 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:02okay so we left off last class by
0:05concluding that information was the key
0:08to uh determining externalities and
0:12externality policy
0:15um assessing externalities requires
0:17collecting the right information and as
0:20we discussed externalities tend to be
0:21cases when the information is either
0:23diffused among many members of the
0:25public or held by an expert who's not
0:28one of the parties directly involved
0:30um and that's what we're going to talk
0:32about today so some of the relevant
0:34information for assessing externalities
0:37is scientific Technical and usually best
0:40accessed by experts and some is
0:42subjective and held by uh dispersed uh
0:47individuals who are directly involved
0:50and this is problematic for policy often
0:52because you have to give these
0:53individuals an incentive to reveal that
0:55information and today we're going to
0:57talk about how to combine these two
0:58types of information and the trade-offs
1:00between ways of uh getting them so we'll
1:03start by talking about benefit cost
1:05analysis as an encompassing you just
1:07dropped
1:08a uh as an encompassing framework for
1:11this type of analysis we actually just
1:12held a conference on benefit cost
1:15analysis at the law school this past
1:17weekend uh on um Financial
1:22regulation um and then we'll talk about
1:26how scientific and Technical studies
1:28combined with economics play into these
1:30types of benefit cost
1:32analyses uh focusing on one particular
1:35ingredient that comes from economics
1:36which is the uh statistical value of
1:39human life and health and we'll go
1:42through an example of a benefit cost
1:44analysis probably the most ambitious and
1:46famous of these the stern review of the
1:48economics of climate change um we'll
1:51then talk about useful types of private
1:53and subjective
1:55information uh contingent valuation as a
1:57sort of naive first pass to addressing
2:00these issues and then we'll talk about
2:02the vicory Clark Groves mechanism as a
2:04way of dealing with the incentive
2:06problems that arise in these surveys
2:08okay so benefit cost analysis is a
2:11common Paradigm for um policy evaluation
2:16and it's supported by economists perhaps
2:18because it lines up so cleanly with the
2:20way that economists think about things
2:22it emphasizes the importance of
2:24tradeoffs pros and cons uh and is
2:27applied to a wide range of decisions
2:29like ins uting regulations building
2:31public projects and for our purposes
2:33determining the right level of caps for
2:35cap and trade uh of
2:38externalities uh for a paguan tax and of
2:41Damages for uh paguan taxes that are
2:44assessed through the legal
2:47system so the benefit I think of benefit
2:50cost analysis is that it provides a
2:52rigorous language to try to make commens
2:54orate different
2:57dimensions of uh of trade-offs in terms
3:01of dollars basically and the different
3:03dimensions of these trade-offs include
3:06direct harms of econom on uh economic
3:09resources how much uh resources are
3:12destroyed and how much the market value
3:14those second can be limits on the way in
3:16which you can use resources and again
3:19you have to use market prices to assess
3:21how valuable different uses of those
3:23resources are and therefore how harmful
3:25it is to limit the use away from its
3:26most valuable
3:28use um facts on human life and health
3:31which we'll talk more about soon
3:33intangible harms on subjectively valued
3:37resources which can be determined either
3:39uh inherently from an ethical Theory or
3:42be derived from how much individuals
3:44value those things in terms of their
3:46willingness to pay to preserve them okay
3:50so benefit cost analysis is often used
3:53to make big public decisions such
3:56as instituting a new regulation on an
3:59indust indry building a public project
4:02in some or building a public project in
4:04some location but the same basic
4:06methodology applies to determining
4:09externalities such as determining the
4:10right level of beian tax you need to add
4:13up all the marginal external cost and
4:16benefits um and it requires considering
4:20the same list of factors that we just
4:22discussed uh another common application
4:25is to assess damages in a court case so
4:28when uh there's there's a big lawsuit
4:30brought that involves uh economic
4:32damages often economists are hired to do
4:35these analyses in fact that's one very
4:37common form of Consulting by
4:39economists and as we discussed in the
4:42last lecture these damages effectively
4:44act as a form of pigouvian
4:46tax uh benefit cost analysis can Al also
4:49often be used to determine the optimal
4:52level of a cap for cap and trade and in
4:55some ways this is the closest to Classic
4:57benefit cost analysis you determine the
4:59best level for such a cap and often the
5:02choice is discretized so it's almost
5:04like making a choice about a
5:07project and these analyses are
5:09undertaken by
5:11economists uh or uh either directly or
5:15that are commissioned by courts or hired
5:17by bureaucracies in the
5:19government so a crucial element in all
5:22of these benefit cost analyses is
5:24technical um what are the physical
5:28effects that actually follow from some
5:30particular action and this typically
5:33requires detailed knowledge of an
5:35interaction with among other disciplines
5:37physicists so for example if you're
5:39trying to think about the benefits and
5:41costs of nuclear power you need to
5:43determine you know what are the chances
5:45of a nuclear reactor having some
5:48disaster uh what are the chances of CERN
5:50I don't know if you guys know what CERN
5:51is it's this particle accelerator in
5:54Switzerland and a long time ago I guess
5:57by your standards but a short time ago
5:59by by my standards when you guys were
6:00probably in middle school there was all
6:02sorts of debate about whether CERN would
6:04blow up the world uh some people said
6:08that CERN would uh tear the universe a
6:11new uh and uh but that didn't
6:15happen so there was but you know there
6:16was some very very small probability of
6:18that and you had to assess how large
6:20that probability was right
6:23um uh chemistry can also be useful so
6:25for example if you put SO2 uh sulfur
6:29oxide into the atmosphere it can cause
6:31acid rain and chemistry can help us
6:33determine how much acid rain and how
6:35severe it will cause or you know it can
6:37help us determine if you pollute a lake
6:38what will be the effect on the chemistry
6:40of that Lake epidemiology can help us
6:42determine if we vaccinate certain people
6:45how uh much will that reduce disease in
6:47the population more generally ecology
6:50can help us determine how fish stocks
6:51are affected by pollution or by
6:53overfishing or what ecosystem services
6:56will be eliminated by
6:58deforestation however of course all this
7:00stuff is useless unless you have the
7:02right economic backend on it right so we
7:04might be able to tell from physics
7:06what's the chance of a nuclear meltdown
7:08but it won't tell us what's the value of
7:10the land that will be lost and the lives
7:12that will be harmed by uh Fukushima
7:15right how valuable are the crops uh you
7:18know we can learn about acid rain and
7:19maybe even how many crops they'll
7:21destroy but to actually put a dollar
7:22value on that we need something coming
7:24from the market prices right or Futures
7:26prices or something like that um um if
7:30we want to know uh how we should treat
7:33the small risk of a catastrophic outcome
7:35like the CERN disaster you need to use
7:38uh the economics of dealing with
7:40uncertain outcomes like concave
7:43utilities for risk from expected utility
7:45Theory right um if we want to try to
7:49Value um the you know production we get
7:53from more pollution today against the
7:55destruction of fish in the future we
7:57need some uh
8:01um
8:03oops
8:07um we need some uh notion of the
8:10discount rate uh to trade off harms
8:13today against values tomorrow um Adam if
8:17we were trying to Value ecosystems
8:20services like things that came from the
8:22rainforest and that we might lose by
8:24destroying that what what are economic
8:26factors that might enter into figuring
8:28that out we probably look at like the
8:32past on how like it has effective
8:34medicine um and then
8:37um and then I guess um account for the
8:41price of the price of those drugs or the
8:44consumer surplus that was associated
8:45with those drugs being created
8:48right
8:51um another question that we might ask is
8:55you know how much our lives or the
8:59health ruined by diseases in the
9:01epidemiological example worth um and
9:04we're going to turn to that in just a
9:05minute um talking about uh the valuation
9:10of of human life so almost all of these
9:12benefit cost analyses involve some
9:14intimate connection of science with
9:16economic
9:18analysis um and perhaps the most famous
9:22input from economics is the value of
9:23human life the reason it's so famous is
9:26because it's so controversial because
9:27most people want to say well life is
9:30just invaluable there's no way to put a
9:31dollar number on life but of course this
9:34is completely inconsistent with the fact
9:36that people drive around in cars which
9:37endanger their lives maybe with small
9:39probability but not that small
9:41probability in exchange for all sorts of
9:43other benefits they get right so you
9:46know when people make decisions they do
9:47make trade-offs they take their life
9:48into their own hands they make
9:50trade-offs between their life and other
9:52benefits they can get and when the
9:54government makes decisions it has to do
9:55the same thing it just has no choice
9:57right um
10:00and in particular the wealthy have a lot
10:02more money to spend on trying to
10:04preserve their lives and we know that
10:06wealthier people get better Healthcare
10:07they do all sorts of things so that
10:09means that they must in dollar terms put
10:11a greater value on their life than poor
10:13people do and so if the government wants
10:16to make consistent decisions it's going
10:17to have to do the same thing so that
10:20leads to the conclusion that a rich
10:21person's life is worth more than a poor
10:23persons in dollars which may seem unjust
10:26but you just have to think about this is
10:28one of the many things that that money
10:29buys money buys you a better house buys
10:32you a nicer car and it buys you the
10:34government putting a greater weight on
10:36your survival than a poor person's
10:39survival right now of course you could
10:40redistribute money so that that's so
10:43that the poor people can have more of
10:45everything that they want but not
10:47allowing people to make tradeoffs
10:49between the weight put on their life uh
10:52in Social calculations and other things
10:54that they want would just be inefficient
10:56right the poor would not want they would
10:59rather have a little bit less weight put
11:01on preserving their life and a little
11:03bit more house or uh TV or car um and it
11:08would be inefficient not to allow people
11:11to make those trade-offs yeah Mike could
11:13you explain more why um like the fact
11:16that rich people can spend more money on
11:18things like healthcare like quality of
11:20life or whatever means that governments
11:22uh value their lives more when they do
11:25policy well government has to Value
11:27their life more or it would be
11:29inconsistent right those rich people are
11:31willing to pay more in order to have
11:33their life preserved so it would be
11:35inefficient for the government not to
11:37put a greater weight on preserving their
11:39lives because the government can make
11:40everyone better off by putting a greater
11:42weight on preserving those rich people's
11:43lives and then taxing those rich people
11:45for it right because they value okay so
11:50imagine that we had two policies one
11:53that could you know with some small
11:55probability kill a rich person or one
11:58that would the same probability could
12:00kill a poor person clearly it's better
12:03to implement the policy that would kill
12:05the poor person why is that uh because
12:08the you could tax the rich person and
12:10redistribute to the poor person and
12:12everyone would be better off because the
12:14poor person isn't willing to pay as much
12:16to reduce that chance of preserving
12:17their life and the rich person is
12:19willing to pay more right so it it would
12:22just be silly as social policy to say
12:25poor people have to have cars that are
12:29uh ugly liveing houses that are falling
12:32down get education that's bad for their
12:34kids but they get to have exactly the
12:38same amount of weight put on them in
12:41public decisions about health right that
12:44just wouldn't make any
12:47sense uh yeah philli because the fact
12:51that rich people also contribute a lot
12:54of tax dollars with their labor and thus
12:57losing a rich worker could be more
12:59harmful to the government than those a
13:02poort worker been also weigh in no this
13:05yeah so that also should be taken into
13:07account and in fact that's a good point
13:09that's a reason why actually the
13:13responsiveness of um that's actually a
13:16very good point so what we're going to
13:18see is that the
13:19elasticity of life with respect to
13:22income is
13:25approximately uh is is going to be less
13:28than one but one reason that that might
13:30be the case is because the government is
13:31taking some of the money uh and if the
13:34government's taking some of the money
13:35then the government should care more
13:37than the individuals themselves care
13:39about preserving their lives um yeah
13:42Mike we have policies that uh like impr
13:44preventing an like an epidemic or
13:47something uh do more for lower income
13:50people because rich people like have
13:52insurance or like things like that that
13:53would keep them or or like you know
13:55lower income people live in closer
13:57proximity to each other in some cases so
13:58like diseas or more likely to spread how
14:00how did those
14:01calculus uh fact I mean I think that in
14:03general you just have to Value people's
14:06lives as they value them themselves in
14:08dollars okay I mean I think that's the
14:11only efficient
14:12policy uh yeah Austin going off of that
14:15um how is like prospect theory and other
14:18behavioral uh corrections I guess worked
14:20into these prect so in fact the origin
14:24of pro of a lot of Behavioral stuff was
14:26studies of these types of things where
14:28they found that there was consistencies
14:29depending on how you frame the questions
14:32um and you know you I I think you get
14:35pretty sensible numbers out of these
14:38analyses um despite all the problems
14:40we'll discuss with them I think that's
14:42one problem but I think there's many
14:43other problems associated with these
14:44analysises that we'll go through but uh
14:47yeah Julian so is this analysis of does
14:51this still hold even if you don't
14:52redistribute we because seems like most
14:55public policies at least that I'm aware
14:57of don't do the kind of like here's how
15:00much benefits will tax them by that
15:02amount no that's not that's it doesn't
15:05require you to really redistribute the
15:06wealth because the truth is that like we
15:08choose the tax system and the amount of
15:12redistribution that we do to like
15:14roughly trade off the benefits of
15:16redistribution against the costs of
15:19reducing like the amount that people
15:21work right and as long as you're doing
15:24that tradeoff correctly then the value
15:27of a dollar to a rich person
15:29from a society's perspective and the
15:31value of a dollar to a per person should
15:33be equal otherwise you're not doing that
15:34tradeoff correctly right so in other
15:37words like the reason why I mean even
15:39though inherently the value to a dollar
15:40to a poor person is greater the thing is
15:43that giving it to the rich person also
15:45leads people to work harder to try to
15:47become that rich person and therefore
15:49generates more taxes for society that it
15:51can redistribute to the poor and as a
15:53result of that these things should like
15:55balance
15:56out uh yeah J sure the value be
15:59calculated based on willingness to pay
16:01and not how much they can pay because no
16:03it is willingness to pay that that
16:05that's the idea but if you ask a rich
16:07person and a poor person how much
16:09they're willing to pay to save their
16:10life they'll probably say the same
16:12amount but how much well if you ask them
16:14if you ask them but like what
16:16willingness to pay is how much you are
16:19actually willing to pay like out of the
16:22income that you could have for your
16:23lifetime right so that's how much you
16:25can pay right not how much you're will
16:27to pay well I don't think that that's
16:29true because the truth is that like
16:31people that would be true if everyone
16:34would be willing to pay their full
16:35lifetime income to preserve their life
16:37but that's not true like most people in
16:39order to reduce by 1% chance the chance
16:42that they die are only willing to pay a
16:44few $1,000 not all their lifetime wealth
16:47so it would if everyone was like running
16:48up against their lifetime budget
16:49constraint what you're saying might be
16:51true but that's not the
16:53case yeah Tanner so can we think of this
16:56as like payment and forance of product
16:58that the uh which person can afford to
17:00pay more for it their their life is
17:02worth
17:03more well I mean Becker would say that
17:07but I'm not sure I would say that like I
17:08I mean I think this is a Distributive
17:10question not a question of like it's a
17:12question of efficiency to make sure that
17:15you like equate these things in terms of
17:18like given someone's wealth what they
17:20get but I don't think that it's that
17:23they're
17:23being I mean the reason the rich person
17:26is making more money is because they
17:28have
17:29greater uh endowments we could
17:31redistribute those endowments from them
17:34but we talked about all the problems
17:35with doing that so given all those
17:37problems with doing that this is the
17:38best that we can achieve and still be
17:42efficient H yeah Philip if you take into
17:45account a typical discounting rate then
17:48is does a value of like a statistical
17:51life does someone seem lower than the
17:53their lifetime income because what from
17:55what I've been seeing the vsl numbers
17:58cost around they seem to be quite a bit
18:01more than at least a lot of the people
18:03on the lower end of the income
18:05distribution actually do make their
18:12lives I think it's pretty comparable
18:14with it but I agree with you okay I
18:17mean someone who makes $50,000 a year
18:21plus they're going to get some amount in
18:23government benefits in some
18:24sense you know times 20 which is roughly
18:29you know that should
18:33be that should be about a million and
18:36people are willing to pay about five
18:37times that to preserve their lives at
18:39that income so yeah that is valuing it
18:43more
18:45but
18:47um okay so there are three um
18:52categories uh of things that need to be
18:54measured for this purposes one is the
18:57statistical value of a life for someone
18:59at some income level the second is the
19:01value of other types of injuries like
19:05diseases and so forth and then you need
19:07to extrapolate this to other income
19:09levels using some income elasticity this
19:12is usually calculated to be about 6 or8
19:15though as Philip said it this should
19:17really be in some sense in post as a
19:19function of their post tax income rather
19:21than their pre-tax income because you
19:24know the government is taking some of
19:27their income and you want the government
19:29to value that part that they're taking
19:32uh greater as well okay yeah Phil just
19:36to be clear those numbers up there are
19:37for pre-tax income elasticity with
19:41respect to pre-tax I think so but I'm
19:42not completely sure okay I'm not
19:45completely um okay so the usual approach
19:49to measuring uh to measuring this is to
19:55assume that um the value
19:59is on uh the life that is like that you
20:05um die or live and the value that you
20:08place has nothing to do with the way in
20:10which you die right that's a crucial
20:12assumption and it's the basis of all the
20:14numbers we'll see though it's not all
20:15that realistic because you know if you
20:18really don't want to die by fire but
20:21you're happy to die by Ice uh then um
20:25you uh should obviously not have the
20:28same will to pay to avoid those two
20:30types of
20:31death um and the way that these things
20:35work is to look for cases where people
20:37have made trade-offs between these
20:39things like they're deciding whether to
20:41take dangerous work like mining or a
20:43safer job that's like along other
20:45dimensions similar or they're deciding
20:48whether to purchase a safe or unsafe car
20:51that are otherwise different and you
20:53look at the price differences as a
20:54measure of willingness to pay and that
20:57measures the willingness to pay have the
20:58marginal person who's just indifferent
21:00between these two choices you then treat
21:02everyone as if they're exactly the same
21:04as this marginal person now you can
21:06immediately see why this is so
21:08problematic imagine that the cost of a
21:10safety measure on a car like starts
21:13really high and gets low over time you
21:16know this would work if there was like
21:18one price at which suddenly everyone
21:20adopted that um that Innovation but in
21:24reality you know as you improve safety
21:27people are going to gradually move into
21:28into buying this car right and if that's
21:31the case then um then the uh marginal
21:37person is not going to be just uh
21:39representative of the whole population
21:41and we're going to get different numbers
21:42for this as the price of this Innovation
21:46comes down which seems pretty
21:48problematic right um in particular like
21:52if it were really true that everyone had
21:53the same value on their life then you
21:55could only get people to be indifferent
21:57to taking this job at like one
21:59particular price and we know that's
22:01almost certainly not the case that as
22:02the wages move or the prices move uh
22:05different numbers of people do it it's
22:07not like it just suddenly jumps and that
22:10um makes all this stuff somewhat
22:13suspect so however I think that what's
22:16beneficial about these numbers is they
22:17give you at least a rough order of
22:19magnitude sense uh that can be used in
22:21government calculations now you can also
22:24try to get more individually at this by
22:26trying to use surveys to get people like
22:29Joan was suggesting to say how much
22:30they're willing to pay but for the
22:32reasons I was talking about with Joan
22:33and we'll talk about a lot more that's a
22:34pretty problematic
22:36process okay so let's take a sample
22:40version of how one of these statistical
22:43life calculations works so imagine that
22:47um every year three of 2,000 coal miners
22:52working in a particular Coal Mine die
22:54and suppose that factory workers have
22:56comparable skills and the job are
22:58comparably pleasant but that only one in
23:00every 2,000 Factory workers dies each
23:02year and suppose that the coal mining
23:05company pays $4,000 more a year and
23:08assume that this is the only relevant
23:09difference between the jobs well then
23:11the difference in the chance of dying is
23:131 over 1,000 and in exchange for this
23:16you get $4,000 and so the value of your
23:19life to you would then be $4
23:22million okay so the real calculations
23:26are no more than detailed EC metric
23:28versions of these same
23:30calculations and to calculate income
23:32elasticities all you do is you look for
23:34a rich man's version of a coal mine
23:36whatever that is um okay so here's a
23:40bunch of studies that have looked at
23:41this they tend to find that statistical
23:44value of life is somewhere between $1
23:45and $10 million for like someone earning
23:48median US income something like that so
23:51this is a little bit dated so my guess
23:53it's a little bit more in present
23:54dollars it's probably closer to 10 to 15
23:56million
23:59okay so perhaps the most ambitious oh
24:02yeah do you have a question yeah sorry
24:03in the previous slide yeah when you said
24:07uh suppose the co mining isay $4,000
24:104,000 more yeah that's the coal mining
24:13so that's the company's cost for
24:14reducing worker risk
24:17right no no that's how much more they
24:20have to pay in order to get people to
24:21work at the co mine and accept the loss
24:24you
24:26see okay so perhaps the most ambitious
24:30uh benefit cost analysis of all time was
24:33the stern review of the economics of
24:35climate change which was commissioned by
24:37the British government along with other
24:38people and wanted to evaluate the
24:41economic benefits and costs associated
24:43with climate change and it did this in
24:45many stages um and Matas is matus here
24:50no um he
24:52dropped anybody else want to talk about
24:55like what the stern review did in order
24:57to Analyze This
25:04okay so it did basically four things so
25:08the first thing it did was to determine
25:09the largest sources of greenhouse gases
25:12and Tanner what did those turn out to
25:14be uh turn to the carbon emissions from
25:18uh energy production animals methane
25:21animals are a surprisingly large Source
25:24that's right so uh so animal have like a
25:29lot cow farts basically account for a
25:31large fraction of all global warming but
25:34also energy from carbon from energy
25:37deforestation reducing the amount of
25:39carbon absorbed
25:41Etc uh second they had to determine the
25:44effects that these had on the climate um
25:48and uh Wesley how does that work what
25:50are the effects of uh greenhouse gases
25:53on the climate well the larger emissions
25:56the more the more the temperature is
25:59going to rise and there's a runaway
26:00effect so if we don't reduce the
26:02emissions enough then it'll still keep
26:04Rising even though we reduce the
26:05emissions yeah so there's a critical
26:07level it's like 80% less than what it is
26:10now at which we could stabilize it but
26:13we have to reduce it a lot and all those
26:15things are based on these climate models
26:16that are filled with uncertainty right
26:19and um so what you really get is a
26:21distribution of potential outcomes
26:22conditional and actions you take rather
26:24than a final outcome so what Wesley is
26:26saying is a little bit like the example
26:28we went through where you know the firms
26:30enter at some particular price but we
26:32don't know what that price is and so it
26:34really looks smooth so even though it
26:35he's right that like you know the
26:37climate there's probably some point
26:39Beyond which we go when things really go
26:41to hell we really don't have a very
26:43clear sense of what that point is and so
26:45in reality it's much more smooth than it
26:48sounds um then you have to try to
26:51determine the economic effects of
26:56um of climate change
26:59Andi what were the primary economic
27:02impacts of climate change that they
27:03determined um one was
27:07that sorry one was that resources are
27:10obviously being affected so uh land
27:12costs say land um if land cost are
27:15proven to be very different that people
27:17are obviously going to suffer different
27:19U another one was U the force were being
27:23armed uhhuh uh so what is uh what is the
27:26word yeah that that's right I mean I
27:29think the biggest one is probably
27:31Farmland like is um harmed and that
27:35makes agricultural production lower so
27:38that's an important thing um then they
27:41had to try to determine the extent to
27:42which adaptation to climate change oh
27:44yeah go ahead um didn't they also say
27:47that one of the economic effect is sort
27:49of this disproportionate
27:51unfair
27:53uh uh effects of uh climate change being
27:57on the poor which didn't contribute to
27:59it at all so yeah though though though
28:02that should from our perspective like of
28:04this analysis actually mean that it's
28:06counted should be counted
28:08less right because they're poor and they
28:11can't afford to well didn't say that
28:13they would be affected by it most
28:16severely and most quickly yeah and I
28:18think that you know I I I know it sounds
28:20unfair responsible for it but they're
28:24paying the price for it and that being
28:26the externality isn't that part of it
28:28yeah I mean that's you know doesn't
28:30sound like a very good thing but if it
28:32was happening to rich people it would
28:33get more weight if you were doing a
28:35standard benefit cost
28:37analysis um so then yeah Mike um I guess
28:43is there a way of like
28:45incorporating ethics
28:47into like it's it's subjective yeah but
28:50like I feel I feel like especially and
28:53and it shows like in the summary of the
28:55report that they clearly believe that
28:57it's unfair these kinds of things that
28:59are created by developed countries are
29:00disproportionated by being not so
29:02developed on yeah so is there a way of
29:05like accounting for those kinds of
29:06things or like changing the calculus of
29:08what we should yeah I mean look you can
29:09incorporate all sorts of other things in
29:11if you want but they have the
29:13consequence that you're not going to get
29:14something that's Proto efficient like
29:17the most efficient thing to do is for us
29:20to you know wait things in the way that
29:23we're waiting them and then compensate
29:25the poor countries like the poor
29:26countries would be happier for us not to
29:28worry about like killing them but to
29:30give them money because they're going to
29:33die from disease or starvation long
29:36before they die from climate
29:39change I mean like I know that it sounds
29:42brutal and whatever but just think about
29:44the logic of it it's like you're taking
29:46people who have life expectancy of 25
29:49years because we don't give a godamn
29:51about them we just let them all starve
29:53and die from
29:55malaria and then we worry about that in
29:5870 years they might have rising sea
30:01levels like how is that generous and
30:06ethical Yeah Austin didn't Larry Summers
30:09get in trouble for email along these
30:10lines yeah I may get in trouble for
30:12saying all these
30:16things
30:18um uh
30:21so then
30:24um we need to determine the extent to
30:27which people adapting to the climate
30:29change can mitigate its harmful effects
30:31and Florence what are some of the ways
30:33people might adapt to climate change to
30:34reduce its harmful
30:36effects um you could uh change your
30:40different Technologies um and reduce
30:44emissions that way well but that's not
30:45that's reducing but but how could they
30:48adapt to the changeing climate
30:56um well they can move of New York City
30:58to Chicago right because Chicago
31:01probably won't be flooded and New York
31:04probably will uh but the problem I think
31:07we have is that um we don't have
31:10institutions that do a very good job
31:12achieving these in the incentives to
31:14migrate so for example a lot of people
31:16like when there's a flood in like New
31:18Orleans they all get bailed out by the
31:19government rather than people being
31:22forced to buy flood insurance though
31:24increasingly the New York Times was
31:26writing about this they've been forcing
31:28ing people to buy more flood insurance
31:29and a lot of people are being forced to
31:30move away from the areas and that could
31:32be very useful in adaptation so how much
31:35of that occurs is really important and
31:37there's an interesting study by Rick
31:39Hornbeck that looked at the Dust Bowl in
31:41the in Oklahoma um when all the people
31:44were forced to move out and tried to
31:46look at how well people adapted to that
31:47and how that mitigated or did not
31:49mitigate the
31:50effects
31:52um now uh Mitch there are some other
31:56impacts that climate change can uh have
31:59sorry that carbon can have other than
32:02just directly on climate change do you
32:03remember what some of those
32:05were
32:10um uh well uh ecologically speaking uh
32:14it can have
32:16changes uh in animals and environments
32:19it can
32:20change
32:23um pH levels in waters it can it can
32:26create high average temperature per year
32:29Rises um yeah yeah so it can saleny and
32:33acidify the ocean right and it can
32:35affect ecosystems by causing certain
32:37things to be able to grow and others not
32:39so those were also Tred they tried to
32:41take into account now corresponding to
32:44the cap and trade versus the uh
32:48pigouvian tax Paradigm um there were two
32:51different analyses that were part of the
32:52stern report and Mike fosco what um do
32:55you remember what those two analyses
32:57they did were
32:59uh there's a huge uncertainty about
33:02where they can cap this that they that
33:05they have to stabiliz between like 450
33:07and
33:07550 um so I think they were going with
33:10the cap so then you can know because
33:13you're not entirely sure of like the
33:14price that need you to do that yeah so
33:17they did two different analyses one was
33:19they tried to say at the business's
33:21usual scenaria what is the marginal cost
33:23of an additional ton of carbon to try to
33:25determine like an optimal tax or
33:27something like that and then they also
33:29tried to calculate what would be the
33:31optimal amount of
33:33carbon um so the benefit of this
33:35approach is that you don't need to do
33:37anything to estimate the technology that
33:40will be arised in the cost of mitigation
33:42so it's simpler but the drawback is that
33:45there might be highly nonlinear effects
33:47of carbon like um like Wesley was
33:51talking about uh that might lead to some
33:54very catastroph catastrophic effects and
33:56this tax really might not do a job good
33:58job capturing that right their estimate
34:01that they came out with was $85 per ton
34:04now the socially optimal level of carbon
34:07for that calculation um they try to deal
34:10with the nonlinearities and try to
34:12figure out beyond what point it might be
34:13very dangerous the drawbacks is this
34:16requires you to have a very good
34:17understanding of the technologies that
34:18are available for mitigation to weigh
34:20those against the costs and they argued
34:22that we should try to stabilize carbon
34:24at 450 to 550 parts per million though
34:27so we're jetting past that at the moment
34:30so I don't think it's very likely that
34:32we'll achieve this
34:34goal this review is quite old by now I
34:37guess right do you guys remember when
34:39the stern review first came out you you
34:42guys were probably in middle school or
34:44something like that
34:46right probably even before that I would
34:49guess I don't know I'm not sure maybe
34:512006 um
34:54so I think what's most interesting about
34:57Stern is that it illustrates how hard
34:59these types of things are like one um
35:03example of this is that many of the
35:06worst costs from climate change come
35:08from the possibility that there's some
35:10very extreme disaster that climate
35:12change causes and if we have a
35:14technology such as geoengineering like
35:17putting uh chemicals into the atmosphere
35:20or into the ocean that would absorb a
35:22lot of carbon or reflect a lot of
35:23sunlight that can eliminate the worst
35:26possible outcomes this can dramatically
35:28reduce the cost not because those things
35:30aren't harmful and bad but because
35:32they're not as harmful and bad as like
35:34the most dire scenarios which in the end
35:36account for most of the costs and and so
35:39uh it's um yeah so whether or not we
35:44think we would end up developing and
35:46using those Technologies could make a
35:47huge difference for whether it makes
35:49sense to uh invest in reducing uh carbon
35:53today okay now the stern review focused
35:56on sort of hard economic costs um and it
36:00leaves out a whole bunch of other stuff
36:04that is clearly pretty relevant so for
36:06example do people like it being warmer
36:09or cooler with rain or you know less
36:11rain or whatever now this might sound
36:13frivolous relative to like you know
36:15swallowing up whole continents but if
36:17you look at the difference in real
36:18estate prices between here and poo Alto
36:20you'll realize that uh the weather makes
36:23a pretty huge difference to how much
36:24people uh value living in a place right
36:28and um you know there's a question of
36:31how you might quantify this in an
36:32objective fashion by using real estate
36:34prices which maybe they could have taken
36:36into account but they
36:38didn't uh yeah t what if like California
36:41that just becomes a desert like well
36:45that would be a harm so you'd have to
36:46you'd have to take that all into
36:48account
36:50um uh yeah all right um didn't like
36:53conman show that people who live in like
36:58Michigan and like California are like
36:59equally happy yeah so the question is
37:02whether you believe or not believe the
37:06happiness numbers versus the prices you
37:09know I don't know uh there's a lot of
37:11debate over those
37:13things I mean I think even Conan is a
37:15little bit less confident now that
37:18happiness is the right measure for this
37:20stuff
37:21so um another question is you know how
37:24much do people care about the animals
37:26that get killed by climate change right
37:28um and you know to quantify this in an
37:31objective fashion you need to use
37:34something like how does the government
37:35make decisions about killing animals in
37:37other contexts what are they willing to
37:38pay to preserve those animals yeah go
37:41ahead Austin so um how does that how
37:44does the second Point play in with like
37:45quantity in variance it's like John list
37:48demonstrated it doesn't matter how many
37:49animals are being saved it's pretty much
37:5120 bucks it would well we'll talk about
37:54the problems with the surveys in a
37:56moment but
37:58uh but I think that in principle you
38:00could be more quantitative about this uh
38:03I'm talking about how you would you know
38:05figure these out in a purely objective
38:07way using a scientist rather than um you
38:10know another question that would be
38:12difficult is imagine that this is going
38:14to destroy some civilizations right uh
38:17the question is how do you put a value
38:19on cultural preservation uh that's tough
38:23another question issue might be that
38:26individual people out there in the
38:28population might know information that
38:30the government doesn't know about um how
38:33costly it is for them to cope with acid
38:35rain and things like that and you know
38:37as we talked about hyek said it's really
38:39important to get information from the
38:41individuals who actually know about
38:43what's um going on and if we ignore that
38:47in these analysis and only rely on
38:49Experts we might get a much less uh
38:51compelling estimate of the benefits um
38:56and because you know for basically all
38:58of these things it becomes important not
39:01just to know um what market prices tell
39:04us but also to know what are people's
39:07personal subjective valuations for
39:09various things and this is true for a
39:11wide range of things when there is
39:13subjective existence values such as the
39:16value of environment of culture of the
39:18place of still living in the place where
39:20you grew up
39:23right
39:26um uh
39:29Lo when there's local information about
39:31Technologies or impacts when you're
39:34thinking about health or life and not
39:35just trying to measure the value of the
39:37marginal people but of inframarginal
39:39people whenever there's local
39:40information about Technologies or
39:42impacts like a farmer may know that his
39:45farm is particularly dry or a New Yorker
39:47may know that uh he's subject to flood
39:50risk because of particular properties of
39:51his house right um so the I think the
39:56question in the things is always to look
39:59at it as an issue of information not
40:02just to say that we inherently need to
40:04go to people in order to get this
40:05information but to ask are there really
40:07concrete pieces of information that we
40:09think that the public might know that
40:10the government might not be able to
40:12figure out using some more objective me
40:14method right um that is it's not the
40:18value to individuals that defines what
40:20we care about uh it's really that there
40:24may be in information that individuals
40:26have which is particularly important so
40:28for example um you know if you try to
40:32figure out how much should we value the
40:33preservation of
40:35cultures uh there's a natural question
40:38of whether this comes from how much
40:39individuals value them or how much we as
40:41a society should value them and what is
40:44it that individuals value about them so
40:45for example imagine that the reason why
40:47we value preserving some cultures is
40:49that they may have a language which even
40:51though we can't currently understand it
40:52holds the key to understanding the
40:53history of human language right now it
40:56would be pretty stupid to go around to
40:58people and ask them how much do you
41:00value preserving the whatever culture
41:02because they would have no idea that
41:03this culture happened to have that
41:05language or not right um instead people
41:09might have some information about how
41:11much they value learning about the
41:12history of the world that the government
41:14doesn't have but the person who has the
41:15information about whether these tribe
41:18holds the key to the history of the
41:19world is not going to be the the
41:21individuals in the public but some
41:22Anthropologist who's an expert on this
41:24right so you always want to ask yourself
41:27who who has what relevant information
41:29and how do I relevantly combine it not
41:32uh you know just assume that whatever
41:33people say must be the correct
41:36number okay so uh when we do need such
41:40information when we really do need to
41:42learn from people how much they value
41:44preserving a culture or uh some
41:47environmental feature that they can
41:48understand the you know a natural
41:50approach is to just ask people how much
41:52they value it and what is that approach
41:54called
41:55Tanner uh Val yeah
41:59exactly um so this is a very simple
42:03approach that is a long history um and
42:06what you do is you basically describe
42:07different scenarios to people and you
42:09ask individuals how much they'd be
42:10willing to pay to have one scenario over
42:13another that's why it's called
42:14contingent valuation because you're
42:15valuing the contingency that something
42:17gets destroyed or preserved and you
42:19impute from this uh you you assume that
42:22this represents their true willingness
42:23to pay um this is most commonly used for
42:27environmental issues but also for
42:29creating
42:30Transportation or other public goods
42:33okay so um Kim what are some examples of
42:37areas where contingent valuation surveys
42:39have been
42:42used uh what value historical Place yeah
42:46absolutely uh can you think of any like
42:48concrete example Royal Palace yeah that
42:53would be good uh another very famous one
42:55is preserving visibility of the Grand
42:57Canyon like someone might say oh
43:00someone's going to build a hotel but
43:01it's going to obstruct your ability to
43:03see the Grand Canyon how much would you
43:05value given the chance that your kids
43:07might go and see it like being able to
43:09go visit the Grand Canyon or pre
43:11payments to preserve SEALS or whales in
43:13the ocean to avoid logging in national
43:16forests the value of the birds killed by
43:19the Exxon Valdes spill the value of
43:22preserving languages of an indigenous
43:24culture or building a new city and these
43:26are real examples that have been
43:28commissioned by regulatory agencies or
43:31courts um and they're very common in the
43:34benefit cost analyses of externalities
43:36regulations and actions they can also be
43:39used to assess damages from oil
43:41spills um and most of these are formal
43:44money based surveys you ask people how
43:46much are you willing to pay how many
43:47dollars but there are also some informal
43:50forms of this such as surveys of public
43:52opinion uh by politicians often
43:54influence the actions as politicians
43:56take and because they're not actually
43:58calibrated you don't actually say how
43:59many dollars they're not the same as
44:01contingent valuations but they have
44:03significant similarities and uh and
44:07impact policy in somewhat similar ways
44:10now these types of surveys are held in
44:12very low repute by economists uh why is
44:15that Adam um well some of the examples
44:18that um the uh study used was like for
44:22example getting like a warm glow the
44:25idea of like you know feeling good that
44:27value these things um also like you
44:30could kind of use as like a political
44:33issue like go zero or really high number
44:36to bump up whatever the average is or
44:39just you're not an expert on the matter
44:41and have like no clue yeah those are all
44:42really good points so I think the
44:44leading point is just people don't take
44:45these things all these seriously like
44:47imagine someone came to you and was like
44:48how much are you willing to pay to
44:51preserve the seals in the ocean you
44:52would say I don't care they just say
44:54whatever like sounds fun to say you know
44:57you know what I mean um the second thing
44:59is people often answer the wrong
45:01question they don't say how much they
45:03personally would be willing to pay but
45:04how much society as a whole should be
45:06willing to pay right and that if you add
45:08it up over all people in society leads
45:10to way way too high of a number right
45:13people even if they you could ask them
45:15usually don't have any idea information
45:17or experience with the question you're
45:18asking them they may not even have heard
45:20of the species of bird that you're
45:21asking them about or the culture that
45:22you're talking about right uh people are
45:25often motivated just to feel good about
45:27the answers that they give so like you
45:29know if you ask them about something
45:31really politically sensitive they'll say
45:32it's like a huge amount like you know if
45:35um how much would you be willing to pay
45:38to stop pollution being dumped on like
45:41poor people in Africa because that
45:43sounds really sensitive people might say
45:45some huge amount even though they would
45:46never actually pay anything like that
45:49right um and a sign of this is people's
45:52responses are highly inconsistent and
45:54apparently irrational they can differ by
45:56orders of magnitude or uh just depending
45:59on how you frame the question or as
46:00Austin was saying they can be completely
46:01invariant to
46:03quantities uh and perhaps the leading
46:05problem is that people have no incentive
46:06to get things right and no experience
46:08that would allow people to get things
46:10right yeah da have there ever been any
46:12um like cases in which people have
46:15actually asked for donations and
46:17actually got monetary contributions well
46:20that doesn't work either because that
46:21that suffers from the free writer
46:23problem oh okay so so the only mechanism
46:28that solves this well we'll talk about
46:29the vicory Clark Groves mechanism as
46:31you'll see that's pretty problematic but
46:33the only mechanism that really solves
46:34this is quadratic voting and in fact one
46:36thing that we're trying to do is set up
46:40uh uh in with the company is to try to
46:42replace contingent valuation surveys
46:45with quadratic voting based surveys
46:47where you're actually forced to make
46:48trade-offs between allocating money to
46:50things and yeah so yeah I don't know
46:53whether this is like a Miss application
46:55of an economic concept but closing that
46:57we could get people to answer the
46:59questions accurately and not motivated
47:01by one or whatever is it still worth
47:03trying to make sense of the answers cuz
47:05I was thinking of like Aros
47:06impossibility theorem there's no
47:08rational way to aggregate all the
47:10preferences and all the wants of these
47:11individuals so a good social again we'll
47:15talk about quadratic voting at the end
47:17but uh that violates errors and
47:19possibility theorem so you're right
47:21given like existing ideas about this
47:24problem there's supposed to be like no
47:26solution to this problem
47:27but I don't think that's true so uh yeah
47:30uh I feel like it might be mentioned
47:32here but isn't the fact that they're
47:34probably not actually paying directly
47:36for whatever caus ER yeah that that
47:39that's sort of what Dake was saying
47:41right he was saying that like you're not
47:42actually forcing people to make
47:43tradeoffs so yeah yeah philli uh arrows
47:46impossibility theorem only says that
47:50purely rank based things don't work so
47:53that's exactly the the reason why QV
47:56isn't rolled out yeah exactly yeah
47:59exactly um okay so let's talk about the
48:04Strategic problem because I think it's
48:05like the biggest problem at the heart of
48:07what's wrong with contingent
48:09valuation um so uh whe what your
48:13incentives are in a Content valuation
48:15survey depend on what it's being used
48:17for right so imagine that no one is
48:19going to take this survey seriously at
48:20all right people are just doing it for
48:23fun or for research or whatever right
48:25then uh what would be your incentives in
48:28a con evaluation survey if you were
48:30responding to one and you thought this
48:31is just going to be used for research or
48:33no one would really care what the
48:35answers were what would be my incentives
48:37to like fill out the survey yeah what
48:39would be your incentive to report that
48:41wouldn't have any I would you know you
48:42would do whatever was sort of fun or you
48:44know whatever like entertained you or
48:46sort of pissed off the survey guy or
48:48pleased the survey guy depending on
48:50whether it's a you know good-look member
48:52of the opposite sex or some uh some
48:55nerdy uh person you know um uh yeah
49:02um and uh that's uh probably true uh
49:08right now because a lot of these things
49:10people don't take seriously because
49:11they're totally stupid um but what if
49:14the survey actually did have a major
49:16impact on public policy Austin what then
49:18would be your incentives in filling out
49:20the survey well I mean you would have an
49:22incentive in the sense of actually being
49:24able to influence policy and what what
49:25would be the right thing for you to
49:27report oh the most extreme version of
49:30believe yeah exactly right so in this
49:33case you have huge incentives so if you
49:34like if you like the Dolphins you'd say
49:36oh you know I think I'd be willing to
49:38pay I don't know7 trillion to deserve
49:41the Dolphins uh you know and if you uh
49:46like imagine that you were a worker on
49:49the Keystone Excel pipeline right and
49:53they asked how much are you willing to
49:55pay to preserve the right of like the
49:58Caribou to go across the plains of
50:01Montana or whatever and you would
50:03probably say zero right uh or negative
50:07because you want to make sure that they
50:08don't block the Keystone Excel pipeline
50:10in order to you know save the Caribou
50:12right so you never have an incentive to
50:15report truthfully and a great example of
50:17this is uh a story from The Economist so
50:21did I tell you guys the story yet
50:24no like the oh it was yeah so there
50:27there was this great story uh where the
50:29economist went around Haiti and it asked
50:33people um you know have you been
50:35receiving uh supplies from
50:38UNESCO and people said oh no we haven't
50:42gotten anything and they went around to
50:44people's houses and they saw these
50:45stacks of like un uh uh you know boxes
50:49with filled with food and they said why
50:51did you lie to us and they said well if
50:52you knew that we were getting food we
50:54wouldn't get anymore right uh so
50:59um so if we want to take seriously how
51:02do we get people to report we need to
51:04consider the incentives that this gives
51:05people right and suppose people know
51:08suppose we know let's go back to the
51:11analysis of externalities we did from
51:13previous class suppose we all know all
51:14the benefits associated with a
51:18particular um project and we're a
51:21particular you know uh thing like
51:23pollution and we're trying to determine
51:25the socially optimal amount but
51:27everybody knows the amount that they
51:28individually are harmed so there's
51:30dispersed information about this right
51:32then the total harm is just the sum of
51:34the individual harms right and the
51:36question is how do we get people to tell
51:38the truth about
51:40um about the um harms that Ur to them
51:45myy how did your reading say that we can
51:48do that uh well they would ask you to
51:51report like the utility you got and then
51:53they would comp they would pay you
51:57all the other people's utilities except
51:58for yourself yeah they would pay you
52:01that amount of money and that would
52:02incentivize you to tell the truth
52:04exactly
52:05exactly um so that's sort of an approach
52:09that pushes this whole thing back a
52:10layer it says how do we figure out what
52:12externalities are well we make everyone
52:14pay their externality associated with
52:15saying that their externalities are a
52:17certain thing right that's paying you
52:19everybody else's utility right so then I
52:22report because the thing is reporting a
52:24harm to the system because it influences
52:26a social decision is really just like
52:28taking an action and if I pay the
52:30externality of that action I will I'll
52:32want to do um what is best for uh
52:36Society right so um Mike uh gin what's
52:41the um externalities of reporting that
52:44your harm is some amount uh it would be
52:48like the benefit you get minus the harm
52:51on other people well the benefit you get
52:53you capture yourself so it's just the
52:55harm on other people but the question is
52:56what is the like why does it create a
52:59harm on other people if you report your
53:03harm to be a certain amount uh because
53:05if
53:07you report because the utility that you
53:10report affects the payouts to other
53:11people under this well it affects the
53:13amount of the action that we take right
53:15right and when once you change the
53:17amount of the action you take it it
53:18harms or benefits other people
53:21right um so the reason is because
53:24without me society would Max maximiz
53:26welfare without looking out for my
53:28interests and with me Society also has
53:30to care about my interests and so my
53:32externality is how much other people
53:35could have achieved if they didn't have
53:36to worry about me minus how much they
53:38achieve when they do have to worry about
53:41me um okay so let's return to the
53:44smoking example and imagine that it's a
53:46small room now and the value of smoking
53:49to One smoker is 100 S - S 2 and there
53:52are two people who are each harmed in
53:54amount s^2 + 5 s and then you are harmed
53:58in amount s^2 + 10s so the question is
54:03um Adam what would be the optimal amount
54:05of smoking if I wasn't in the
54:08room
54:10um would be the marginal
54:14cost or the marginal benefits of the
54:16smoker should be a
54:38um it be
54:41uh
54:4315 okay so uh the marginal benefit to
54:46the smoker right is 100 minus 2s the
54:49marginal cost to each of the
54:52nonsmokers is 2s + 5 4 4 S +
54:5810 together so yeah uh so exactly uh so
55:04it's yeah so it's 4S + 10 and then you
55:08set that equal to 100 - 2 s and you get
55:10s =
55:1315 now the question is what's the
55:16optimum including you Mitch what would
55:19be the optimal amount if we include the
55:22harm that occurs to you as well
55:34um sorry was what so what is the optimal
55:39amount okay so the smoker gets benefit
55:43100 S -
55:44s^2 uh from smoking and each of the two
55:47other nonsmokers get this harm and you
55:50get a harm s^2 plus 10s from smoking
55:54taking place so if we take into account
55:55your U ility as well as the other
55:57people's what would be the optimal
55:59amount of
56:08smoking sending the marginal benefits to
56:11all people in society to the marginal
56:15cost you want to pass on this yeah
56:17somebody else want to try it uh uh yeah
56:21Dave um it would be S equal 10 because
56:25you would take the derivative of um your
56:29marginal utility which is or F which is
56:322 s + 10 yeah and then you add that to
56:34the marginal arm of the other two people
56:37yeah so that's 6X s plus 20 in total is
56:41equal to 100s oh sorry 100 - 2
56:45s uh 8s or yeah yeah 8s sorry 8 S Plus 8
56:51s equal is equal to n uh 80 80 exactly
56:56yeah
56:59um so uh then uh Barett can you uh
57:04figure out what the externality is that
57:06you create on other
57:08people
57:16uh so it would be
57:19um five times marginal cost
57:24or Okay so um the question is how much
57:27welfare would would people have gotten
57:30if they had been able to smoke
57:3415 so that's just the total benefit to
57:37the smoker if they smoke 15 minus the
57:39total cost to the other two guys
57:56BR you can do it on the border on I
57:58guess you're going to do it on
58:22calculator uh I got 675
58:26oh no it's so it's 75 so if
58:30um the net harm to everybody else is net
58:34benefit to everybody else is
58:35proportional to
58:3790s minus uh 3
58:41S2
58:44uh uh sorry
58:48minus
58:5290s minus I guess this should be
58:54multiplied by three no
58:57I I just added up the totals I also get
59:0075 you get 75 yeah you have 2 s squ - s
59:05squ so you should have just s Square
59:07which is oh no no no it's because that
59:10the lower one is negative so it should
59:12be - 3 S 2 yeah it should be - 3 s^2
59:15right so this should be 90 * 15 - 3 *
59:2015 and then we have to minus however
59:23much value is created afterwards which
59:25is 90 *
59:2710 and then plus because minus minus
59:31again so * 3 * 10 so can anyone do that
59:34does it give still
59:4575 45 -
59:51375 is 3 should be 335 I think I'm going
59:58to get up Excel and try to do it there a
1:00:00really large
1:00:02number
1:00:072775 90 * 15 + 3 * 15^ 2 - 90 * 10 + 3 *
1:00:1910
1:00:21squared 1425 woo totally wrong
1:00:27can't believe I've been had having this
1:00:29thing going for three years and we
1:00:30didn't figure that out before
1:00:34um okay so 1425 is the is the loss that
1:00:38you
1:00:39create um but you're glad to pay that
1:00:43why are you glad to pay that
1:00:461425 wait sorry can you can you I think
1:00:49I think you added * 75 way more credible
1:00:54to me
1:00:58uh yeah you added
1:01:01three 75 it actually was right how the
1:01:05heck did that
1:01:08happen okay it was actually 75 what do
1:01:11you
1:01:12know
1:01:14um okay so why are you willing to pay uh
1:01:18that um so imagine that you pay your
1:01:21externality right as we said under vcg
1:01:24why are you willing to pay that well
1:01:26because you're going to save an amount
1:01:28which is equal to 15
1:01:32squar times + 10 *
1:01:3515 minus how much smoking happens
1:01:38afterwards which turns out to be equal
1:01:39to
1:01:41125 okay so why would did it have to be
1:01:44that
1:01:44way well it had to be that way because
1:01:48Society was maximizing the total sum of
1:01:50everyone's
1:01:52utility given what you report and you
1:01:56had to pay everybody else's loss so once
1:01:59you add up everybody else's loss plus
1:02:01your
1:02:02gain either society's not going to move
1:02:05the amount right or you're going to gain
1:02:08more than everybody else loses because
1:02:10it's maximizing the total welfare right
1:02:12and so once you're forced to pay what
1:02:14you the harm you cause everyone else
1:02:16you'll always want to tell the truth
1:02:18because your interests are aligned with
1:02:20societies it's really the same principle
1:02:22of payments in accordance with product
1:02:24that we've been studying before
1:02:26right now you can look at what this
1:02:29thing was visually so imagine that this
1:02:32is the benefit to everyone else let's
1:02:34just say it's flat I mean before we had
1:02:35it being uh downward sloping but let's
1:02:38just say it's flat and imagine the cost
1:02:40everybody else is this the cost to you
1:02:42is this then the cost of society is this
1:02:44the harm that you create for everybody
1:02:46else is that you move from this point
1:02:48which was optimal for them to this point
1:02:50which is optimal including you and that
1:02:52creates a dead weight loss on everybody
1:02:53else of this harburger triangle here
1:02:57now notice that that triangle as all
1:03:00triangles is proportional to the square
1:03:03of the length by which it moves
1:03:06right that's the reason why you have
1:03:09quadratic voting is
1:03:10quadratic because the externality that
1:03:12you create on everybody else is
1:03:14proportional to the square of the amount
1:03:16that you influence the
1:03:19decision okay now the most common
1:03:23application of um this in
1:03:28practice um is the second price auction
1:03:32does anyone want to describe what the
1:03:33second price auction is to me and and
1:03:35say why it's an example of this yeah
1:03:37reishi um the second price auction is
1:03:41where whereas a conventional auction is
1:03:43where you know you if you have two
1:03:45people bidding the the first person ends
1:03:48up paying what the second person pays
1:03:50plus slightly more in a second price
1:03:52auction I think the first person pays
1:03:54what the second person
1:03:57right so yeah so in a traditional
1:03:59auction people bid and you pay whatever
1:04:01it is that you bid but here you don't
1:04:03pay what you bid you pay what the second
1:04:05highest person bid right that's why it's
1:04:06called the second price auction so
1:04:08suppose everyone bids for an object if
1:04:10the winner pays the bid of the second
1:04:11highest that's his externality can
1:04:13anyone figure out why that's his
1:04:16externality uh yeah was so that's amount
1:04:20that the second person would have valued
1:04:22the object and so by you getting it and
1:04:23them not getting it that's how much you
1:04:25CA them to exactly right yeah now um you
1:04:30always want to bid how much it's worth
1:04:32to you so let's let's see why that's the
1:04:34case conditional on you winning or
1:04:37losing you don't affect how much you pay
1:04:39by changing how much you bid right so
1:04:42the only thing you can affect is whether
1:04:43you win or not and the question is would
1:04:45you ever want to
1:04:47lose when by bidding below what you
1:04:50value it at or would you ever want to
1:04:52win by bidding above well no why imagine
1:04:55that you bid a little bit below what you
1:04:57value it at and someone ends up bidding
1:04:59in between what you value it at and what
1:05:02you've bid then you lose even though you
1:05:05could have gotten it for a price that
1:05:07was below what you valued it at now
1:05:09imagine that you bid above and someone
1:05:11bids in between uh what you bid and what
1:05:14you value it at then you end up winning
1:05:16it but you have to pay more than it's
1:05:17worth to you and you cost them losing
1:05:19more than yeah exactly so um that's the
1:05:24second price auction and some times a
1:05:26general more General version of this is
1:05:28used for what's called package auctions
1:05:30imagine you have a bunch of different
1:05:31goods and you could get different
1:05:33combinations of them you could submit
1:05:35bids for all different packages and then
1:05:37what you're forced to pay is so they we
1:05:39they choose the arrangement the awarding
1:05:41of goods that maximizes the total amount
1:05:43of bid that they got right and then they
1:05:46make people pay how much uh if they
1:05:49hadn't bid uh would have been created uh
1:05:53as value otherwise minus how much value
1:05:55was created right and that's basically
1:05:59the generalization of this notion of uh
1:06:01second price auction you could also use
1:06:03this in course selection so imagine that
1:06:05you were trying to choose courses you
1:06:07could put in bids for all of the
1:06:09different courses that you could get and
1:06:11then you could have some artificial
1:06:13currency and you could be forced to pay
1:06:15however much you would otherwise have
1:06:17bid right now there are many proposals
1:06:19to use this mechanism but it's very
1:06:21uncommon in practice um for a number of
1:06:26reasons uh let me give you one sort of
1:06:28creative example to use this so you know
1:06:30eminent domain is a process where the
1:06:32government comes in and takes people's
1:06:34land uh and then pays them whatever it
1:06:36thinks is a fair compensation for that
1:06:39imagine you had a system where someone
1:06:41had to make an offer for how much
1:06:42they're willing to pay to get the whole
1:06:43package of land if they want to
1:06:45redevelop it and each person there has
1:06:47to say how much the land's Worth to them
1:06:49and then if the offer is greater than
1:06:51the sum of the values that people have
1:06:53for their lands the land gets sold off
1:06:55to the guy and then if anyone had caused
1:06:59the sale to go a different way they're
1:07:01forced to pay however much it was worth
1:07:04to everybody else um to go the other way
1:07:09so for example the person who bids for
1:07:11it if he wins he has to pay the sum of
1:07:13the values to the of the people and if
1:07:16anybody um bids high enough so that it
1:07:20would have been sold had he not bid but
1:07:24given that he bid it doesn't get sold he
1:07:25has to pay the amount by which in total
1:07:28everybody else would have
1:07:29benefited now this this mechanism vcg is
1:07:33very elegant and kind of interesting but
1:07:35it has a huge number of problems we're
1:07:37going to talk a lot more about these in
1:07:39the last class but let me just raise a
1:07:41couple of them now so one problem that
1:07:44it has is that the money that gets
1:07:46raised has to be given not back to the
1:07:49people who participated but to some like
1:07:51central government Authority or to some
1:07:53third party why is that well otherwise
1:07:57um each of the individuals would have an
1:08:00incentive to uh bid in a way that would
1:08:04make other people be forced to pay
1:08:07more so for example imagine that I
1:08:09thought I were going to lose an auction
1:08:11but that you took the proceeds from the
1:08:13auction and gave it back to me I'd have
1:08:15an incentive to raise my bid just to get
1:08:17the other guy to pay more so that I
1:08:19would get some of that money back right
1:08:22so that money has to basically go
1:08:24outside of the system or be destroyed in
1:08:26order for the system to work which is
1:08:29pretty problematic right because even
1:08:31though it might get the efficient
1:08:32outcome you're going to waste a huge
1:08:33amount a second big problem with
1:08:36vcg uh probably the bigger problem is in
1:08:39these collective decision settings uh
1:08:41it's incredibly susceptible to
1:08:42manipulation imagine we were deciding
1:08:44what day to have the final exam for the
1:08:47course and imagine there were two
1:08:48choices of days I could go through the
1:08:50audience and ask everyone you know how
1:08:52much you were willing to pay and subject
1:08:53you to the vcg mechanism but imagine
1:08:56that Mory and Bart were friends and they
1:08:58both said I'm willing to pay
1:09:00$1,000 uh to have it on this day rather
1:09:03than that day well then we would we
1:09:06would definitely have it on that day it
1:09:07would overwhelm everybody else's but
1:09:09neither of them would pay anything
1:09:10because even without their report
1:09:12because of the other one uh it would
1:09:14have still gone that way right so
1:09:16because neither of them is individually
1:09:18pivotal uh it's incredibly incredibly
1:09:21susceptible to
1:09:23collusion so these are two of many
1:09:25problems with the vcg mechanism which
1:09:27make it fundamentally impractical but
1:09:30what is quite practical is this basic
1:09:32Insight that these in vcg these
1:09:35externalities are always going to look
1:09:37like these squares right we saw that
1:09:40here these triangle sorry we we saw that
1:09:43here um but that's a very general
1:09:45property and basically the origin of QV
1:09:48was realizing that um that even though
1:09:51uh you know vcg makes no sense for these
1:09:53reasons I was just saying that it always
1:09:56ends up looking like this uh thing
1:09:58that's a quadratic function of however
1:10:00much influence you have okay let you
1:10:03guys out early today good luck on the
1:10:04problem set

About this transcript

This page contains the full transcript of ECON20110 Fall 2013 - Lecture 8: Assessing Externalities by Glen Weyl, generated from the public captions YouTube serves with the video. The transcript has 11,555 words across 1,696 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.