Market Failures I: Externalities

DEVRAJ Today, we're going to talk about externalities. And we'll start with some simple theory--

so the theory of externalities. Now, as I've said before, the fundamental mantra

of most this course is that markets know best. Indeed, as I said, the first fundamental theorem of welfare economics-- that powerful-sounding name-- is the conclusion that the perfectly competitive private market will deliver

the welfare-maximizing outcome. So standard economic starts from the proposition the market

knows best. In that, world the government can only screw things up, OK?

All the government does is sort of mess up the first best. Now, either, despite this, the government

is, in the typical developed economy, at least, typically, 25% to 30% of the economy is the government.

So either these developed economies have it totally wrong, or there are some reasons

why we might actually depart from the standard 14.01 model of thinking for reasons for government. Now, we didn't cover one, which was imperfect competition.

We talked about it in perfectly competitive markets, government policy can make things better. What I'll do in the next two lectures

is talk about two other kinds of ration-- Oh, I'm sorry. So for [INAUDIBLE],, we talked about imperfect competition,

and obviously, redistribution is another rationale for government. We talked about that last time, too. If we think that social welfare is

increased by redistributing from rich to poor, despite some leakage-- [AUDIO OUT] So we've covered two roles.

There's fixing imperfect markets, and there's redistribution. In the next two lectures, we'll talk about two other roles

for government. And if this whets your appetite and you want to learn more, you could take 14.41, which is a course all about public policy.

You can learn a lot more. But we'll talk about two other reasons. Today, we're going to talk about externalities,

which is another justification for government involvement. And we're going to dive right in with an example.

So what's an externality? An externality occurs whenever the actions of one party

impact another party in a way for which the first party does not receive the costs and benefits.

So if my actions impact you, but I don't bear the cost if they're negative actions

or their benefits if they're positive actions, then that's an externality. So any time one party's actions affect another party,

but the first part does not receive either the costs or the benefits of that impact, that's an externality.

Let's start with the standard theory example, OK? The standard example we always teach now

is the example of the steel plant dumping sludge in a river. So imagine somewhere in America, there's

a steel plant located on a river, and they produce steel. And the production process has a certain production function.

And one feature of this production function is for every unit of steel produced, a unit of sludge,

a unit of waste, is produced. And the steel plant dumps that into the river. So just given the nature of their production process,

one unit of steel equals one unit of sludge is dumped in the river. Now, let's imagine that down this river,

there's a set of fisherman. And the fishermen make their living fishing in the river,

and that when the steel plant dumps sludge in the river, it flows downriver and kills their fish,

kills some of their fish, OK? Now, this is a classic exempt of what

we call a negative production externality.

It's production in the sense it's coming from the supply side. It's coming from the production of the steel.

It's producing this externality. It's negative because it's negatively impacting

the fishermen down the stream. And it's an externality because the steel plant doesn't give a shit.

They dump their sludge. It goes down, and it kills some fish. And they care, but they don't bear the cost of doing so. The fishermen bear the cost of doing so.

It's an externality, because the plant dumps the sludge, but it's the fishermen who bear the consequences.

It's a classic negative production externality-- "negative" because it's bad, "production" because it comes from the supply side,

"externality" because the costs are borne by another party, not by the party doing the activity.

Graphically, we can analyze that in something like figure 23-1. This is a classic externality diagram.

Here we have the market for steel. We look at the market for steel, because it's the production of steel that generates

this negative externality. We look at the market for steel. You've got a quantity of steel being produced,

and a price of steel-- standard market. You've got a supply curve and a demand curve. So in our analysis to date--

and let's be perfectly competitive. Our analysis to date, we say where supply equals demand is at point A,

with a quality Q1 and a price P1. And we would say that is the welfare maximizing outcome, perfectly competitive market, OK?

What this misses, however, is that the supply curve f is the private marginal cost.

That's the cost of the machines, and the building, and the workers that make the steel. But we as a society don't just care about the private cost.

We care about the social cost. We care about the fact that-- think of a society with the social welfare function.

That social welfare function incorporates both the steel plant and the fishermen. So we don't just care about the steel plant's costs--

the machines and the workers. We care about the costs and the fishermen, too. We care about social marginal costs.

What are social marginal costs? That is private marginal costs plus the damage done

to the negative externality. And what we're going to do is we should assume that's $100 per unit of steel produced.

We're just making that up. We're just going to assume that every steel unit produced delivers $100 in negative externality in terms of fish

killed downstream. Now, that does not have-- I haven't seen this constant. It doesn't have to be constant.

You can imagine a little sludge doesn't do much damage, but if you dump more sludge, it does more damage. That would be a non-linear curve that wouldn't [INAUDIBLE]..

But to make life easy, imagined sort in this region, every unit of steel

produces a unit of sludge that kills $100 worth of fish, OK?

And meanwhile-- now, on the other hand, the demand curve-- so the supply curve is the private marginal cost.

The demand curve is the private marginal benefit. But in this case, that is the same as the social marginal

because, there's no externalities from the consumption of steel. We'll come back in a minute to consumption externalities.

But let's assume for a minute that my private benefit from consuming steel is society's benefit from consuming steel.

It's not like by consuming steel, I create some positive or negative externality, OK?

So the demand curve is both a private and social marginal benefit, but the private marginal cost

is below the social marginal cost-- the social marginal cost, the private marginal cost, plus the externality damage.

Now, what does this mean? This means that the optimal level of production from society's perspective is not point a.

It's point c. From society's perspective, incorporating the damage

to the fishermen along with the cost of the steel plant, then the optimum point is where the social marginal cost

equals social marginal benefit, which is 0.2. So from the society's perspective,

there should be less steel production, OK? Add an amount to 2.2, which is below 2.1.

But the steel plant doesn't care. The steel plant produces T1. They don't care about some fishermen downstream.

So you have an overproduction and a deadweight loss. Now, here's the key thing about deadweight loss.

We've talked about deadweight loss before, but here's the new trick with deadweight loss. Deadweight loss is always drawn with reference

to social marginal benefit and cost curves, not private. So deadweight loss is relative to what's

best for society, not what's best for one particular actor in society. So the deadweight loss is all the units between Q2 and Q1

where the social marginal cost exceeds the social marginal benefit.

So for that first unit to the right of Q2, there's this tiny deadweight loss.

But as you move further and further from Q2 to Q1, the deadweight loss grows. Because those are units where there's

a bigger and bigger gap between the social marginal cost and the social margin benefit.

The bottom line is, we're back to chain of logic. The private steel plant does what's privately optimal.

That's with private marginal cost equal to private marginal benefit. That's at point a.

From a social perspective, we care about social marginal cost and social marginal benefit. Social marginal benefit is the same as private

marginal benefit in this example. But social marginal cost includes the cost to the fisherman, so it's higher.

Given that it's higher, we'd socially optimally produce less steel, OK?

And therefore, we would want less production of steel.

Now, once again, there's a lot of complications here. For example, what if we, instead of using less steel,

we just stop them from polluting? Well, that would be another end. There's lots of complication here, but for now, just assume that's the way the production

function works, If you're going to make steel, you're going to pollute, OK? So that's the outcome.

That's what we call-- that's how we think about a negative production externality. Now, externalities can also happen on the consumption side.

And the classic example of negative consumption externality is smoking.

So let's go to figure 23.2. Now we have the market for cigarettes.

There's the quantity of cigarettes on the x-axis, the price of cigarettes on the y-axis. Now, here, we're going to assume the private marginal costs

curve is equal to the social marginal cost curve. That is, there's no-- maybe, when they make tobacco for cigarettes, they pollute.

But we'll ignore that for now, OK? We're just going to assume that there's no pollution from the production side.

But on the consumption side, when I smoke a cigarette, I do damage to other people, OK?

Most perniciously, most obviously, through secondhand smoke. I'm very sensitive to this right now, because my wife's mom smoked three packs a day.

My wife now has COPD, Coronary Obstructive Pulmonary Disease, and was just in the hospital with pneumonia,

because her mom smoked three packs a day, and so her lungs are weak. So this is the example of a negative consumption

externality. Her mom smoked, and smoked to the point of point a,

where her private benefit is equal to private cost but the social cost includes my wife, I like to think.

And therefore the benefits include the negative impacts

on my wife. So this led my wife's mom to over-smoke. Or in general, people over-smoke because they

don't account for the fact that there's damage on others. Therefore, there is too much smoking.

Too much smoking, because it doesn't account for the damage on others. But in this case, this is not a change in supply curve.

It's a reduction in the benefits. Because the production of cigarettes hasn't changed. It's just the value of consuming cigarettes has changed.

Cigarettes are less valuable to consume because you get enjoyment from smoking-- at least some people think it's enjoyable.

But offsetting that is the damage you doing to others. So that the shift downward of the marginal benefit curve.

Now, the social marginal benefit is below the private marginal benefit.

And the new equilibrium is where the social marginal benefit equals the social marginal cost, OK?

And therefore, the optimal level of smoking is lower than what individuals would privately choose.

Yeah?

STUDENT: Where did you get the margin at the end? Because there is still profit margin in the business. JONATHAN GRUBER: Yeah, so that's a very good question.

I'm going to come back to what happens when you depart-- we're assuming perfectly rational people

with perfect information. I'll come back to this. This gets more complicated [INAUDIBLE].. This is a perfectly rational person [INAUDIBLE]

even before they know it. I'm not saying it's-- I'm not saying they're bad people. I'm not saying smokers are bad people, because they're

doing what's privately optimal. There's nothing bad about doing what's privately optimal. But it doesn't matter if they knew or not.

The point is, they're doing damage. From a social perspective, there's too much smoke. Yeah?

Yeah. STUDENT: [INAUDIBLE] JONATHAN GRUBER: Right. STUDENT: [INAUDIBLE] example [INAUDIBLE]..

JONATHAN GRUBER: Where does $0.40 come from? STUDENT: $0.40. JONATHAN GRUBER: OK, so $0.40 cents per pack is actually an empirical estimate of the external cost,

the external damage done by smoking. So it's actually quite interesting, OK? So basically-- now, this is without secondhand smoke.

Think about the exercise in smoking. So for example, you get sick, and your medical costs are higher.

Actually, if you get external costs, because smokers start a lot of fires. Thousands of people die every year in America

in fires started by smokers. That's an external cost, because other people die, and the fire department-- things like that.

Smokers are less productive, so firms make less money, because they have to take smoke breaks and things like that. You add all that up, and you get $0.40 a pack.

Now, that's without secondhand smoke. The problem with secondhand smoke is, we don't know how big that affect is.

The estimates are someone between $0.01 a pack and $2.00 a pack. So it's just really, really hard to measure.

And it's really hard to measure partly because of a really interesting phenomena. Which is, you might have come back to me and said,

wait a second. What happened to your wife is not an externality. Why would-- under what situation-- here's

a hard question. Under what set of assumptions would the damage done to my wife by her mom's smoking not be an externality?

Let's say her mom was still alive. You could see her in the hospital.

Yeah. STUDENT: [INAUDIBLE]

JONATHAN GRUBER: Exactly, if only [INAUDIBLE] her mom is going to bear the cost. But if her mom knew that by smoking--

this is where information comes in-- knew that by smoking, she'd make her daughter sick,

and chose to smoke anyway, then it's not an externality. In that case, in the marginal benefits for the mom

would be included the damage done to the daughter. So what's really interesting, externalities

is not a strictly easy thing to measure. It depends on-- now, the fisherman [INAUDIBLE]

floats down the river. That's different. But here, the externality is only true if the mom didn't account.

So the key thing, the key question that you have to ask is not, is do families maximize a family utility function

or an individual utility function? If her mom was maximizing her individual [INAUDIBLE]---- sorry to rag on my mother-in-law-- may she rest in peace-- but it's a good example.

If she was measuring-- if she was just focusing on her individual utility--

then she should just consider the benefits to her of smoking versus the cost of a pack of cigarettes.

But if she was maximizing family utility, she should take the benefits to her smoking minus the damage it does to her daughter versus the cost

of a pack of cigarettes. Yeah? STUDENT: [INAUDIBLE]

JONATHAN GRUBER: Ah, well, no, But that's-- you're assuming the family utility function. That's what I mean by family utility function.

I mean you care about people in your family. Now, basically, the interesting question is, in other words,

it's just about weight. Let's say at one extreme, a purchase is a function, I don't care about anyone in my family. A full family utility function, I treat them

as if they are myself. I care about the [INAUDIBLE]. Where do people line in between? Now, in fact, there's some evidence on this.

And the evidence is basically that women care about their families more than men. So there's actually a number of studies

in developing countries. There's a cool way to test this. And the way you test this is you say,

if family utility maximization holds, then it shouldn't matter who controls the resources in the family.

If I take $1 from a wife and give it to a husband, that shouldn't change anything, because they're both maximizing the same family utility function.

But if there's individual utility functions, then who controls the resources will determine how the money gets spent.

And there was a quite striking example from the UK where they had a tax break for children.

And it used to be what happened-- this is back in the '70s-- it use to be what happened was, for every child in the household,

at the end of the year, a check got sent to the house. Well, typically, women were home and controlled the money

in the house. So the women got that check. They then changed the law and said, no, instead, we're just going to let take-home pay be a bit higher.

So instead of sending a check, let me see your check each week, we'll include it in your take-home pay. Well, the take-home pay went to the men.

Men were working, women weren't. So it's shifted from women controlling the money to men controlling the money.

And what they found was this shift led to much less being spent on children's clothes and books, and much more being spent on alcohol and tobacco,

which is inconsistent with a family utility function. It's consistent with men caring less about their kids than women, OK?

So the bottom line is, we don't have a perfect family utility function. Probably we don't have a perfect individual utility function

either. Probably, people care somewhat about their family. But as long as it's not a perfectly family utility function, then there's some externality on family members

from secondhand smoke. So $0.40 is clearly too low. We just don't know how low, OK?

Now, externalities don't have to be negative. They can also be positive, OK?

Let's think about a positive consumption externality. And let's come to one of my favorite people

I like talking shade of, or throwing shade to, which is my neighbor, OK?

I don't get along with my neighbor at all, and basically, part of the reason I don't get along

with my neighbor is because my neighbor has a tendency to start massive landscaping projects,

and then leave them half done, and just not do anything for 10 years. And it's unattractive.

And so basically, let's imagine that my neighbor is looking outside his yard.

And he's thinking about the landscape. And he's thinking, what's it worth to me to fix this dirt pile in my yard--

which is, now, finally, after 20 years, fading away. But it was there for about 20 years. What's it worth to me to fix this dirt pile in my yard?

OK, well imagine that the dirt pile will cost $1,000 to haul away, but he only gets $800 of this utility

from looking at it. So let's say I get $500 of utility from not looking at it.

And let's say-- this is totally true-- he doesn't care about me. Then he'll say privately, I shouldn't move the dirt pile.

My benefit from doing so is $800. My cost of doing so is $1,000, OK?

So the cost is a [INAUDIBLE] if I don't do it. But socially, he should get rid of the dirt pile, because the benefit to society is the $800 to him

and the $500 to me. So if he removed the dirt pile, there would be a positive externality.

For him to consuming landscaping services, there would be a positive externality, which is my improved view.

So the positive consumption externality, OK? So that is-- you could think of this thing as figure 23.2,

but just flip it now. Think of my consumption having a marginal social benefit.

In that case, unlike the case of cigarettes, where there's too much consumption,

now there's too little consumption of landscaping services by my neighbor, Because he's not accounting for the benefit to me, OK?

Yeah? Let me come to that in one minute, OK?

Now, finally, there's one last example, which is you could have positive production externality.

And the classic example of positive production externality is research and development by companies.

When companies do R&D, it benefits them, but it also produces general knowledge that benefits everyone, including their competitors.

As a result, if you go back to figure 22, 23.1, my social marginal costs of doing R&D are actually

lower than my private marginal costs, because I'm producing benefits for other produces.

It's the opposite of dumping sludge in the stream. I'm dumping knowledge in the stream, OK? And that knowledge benefits everybody downstream.

Indeed, the estimates are that $1.00 of R&D is worth about twice as much to society as it is to the individual firm to do

it. Indeed, if you'd like to learn more about this,

you can pre-order my new book, Jump-Starting America, which talks about how we're going to make America run better again by investing in public R&D.

So I talk all about it. Very excited about that right now. So those are the kinds of externalities,

positive and negative. Now, the question many ask is, so what?

So these things exist. Why can't people just figure it out? Take me and my neighbor, OK?

There's an easy solution here. I just go to my neighbor, and I pay him 200 bucks--

or $201. He's $1 better off. I'm $299 better off.

Problem solved. Indeed, there's a famous economist named Ronald Coase--

Coase-- who proposed that there is no such thing, really, as an externality, because all externalities

can be internalized through private negotiation.

So as long-- here's the key-- as long as there's well-defined property rights.

So in the case my neighbor, there are well-defined property rights. He owns the dirt pile. So I go to him and offer him money to fix it,

to get rid of it. Likewise, imagine the fisherman owns the river. They could then go to the steel plant and say,

we are going to charge you for dumping sludge into the river. And we'll charge you enough so that your private costs go

to the social costs, OK? So basically, Coase said there's no problem here.

There's no need for government. The private market can take care of it. Well, that's a really cute argument in theory,

but in practice, it breaks down. Because think about that argument for the big biggest [INAUDIBLE] in the world today,

which is global warming, OK? Global warming happens because I drive my car and people in Bangladesh drown.

How are they supposed to come to me and say, John, you should drive less. I'm going to pay you to drive less so I don't drown.

And that's just crazy, right? It doesn't work for big externalities. It doesn't always work for little externalities.

So think about yourself, OK? Think about your neighbor playing music too loud.

You have a 14.01 test the next day. Your neighbor's done with finals. They're partying, playing their music too loud.

That's a classic negative externality on you. Now, in principle, you should say, well,

I can figure out how much it's worth to me and go pay them to turn their music down. But in practice, you're not going to do that, right?

You'd be, like, a social outcast. OK? In fact, people don't work this way. There's a famous story about an economist-- it can't be true,

but it's a wonderful story-- who was on a plane trying to get work done. The person next to him wouldn't stop talking,

so he offered them 5 bucks to shut up. Now, that would be the Coasean solution.

But in reality, life doesn't work that way. So in fact, the private market is unlikely to solve these kinds of problems.

Yeah? STUDENT: [INAUDIBLE] what does it cost [INAUDIBLE]??

JONATHAN GRUBER: That's another problem. There's lots of reasons why you can't-- there's no way.

Even Coase bless his soul, would not take the private negotiation to solve the global warming problem. There's no way, OK?

So that means we need the big bad government.

We need government solutions, because the private market has failed. The private market has failed because the private equilibrium

is not welfare-maximizing, OK? So if you go back to 23.1, the private market

has failed because the private market overproduces steel. Therefore, there is a role for the government to,

potentially, fix this market failure. It doesn't mean the government will, OK? This is the key thing. I emphasize this in my course.

The way to think about the logic of government involvement in the economy is with two steps. One is, is there a market failure?

That is, will the private market fail to maximize welfare? And two is, can the government actually make it better as opposed to making it worse?

So just because if I have a market failure, it doesn't mean there's a role for the government. It means there's a potential role for the government. Well, let's talk about, theoretically,

what the government might do, and, in fact, what the government actually does do. So theoretically, the government has two options.

The first option is regulation.

So going to figure 23.1, the government could summon and say, hey, steel plant, we've done the math,

and you're optimal level of steel production is 2.2, so we're going to regulate you so you

can't produce more than 2.2. And if the government did that, we would get to the optimal outcome.

If the government said, I'm setting a max at Q2, then we get the optimal outcome, which is 2.2 will be

[INAUDIBLE]. The problem with that is exactly the problem we talk about when we talk about regulating monopolies,

which is that requires a huge amount of government information. They need to know the shape of the supply-- they need

to know the shape of the supply curve, the shape of the demand curve, and the size of the externality. They need to know all three things--

supply, and the size of the externality-- in order to get that right.

And if they get it wrong, they could lead to-- if they get it too wrong, they could shut down a steel plant

that's still totally valuable to have running. So we have this conversation at length with monopoly regulation-- all the same problems of life.

That's why economists typically prefer a simpler solution, which is taxation,

or, what we call in economics "corrective taxation."

And this is exciting, because when I taught taxes last time and the time before, taxes caused deadweight loss.

Taxes caused inefficiencies. But actually, in the case of externalities, taxes can reduce deadweight loss.

How's that? Let's go to figure 23.3. Back to the steel market.

Imagine I don't know supply and demand curves. I just know the size of the externality.

It's $100 a unit of steel. So I simply go to the steel plant and say, look, for every unit of steel you produce,

I'm going to charge you $100. You do what you want. I don't know where you're going to end up, but I'll charge you $100.

That means that the steel plant-- the steel plant's private marginal cost curve will now align with the social marginal cost curve.

You've essentially done what Coase wanted us to do. You've internalized the externality, but done through government action rather than

private negotiation. You said, you are imposing a social cost

above and beyond your private cost. We are going to essentially make you consider-- it's like moving to a family utility function.

We're going to make you consider all of society in your decision. We'll make you internalize the damage you're imposing in your decision.

And the way we're going to internalize that is by charging you for the extra costs you're imposing on society, OK?

So that is what we call "corrective taxation," OK?

And corrective taxation leads to the optimal outcome. If we say to your steel plant, for every unit you produce,

we are now going to tax you by amount MD, which is the marginal damage, then it leads them to actually choose

point B-- which is confusing, because it was point C on the other graph, OK? To choice point D--

OK, it was D on the other graph-- which is the optimal outcome. You could say that the system is that [INAUDIBLE]----

which is the optimal outcome. It will cause them to choose the optimal outcome, OK?

So we've essentially got the private market to do the right thing by forcing them to internalize

the externality, much as Coase would hope it would happen through private negotiation, we've caused to happen through corrective taxation.

Likewise, we can think of the same thing with smoking. So go back to figure 23.2 Imagine I levied a tax of $0.40

a pack on cigarettes. Then the marginal benefit of smoking would fall by $0.40.

My willingness to pay-- remember, we [INAUDIBLE] analyze taxes [INAUDIBLE]. When we say you have a tax, it essentially

shifts down the demand curve, tax on consumers. It lowers my willingness to pay. I'm only now willing to pay $0.40 less a pack,

because I have to pay $0.40 in tax every pack. That shifts my demand curve down. Plus, the private marginal benefit

now equates to social marginal benefit. Boom-- no more externalities of smoking.

So here's the crazy thing. Taxes have now removed the deadweight loss. Taxes haven't created the deadweight loss.

Taxes have removed the deadweight loss. So in a perfectly competitive market with no market failure,

taxes create deadweight loss. But once there is market value, taxes can actually

offset deadweight loss, OK? And that's the theoretical case for government intervention.

Theoretically, the government can make things better through opposing corrected taxation.

OK? Questions about that? Yeah? STUDENT: Elasticity plays into it.

If somebody is really inelastic, has really inelastic consumption of cigarettes, would this tax--

JONATHAN GRUBER: You just need a bigger tax. STUDENT: OK. JONATHAN GRUBER: Yeah. So basically, the way elasticity plays into it-- now, if they're totally inelastic, you'd

have to wear something else, Nothing's ever perfectly. But essentially, you could take the diagram, and essentially--

no, I'm sorry. No, let me back up. I said that incorrectly. Elasticity doesn't matter for the actual government policy.

All I want to do is impose the size of the externality, OK?

If, demand is inelastic, that could mean people love smoking.

And as a result, you won't reduce smoking that much. But that's OK, because basically,

all we care about is correcting the externality, OK? We're not trying to punish smokers.

We're not trying to achieve an optimal-- and that's the difference between regulation

and taxation. Regulation says, no, I know what the right level of smoking is. I'm going to impose this. Taxation is, look, I don't know the right level of smoking is.

Many people frickin' love smoking. Imagine if there was externalities to consumption of insulin, OK?

Who wants to say, no, less insulin, because-- you want to say, no, we have to have a ton less insulin.

We should say, no, we should tax for the externalities they impose. But if they still choose to do a lot of insulin, then they're still doing the socially optimal decision.

It's a really good question, because it highlights the key nature here, is this is why this framework, traditionally, was very controversial

with environmental advocates. Environmental advocates hated this framework, but they're like--

they called it a "pollution absolution." Guess what absolution is. It's this thing that these corrupt monks

used to give before Martin Luther came along to reform the church back in the 1400s. Monks would come along and say, yeah, I know you sinned,

but if you give me one coin, or whatever, I'll say your sin's absolved.

So the monks would sort of bribe people to get into heaven, OK? It's called absoltuion. So for many, this was called pollution absolution.

Environmentalists hate this way of thinking about it. They're like, wait, you're letting people pay to get out of the damage they're doing?

And the answer is, yeah, it's the right thing to do. Because in fact, if people pay and don't change their behavior, that means they just

love doing that activity. And why should you penalize them just because they love doing that activity? We should penalize them to the extent they're hurting others,

not care about the enjoyment they get from it. So if people, for some reason, just love smoking,

let them smoke. But make them account for the damage they're doing to other people, OK?

It's a great question. Other questions? Yeah?

STUDENT: [INAUDIBLE] if a person does [INAUDIBLE]---- JONATHAN GRUBER: Great question. I'm going to come back to it at the end of the lecture.

Great question. OK, let me talk-- that's a great segue to what do we talk now, which is, let's move from theory to practice.

Let's talk about actual government policy, and what governments do, and how it's informed by this kind of framework.

Let's start with environmental externalities-- the classic piece.

The classic case of negative externalities is the pollution in the environment, OK?

G environmental science, for example, we produce electricity. The emissions from that electricity, the sulfur,

and the nitrous oxide-- sulfur dioxide, nitro us oxide-- go into the air. They form particulate when get into our lungs

and causes us to get sick. They cause acid rain, which falls and erodes [INAUDIBLE] cities, and buildings,

and things like that. And then, of course, we have global warming, which is the single biggest issue facing--

maybe not you, almost certainly your kids, and definitely their grandkids, OK?

Depends how long you're going to live. You guys aren't going to make it to 2100. Yeah, probably not you.

Your kids when they're old, and your grandkids for sure-- which is, basically, the amount of fossil fuels

we burn if it's loaded. We now-- and the amount of carbon dioxide in the atmosphere is now the highest

in at least 200,000 years. And by estimates, including from the Trump administration,

this will lead to massive warming of the earth. And by the year 2100, global temperatures could rise as much as 10 degrees Fahrenheit.

Now, for those of you from North Dakota, that might not sound so bad. Even right now in Boston, it might not sound so bad.

Remember the great line from The Simpsons when it snows and Homer goes, Hey, Lisa, said there's 10 feet of global warming.

But basically, the truth is it's very bad, because basically, it leads to enormous rises in sea level by as much,

essentially, as 3 feet. And just think about that.

That means, for example, Bangladesh is gone. Bangladesh has 100 million people,

150 million people-- gone. Cape Cod's gone. New York City, largely gone. Florida, gone.

Basically, America becomes the Midwest. Basically, essentially, you've got--

and to put it in economic terms, the Trump administration's calculations suggest the US will be 10% poorer by 2100

as a result of global warming. This is real. The conflicts are real.

But it's a classic externality, because it's the ultimate externality. Because every single time you drive, you contribute to it.

And every single time someone in Australia drives, they contribute to it. Because it all just goes in the air,

and mixes together, and create one level of global warming. It's a completely global externality.

Any pollution of emissions anywhere has the same effect. And the effect is very distant in space.

So if you're in Minnesota, the effect is not nearly as bad as if you're in Bangladesh. It's the city is time, because the effect is going to be way

in the future, not on you. It's a classic Externality, because the effects clearly don't affect you.

And they arise from your consumption. Yeah?

STUDENT: [INAUDIBLE] JONATHAN GRUBER: Oh, yeah. Well, it's not-- the time part is what

makes it a bad externality. As long as when I drive, I do not

pay any extra cost for the marginal damage I'm going to the environment, then It's an externality. It doesn't matter if it's going to affect now or in the future.

OK, the bottom line is, when you drive, you are likely making the environment worse.

Ah, OK, you live in that environment.

But the problem is you, particularly here in Boston, aren't going to feel, necessarily, the consequences of global warming, at least initially.

So you're right. If you lived in your own little world, that's a great one. Remember, externalities is you bear it.

If you'll be in your own little world where you drove, and made your own little rain clouds like Charlie Brown, and you covered all the conflicts,

then it wouldn't be an externality. But the truth is, that's not true. Many, many people suffer the consequences.

Now, what do-- yeah. STUDENT: [INAUDIBLE]

JONATHAN GRUBER: That's right. The key issues is discount. There's a ton of really deep and hard issues here.

How do you think about discount when the damage is 100 years from now? So let's say you took a standard interest rate today.

Today, interest rates are low-- 3%. Well, it's still true that 100 years from now, that means damages are worth 20% of what they're worth

today. At a higher discount rate, they're worth effectively 0. We're basically saying, if you have any form of discounting

of any reasonable magnitude, you're basically saying your great-grandkids don't matter today. Well, that seems implausible.

So there's a lot of deep work on multi-generational discounting. How do we think about discounting for hazards which

are way, way off in the future? Now, it turns out, even with reasonable discount rates, the cost is still huge today, because they're

so enormous in future. Even if we discount them back, they're still big today. But that's the kind of difficult issues

that is really exciting and interesting stuff. So what do we do? Well, in Europe, they've gone the taxation route.

A gallon of gas in France is $7. Macron has just proposed increasing the tax,

and it's causing Paris to burn-- ironically, increasing global warming. But typically, in Europe, gasoline

is many, many times more expensive than in the US. And basically, they have carbon taxes that try to control this.

But it proved very difficult to do that in the US. Indeed, when Bill Clinton proposed a $0.03 tax on gasoline when he first elected,

it basically cost him the the Congress. Think about that, $0.03. That's a daily swing in the price of gas.

Gas taxes are very hard to impose in the US. So that's why there's been a move away from a taxation

approach towards a regulation approach. Even though regulations [INAUDIBLE] have other weaknesses, there's been

a move toward trying to cap global emissions, trying to regulate a cap on global emissions.

And this began with the first such agreement negotiated in Kyoto in 1990--

in December '97. I was at that. That was pretty cool. I was part of the Clinton administration delegation

that went over there. And it was crazy, because basically, you had all these--

think about how hard this is to really negotiate. So let's say you're India. And you come, and you say, wait a second.

Let me be clear. In 1997, of all the stuff in the air, the US was responsible for about 30% of it,

and the rest the world combined with 70%. Then you come in and say, wait a second. We haven't done anything to the environment, OK?

We're ready to grow. We want our air conditioning. We want our cars. We want our-- and you say, you say, no, wait a second, you can't have that because we screwed up the environment.

Screw you! OK, you guys, screwed it up? You guys fix it. It's a pretty hard negotiation you have to do.

Plus, the costs are quite large. So it was actually pretty cool. I was there for four days. I slept four hours.

In Japan, they have this cool coffee in a can that you just sort of live on.

And we negotiated an agreement which was not going to solve the problem, but it was going to slow the growth of emissions.

But America was never actually going to ratify the agreement, so it never went anywhere.

Fast forward to 20 years later to three years ago in Paris.

We signed another climate agreement called the Paris Climate Accord, which was similar to Kyoto in that it set targets

that countries could meet. And we were successful that more countries participated in Kyoto. Countries like India and China didn't participate.

But now, most countries that participated in the Paris Agreement limiting emissions are-- the US.

In that case, signed it, but now, the Trump administration has backed off it. So we're were not enforcing the Paris Agreement.

Now, you can completely understand politically why a leader would not want to enforce an agreement like this.

You're imposing pain today for the benefit the people you'll never even meet, OK?

First of all, they live in other countries, by and large. Second of all, they're going to be 100 years from now. So politically, it is unbelievably challenging.

But the problem is, by the time those people are drowning, it's too late.

Basically, even if today, if we just suddenly stopped using any fossil fuel today-- which is impossible--

but let's say that we stop using any fossil fuels today, temperatures would still rise for something like 300 years

before they start to fall again, OK? So basically, this is an incredibly hard problem

because the costs are super distant-- the benefits are of fixing it are super distant if the costs were imposed today, And that's

a very, very hard problem. That's why this is probably the single biggest social problem facing the world over the next 100 years,

is how do we deal with-- the economics of the answer is easy. I mean, it tells you what to do.

It's easy. You post taxes or regulations. You essentially phase out of being a fossil fuel-based economy.

OK so theoretically, it's no problem. The problem is all political, which is basically

how you get people who are going to to bear the cost in the near term without [INAUDIBLE]?? And that's it's an enormous challenge.

Yeah. That's a deep question about the nature--

essentially, there's a different equilibrium where they have more faith in the government in New York City than in the US.

Maybe that's because you're settled by Frontier, Venture, people who never never like to be told what to. Do I didn't quite know what that is.

But in any case, this is different attitudes towards the government in general. And I think, essentially, if you're going to deal with this,

you're going to have to essentially trust-- think of a social welfare function, which is not everybody today, but everyone

everyone in the whole future with some form of discount. You've got to trust that social welfare function, which

is the governments. In Europe, they trust that more than in the US. But look, I don't want to minimize how hard this is, OK?

There's enormous disagreement about how fast we should go. In fact, I was originally--

one of the reasons I went to Kyoto was because the economists in the Clinton administration were opposed to how strict the Kyoto targets were.

We thought it was too severe. And I actually, in 1997, was the subject, I believe, of one of the very first spam attacks.

I got spammed by hundreds-- which, at that point, was enormous-- of environmental activists saying how terrible I was because I didn't want to restrict emissions more, OK?

Basically, there's huge disagreement in this. But it's a topic we can't ignore.

It's a critical topic. Likewise, another critical source of externalities.

Is health externalities. Now, we've talked about smoking, but there's

lots of externalities. So smoking, the externalities are actually not that large.

$0.40 a pack is not that large. Depends on if you use secondhand smoke. Compare that to drinking.

The externalities for drinking are enormous, because drunk drivers kill people, OK?

Indeed, there are 13,000 deaths per year. To be politically incorrect, let me say it--

that is our 9/11s every year from people drunk driving.

400,000 people are injured every year from drunk driving. That's a classic externality.

If I kill someone in my car, maybe my license will get taken away. But m I'm not bearing the full cost of killing, OK?

Yeah. STUDENT: [INAUDIBLE] JONATHAN GRUBER: That's a great question. This is-- I'm trying to compress my whole course into one

lecture, which is hard. But there are estimates of value of a life. Typically, we say a life is worth about $10 million.

We can talk another time about where that comes from. But basically, if you take those valuations,

the externalities from drunk driving are enormous. And yet, we have very, very low taxes on alcohol [INAUDIBLE]..

Yeah? This is a great question.

We could go all day on this. Absolutely, they are. And this is why nobody likes economists.

For example, we think babies' lives aren't worth as much as your lives. Why? Because you've got more embedded human capital.

You've formed more connections. The value of your life is worth more. Tell that to the parent of a newborn, OK?

So there's all sorts of interesting and difficult issues in valuing life. That's a whole separate field.

But in any case, the bottom line is, since life's valuable,

there is an important externality. Now, let's come to the interesting case, a very hot

topic, which is illegal drugs. Now, let's talk about cigarettes for a second. Cigarettes impose externalities, but I never

said the answer with to ban cigarettes or make them illegal. I said the answer was to tax them

to the extent of their externalities. So let's say we could properly measure secondhand smoke or whatever.

We do it. So indeed, cigarette taxes today probably exceed

any reasonable expert of externality. I'm one of the biggest anti-smoking advocates out there, and I would admit cigarette taxes today,

which average nationally several dollars a pack, probably exceed the externalities of smoking.

Moreover, then, illegal drugs, we don't just tax them. We ban them. Until recently, in all states, we banned marijuana.

We still banned other drugs everywhere. Why? Think about it.

The framework I taught you so far does not say you should ban anything. If I love doing cocaine, you should

tax the externalities I do. Indeed, let's go further. There aren't many externalities to most illegal drugs, OK?

Guys who do cocaine, they may get in car accidents-- or heroin-- but they largely just kill themselves.

They don't do a lot of social damage. They largely just kill themselves. Indeed, what are the externalities from illegal drugs?

They come from fighting illegal drugs. It's the crime that comes since the drugs are illegal, OK?

That should be externalized from drugs. So for that reason, economists, for decades, have argued drugs should be legal-- all drugs.

Because if you think about it, they should be legal and taxed. Because the externalities largely come to the fact that they're illegal.

You think about all the deaths in Mexico. Think about what would happen in Mexico if drugs were legal in America.

Suddenly, the price would collapse, the cartels would have no money to make, and they'd break up, and there wouldn't be all these deaths in Mexico.

Those deaths in Mexico are our on us. It's on the US consumers who demand these illegal goods from Mexico that create

the demand for crime in Mexico. It's probably-- if drugs were legal, it wouldn't be a problem.

So why not legalize drugs? Well, we only have one minute. So the answer is because we don't think this model applies

in all [INAUDIBLE]. In particular, we don't think people always make rational decisions. And this framework computes rational decisions.

We think that people make mistakes. And if the mistakes are severe enough, then taxation is not the right answer.

You need an outright ban. And that's why we do. That's why we say, look, we think that people,

most people try illegal drugs before the time they're 18. We think that people before they're 18

don't have fully formed prefrontal cortexes. They're not fully rational. They get addicted, and that has lifelong consequences.

Therefore, we make the drugs illegal, because we think that the costs of those consequences are so large that we don't want to take a chance.

But that raises the whole debate we should all be having about marijuana legalization. Is that true of marijuana?

The externalities of marijuana are really small, OK? Indeed, the damage you do to yourself is actually pretty small-- much less than smoking.

So how do we think about whether marijuana should be legal or illegal, versus cocaine and heroin, and how much it should be taxed?

So that's the kind of set of interesting issues you raise. We talk about those. So anyway, if you find that more interesting, take 14.41.

We'll spend many lectures talking about it.