Created with Coconote - [Link]

app

Environmental economics - Week 9 lecture

Good morning, everybody. Same faces. Okay. So how are you today? Okay.

At least it's not raining. Okay. So, today, what I'm gonna do is finish off
looking at the comparative analysis of instruments, and then I'm going to
start actually, in preparation for tomorrow for the seminar tomorrow, I'm
going to start the environmental the decision making under, irreversibility,
which is actually the last topic in the line. And then next week, I'm gonna do
stock pollution, and then I'm gonna go back to look at unobservability. So
looking at controlling, polluters' actions ex ante.

K. Sorry. Expose because ex ante is when we apply the tax, and then they
undertake the appropriate action. So ex post, so liability. Okay.

And tomorrow, for the seminars, you've got seminar 9, which is the final one
in the list. And then next week, we'll look at the questions from the
coursework, and we can discuss then if you have any questions about the
policy brief. K. Okay. So, any questions from last week?

And please do, I haven't met seen hardly anyone, I'd say, maybe 1 person
over the time in terms of office hours. If you have any questions in terms of
the lecture material, please or maybe 1 or 2, please drop by and ask. Okay.
Okay. So just to recap, let me just start just to recap the points, the key
points here.

First of all, I make a point that irrespective of these criteria that we're going
to use to, decide whether one approach is better than another, The political
the policymakers or the political environment or whatever it is has tended to
prefer to control quantitatively, so pollution quantitatively rather than
through prices. So policymakers tend to use command and control, which is
like European directives, okay, or standards rather than prices to control
pollution. So irrespective of of, you know, these pros that we go through so
the advantages of using a price instrument, they've tended to go down the
route of using quantity control. And maybe some of what we look at,
particularly the last one that I left the last day, might explain why. So from
our perspective, I would say, as environmental economists, I would say
probably the main criterion or the key one that but it depends on the
context.

Right? But if we were just to pick out one important criterion, we would want
to be cost effective because we're in the business of maximizing welfare, so
we don't wanna waste resources. So we want to, control pollution at least
cost, and market based instruments achieve that. Okay. And the key thing
that I showed you the last day is that even without knowing the cost
structure of firms, we can use a tax, for example, or we can use a permit
scheme, and we can still achieve cost effectiveness.

Whereas when we use standards, if the cost not knowing the cost structure
in advance, it's unlikely that standards will be cost effective. And remember
what the definition of cost effectiveness is. It's equalization of marginal
costs. K? So taxes and permits achieve that, whereas standards are unlikely
to do so.

And I also showed that it does hold when we have nonuniformly mixing
pollution as well. So another, criterion that we might want is that over time,
we perhaps we would think that it would be a good thing if pollution fell over
time. K? Or it might be a good thing if we can incentivize, polluters to
innovate in pollution control technology. So, again, when you charge
polluters, when you make them pay for the pollution that they produce, it
they may have the incentive to innovate.

And through that innovation, pollution control because innovation here is

reducing the marginal abatement costs. And remember, the optimal level of
pollution is a comparison of the costs and the benefits of pollution control.
So a comparison of environmental damages, for example, and the benefits
of pollution control, okay, or the cost of pollution control and the
environmental cost. But if we reduce emissions, it's a ban environmental
benefit, so we're weighing the cost and the benefits. If we reduce the costs
on one side, then the optimal level of pollution can fall over time.

And this is the diagram that we used to explain. So here, MC1 is the marginal
cost curve before innovation. MC 2 is the marginal cost after innovation. A is
the initial point. We're talking about a per unit tax here.
So the tax payment here at a is t times e star. K? Then if they innovate
before the policymaker reoptimizes, so the socially optimal level, once
they've reoptimized will be c, which is e double star. But before that, what
do we know? We know that polluters always go to the point at which the
price instrument equals the marginal abatement cost.

K. So the true marginal abatement cost after innovation here is MC 2. So

they'll equate t equal to MC 2 and go to point b or e hat, and then the
policymaker will be optimized. And you can see here, there'll be tax payment
before innovation and tax payment after innovation, and the difference is the
savings in terms of tax payment. And they'll compare those savings to the
cost of innovation.

And if it's not a one off expenditure, if it's expenditure over time or the
benefits occur over time or the costs occur over time, then we have to talk
about present value. So if the present value of savings, of tax savings,
exceeds the present value of the cost of innovation, then they will innovate.
And you can see that the optimal level of pollution will go from e star to e
double star. So over time, the optimal level of pollution will fall. And that's
what we mean by saying that taxes or permits are dynamically efficient.

The optimal level of pollution will fall over time. I used a 1st share diagram
just to explain that subsidies don't are unlikely to share this, characteristic.
So when we introduce subsidies and remember, this is these are not margin
these n c's here are not marginal abatement cost curves. K? These are your
usual costs of output, okay, which you would have had in your, because you
can see that they're upward sloping as y increases.

Okay. The MC is upward sloping. Okay. And the MC and the AC here are
before the subsidy, and the MC prime and the AC prime are after the
subsidy. And we can see here that a subsidy increases marginal costs but
reduces average costs.

So reducing average costs is probably intuitive, and why it increases

marginal costs is that in the presence of a subsidy, it's more expensive to
expand output because you forego the subsidy by expanding output. Okay?
And here you can see the individual level, y goes to y prime. Okay. So
initially, in equilibrium, we're at a.
Okay. So before the price has fallen due to the right hand side and
expansion in supply because firms have come into the industry. Okay. So
price will fall as a result of firms entering the industry. Before that happens,
when the subsidy has been implemented, so at a p is equal to ac which
means normal profits are being earned which is 0.

When p is above ac, supernormal profits are being earned. So the AC has
fallen, but the p, before it adjusts, p exceeds average costs, the market
price. So their supernormal profit's being earned. Entry is encouraged. The
shift in the supply curve will push down this the price until p is equal to the
new AC.

P is equal to ac prime, at which point there's no further entry into the

industry. So at the individual level, we can say the subsidy encourages a
reduction in output, and output is linked to emissions. Whereas on the
industry side, if you have entry, you can have an increase in aggregate
emissions. So overall, aggregate emissions could increase. And and looking
at this, it's saying yes.

They can they will increase. So subsidies are dynamically inefficient if we're
linking output and emissions. Environmental effectiveness, right, is where ex
ante, I think that there should be a level of pollution. If I introduce an
instrument, tax, okay, and depending on what I think the cost curves are, for
example, that says to me what level of emissions will be forthcoming, and
then I expect, right, I expect that level of emissions. And then aft after the
regulation is introduced to tax, that's actually the level of emissions that is
realised.

That's what environmental effectiveness is. So what I expect in terms of

emissions is what is actually realised. K? And when there's uncertainty with
respect to marginal date and costs, taxes if I lose taxes, they're not
environmentally effective. And that's we're gonna explore that a bit more
today.

Okay. So what I mean by not environmentally effective and environmentally

effective? So if I'm looking at, this oh, K. This hasn't let's hold on for a
second. It's not.

Okay. Okay. So here, if I use the tax right? So here, this is the marginal sorry.
This is the expected marginal cost.

This is the true marginal cost. So here, the policymaker has overestimated
the marginal costs of abatement. And, also, here, you can think of as I said,
you can either think of a single firm in the industry, only 1 firm, or you think
of aggregate. So this is aggregate marginal costs. K.

So the policymaker has overestimated the marginal abatement costs and

bases regulation on the basis of their expectations. So they set t equal to
the marginal damage at what they think is the socially optimal level of
pollution, which is eexp here. But what do firms do? They react to the tax
along their true marginal abatement cost. So they're going to produce
eactual.

So the discrepancy or the difference between Eactual and Eexpected says

that taxes here in this case are not environmentally effective because
Eactual is not equal to e EXP. Whereas if they use a quantity control, right,
so this would be a quantity control, which is either permits. So distributing
the number of permits equal to what they think is the socially optimal level
of pollution or placing a standard there so the overall amount of pollution is
fixed at EXP, standards are environmentally effective because eactual will
equal EXP be assuming that they don't pollute more than the permits that
they hold, and they comply with the standard. So quantity control is
environmentally effective under uncertainty, whereas a tax may not be, and
it isn't when the uncertainty relates to marginal pigment costs. K.

And that's what we went through that. So other issues to think about, for
example, if there's high inflation, a tax the real tax will be different to the
nominal tax, so you have to adjust taxes for inflation. Whereas if we use
permits, the price of the permit will tend to rise with inflation. K. And this is
where we left off the last day.

So this explains, to some degree, why policy makers use standards rather
than taxes. So with a tax, there's a double burden on firms. So there's a
resource cost in terms of reducing abatement, and then there's a tax
payment, which is the other, burden. So there's 2 burdens on firms when
they pay taxes or pay a charge on emissions, whereas there's only one
burden on them when they face a quantity control, and that's the resource
cost. Obviously, they're not paying any charge.
And remember, firms are going to lobby policymakers. Right? So they're
going to resist taxes. And policymakers tend to want to get firms on board
because they want to increase the likelihood of compliance. Okay?

Because remember, you can always try to deviate and particularly if your
actions are unobservable. K. So firms don't like taxes for this reason, and
then policymakers tend to adopt the approach that they think may be more
successful in terming in terms of controlling their actions. And then the
other thing is I I left off the last day was that firms if you use a uniform
standard and when standards are used, they tend to be uniform. So policy
makers do not tend to use differentiated standards.

So they're facing different firms. They don't tend to use differentiated

standards because they're too difficult to implement. They tend to use
uniform standards, and firms perceive uniform standards to be fair. Okay?
Even though they're not efficient if they're face if they're different in terms
of their cost structure.

So they perceive them to be fair. And, again, they may be more likely to
comply if they think the approach is fair. Okay. So the arguments for market
based are that they're cost effective. They can incentivize innovation and
therefore and thereby reduce the optimal level of pollution over time.

They comply with the polluter pays principle, which is the, underlying
principle of all environmental policy is that the polluter should pay. Okay? If
they produce the pollution and create the externality, they should pay for it.
And that's coming from the OECD in the 19 seventies, and it underpins UK
environmental policy and European environmental policy and policy, in fact,
all over the world. K?

So that's a principle coming from the OECD in the 19 seventies. And as we

saw in our seminar, the subsidy doesn't comply with this. The subsidy does
not comply with polluter pays. So if you thought that that was a really
important principle to comply with, you wouldn't choose subsidies. In the
case of permits, as I've said, price reflects the market, and so is potentially
automatically indexed for inflations, which means you have to intervene less
frequently into the market.
And you don't really want to be intervening as a government too much into
the market, Although you do have to make sure that it, minimize market
abuse. And the arguments against using market based instruments are the
existence of uncertainty. And I've as I've shown, if the uncertainty relates to
marginal costs, taxes will be environmentally ineffective. And then the
question is, why might this matter? So why does it matter if EXP is different
to eactual?

Okay. I'll leave that with you and come back. The long run impact of
subsidies is that it might actually encourage more pollution over time
because it encourages entry into the industry. As I said, subsidies are
against the polluter pays principle. You can have erosion of the incentive
effect of taxes and potentially subsidies if you use subsidies.

And as I said, it depends on the institutional framework of the country in

which you are If the tax collection system or the subsidy disbursement
system is not very well developed, then you might not want to use that. You
might want to use standards rather than prices. Okay. So comparing taxes to
standards under uncertainty. And I did this now in, the seminar.

So, I'm just going to explain this with the diagram, and then you can I've
written it all here, so I'm not gonna go through this here. But it'll be the same
thing. So just I'll highlight. So when you're comparing taxes versus
standards, you also have to identify the optimal level of pollution. And the
optimal level of pollution will be the intersection, in this case, of there's a
single MD curve.

So there's no uncertainty with respect to marginal damages here, and the

only uncertainty is with respect to marginal costs. Of course, I could it's
more likely that there'll be uncertainty with respect to both. And the
policymaker won't know damages, and they won't know costs. But then it
gets very complicated in the diagram. K?

So I'm I tend to use uncertainty with respect to one aspect at one time. So if
I'm comparing taxes here versus standards, I need to identify the optimal
level of pollution, which is the intersection of MD and the true marginal cost.
So this is the true marginal cost. So here now, what's happening is the
reverse of what we had before. The policymaker is underestimating the
marginal abatement costs.
Okay. The cleanup costs. They're underestimating how expensive it is for
the firms to clean up pollution. Okay. So what does that mean?

So as I said, here, first of all, we can identify. So you go to the point of

intersection here, and this is the optimal level. If they use a tax, remember,
they base the tax or the standard on their expected marginal cost. So
they're going to set the tax equal to this point here. Right?

So they're basing it on MCE. So this is going to be the expected level of

pollution using a tax. And if they use a standard, again, they base it on MCE,
so this is the standard. And this will be the expected level of pollution if they
use a standard. When they use a tax, we know, which I've already said, the
firms go they react to the tax along their true marginal abatement cost.

So you they're going to go to this point here. So with the tax, you're going to
get eactual. That's with the tax. Right? But with the standard, eactual is
equal to EXP.

So now there's a divergence between the actual level of emissions that you
get when you use a standard versus a tax. Okay? So the amount of
emissions that you get when you use a tax here will be higher than what you
expected, which will mean will mean there'll be more damages associated
with those greater level of emissions. And now to compare. Right?

So remember, in the absence of any intervention, I'm going to be here.

Okay? In the absence of any intervention, I'm going to be here. I minimize
my costs. I'm talking about the true costs.

This is the the firm's. Okay. So what are the total marginal sorry. The total
costs, right, under the tax? So which actually will do the efficient.

So if there was full information, right, the total abatement cost here that the
the policymaker would expect would be here. So you just go from the Emax
up to the Estar, and then the damages would be here, the total damages. So
that's under efficiency. The socially optimal level of pollution, the total
abatement cost will equal the triangle on the right hand side between E max
and E star under the true marginal cost curve, and the total damages will be
from 0 to E star under the marginal damage curve. And then you compare
the 2 instruments to this.

So for the, if they use the tax, k. You just go up to eactual for the tax, and
you go up to eactual under the damage, marginal damage. So the excess so
so the loss or the excess cost over and above what's efficient is that
triangle. So from using taxes, that's the social welfare loss associated with
using taxes. K.

And if I do it for the standard okay. So let me look at the so for the standard,
it's going to be up to eactual for the standard and under the marginal cost
true marginal cost between Emax and Eactual for the standard, which is
equal to EXP. So this is a social welfare loss from using the standard, and
this is a social welfare loss from using the tax. So in this case, if I use a tax, I
have a much greater social welfare loss, which would mean as a
policymaker, I would prefer to use a standard in this case to control
pollution. Okay?

And in the seminar, what we did was that there was uncertainty with respect
to marginal damages, not with respect to marginal abatement costs. And in
that case, the policymaker is indifferent because you get the same social
welfare loss attached to taxes and the standard because EXP is equal to
eactual for both instruments. Whereas in this case, there's a divergence
between eactual for taxes and eactual for the permit or for this or for the
standard. Now what determines this? Right?

What determines whether the social welfare loss is bigger or smaller for
taxes versus quantity control? What determines it is so let me just draw.
Let's go back. And I'm gonna do it just with, let me see. Yeah.

So that might be easier. Let's say this is MD. I'm just going to look at a tox
here. Brilliant. Okay.

Probably, this is not because I have to I want to look at the difference

between the 2. I have to do it. I'm just trying to make it. K. And the I have to
do for both.

K. M c true, m c e. Okay. So this is the social welfare associated with the tax,
and this is with the quantity control. So what determines whether a, b, c, d,
e?
So what determines whether a, b, c is greater than or less than b, d, e, do
you think? It'll be the relative slopes of the 2 curves. K? So if I make MD
flatter relative so, okay, so MD becomes flatter. You can see that what was
ABC becomes a smaller, triangle.

And what becomes and this will be the red one will be the new social welfare
loss attached to using the standard. K. If you shift so I'm just doing it in
terms of MD because it's easier. So if you can visualize, flattening the MD
curve means that the social welfare loss attached to using standards
becomes bigger, and that with respect to using taxes or price control
becomes smaller. So, intuitively, what's behind this?

And this relates to environmental effectiveness. If I have uncertainty with

respect to costs, the steeper the marginal damage curve, the lower the
social welfare loss attached to using the quantity control relative to using
the price control? Why do you think that is intuitively? The steeper the
marginal damages. I would want so the more toxic the pollutant or
hazardous the pollutant, I would want to be more environmentally effective,
wouldn't I?

I would want to be sure of the amount by which I control the pollution, the
more dangerous the pollutant. Here, we're comparing the damages cost
versus the cleanup costs. Right? And there's a trade off. So if it's highly
damaging the pollutant, which means a steep MD curve I could characterize
it as a steep MD.

So marginal damages are rising more than marginal abatement costs for
each unit of pollution. Okay. So there's costs are rising more on the damage
side rather than on the cleanup cost on the cleanup side. And I want to
minimize. Right?

That's what I'm minimizing total cost. That's what I'd like to do. I would
prefer if it's highly damaging, I would prefer to be more sure about the
amount of emissions that's realized after I introduce the instrument, which is
the standard. So it's more important to be environmentally effective if
marginal damages are rising faster than marginal cost of cleanup, and that
relates to how damaging the pollution is. If marginal costs are rising faster,
so the steeper the marginal cost curve is relative to marginal damages, then
cleanup costs are rising faster, and I would prefer to use the tax.

Because, remember, what does the tax do? Which standards don't? Taxes
are cost effective. They equalize marginal abatement costs. So they're cost
so I would prefer if cleanup costs are rising cleanup, I put more emphasis on
the environmental embark the environmental I put more emphasis on the
environmental the environmental effectiveness attribute of the quantity
control versus the lack of cost effectiveness attribute of quantity control.

Okay. So the more damaging the pollutant, let's stick with the MT curve, and
then I would choose a quantity control. And then, obviously, the corollary of
that is what if the slopes are equal. So if the relative slopes are equal, the
social welfare loss attached to using a tax is the same as the social welfare
loss to using a quantity control. So when the relative slopes are equal, that
means the policymaker is indifferent between using either a price control or
a quantity control.

And that's Weitzman's theorem. And that's so in the seminar, we the

question was 3 cases where the policymaker would be indifferent between
using a price control and a quantity control, taxes versus standards. Okay.
So full information, let's say, was the first case under certainty. So you knew
all the costs, and you knew all the damages.

You knew the king of costs. You knew the damages. So under under
certainty, you would be indifferent because the say the actual level of
emissions would be the same in each case. The second case was under
uncertainty, but that uncertainty relates to the marginal damage curve only.
Then the social welfare loss attached to both will be the same, so the
policymaker is indifferent.

And the final one is where the uncertainty relates to marginal cost of
cleanup, but the relative slopes are the same. So in those three cases, the
policymaker is indifferent. Okay. Any questions? So the I've written it all
there on the slides, and that's what I said.

And that's okay. And that's when the relative slopes are equal. Any
questions? So yeah. You draw it and see and have a look what you think.

So draw it first and then come back to me. Okay. So the point here is right?
So I've said to you at the opening, right, we'll always use a price instrument
as environmental economists. But, actually, it's not necessarily true.

Under uncertainty, there may be an argument for using a quantity control

over a price instrument because marginal damages are relatively steep. K?
So, you know, depending on the context, we don't necessarily always want
to opt for price control. So the cost effectiveness criterion doesn't always
dominate. K.

So if there are no questions any questions? Yeah. Say that again? Do you
have people with the Yes. You can.

Yeah. The way I've done it is purely diagrammatically, not the way but but
what you did in the seminar, you did it just looking at the objective function,
costs, and benefits plugging in. And then it's a bit more complex. Well, not
really not actually analytically, probably. We just see exactly that approach
now.

Yeah. It could be. And and if even if I did it nonlinearly linearly

diagrammatically. Okay. It might make it a bit more in terms of relative slope.

So I've tended to do linear, to be honest, in terms of representing it

graphically. Yeah. But, of course, you can do it analytically. And if it's
nonlinear, it's it's it's no harder. Okay.

To be honest, whether the marginal, yeah, marginal damage or the marginal

abatement cost is linear or nonlinear, I would say, is immaterial in the way
I've done this. So it's just easier to do it linearly, graphically. Okay. So
environmental decision making under uncertainty and irreversibility. So this
will also relate to, in some way, the analytical approach to determining the
social welfare loss.

So we're gonna just do exactly. So for example, in the seminar that we had
last week, when I asked you to estimate the social welfare loss, I did it
graphically. Right? So estimating that area under the curve. Whereas, what
you could have done was just plug in the re the relevant emission levels into
the damages and the costs.

Okay? So, it's difficult to explain that. Just let me so just let me go back for 1
minute to explain. So when we got here so one other way, I've done
graphically. So here, they're linear.

Right? So you can just estimate that, the area of the triangle, half the base
multiplied by the perpendicular height. When they were on nonlinear, we
have to integrate. Alternatively, you could so if you are so you're minimizing
damages plus costs. Right?

And so you would find out what total cost so that's equal to your total costs.
So you would plug in eactual for that, and you would get total cost
associated with emissions equal to eactual, which is under the tax. And,
alternatively, you would minimize d plus c, which are equal to total costs,
and you would plug in EXP here, which would give you for the standard.
Okay? And then you could take one from the other to see the difference
between the 2.

Okay? So if the total costs were greater so if the tote so this is under the tax.
This one here. And this is under the standard, this one here. If the total costs
using the tax was greater than the total cost using the standard, then you
would opt for the standard.

So this is an alternative way of deciding which instrument you should could

you should go for. So in that way, rather than doing it graphically, you can
also go back to the objective functions and plug in the relevant level of
emissions and find out what the total costs are and compare. Okay? And
we're gonna use that approach here. Obviously, it's gonna be in a different
context.

Okay. So if you remember way back to the second lecture, we looked at the
potential for irreversibility. Like, how would that affect decision making and
uncertainty? Now uncertainty here can be defined in different ways, and
we're going to define it. So risk, okay, is where you can determine the
probabilities of different outcomes.

So for example, when we have the toss the coin tossing example in micro.
Right? So an unbiased coin, 5050. Okay? I can determine the probabilities.

And if it was deciding a a gamble, whether I would enter into the gamble or
not, I would know that if it was heads, I'd get something. And if it was tails,
I'd get something else. So I can determine the outcomes, and I can
determine the probability. So that's a problem with respect to risk. So it's
risk characterizes that problem.

But I may not be able to to determine the, probabilities or even the

consequences. And you could think of that as radical uncertainty. So I don't
know the outcomes, and I don't know the probabilities. So in this case, my
type of uncertainty is actually risk. Okay?

So I'm going to assume I can determine the outcomes, and I can determine
the probabilities. But I still have uncertainty because, yeah, there's a chance
I could end up in one state of nature versus a a different state of nature. And
the irreversibility here is that if I undertake a particular course of action, then
I cannot undo the outcome from that course of action. So if I drain a wetland
to construct housing, I can't undo that within any reasonable human time
scale. I can't return the wetland to its original level, original state.

K. So that's the irreversibility. And a lot of decisions that we make within the
environmental arena have irreversible consequences. And then the question
is and if I was just to use a standard cost benefit analysis without taking into
account this potential for irreversible actions, so traditional cost benefit
analysis assumes that actions are reversible. So I don't take into account
this potential irreversibility.

I underestimate the benefit of the irreversible action, which is the

development. I over sorry. I overestimate the benefit of the development, or I
underestimate the cost of not preserving. Okay? So let me say that again.

So if I don't so in the traditional cost benefit analysis, let's say bt minus ct,
okay, multiplied by the discount factor or divided by 1 plus r, t equal to 0 to
n. Okay. So in this case, if, let's say, I'm talking about preservation. So in the
case that I looked at, preservation of a wetland versus developmental
housing. Okay.

So you could think here that would be the cost. So it would be the foregone,
the loss of biodiversity, the loss of ecosystem services. Okay. So I
underestimate, right, the cost of development, which is the loss of
biodiversity and ecosystem services, or I overestimate the benefit of
development. So if I underestimate the cost or I overestimate the benefit,
and remember, this is this is equal to net present value, it means that the
likelihood of net present value being greater than 0, which is the decision to
go ahead with the investment, is more likely to be positive.

So if I don't account for irreversibility of actions, the net present value of

development will more likely be positive, which means I'll be more likely to
go ahead with the development. Will more likely be positive, which means I'll
be more likely to go ahead with the development. So traditional cost benefit
analysis does not account for this potential irreversibility issue. So I need to
do I need to add in something else. Right?

And that adding in something else, right, to account for, let's say, the loss of
resources. So, again, as I said, if I develop something, I lose that resource
forever. Right? But and we can think here. What is the benefit of the
resource?

Okay. So what is the loss of benefits? Okay. I can characterize the value of
resource as its use value. So the use I get from the resource of the
ecosystem services of what it supplies to me.

Okay? So in terms of flood flood regulation, filtering of water, etcetera. The

nonuse value, for example, the biodiversity that it supports. So I believe in
that the biodiversity should exist independent of what use I will derive from
it. And the option value, and this is where this is relevant in terms of
irreversibility.

The option value is the value I put on a resource when I don't know so I'm
not using the resource now, and I don't know what my preferences will be in
the future. But I would be willing to pay some amount to keep that resource
available if for the future for me to use if I so want it. Okay? So we, as
individuals, might be prepared to pay something to keep a resource in
existence in case we would want to use that resource in the future, and
that's called an option value. Okay.

So that's another type of value, and we should be we should, in some way,

try to incorporate that into our decision making that we would actually be
willing to pay something to preserve a resource because if we and that will
account. So the benefits of preservation should be adjusted upwards by
this, option value. Because if it's not preserved, it'll be lost forever. Okay. So
that's the type of uncertainty I'm looking at.

I'm looking at it in terms of risk. I, yeah, can assign probabilities, and I know
the consequences, which is, you know, probably a little bit it makes things
easier analytically, but not maybe not so characteristic of environmental
issues that we face. We don't always know the outcome, and we find it very
hard to put probabilities, but we may look historically to try and inform the
types of probabilities we might put on different algorithms. And
irreversibility is so this is where the precautionary approach comes in. If If an
action has irreversible consequences, maybe I should adopt a precautionary
approach.

Okay. So I'm gonna leave it there for and we'll take a break, and we'll start at
the courthouse. And if you have any, questions in the meantime, please ask.
Okay. So what I go through here, you do these steps for the seminar
tomorrow.

So you have 2 questions based on this topic tomorrow. So, different

functional forms, but, essentially, it's the same approach for both of the
questions. Okay. So to understand the effect of irreversibility on decision
making, we're going to compare 3 scenarios. So know that we're gonna
assume first no irreversibility so that actions are reversible.

2, that there is irreversibility, but it's ignored by the decision maker. And 3,
there is irreversibility, but it's accounted for by the decision maker and just
go through what each of these scenarios would look like and what, results
they give. So the assumptions we're gonna make is that you can think of, as
I said, an area of land, hectare of land. In this case, it's forest. So a one is the
area of forest conserved in period 1.

So the subscript denotes the period of time, so we only have 2 periods of

time here. So we need the future here because we need to have uncertainty.
Okay. So we know what happens today, but we don't know what happens in
the future. And we're going so and the thing is if you reverse the
irreversibility here is the development.

So if I develop, if I develop the hectare, that means I can't conserve it. Okay.
So it's a zero sum game. So they're opposite to each other. Conserve,
develop.
So whatever I develop, I cannot conserve. Okay. So if I develop half of the
hectare, I conserve half of the hectare. The benefit of conservation, right, in
period 1 is given by b 1, and the benefit of conservation in period 2 is given
by b 2. And then I just give it functional forms.

So this is to give me a downward sloping marginal benefit function. So

benefits of conservation increase, but the marginal benefits so the marginal
benefit is positive, but it decreases. K? And the only difference between the
two is this k. Okay?

So, obviously, the benefit of a one a one is here. That's obviously going to
be the difference. And the benefit of a 2 I'm gonna have a 2 here. But in
terms of functional form, the only difference is this, and k is greater than 1.
So here I'm saying the benefits.

So if a 1 is equal to a 2, the benefits are greater in the 2nd period versus the
1st period. So for the same level of conservation, the benefits are greater
when k is greater than 1. So that just gives me a higher intercept. But the
slope of the marginal benefit curve, which will be beta, is the same in both
cases. Okay?

So, remember, these are total benefits, but the derivative is giving me the
marginal. And I'll draw this. When I when I do this, at the end, I'll draw it.
Okay. So with case 1 is no irreversibility.

So this just means that if I conserve, an amount in period 1, I can conserve

more in period 2. Okay? Because remember, whatever I don't conserve is
development. K? So I can conserve more in period 2 if there is no
irreversibility.

If there is irreversibility, I can never conserve more than I conserve I can

never conserve more in period 2 than I did in period 1 if there is
irreversibility. So it's capped. In period 2, my level of conservation will be
capped by whatever I did in period 1 if there is irreversibility. If there's no
irreversibility, I can conserve more in period 2 than I did in period 1 because I
can undo the development. K.

So in period 1, so no irreversibility. That just means that I can look at 2 of the

both periods in terms of decision making independently. So I in period 1, I
just maximize my total benefits. And then when I get to period 2, I maximize
my total benefits. So I choose a 1 and a 2.

Okay? So and it's a I mean, the math here is extremely simple. Okay. So I just
different and here, actually, what you sent me to to differ from what I asked
you to do tomorrow on, interpreting either 2 things. I can interpret this as
net benefits, so the the costs are implicit, or I could assume this I could
assume that this is benefits and costs are 0.

Okay. So it's capturing both scenarios. So tomorrow in one of the questions,

I've got benefit curve and a cost curve. Okay. So we can assume for that
simplicity that this is net benefits of conservation in period 1 and net
benefits of conservation in the next period, net benefits of conservation in
period 2.

Okay. So I just to find the optimal level of conservation, I just differentiate

with respect to a 1, and I get a 1 is equal to alpha over beta. And then no
surprise that when I differentiate with respect to a 2, I get the optimal level
of conservation of a 2 is k alpha over beta. So given that k is greater than 1
means that a 2 is greater than a 1, which is permitted because there's no
irreversibility. In case 2, there is irreversibility.

So this is like a physical limit. This is like, as I said, if I develop, I can't undo
development. Development has irreversible consequences. So I can't
decrease the amount of land developed. So I can't return that wetland to its
pristine state.

Whatever I've developed, whatever portion I've developed, I can't return it to

its pristine state. So here, I take again because in this case, you can think of
it as being myopic. Right? The policymaker is not looking ahead. K?

So they maximize their net benefits in period 1. They choose the optimal
level of pollution optimal level of conservation, and we don't even have to do
the maximization for period 2. But we'll do it anyway because we already
know what a 2 will be. Okay? How do we know what a 2 will be?

Without doing any maximization, how do we know? Or what will a to b and

y? Okay. If I draw okay. And then you tell me.
So this is period 1, marginal abatement. K. The intercept is alpha, and the
slope is beta. And this is a 1. And then period 2.

The intercept is k alpha. Okay. This is a2. This and the again, the slope is the
same, so they're parallel. And so this is b1 prime, and this is b2 prime.

So in the first case that I've just done, this would be my alpha over beta, and
this is my k alpha over beta. And so that that was case 1, and there's no
irreversibility. And now I'm asking you, given that for case 2, I've got the
optimal level of conservation in period 1 is alpha over beta, how do I know
what the optimal level of conservation in period 2 is and why? You know?
Exactly.

So given that the marginal benefit is greater here, I'd like to be here, 5 k
alpha over beta. So I I I want to do as much conservation as I can up to this
point. Right? So that just means that given alpha over beta is less than k
alpha over beta, I want to do as much as I can up to k alpha over beta, but
I'm constrained by purely one conservation. So the socially optimal level of a
bay of conservation in period 2 is alpha over beta.

So without doing any, maximization, you already know that it's alpha over
beta. So in case 2, a 1 is equal to a 2 is equal to alpha over beta, But you can
do it if you like. K. So we can just here so I'm maximizing benefit 2 subject to
the irreversibility constraint. So a 2 less than or equal to a 1 is the
irreversibility constraint, and it's going to be binding.

So a 2 will equal a 1. K. And what about case 3? So in case 3, I'm long

sighted. So I'm not myopic as a decision maker, but I still face the
irreversibility constraint.

So long sightedness just means that I'm gonna maximize total benefits over
time. Okay? So I maximize b 1 plus b 2 subject to the irreversibility
constraint. So I set up the Lagrangian. Okay.

So this is the Lagrange multiplier, and this is the constraint. And the
constraint will be binding. So I already know that a one will equal a 2. Okay.
And then I differentiate with respect to a one and then with respect to a 2.
Okay. So this is the FOC with respect to a 1, and this is the FOC with respect
to a 2, and this is the FOC with respect to lambda. K. And then taking this, I
know that a 1 is equal to a 2, so I can plug it in here. So then all I've done in
the this is that I've rewritten.

So this is 1, 2, and 3. So given 3, a one is equal to a 2. This is going to be 2

prime. I plugged in for a 1 into a 2, so it's a 2, and this is just 3 repeated. And
then what I do is just eliminate the lambda and solve for a 2.

And I get a 2 is equal to a 1 is equal to alpha times k+one over 2 beta. Okay.
So if I wanted to draw this, I would need to know how does this compare to
alpha over beta and k alpha over beta. Okay. And that's what I do here.

And then I go back to the drawing. So all I'm doing here is I want to know
where to put the optimal level of conservation when I'm not myopic and
there is irreversibility. So in that diagram, I need to know where to put it. K.
So I need to know whether this amount under case 3 so this is case 3, and
this is case 2.

So I'm taking the fact that there is irreversibility, and I'm comparing the two
cases where I have myopia and where I don't have myopia. So where I'm
taking irreversibility into account and where I'm not in terms of, or let me put
it that where I just leave it as when I in both cases, there is irreversibility.
Okay? And in the first case, on the left hand side, I don't take alongside of
you, and on the right hand side, I do. So I maximize total benefits on the
right hand side over time.

And on the left, I'm just maximizing benefits at each point in time when I get
to that period. So to compare then, you just is it greater, equal to, or less
than? So all you do is, get rid of all the common terms, which is what I've
done here in the second line, and then I just cross multiply and then just
bring the numbers together. So I'm bringing the one over here. And I know
by assumption that k is greater than 1, which implies that the right hand side
is greater than the left hand side.

So the optimal level of conservation under case 3 is greater than the optimal
level of conservation under case 2. So when I go back to the diagram, what
does that mean? It means that so if I want to know actually so you can tell
me can you just do it now and compare? Is it, less than or greater than k
alpha over beta? So I want to know can you just check?

So instead of alpha over beta, is alph is alpha k+one over 2 beta greater than
equal to or less than k alpha over beta? Can you just check that for me now?
Doing exactly what I did. Are in any way that you your approach. Okay.

I'm gonna assume that it's less, but you have to check yourselves. So just
for the purpose of the drawing. Okay. So what that means is that case 3 is
going to be here. So a 1 for case 3 so this is a 1.

Case 3, a 2 case 3, and this is a one case 2, and this is a 2 case 2. K. So I'm
comparing case 2 and case 3. Okay. So what's happening and this let's say
that we extend the marginal benefit curve below the x axis.

When I take irreversibility into account and I what that means is I'm taking
the long view. So I'm maximizing total benefits over time. I'm willing to
conserve more in period 1 at a cost. So there's a cost here. So the benefit is
negative.

Right? So I'm willing to conserve more in period 2 at a cost so that I can

conserve more in period. So so I say this. I'm willing to conserve more in
period 1 at a cost so that I can conserve more in period 2. And overall, I'll be
better off.

So what how how much better off am I? If it's a 3 versus a 2, this is the extra
total benefit I get from conserving a 2 under case 3 versus a 2 under case 2.
And, diagrammatically, you can see as well that you could also check this
analytically, which we will do. But, diagrammatically, you can see that the
total benefit is greater than the total cost. So the net benefit is positive.

So by conserving more in period overall, over both periods, I'm better off.
Okay. By conserving more in period 1, it leaves open the option, right, to
conserve more in period 2 relative to case 2. And by doing so, my overall net
benefit will go up. Okay.

So case 3, I'm saying the total benefit associated with case 3 remember,
costs are implicit here, or we're thinking of it in terms of net benefit. So if we
just think of it in terms of benefits, overall, the benefit in case case 3 is
greater than the benefit in case 2. So case 3 is better than case 2. So case 2
is suboptimal. So by being myopic, I'm actually, not maximizing welfare.

Okay. Okay. So I've just summarized there. So case 1 is where there is no

irreversibility. So I can I can conserve more in the future than I did today?

Case 2 is where I am myopic. So that's why I have IRM. So m stands for

myopia. IR stands for irreversibility. So that denotes the conservation under
irreversibility and myopia.

And then case 3 IRO, it's irreversibility and optimal optimal decision making
or optimality. K. And I've just compared these. Okay. So what is the effect
then?

And so I've shown you diagrammatically. So what is the cost of being

shortsighted? So that would be what would be if I was to do that? As I've
just said to you, you'd work out the benefits of case 3, work out the benefits
of case 2, and then the cost would be the difference between the 2, the cost
of, case 2 or the benefit of case 3. You'd be be would be associated with
higher benefits for case 3.

So what is the cost of being shortsighted? That's just the reduction in total
benefits of case 2 relative to case 3. So how would we answer this? So we
can look at it in terms of the diagram, but we can also look at it in terms of
total benefits. Okay.

So all we do is we plug in the relevant a a one and a 2 for case 2. So that was
alpha over beta. So you just plug in alpha over beta into the benefits for
period 1 and the benefits for period 2. So this is the benefit function for
period 1, and this is the benefit function for period 2. So remember, it was
alpha a 1 minus beta over 2 a 1 squared, and it's k alpha a 2 minus bet b
beta over 2 a 2 squared.

So all I'm doing is plugging in for a 1 and a 2 here. So in this case, plugging
in for b 1 and plugging in for b 2. Okay. So just let me explain. If you, it's not
that easy to look at straight away, but let me just ex how I simplify it.

So these two terms are the same. Okay? So minus a half minus a half gives
me 1, okay, which is this. Okay. So these 2 give me this term.
Okay? So it's a half beta times the square bracket squared minus a half beta
times the square bracket squared, which is, excuse me, minus beta times
the square bracket squared. And in terms of this one, the other term, and I'll
do exactly the same in the next one, is I'm just factoring out the common
term, which is a times alpha divided by beta. So it's I'm factoring out this,
and I get 1, and I get k here. So it's 1 plus k.

So I'm just factoring out the common term. And then, one of these betas will
cancel with that with that beta, and so I'm gonna end up so that cancels. So
this one will cancel with this, and I end up with k alpha squared over beta,
which is these 2. K. So that's the total benefits from case 2.

I do exactly the same for case 3, and it's the same approach. These two
terms are the same. Sorry. Not that one. These two terms are the same.

It's a half times minus a half times x minus a half times x, where x is the
what's the beta times what's in the bracket, which gives me one time so it
gives me x. Right? And then this, what I'm doing, is factoring out the
common term again. Okay? And I get k plus 1, and then I simplify these 2 to
get this.

So the total benefits from case 3 are that. And then, again, I can either
subtract 1 from the other, or I can see whether one is bigger than the other.
So in this case, I'm subtracting to see whether it's positive or negative, the
difference. Okay. So I assuming that case 3 optimal decision making is
better, I should get a positive, and that's what I get.

K. So in this, I'm just getting a common denominator. I'm combining these 2

two terms in one. And because even though k is greater than 1 because it's
1 minus k all to be squared, that's a positive. So the total benefits from not
ignoring irreversibility are higher, and the cost of ignoring irreversibility is the
difference between the 2.

K. So introducing then uncertainty. So suppose future benefits are

uncertain, how might this affect the decision to conserve? And it depends.
So intuitively, you might think I mean, and I'm just giving you a I'm just trying
to give you something that counteracts your intuition.

Okay? So we said already, you know, if we're adopting a precautionary

approach and we're uncertain, we would be very slow to undertake an
irreversible action because that's going to foreclose future options for us. K?
So but I'm just gonna show you an example where, actually, you're going to
conserve less when there's uncertainty with respect to future benefits. And,
if expected benefits are less in the future, then, obviously, you would
conserve less.

So this is just an outcome of how I characterize future benefits. Okay. So

remember I said I'm dealing with the situation of risk, so I know what the
probabilities are. So I've only got 2, situations, so 2 possible states of
nature. And those states of nature, the outcomes or the consequences of
those are the payoffs.

What I'm saying is that the future benefit will be b 2. So the b 2 function with
a probability of p, or I'm saying it will be the b 1 function with a probability of
1 minus p. But remember that the amount of conservation in the future is
called a 2. Alright. So to derive one of the benefit functions, I'm just so I'm
using the 2 benefit functions I have just for exam demonstration.

So I'm using the b one function, which was alpha a 1 minus beta over 2 alpha
1 squared sorry. A a one squared. But remember, I'm talking about the
future. So I'm using the b one function, but the a is a 2. Okay?

And I'm saying that that is equal to that, and there's a probability of 1 minus
p that they are the future benefits. And there's a probability of p that b 2 are
the future benefits. K. So here, I'm going to have expected, net benefits,
which are all of this. Sorry.

All of all of this. K. These are my benefits in period 1, which I'm assuming are
certain. K. So my future benefits are uncertain.

So I have expected. These are the expected benefits of a 2, and this is the
benefits of a 1. Okay. And this is the constraint, the irreversibility constraint,
and this is the Lagrange multiplier. Okay.

So differentiating with respect to Lambda, I get a one equal to a 2. So the

irreversibility constraint is binding. Differentiating with respect to a one, I get
this, and I'm going to use this, this, and this. So this is just differentiating
with respect to a 2. Okay.
So what I'm going to do is here, bring this to the right hand side. So it's
minus lambda is equal to alpha minus beta a 1. And then I'm going to use
this and write it as alpha minus beta a 2, And then I'm going to substitute it
in there. K. So just repeat.

So I'm using these 3 for stored conditions, and what I do is I rewrite this. So
this is 1, 2, and 3. Okay. So I rewrite 2, which I've called it here. I rewrite 2 in
terms of lambda, and then I replace a 1 with a 2 according to equation 1.

And then I substitute all of this into equation 3. And then I just, determine a
2. Okay. Okay. Solve for a 2.

And I get this expression here, and I know that a 2 is equal to a 1. Okay. And
then the question is, do I conserve more under uncertainty or less under
uncertainty? So this is what I conserved under certainty. This was case 3,
and this is case 3 prime, which is under uncertainty.

So this is uncertainty case 3 prime, and this is certainty. And I just do the
same approach as I did before. Cancel out all the common terms, break it
down, and, eventually, you will get to k is greater than 1 by assumption. So
you'll get 1. You'll end up with k is greater than, equal to, or less than 1.

And we know that by assumption, k is greater than 1. K. So the right hand

side so the level of conservation under case 3 exceed exceeds the level of
conservation under case 3 prime. So under certainty, I can serve more than
under uncertainty. And then why is that the case?

Let me go back here. Why is that the case? Why do I conserve more
uncertainty? Okay. So under certainty, I was maximizing b 1 plus b 2.

And under uncertainty, I'm maximizing b 1 plus expected benefits 2. So

these 2 are the same. So the only difference is this. So how does expected
benefits 2 compare to benefits 2? Expected benefits 2 is less than benefits
2.

Expected benefits are lower. Right? Because I'm combining it with the
benefit function that relates to period 1, which is lower than the benefit
function in period 2. So e b 2 is less than b 2. That that's what drives the
result.

Okay. And that's it. So so I've got 2 questions. Please attempt them because
it's exactly the same for tomorrow. Just attempt them and go through the
case 1, case 2, and case 3 for different functional forms and also draw them.

And going back to your question, one of them is nonlinear. Okay? One of the
cost functions, I think, is going to be nonlinear. Yeah. And, actually, the
benefit as well.

Yeah. Both maybe. I can't remember that. But so in drawing, it just keep that
it keep when you're drawing, think about what you're representing. Okay.

So that's it. If you have any questions, I'm happy to answer them or about
this or about anything preceding this. Or if you have any doubts at all about
any aspect of the course now, I can, answer.