energy, pollution control and economic growth

7/29/2019 Energy, Pollution Control and Economic Growth

1/28

Eeg, P C

Ecc G

Ross McKitrick

OccasionalPaper 52


2/28

Imprint:

Published by the Liberal InstituteFriedrich-Naumann-Stitung r die FreiheitReinhardtstrae 12D10117 Berlin

Phone: +49 30.28 87 78-35Fax: +49 30.28 87 [email protected]

COMDOK GmbHOce Berlin

First Edition 2012

I you wish to support our work:Commerzbank BerlinBIC 100 400 00Donations account: 266 9661 04

Donations receipts will be issued.


3/28

EnErGy, Pollution Controland EConomiC Growth

Ross McKitrick

Prepared or the Colloquium Energy: Friend or Foe? Energy Use, Economic

Development and the Environment organised by the Friedrich-Naumann-Foundation or Freedom (FNF), Truman House, Karl-Marx-Str. 2, 14482 Pots-dam, Germany.


4/28


5/28

Contents

1. Introductory summary 5

2. Energy as a Causal Factor 5

2.1 Engine o Liberation 5

2.2 Cause or Eect? 6

3. Energy and Pollution 7

3.1 Three Types o Air Contaminants 7

3.2 Prospects or decoupling rom economic growth 7

4. The Energy Mix 12

4.1 Price and Production Trends 12

5. Economic principles 14

5.1 Using the price system 14

5.2 Taking into account externalities 16

5.2.1 Emission pricing 16

5.2.2 Pricing versus quantity targets 16

6. Concluding remarks 23

7. Reerences 23

About the author 24


6/28

4


7/28

5

1. Introductory summary

Growth in energy consumption is not just a by-product o economic growth,

it is a causal actor. Attempts to constrain energy consumption can thereorebe expected to impede economic growth. Fossil-based energy has an economicadvantage but gives rise to pollution concerns. But conventional air pollutionis not a necessary by-product o power production and has been successullydecoupled rom it in western countries through development o scrubbers andother end-o-pipe treatments. Indirect air contaminants, such as ozone andaerosols, are not trending upwards and peak level episodes have allen, butthe complexities o the atmospheric processes in these cases make progressinherently slower. Carbon dioxide is not controllable by conventional scrubbers,instead it is tied linearly to energy consumption. CO

2emission reductions will

thereore be much more expensive to achieve under current technology. Eco-nomic reasoning provides a ew basic ideas or guiding energy and pollutionpolicy. Regulations should be ocused directly on the variable o interest, noton indirect market transactions. The principle o pricing states that the mostecient outcome o well-designed regulations will match that which wouldhave resulted rom using price instruments directly applied to the emissions oconcern. The pricing principle helps explain why some interventions, like the

subsidy-driven expansion o the renewables sector, are inecient.

2. Energy as a Causal Factor

2.1 Ege f le

Electricity and the spread o household appliances results in increased conve-nience, reduced indoor air pollution and gains in population health throughbetter ood handling and indoor climate control. The availability o inexpen-sive electricity has truly been a miraculous source o economic growth, socialprogress and equality in the west. Greenwood et al. (2005) showed that, inAmerican households, the prolieration o household appliances, made possibleby the spread o electricity, reduced required average housework time rom 58hours per week to 18 hours per week between 1900 and 1975. They ound that

adoption o durable goods in the home (major appliances like washing machinesand stoves) accounted or over hal the increase in emale participation in theUS labour orce over the 20th century. A similar nding can also be expected


8/28

6

or other countries,, namely that increased productivity in the home leads tomore time or participation in the paid workorce.

2.2 Ce Effec?

When examining the link between energy consumption and growth we needto address the question o whether increased energy consumption causes GDPgrowth, or is caused by GDP growth. The distinction is important. I increasedenergy consumption is merely a by-product o growth, it could potentially becapped and reduced without dampening economic growth; in other wordsgrowth and energy consumption could be decoupled. But i increased energyconsumption is an input to growth, the two cannot be easily decoupled andeorts to cap or limit growth o energy consumption may impede economicgrowth.

Detecting the direction o infuence between two co-trending variables is doneusing a time series method called vector autoregression (VAR). Econometri-cians reer to an empirical causality nding as Granger-causality, ater thepioneering econometrician Clive Granger. I knowing the value o one vari-able x at time t permits more precise orecasts o another variable y at timet+1, but not vice-versa, x is said to Granger-cause y, since it is a reasona-

ble inerence that the direction o infuence is only in the one direction. VARanalysis has been applied to US data (Stern 2000), Canadian data (Ghali andEl-Sakka 2004) and others. The results show that energy consumption causeseconomic growth, and in some cases the causality runs both ways. This indi-cates that energy availability is a constraint on economic growth, rather thanenergy consumption being an unnecessary by-product o economic growth.Stern (2000, p. 281) concludes as ollows:

The multivariate analysis shows that energy Granger-causes GDP eitherunidirectionally as indicated by the rst o the three models investigated orpossibly through a mutually causative relationship The results presentedin this paper, strengthen my previous conclusions that energy is a limitingactor in economic growth. Shocks to energy supply will tend to reduceoutput.

The phrase energy is a limiting actor in economic growth is an importantstatement o conclusions. Energy consumption is not merely a by-product that

can be decoupled rom GDP growth. Deliberately reducing energy consump-tion will likely reduce economic growth, thereby increasing the reluctance opolicy makers to attempt a reduction. For this reason we need to nd out in


9/28

7

which circumstances policy makers might be tempted to interere with energyconsumption. In the 1970s the typical explanation was a ear o energy shor-tages. This ear abated as supplies proved ample over the ollowing decadesand prices stayed low. Now the typical explanation is ear o pollution. So we

turn to an examination o air emissions.

3. Energy and Pollution

3.1 tee tpe f a C

It is helpul to distinguish between three types o air contaminants: direct,indirect and CO

2. Direct contaminants are emitted to the air by stationary or

mobile sources in the orm in which they persist as contaminants. Examplesinclude sulphur dioxide (SO

2), nitrogen oxides (NOx), particulate matter (PM),

and carbon monoxide (CO). Indirect contaminants are ormed in the air romother compounds. For instance, ground-level ozone (O

3) is not emitted direct-

ly, instead it is ormed as a chemical reaction under intense ultraviolet lightinvolving volatile organic compounds (VOCs) and NOx, which are themselves

emitted. Likewise ne particulates, or aerosols, are ormed rom precursor com-pounds such as SO

2and NO

2. CO

2was not historically considered a pollutant

because it is not a direct threat to human health. It has only become a ocuso regulatory interest as an agent involved in climate change.

3.2 Ppec f ecpg f ecc g

The three types o air contaminants can be distinguished based on the ease

with which they can be decoupled rom economic growth. Briefy, the dicul-ty ranges rom easy or the rst type to dicult or CO2. Emissions o direct

air contaminants like SO2

and particulates have been sharply reduced to theorder o 80 - 95% in western countries despite decades o concurrent growthin energy use and output. The availability o scrubbers and other end-o-stackemission controls, more ecient uel burning systems, higher quality uel sour-ces and similar technical innovations have been extremely eective in allowingwestern countries to increase their energy use while cutting emissions.

Plotting U.S. total particulate emissions against real income (Gross DomesticProduct per capita) over the period 19451998 shows the relationship is clearlydownward-sloping.


10/28

8

Fge 3.1: U.S. Real GDP per capita and total particulate emissions.

Source: McKitrick (2010 Fig. 1.7) using data rom the US Environmental Protection Agency.

On the other hand, American carbon monoxide (CO) emissions over the sameinterval show an upside-down-U shaped pattern (Fig.1.8).

Fge 3.2: U.S. Real GDP and total carbon monoxide emissions.

Source: McKitrick (2010 Figure 1.8) using data rom the US Environmental Protection Agency


11/28

9

Fge 3.3: U.S. Emissions by type, 19451998.

Source: McKitrick (2010 Figure 1.9)

In both cases, it is clear that income and pollution are not positively corre-lated, and at current income levels are negatively correlated. By the end othe 1990s most direct US air contaminant emissions were at or below wherethey were at the end o the Second World War, and in some cases ar below,as shown in this gure:

The data in Figure 3.3 have been scaled so the 1945 value equals 100 so thatcomparative changes are easy to visualize. NOx grew until about 1975 and

leveled o thereater, while all other air pollution emissions ell, and mean-while real US economic growth continued accelerating. Clearly the decouplingo economic growth and conventional, direct air pollution has been proven tobe technically easible.

On the other hand, due to the complexities o the linkage between specicemission types and the ormation o indirect contaminants, there has beenless success dealing with ground-level ozone and aerosol pollution. For exa-mple, the rate o ormation o O

3is not a simple linear unction. Under some

circumstances, reductions in NOx emissions may increase ozone ormation,rather than decrease it (Adamowicz et al. 2001), depending on the level o UVlight and the abundance o VOCs. O

3only alls i VOCs and NOx are reduced


12/28

10

together in the correct proportions. NOx emissions themselves are controlla-ble, though not as easily or cheaply as SO

2or CO.

Figure 3.4 shows data rom Toronto, Canada, and may be taken as typical or

many North American urban locations. The monthly average ozone level showsstrong seasonality, peaking in summer and reaching a minimum in winter, andthere is relatively little trend over time.

Fge 3.4: Toronto monthly average ozone concentrations 1974-2003.Source: McKitrick et al. (2005). Data rom Environment Canada NAPS archive.

However, one eature o the data not apparent in the monthly average readingsis that the summer ozone spikes have tended to become less intense over time,leading to a reduction in local violations o ozone standards in North Ameri-

can cities. Ozone standards in North America are not based on average levelsover time but on peak levels reached during episodes in which meteorologicalconditions avour rapid ozone and aerosol ormation (smog episodes). Theseusually occur on hot, muggy days in summer. Figure 3.5 shows the reductionin monitoring site violations in US cities rom 1975 to 2000. As is clear romthe data, these events now occur less requently, even though the averageozone level has not declined a great deal. Also note that, by comparison, COviolations have allen much arther, eectively to zero by 2000.


13/28

11

Fge 3.5: U.S. Air Monitoring Violations per station per year, CO andOzone.

Source: McKitrick (2010) Figure 1.10.

The third type o contaminant is CO2. In this case, there are no scrubbers, nor are

emissions reducible simply by increasing the eciency o the burn. The emis-sions are strictly determined by the carbon content o the uel. Consequently,CO

2emissions are eectively an index o ossil uel use, with the carbon/joule

content highest or coal, next or oil and least or natural gas. While it is pos-sible to remove the CO

2rom a smokestream, it emerges as a gas and cannot

be disposed o except by pumping it deep underground, which is costly both

nancially and in energy terms, as well as being easible only in limited loca-tions. So, in eect, CO2emissions cannot be decoupled rom economic growth

except by the slow process o increasing energy eciency or switching romcoal to oil or gas or electricity production. Since coal has, historically, beenthe cheapest orm o ossil energy or electricity production, as well as theeasiest to transport and the one most widely available, it has an advantage interms o cost and convenience. Since it is also the most CO

2-intensive orm

o uel, this means that CO2-intensive power generation has a relative cost

advantage over other orms. Coupled with the absence o scrubber technology

we understand the reason why attempts to sharply reduce CO2 emissions canbe expected to have serious economic consequences.


14/28

12

4. The Energy Mix

4.1 Pce Pc te

The reason or the long term position o coal as a leading energy source isshown by the ollowing chart, which tracks real energy prices rom 1949 to2010. Coal is not only the cheapest energy source, but has the least price vo-latility. Natural gas remains relatively expensive even ater its recent largedrop in price.

Fge 4.1: Real ossil energy process, 1949-2010. (1949=100).

Source: http://www.eia.gov/totalenergy/data/annual/index.cm#nancial.

On the production side, ossil sources continue to dominate electricity produc-tion, as shown in Figure 4.2. In 1973, 86.6% o world energy came rom coal,oil and natural gas. In 2009 these sources contributed 80.9%, with nuclearnow providing 5.8%. The share attributable to biouels, waste, geothermal andother renewables went rom 10.7% to 11%, in other words, it hardly changed.Even in countries that engaged in signicant policy eort to increase use orenewables, they did not gain much market share. American data are shown inFigure 4.3. Up to around 2006 renewables remained at or below 5% while coal

grew steadily to just over 30%, ater having allen rom 40% to about 20%in the 1950s and 60s. Ater 2006, heavy policy intervention led to renewables


15/28

13

growing rom 5% to 7%, displacing some coal and nuclear, while natural gasgrew rom 30% to 33%.

Figure 4.2: World energy production by uel source.

Source: http://www.iea.org/publications/reepublications/publication/key_world_energy_stats-1.pd.

Fossil sources (coal, gas and oil) accounted or 91% o US energy in 1949, and78% in 2010. Most o the dierence was made up by the introduction o nucle-ar, which now provides just over 10% o US energy. The continued dominanceo ossil sources, and the ailure o renewables to advance, are an indicationthat there are considerable technical and economic obstacles to their use. The

economic obstacles are indicated by the act that they only advance when sub-sidies are oered, and they decline when the subsidies are removed.1

1 See, or instance, http://www.bloomberg.com/news/print/2012-05-29/spain-ejects-clean-power-industry-with-europe-precedent-energy.html.


16/28

14

Fge 4.3: US Energy production by uel source.

Source: http://www.eia.gov/totalenergy/data/annual/index.cm#summary.

5. Economic principles

5.1 ug e pce e

Ignoring pollution externalities and natural monopolies or a moment, i weask what would be the best way to organize a countrys electricity generation

system, economists will typically turn to the price system to nd the best wayto proceed. The price system does not dictate how the energy will be provi-ded, it only provides a mechanism or the market to do the job in the mostecient way possible.

On the supply side, investors need to make long term commitments to a phy-sical capital stock and accompanying uel sources. The price system naturallyencourages the processing o large amounts o inormation necessary to ndand select the options that minimize the long term cost o production, given

the relative endowments o resources available in each location. The result isa supply schedule that presents consumers with electricity quantities and theircorresponding unit prices. I the supply system is competitive, these prices will


17/28

15

converge on the marginal value to society o the resources used to provide theelectricity. Then consumers can make decisions about how much electricity touse, presumably on the basis that the value they derive rom each kilowatt oconsumption exceeds the cost o purchasing it.

This is standard microeconomic reasoning, and applies to any energy market,including uels. The point to be emphasized is that there is nothing specialabout electricity. The price system provides incentives or ecient supply anddemand decisions without the need or a central coordinator. However, twoissues can arise that require modications to the market structure. First, be-cause electricity generation can involve large capital expenses there is a risko natural monopolies orming. I one envisions a situation in which a marketis small relative to the size o any power source, the rst operator to enter themarket can achieve economies o scale so quickly that any potential entrantmust do so at higher supply costs than the incumbent. In this case no rm willbe able to contest the market and a monopoly will result. The incumbent rmcan then over-charge consumers, setting the price up above the competitiverate but just below the rate that would attract competitors. This issue is o-ten presented in textbooks, but in practice is not a great concern. I a singlemonopolist emerges, it is relatively straightorward to implement price regu-lations to limit its power to overcharge consumers. Moreover, technology is

such that continental electricity markets have ormed in which suppliers mustcompete with each other to an extent sucient to alleviate concerns aboutnatural monopolies. I supply monopolies exist today it is likely that they arethe result o government policy, not the rationale or them.

Second, and more relevant to this discussion, some power generating sourcesgenerate external eects in the orm o air emissions. The question then beco-mes, how should policy makers intervene in the power sector to address this?

The next section will examine this more closely using microeconomic reaso-ning, but the key concept can be stated simply as the principle o targeting.Whatever the matter or regulatory concern, eciency is aided by targetingthe intervention on the specic issue itsel, rather than on secondary issues ormatters only partly connected to it. I the issue o concern is the sulphur comingout o a smokestack, implement policies that regulate the sulphur coming outo the smokestack. Do not implement policies that tell customers hundreds omiles away rom the smokestack how they should boil their eggs, in the hopesthat this will aect how much electricity they use and, eventually, how much

sulphur comes out o the smokestack. Any reductions in sulphur emissionsachieved by rules about how people hundreds o miles away should boil theireggs will be achieved with ar greater cost and inconvenience to industry and


18/28

16

the public than a rule directly limiting sulphur emissions. It would be more e-ective and ecient simply to regulate the sulphur emissions themselves andleave the remaining decisions about electricity supply and demand to marketparticipants. The idea, once grasped, is intuitively obvious, yet modern envi-

ronmental policymaking seems to have completely missed it, as evidenced bythe prolieration o rules about appliance standards, light bulbs, windmills,etc., all o which are done in the name o controlling emissions coming outo power plant stacks.

5.2 tkg cc eee

5.2.1 Emission pricing

No conceptual diculties emerge when incorporating pollution externalitiesinto the standard microeconomic model o the market. At the theoretical levelit is possible to solve the equations describing a welare-maximizing competi-tive market outcome in the presence o pollution externalities.2 It is a seriousmistake to suppose that the mere existence o externalities renders the eco-nomic approach to energy markets invalidar rom it. The economic modelhandles externalities quite readily and provides sound guidance relating to thereasons or current policy diculties.

I external costs were dealt with using price instruments, the basic insight othe Sandmo analysis is that, rather than subsidizing indirect control measuresor regulating quantities in related markets, emitters should pay a ee per unito emissions that refects the marginal social cost o their actions. This is anold idea, traditionally attributed to Pigou in the 1920s, but the Sandmo ana-lysis allows us to draw more detailed conclusions.

First, the optimal policy applies to the emissions themselves, and nothing else.Any additional price or regulatory intervention must, by necessity, make it cost-lier to achieve the same outcome. This goes against the instinct o policyma-kers, who oten want a portolio o measures to deal with what sounds likea complex issue. In the case o greenhouse gases, since so many sectors o theeconomy are involved, and the industrial processes are oten very complex,policymakers have apparently taken the view that many dierent complexrules and procedures are needed. But this is a allacy, and refects a conusionin thinking. Despite the number o economic sectors and complex activities

2 Sandmo (1975) is the standard reerence on this. My textbook (McKitrick 2010, chapter 8)provides a detailed derivation and explanation o Sandmos result.


19/28

17

involved in greenhouse gas emissions, only one price instrument is needed,namely a price on CO

2. The rest can ollow rom decentralized decision-making

without the need or central economic planning.

Second, the optimal policy should take the orm o a price on emissions setsomewhat below the estimated marginal damages, with the reduction propor-tional to the inverse o the marginal cost o public unds. The reason or thiscondition has to do with the distortions created elsewhere in the economy bythe necessity o unding the government using a tax system, which infates thecost o providing public goods, including pollution control. I the unds raisedby a carbon tax are treated in a revenue-neutral way, namely to pay or re-ductions in other taxes, and the carbon tax is scaled to take into account theoverall distortionary burden o the rest o the tax system, then the resultingemissions level will refect the optimal tradeo between costs and benets opollution control and the emission reductions will have been achieved at theminimum cost to society.

5.2.2 Pricing versus quantity targets

The integration o pollution control into general optimal policy making can beachieved naturally in a ramework in which the policy targets the emissions

price, rather than the quantity. In other words, economists have worked throughthe optimal conditions or pollution policy by thinking in terms o emissionees (prices), rather than emission standards (quantities). This is partly orconvenience, since the math is easier when working in terms o prices ratherthan quantities. But there are deeper reasons or avouring price instrumentseven when it is possible to devise a policy intervention that targets quantities.Only in very simplistic rameworks is there a symmetry between regulatingprice and regulating quantity. Some policy analysts mistakenly assume that

the two are always interchangeable. In the carbon policy world this meansthat policymakers are sometimes told that it makes no dierence whether theyopt or a carbon tax or a tradable permit (cap-and-trade) system. But this isincorrect or two reasons.

First, pricing instruments raise revenue by capturing all the scarcity rents. Anypolicy that limits a valuable activity (such as by imposing emission reductions)creates gaps between the cost o providing a good or service and the returnon it at the regulated margin. These gaps create windall gains, or rents as

economists call them. A good example is the taxi industry. Cities issue a li-mited number o licenses or taxis, and in most cases the number o licenses,and hence competitors, is limited to such an extent that taxis can charge more


20/28

18

than the actual cost o providing the service. Knowing that these rents arecreated by the licensing policy, buyers are willing to pay considerable sums toget a taxi license. The license itsel costs nothing or the regulator to create,so the value o the license represents the capitalization o the expected rental

fows. It is not new wealth, however, instead it is a transer rom householdsto license owners, and the rents transerred are always smaller than the ad-ditional costs created by the license system. The dierence is reerred to asthe deadweight loss.

Caps on emissions also create rents, but they are hidden in layers o economicactivity and accrue to industries and investors in ways that can be dicult topredict. Just as with taxi rents, they are not new wealth, instead they representtransers o wealth rom consumers to producers. I the regulator uses cap-and-trade, issuing permits and then allowing rms to trade them, the valueo these rents becomes visible as rms bid or the permits.

For instance, i a regulator issues permits or 100 megatonnes o carbon emis-sions, and they trade on the market or $20 each, the total capitalized valueo the rents created by the policy is thereore 20 x 100 million = $2 billion. Thetotal cost o the policy to households is thereore $2 billion plus the actualcost o emissions abatement which will be passed onto households through

higher prices and reduced rates o return or labour and capital as well asthe second-order distortions resulting rom applying the existing tax systemto markets where prices are now higher and quantities are lower. Emissionees capture the $2 billion or the public purse. I this is treated as new re-venue then households are no better o. But i the revenue is used to undreductions in other taxes, then the social cost o the policy drops by $2 billionplus the value o the reduced excess burdens o the tax system. It is possibleto capture the rents rom tradable permits by auctioning them rather than

giving them away ree, but this practice has been rare. It is not possible tocapture the rents under a system o regulatory standards. Hence quantity andprice controls are not symmetric in their overall macroeconomic costs, sincequantity controls typically ail to capture the rents created by the policy andthis adds to the costs to households.

Second, the symmetry between price and quantity breaks down when thereis uncertainty. A regulator can target the emissions price, and let the marketdetermine the resulting quantity, or he/she can target the emissions quantity,

and let the market determine the price, but he/she cannot determine both. Theprice o emissions corresponds to the marginal cost o emission reductions,or the marginal abatement cost. Having targeted the quantity o emissions,


21/28

19

the market will determine the marginal abatement cost (MAC). I the regulatorintends to target the marginal cost, the market will determine the resultingquantity o emissions.

Fge 5.1.: Price regulation versus quantity regulation

Source: Taken rom McKitrick 2012, g. 4

Figure 5.1 illustrates what this means. The horizontal axis shows the quanti-ty o emissions and the vertical axis shows the marginal abatement cost. Asemissions go down, the cost o each urther unit o emission reductions goesup, which is why the line slopes up as you read it rom right to let. There aretwo rms. Firm A has a relatively shallow MAC line and rm B has a relativelysteep one, meaning that it is relatively more expensive or rm B to cut emis-sions than or rm A. Now suppose the regulator orders each emitter to red-uce emissions to the level E1. For emitter A this implies that the last unit oemission abatement costs $25, while or emitter B it costs $100. Now suppose

the regulator did not intend emitters to pay more than $25 per unit to reduceemissions. By imposing a target on the emissions quantity axis (namely E1),the regulator orces the marginal costs to go above the intended level.


22/28

20

I the regulator instead imposed an emission ee o $25 per unit, both rmswould cut emissions, since, over a limited interval, they save more in taxes bydoing so than they incur in abatement costs. The point each rm would aimor is where their MAC line crosses the tax rate. Thus rm B will not reach the

emissions level E1; it will reduce emissions only to the level E2. Below thispoint the rms marginal abatement costs are above the cut-o.

The regulator cannot cap the marginal cost at $25 while asking both rms tocut emissions to E1. There is an inverse relation between the marginal cost cut-o and the volume o emission reductions that can be achieved. The higher theacceptable marginal cost o abatement, the lower the resulting emissions levelwill be. The lower the acceptable marginal cost, the higher the resulting emis-sions level will be. In this way the marginal abatement cost curve resembles ademand curve, since it shows the emissions level at each marginal cost level.

What happens i a regulator only intends or rms to incur a maximum cost o,say, $25 per tonne or abatement, but imposes a target that implies a margi-nal cost o, say $100 per tonne? Figure 5 shows that the result or rm B is agap between the intended ($25) and the actual ($100) marginal cost. The sizeo this gap is a measure o the uncertainty cost arising rom the policymakingerror. Would it have been better or the policymaker to set a price target in-

stead o a quantity target? The answer depends on two things.

First, i there are known danger thresholds associated with emissions, theregulator may decide that a quantity target is necessary, regardless o thecosts. This may be the case with highly toxic local emissions, or instance. Itis unlikely to be the case with general air pollutants that mix over large areassince emissions rom any one source have relatively small eects on the ove-rall concentration. In the case o carbon dioxide, since marginal damages o

emissions do not change over the entire range o a countrys emissions, thisconsideration does not apply.

Second, i the marginal abatement cost curve is relatively steep, it is morelikely that targeting an emissions quantity will lead to relatively larger uncer-tainty errors than would an emissions price target. The steeper rm Bs MACin gure 5, the larger the gap will be between the intended and the actualmarginal cost o emission reduction.

Since there are no scrubbers and no inexpensive abatement options, the mar-ginal cost o reducing CO

2emissions rises quickly. Attempting to pick a target

on the quantity axis guarantees large swings and errors on the price axis. But


23/28

21

or the same reason, when the MAC is steep, picking pretty much any emis-sions price will tend to yield only small errors in the resulting emissions level.Economists have shown that using price-based emission reduction policies orglobally-mixed pollutants with steep MACs which describes the CO

2case

leads to lower uncertainty costs, or in other words lower costs associated witherrors in choosing the right level o policy intervention.

In practice it is much more common or regulators to target the emission quan-tity than the emissions price, or marginal cost. But this is only due to habit,and in reality there is always an implied emissions price. I the regulator picksa target like E1 and insists that it be enorced, this implies the marginal costcut-o is at least $100; in other words we are willing to orce emitters toincur costs o $100 per tonne in order to reduce emissions.

It is important to recognize that the success o a price-based instrumentis not measured by whether the emission reductions achieve some arbitraryquantity target. In Figure 5, it is not the case that a $25 tax is successul inreducing rm As emissions but unsuccessul at reducing rm Bs emissions.The tax was successul at reducing each rms emissions up to the point whereurther emission reductions would be inecient. The act that Firm B undertookrelatively little abatement is not a ailure, it is the ecient outcome.

5.3 te pcg pcpe

Even i, or some reason, a policymaker insists on regulating the quantity oemissions rather than the price, the analysis herein can still provide guidanceon how it should be done. Emission pricing works in practice, but it also worksin theory. So at a simple level we can conduct an economic thought experi-ment along the ollowing lines. Suppose we are concerned about particulate

emissions rom power plants. In line with the principle o targeting a per-tonneprice is placed directly on particulate emissions. How would emitters respond?Analysts amiliar with the power sector can likely oer some reasonable spe-culations. Some rms will have older plants that they simply close down andreplace with newer more ecient units. Some will install scrubbers and otherend-o-pipe treatment. Some will nd all their options too expensive and will

just pay the tax instead. Once these responses have been initiated, the priceo electricity rom power plants would probably go up slightly. Other rmsmight then look at whether nuclear, wind or solar energy is now cheaper to

provide than coal-powered electricity. From the North American experiencethe answer is probably not. We can see rom Figure 3.3 that US particulateemissions (PM10) had already allen 50% by 1975 beore nuclear power beca-


24/28

22

me a major energy source, and nuclear required heavy government subsidiesto make it viable. Particulates had allen close to 90% by 2005, ater whichwind energy began to grow only because it was pushed by government policy.So even ater incurring the costs o heavy PM10 reductions, the private sector

was still reluctant to invest in nuclear and wind sources without governmentsubsidies. This means that even ater pricing in pollution control or conventi-onal sources, it remains relatively inexpensive compared to alternative energysources, so the production sector would not respond to an emissions tax by,or instance, building windmills.

Likewise, we can ask how households would respond to the slight increase inelectricity cost once a tax on particulate emissions was implemented. Chancesare most would make only slight behavioral changes, such as reducing powerconsumption a little bit. There might be a ew changes at the margin in ap-pliance and light xture choices, but it is unlikely households would make anextreme switch towards one type o lightbulb, or instance.

I the regulator wants to use quantity controls rather than pricing instru-ments, the pricing principle states that the best way to do so is to gure outhow rms would have responded to a price, and then aim or that outcome inthe quantity space. So in our thought experiment, concerns about particulate

emissions would mainly lead to adoption o scrubbers and accelerated turnovero old plants, but not to investment in wind energy or complete rejection oexisting household appliances and light bulbs. The contrast with actual expe-rience then becomes evident. Regulators concerned about air pollution haverequired installation o scrubbers but also oer lavish subsidies or wind andsolar energy and have issued bans and directives on various household appli-ances. These are wasteul and unjustied.

In the same way, we can ask how rms would respond to a levy on carbondioxide emissions? Many power generating rms would simply pay it and onlyslightly adjust their uel consumption, because o the steepness o their mar-ginal abatement cost curves. Others would act more aggressively, perhaps byreplacing coal plants with natural gas ones. But under any reasonable car-bon price it is unlikely German power producers would have invested much,i anything, in solar arms or wind turbines. Hence the push to develop theseindustries is not justied as a CO

2emissions policy, since a policy targeted on

ecient control o emissions would not have led to them being developed.


25/28

23

6. Concluding remarks

Growth in energy consumption is not just a by-product o economic growth,

it is a causal actor. Attempts to constrain energy consumption can thereorebe expected to impede economic growth. Fossil-based energy, especially coal,has a stable long-term economic advantage as a uel source. Concerns aboutpollution have led to considerable policy intervention in the energy sectorover the past decade, but much o it has been based on misunderstandings othe relevant pollution issues. Conventional air pollution is not a necessary by-product o power production and has been successully decoupled rom it inwestern countries through development o scrubbers and other end-o-pipetreatments. Indirect air contaminants, such as ozone and aerosols, are not tren-ding upwards and peak level episodes have allen, but the complexities o theatmospheric processes in these cases make progress inherently slower. Carbondioxide is not controllable by conventional scrubbers, instead it is tied linearlyto energy consumption. CO

2emission reductions will thereore be much more

expensive to achieve under current technology.

Economic reasoning has led to a ew basic ideas or guiding energy and pollu-tion policy. The principle o targeting states that regulations should be ocused

directly on the variable o interest. I we are concerned about air pollutioncoming rom power plant smokestacks, then we should regulate the emissionscoming rom the smokestacks, not kitchen appliances in households hundredso miles away who buy their electricity rom the power plant. The principleo pricing states that the most ecient outcome o well-designed regulati-ons will match that which would have resulted rom using price instrumentsdirectly applied to the emissions o concern. In practice, achieving that out-come is easiest simply by using price instruments. But i this is not easible,

the pricing principle can still be used as an analytic tool to understand whysome interventions, like the subsidy-driven expansion o the renewables sec-tor, is inecient.

7. References

Adamowicz, W. et al. 2001. A Review o the Socio-Economic Models and Rela-ted Components Supporting the Development o Canada-Wide Standards orParticulate Matter and Ozone. Royal Society o Canada, Ottawa.


26/28

24

Ghali, K.H. and M.I.T. El-Sakka (2004) Energy use and output growth in Canada:a multivariate cointegration analysis. Energy Economics 26:225 238.

Greenwood, J., A. Sheshadri and M. Yorukoglu (2005) Engines o Liberation.

Review o Economic Studies 72 109 133.

McKitrick, Ross R. (2010) Economic Analysis o Environmental Policy Toronto:University o Toronto Press.

McKitrick, Ross R. (2012) The High Price o Low Emissions: Benets and Costso GHG Abatement in the Transportation Sector. Ottawa: MacDonald-LaurierInstitute, March 2012.

McKitrick, Ross R., Kenneth Green and Joel Schwartz (2005) Pain WithoutGain: Shutting Down Coal-Fired Power Plants Would Hurt Ontario Fraser In-stitute, January 2005.

Sandmo, Agnar (1975). Optimal Taxation in the Presence o Externalities.Swedish Journal o Economics 77: pp. 86 98.

Stern, D.I. (2000) A multivariate cointegration analysis o the role o energy

in the US macroeconomy. Energy Economics 22:267 283.

About the author

r mcKck is a Proessor o Economics at the University o Guelph. Hespecializes in environmental economics. He has published many studies on

the economic analysis o pollution policy, economic growth and air pollutiontrends, climate policy options, the measurement o global warming, and sta-tistical methods in paleoclimatology.


27/28


28/28

Occasional PaperLiberal Institute o the Friedrich Naumann Foundation or FreedomOther publications under www.libinst.de

[90] Stean MelniklibEralism, libEral PoliCy and PrivatE ProPErty

[87] Robert Nenon-CEntralism

thE swis ExPErimEnt basEd on loCal GovErnmEnt, dirECt dEmoCraCy and taxComPEtition

[86] Fred E. FoldvarythE PrivatE Provision oF PubliC Goods:thE histor y and FuturE oF Comunal libEralism

[84] Thomas CieslikthE amEriCan drEam Job miGration into thE unitEd statEs

[82] Philippe LegrainmiGration mattErshow GErmany and thE world Can bEnEFit From a FrE movEmEnt oF PEoPlE

[78] Indur M. GoklanyadrEssinG ClimatE ChanGE in thE ContExt oF ot hEr ProblEms a PlEa For rEalism ovEr idEoloGy

[56] Thomas LenkrEForm oF thE FinanCial Equalisation sChEmE in GErmany:a nEvEr-EndinG story?

[55] Julie SimmonstrEaty FEdEralism: thE Canadian ExPEriEnCE

[54] George HerbertthE trEaty oF lisbon a thrEat to FEdEralism?

[53] Dan Lewiswill wE EvEr run out oF EnErGy? thE history and FuturE oF tEChnoloGiCalProGrEss and EnErGy suPPly

[51] Lincoln AllisonsPort as virtuE ... as lovE ... as CommErCE

[50] Jutta BraunFootball and PolitiCal FrEEdom: thE historiCal ExPEriEnCE oF dividEd GErmany

[49] Raymond D. SauersPort, thE statE, and thE marKEt

[48] Tom G. PalmerFrEEdom ProPErly undErstood

[47] Temba A. NolutshunguCivil soCiEty? nGoism at worK

[46] Martin Fronek and Jir SchwarzKyo to sEvEral yEars aFtEr by

[45] Doug Bandowhuman riGhts in danGEr? myths and rEalitiEs in thE un

[44] Parth J. ShahCsr: CaPitalism at its bEst or an anti-CaPitalist mEntality?

[43] Frank VibertsoFt PowEr and intErnational rulE-maKinG

[42] Erich Weedenation-buildinG in thE middlE East: thE nEw imPErialism?[41] Dr. Wol-Dieter Zumport

energy, pollution control and economic growth

Documents