Market Failure and the Structure of Externalities · 2018-09-26 · Although markets are not...

JOBNAME: Earthscan−renewenerg PAGE: 1 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 51A949C3/production/earthscan/books/renewenergy/chap05

5

Market Failure and the Structureof ExternalitiesKenneth Gillingham and James Sweeney

Policy interest in renewable energy technolo-gies has been gathering momentum for the

past several decades, and increased incentives andfunding for renewable energy are often describedas the panacea for a variety of issues ranging fromenvironmental quality to national security togreen job creation. Sizable policies and programshave been implemented worldwide to encouragea transition from fossil-based electricity genera-tion to renewable electricity generation, and inparticular to fledgling green technologies such aswind, solar, and biofuels.

The United States has a long history of policyactivity in promoting renewables, including state-level programs, such as the California Solar Initia-tive, which provides rebates for solar photovoltaicpurchases, as well as federal programs, such as taxincentives for wind. Even in the recent stimuluspackage, the American Recovery and Reinvest-ment Act of 2009, $6 billion was allocated forrenewable energy and electric transmission tech-nology loan guarantees (U.S. Congress 2009).(See Chapter 11 for further discussion of the U.S.experience.) Moreover, such policies are notrestricted to the developed world. For example,China promulgated a National RenewableEnergy Law in 2005 that provides tax and otherincentives for renewable energy and has suc-ceeded in creating a burgeoning wind industry(Cherni and Kentish 2007).

Advocates of strong policy incentives forrenewable energy in the United States use a vari-ety of arguments to justify policy action, such asending the “addiction” to foreign oil, addressingglobal climate change, or creating new technolo-gies to increase U.S. competitiveness. However,articulation of these goals leaves open the ques-tion of whether renewable energy policy is a sen-sible means to reach these goals, or even whetherparticular renewable energy policy helps meetthese goals. Furthermore, many different policyinstruments are possible, so one must evaluatewhat makes a particular policy preferable overothers.

Economic theory can provide guidance andmore rigorous motivation for renewable energypolicy, relying on analysis of the ways privatelyoptimal choices deviate from economically effi-cient choices. These deviations are described asmarket failures and, in some cases, behavioral fail-ures.1 Economic theory indicates that policymeasures to mitigate these deviations can improvenet social welfare, as long as the cost of imple-menting the policy is less than the gains if thedeviations can be successfully mitigated.

Under this perspective, policy analysisinvolves identifying market failures and choosingappropriate policy instruments for each. While analmost unlimited number of different possible

Kerrypress Ltd – Typeset in XML A Division: chap05 F Sequential 1

JOBNAME: Earthscan−renewenerg PAGE: 2 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 6A43C373/production/earthscan/books/renewenergy/chap05

policy instruments can be envisioned, an analysisof relevant market failures allows us to identifywhich instruments are most likely to improveeconomic efficiency.This endeavor is complicatedby the complexity of some market failures, whichmay vary intertemporally or geographically.

This chapter explores these issues in the con-text of renewable energy, with a particular focuson renewable energy used for electricity genera-tion. It first sets the stage with a brief backgroundon the fundamental issues inherent in renewableenergy. Next, it elaborates on the concepts ofcompetitive markets and resource use, and howthe deviations found in reality from the assump-tions of perfect markets may result in market fail-ures. This leads naturally to articulating the classesof possible deviations from perfect markets. A dis-cussion follows of the use of policy instruments tohelp mitigate or correct for these market failures,with a particular focus on how the structure of thefailure influences the appropriate policy approach.

Fundamental Issues inRenewable EnergyRenewable energy, including wind, solar, hydro,geothermal, wave, and tidal, offers the possibilityof a large, continuous supply of energy in perpe-tuity. Analysis of the natural energy flows in theworld shows that they provide usable energymany orders of magnitude greater than the entirehuman use of energy (Hermann 2006). Forexample, the amount of sunlight reaching theearth is more than 10,000 times greater than thetotal human direct use of energy, and the amountof energy embodied in wind is at least 4 timesgreater (Archer and Jacobson 2005; Da Rosa2005; EIA 2008). In principle, renewable energyoffers the possibility of a virtually unlimited sup-ply of energy forever.

In contrast, most of the energy sources we relyon heavily today, such as oil, natural gas, coal, anduranium, are depletable resources that are presenton the earth as finite stocks. As such, eventuallythese stocks will be extracted to the point thatthey will not be economical to use, because of

either the availability of a substitute energy sourceor scarcity of the resource. The greater the rate ofuse relative to the size of the resource stock, theshorter the time until this ultimate depletion canbe expected.

These simple facts about the nature ofdepletable and renewable resources point to aseemingly obvious conclusion: the United Statesand the rest of the world will eventually have tomake a transition to alternative or renewablesources of energy. However, the knowledge thatthe world will ultimately transition back torenewable resources is not sufficient reason forpolicies to promote those resources. Such transi-tions will happen regardless of policy, simply as aresult of market incentives.

The fundamental question is whether marketswill lead the United States and the rest of theworld to make these transitions at the appropriatespeed and to the appropriate renewable resourceconversions, when viewed from a social perspec-tive. If not, then the question becomes, why not?And if markets will not motivate transitions at theappropriate speed or to the appropriate renewablesupplies, the question becomes whether policyinterventions can address these market failures soas to make the transitions closer to the sociallyoptimal.

The question of why not may seem clear tothose who follow the policy debates. Environ-mental and national security concerns are fore-most on the list of rationales for speeding up thetransition from depletable fossil fuels to renewableenergy. Recently there have also been claims thatpromoting new renewable technologies couldallow the United States, or any country, tobecome more competitive on world markets orcould create jobs.

But much national debate often combinesthese rationales and fails to differentiate amongthe various policy options, renewable technolo-gies, and time patterns of impacts. The rest of thechapter explores these issues in greater detail inorder to disentangle and clarify the arguments forrenewable energy policy.

70 Kenneth Gillingham and James Sweeney


JOBNAME: Earthscan−renewenerg PAGE: 3 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 74887430/production/earthscan/books/renewenergy/chap05

Resource Use and Deviationsfrom Perfectly FunctioningMarketsWelfare economic theory provides a frameworkfor evaluating policies to speed the transition torenewable energy. A well-established result fromwelfare economic theory is that absent market orbehavioral failures, the unfettered market out-come is economically efficient.2 Market failurescan be defined as deviations from perfect marketsdue to some element of the functioning of themarket structure, whereas behavioral failures aresystematic departures of human choice from thechoice that would be theoretically optimal.3

A key result for analysis of renewable energy isthat if the underlying assumptions hold, then thedecentralized market decisions would lead to aneconomically efficient use of both depletable andrenewable resources at any given time. Moreover,the socially optimal rate of transition fromdepletable energy supplies to renewable energycan be achieved as a result of decentralized marketdecisions, under the standard assumptions thatrational expectations of future prices guide thedecisions of both consumers and firms (Heal1993).

Although markets are not perfect, the conceptof perfectly competitive markets provides abenchmark for evaluation of actual markets. Iden-tification of market imperfections allows us toevaluate how actual markets deviate from the idealcompetitive markets and thus from the economi-cally efficient markets. Hence with economic effi-ciency as a policy goal, we can motivate policyaction based on deviations from perfectly com-petitive markets—as long as the cost of imple-menting the policy is less than the benefits fromcorrecting the deviation.4

For renewable energy, market failures aremore relevant than behavioral failures, as mostenergy investment decisions are made by firmsrather than individuals, so some of the keydecisionmaking biases pointed out in thebehavioral economics literature are likely to playless of a role. However, behavioral failures mayinfluence consumer choice for distributed genera-

tion renewable energy (e.g., residential solar pho-tovoltaic investments) and energy efficiency deci-sions.5 These could imply an underuse of distrib-uted generation renewable energy—or an overuseof all energy sources (including renewables) ifenergy efficiency is underprovided.

Both market failures and behavioral failurescan be distinguished from market barriers, whichcan be defined as any disincentives to the use oradoption of a good (Jaffe et al. 2004). Market bar-riers include market failures and behavioral fail-ures, but they also may include a variety of otherdisincentives. For example, high technology costsfor renewable energy technologies can bedescribed as a market barrier but may not be amarket failure or behavioral failure. Importantly,only market barriers that are also market orbehavioral failures provide a rationale based oneconomic efficiency for market interventions.

Similarly, pecuniary externalities may occur inthe renewable energy setting and also do not leadto economic inefficiency. A pecuniary externalityis a cost or benefit imposed by one party onanother party that operates through the changingof prices, rather than real resource effects. Forinstance, if food prices increase because ofincreased demand for biofuels, this could reducethe welfare of food purchasers. However, the foodgrowers and processors may be better off. In thissense, pecuniary externalities may lead to wealthredistribution but do not affect economic effi-ciency.

Nature of Deviations fromPerfectly Functioning MarketsIt is a useful to consider deviations from perfectlyfunctioning markets based on whether the marketfailure is atemporal or intertemporal.

Atemporal deviations are those for which theexternality consequences are based primarily onthe rate of flow of the externality. For example, anexternality associated with air emissions maydepend primarily on the rate at which the emis-sions are released into the atmosphere over aperiod of hours, days, weeks, or months. Such

Market Failure and the Structure of Externalities 71


JOBNAME: Earthscan−renewenerg PAGE: 4 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 7F764FA3/production/earthscan/books/renewenergy/chap05

externalities can be described statically. They maychange over time, but the deviation has economicconsequences that depend primarily on theamount of emissions released over a short timeperiod (e.g., hours, days, weeks, or months).These may have consequences that are immediateor occur over very long time periods.

Intertemporal deviations are those for whichthe externality consequences are based primarilyon a stock that changes over time depending onthe flow of the externality. The flows lead to achange in the stock over a relatively long periodof time, typically measured in years, decades, orcenturies. The stock can be of a pollutant (e.g.,carbon dioxide) or of something economic (e.g.,the stock of knowledge or of photovoltaicsinstalled on buildings). If the flow of the external-ity is larger (smaller) than the natural decline rateof the stock, the stock increases (decreases) overtime. Intertemporal externalities can best bedescribed dynamically, for it is the stock (e.g., car-bon dioxide), rather than the flow, that leads tothe consequences (e.g., global climate change).

For some environmental pollutants (e.g.,smog), the natural decline of the stock is rapid—perhaps over the course of hours, days, weeks, ormonths. For these pollutants, the stock leads tothe damages, and the stock is entirely determinedby the flow over this short time frame. These canbe treated as atemporal deviations, as the dynamicnature of the externality is less important withsuch a rapid natural decline rate.

For atemporal externalities, the appropriatemagnitude of the intervention depends primarilyon current conditions. Thus, because conditionscan change over time, the appropriate magnitudecould increase, decrease, or stay constant overtime. For intertemporal externalities, the appro-priate magnitude of the intervention dependsmore on the conditions prevailing over manyfuture years than on current conditions or those atone time. As time passes, the appropriate magni-tude of the intervention changes but, more pre-dictably, based on the stock adjustment process.Therefore, the appropriate price or magnitude ofthe intervention will have a somewhat predictabletime pattern.

Atemporal (Flow-Based) Deviations fromEconomic Efficiency

Atemporal deviations from economic efficiencyfall into several categories: labor market supply–demand imbalances, environmental externalities,national security externalities, information mar-ket failures, regulatory failures, market power,too-high discount rates for private decisions,imperfect foresight, and economies of scale.

Labor Market Supply–Demand Imbalances

Unemployment represents a situation in whichthe supply of labor exceeds demand at the prevail-ing wage structure, perhaps because of legal andinstitutional frictions slowing the adjustment ofthe wage structure. In the United States, suchunemployment does not occur very often, typi-cally only during recessions. At times of fullemployment,6 abstracting from the distortionaryimpacts of income or labor taxes,7 the social costof labor (i.e., the opportunity cost and other costsof that labor to the employee) would be equal tothe price of labor (i.e., the wage an employer mustpay for additional labor), and hence there is noroom to improve economic efficiency throughgreen jobs programs.

With unemployment, however, the price oflabor exceeds the social cost of that labor. Thisdifference represents a potential net economicefficiency gain, and thus any activity that employsadditional workers may improve economic effi-ciency. For example, if an additional amount ofsome economic activity produced no net profit,and therefore would not be privately undertaken,the net social economic gain would be equal tothe differential between the price of labor and itssocial cost.

With unemployment, the opportunity cost(and other cost) of labor to the person beingemployed could be expected to vary substantiallyacross individuals. Some unemployed persons mayuse their free time productively to perform workat home or improve skills, so that the opportunitycost of labor might be only slightly below thewage. Others may not be able to make such pro-ductive use of their time, so that the opportunity



JOBNAME: Earthscan−renewenerg PAGE: 5 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 75B0D3EF/production/earthscan/books/renewenergy/chap05

cost might be virtually zero, significantly belowthe wage. Thus the potential net social gain fromadditional employment could range from nearlythe entire wage to zero.

Little evidence exists to suggest that additionalemployment in renewable energy can providelarger net social gains than any other industry,including the fossil-fuel industry. Moreover, suchgains must be seen as transient possibilities in aneconomy such as that of the United States, whichregularly is near full employment.

Environmental Externalities

Environmental externalities are the underlyingmotivation for much of the interest in renewableenergy. The discussion here focuses on generalissues in environmental externalities; specificissues inherent in intertemporal environmentalexternalities are addressed below in the sectiontitled “Stock-Based Environmental Externalities”.Combustion of fossil fuels emits a variety of airpollutants, which are not priced without a policyintervention. Air pollutants from fossil-fuel com-bustion include nitrogen oxides, sulfur dioxide,particulates, and carbon dioxide. Some of thesepollutants present a health hazard, either directly,as in the case of particulates, or indirectly, as in thecase of ground-level ozone formed from high lev-els of nitrogen oxides and other chemicals.

When harmful fossil-fuel emissions are notpriced, the unregulated market will overuse fossilfuels and underuse substitutes, such as renewableenergy resources. Similarly, if the emissions arenot priced, firms will have no incentive to findtechnologies or processes to reduce the emissionsor mitigate the external costs. The evidence forenvironmental externalities from fossil-fuel emis-sions is strong, even if estimating the precise mag-nitude of the externality for any given pollutantmay not be trivial.

In some cases, there may also be significantenvironmental externalities from renewableenergy production, such as hydroelectric facilitiesthat produce methane and carbon dioxide emis-sions from submerged vegetation, or greenhousegas emissions and nitrogen fertilizer runoff fromthe production of ethanol biofuels. In many other

cases, these environmental externalities are rela-tively small. Whether renewable energy resourcesare underused or overused relative to economi-cally efficient levels depends on which of the twoenvironmental externalities is greater: those fromfossil fuels or from the renewable energyresources. In most, but by no means all, cases, theexternalities from the fossil fuels are greater,implying that the market will underproviderenewable energy.

Unpriced environmental externalities fromeither fossil fuel or renewable energy use wouldimply either an overuse of energy in general or anunderuse of potential energy efficiency improve-ments.

National Security Externalities

Oil production around the world is highly geo-graphically concentrated, with the bulk of the oilreserves in the hands of national oil companies inunstable regions or countries of the world, such asthe Middle East, Nigeria, Russia, and Venezuela.Oil-importing countries, such as the UnitedStates, European nations, and China, have seenlarge security risks associated with these oilimports. In response, they have laid out substantialdiplomatic and military expenditures in theseregions, at least partly in order to ensure a steadysupply of oil. If increases in oil use lead to addi-tional security risks, these risks represent an exter-nality associated with oil use. Moreover, if theadditional security risks are met with increases indiplomatic and military expenditures, then theseadded expenditures can be used as an approximatemonetary measure of the externalities.

However, it appears unlikely that a modestincrease or decrease in oil demand will influencethese expenditures due to the lumpiness of theexpenditures, even though the increases in oil usecould lead to additional security risks. Conversely,long-term large changes in oil demand mayreduce national security risks and the correspond-ing military and diplomatic expenditures.

In many countries around the world, such asthose in Europe, the use of natural gas may havenational security externalities because of similarissues. Quantifying the national security exter-



JOBNAME: Earthscan−renewenerg PAGE: 6 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 76FD9E20/production/earthscan/books/renewenergy/chap05

nalities associated with oil or natural gas con-sumption is more fraught with difficulties thandoing so with environmental externalities, yetsome analysts have suggested that the magnitudemay be substantial (Bohi and Toman 1996). Oth-ers are more sanguine and believe that globalenergy markets can substantially buffer nationalsecurity risks.

In the U.S. context, natural gas and somerenewable energy resources, such as biofuels, aresubstitutes for oil with few or no energy securityexternalities and thus would be underused relativeto the economically efficient level. Improving theenergy efficiency of vehicles and furnaces is also asubstitute for oil and would also be underused.Most renewable energy resources produce elec-tricity, so until electric vehicles are a viable large-scale substitute for conventional vehicles fueled byrefined oil products, national security externalitiesapply only indirectly to such renewable energyresources. However, these national security exter-nalities, although indirect, can be important. Forexample, the production of electricity fromrenewables could lead to reductions in natural gasused for electricity production. This reductionwould lead to more availability of natural gas forother purposes, such as heating, where it couldsubstitute for oil in some locations. For biofuels,national security externalities are of foremostconsideration. Moreover, in the European con-text, renewable energy directly substitutes fornatural gas.

Information Market Failures

Information market failures relate most directly tothe adoption of distributed generation renewableenergy by households, such as solar photovoltaicsystems or microgeneration wind turbines. Ifhouseholds have limited information about theeffectiveness and benefits of distributed genera-tion renewable energy, an information marketfailure may occur. In a perfectly functioning mar-ket, one would expect profit-maximizing firms toundertake marketing campaigns to inform poten-tial customers. However, for nascent technologiesthat are just beginning to diffuse into the market,economic theory suggests that additional infor-

mation can play an important role (Young 2010).Information market failures are closely related tobehavioral failures. Reducing information marketfailures would also be expected to reducebehavioral failures associated with heuristicdecisionmaking.

Imperfect foresight by either firms or con-sumers (or investors in the stock market whoinfluence firms) suggests an inability to predictfuture conditions accurately, which may lead to anunderestimate or overestimate of how energyprices may rise in the future. If firms systemati-cally under- or overestimate future energy prices,then there may be an underinvestment oroverinvestment in research and development(R&D) for renewable energy technologies relativeto the economically efficient level.

Although it certainly seems plausible thatfirms have imperfect foresight, it is less plausible tobelieve that this imperfect foresight will systemati-cally lead to an underestimate of future energyprices, rather than random deviations that aresometimes underestimates and other times over-estimates. Even if firms have imperfect foresight,as long as the firms’ estimates of future prices arenot systematically biased, then on average invest-ment in renewable energy technologies wouldstill follow the economically efficient path. In thissituation, errors leading to overinvestment wouldbe balanced by those leading to underinvestment.At present, there is little evidence either for oragainst the hypothesis that firms systematicallyunderestimate future price increases.

Another information market failure is theclassic principal-agent or split-incentive problem,which may influence renewable energy adoptionin two ways. First, in many cases for rental prop-erties, landlords make the decision about whetherto invest in distributed generation renewableenergy, while tenants pay the energy bills (Jaffeand Stavins 1994; Murtishaw and Sathaye 2006).Second, if landlords are not compensated for theirinvestment decisions with higher rents, then theywould tend to underinvest in distributed genera-tion renewable energy. This market failure hasbeen most carefully examined in the context ofenergy efficiency (e.g., see Levinson andNiemann 2004), but the extent to which this



Kenneth Gillingham

Inserted Text

and heating

JOBNAME: Earthscan−renewenerg PAGE: 7 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 73BD4831/production/earthscan/books/renewenergy/chap05

market failure is important for renewable energyhas not yet been empirically examined.

Finally, there may be a principal-agent prob-lem relating to managerial incentives. In manycases, managers have their compensation tied tothe current stock price, rather than the long-termperformance of the company (Rappaport 1978).However, investors may have difficulty distin-guishing between managerial decisions that boostshort-term profits at the expense of long-termprofits and those that boost both short- and long-term profits. In the context of renewable energy,the emphasis on short-term performance maylead to underinvestment in R&D for renewableenergy technologies, for the benefits of develop-ing such technologies are likely to be receivedover the long term, while the costs are borne inthe short term. Of course, this issue may occur inany industry and is not unique to renewableenergy resources.

Regulatory Failures

In some cases, the regulatory structure can createperverse incentives. For example, average costpricing of electricity implies that consumers oftenface a price of electricity that does not reflect themarginal cost of providing electricity at any giventime. This may influence the adoption of distrib-uted generation renewables, such as residentialsolar photovoltaic (PV) systems. In many loca-tions, electricity output from a solar PV unit tendsto be higher during the day, corresponding totimes of high electricity demand. To the extentthat the solar PV output is correlated with highwholesale electricity prices, consumers and firmsdeciding whether to install a new solar PV unitwill undervalue solar PV absent tariffs thataccount for the time variation. Borenstein (2008)quantifies this effect in California, finding thatsolar is currently undervalued by 0% to 20%under the current regulatory framework, and thatthis could rise to 30% to 50% if the electricitysystem were managed with more reliance onprice-responsive demand and peaking prices,because solar output would be concentrated attimes with even higher value.

Too-High Discount Rates

In some cases, the discount rate for private invest-ment decisions may be higher than the social dis-count rate for investments with a similar risk pro-file. For example, the corporate income tax dis-torts incentives for firms to invest, effectivelyimplying that they require a higher rate of returnon investments than they would otherwise. Alter-natively, credit limitations may also occasionallylead to a higher rate of return required for invest-ments. These credit limitations may be due tomacroeconomic problems, such as the recentliquidity crisis in the United States, or individuallimitations on the firm involved in the renewableenergy investment. Individual credit limitationsmay also apply in cases where consumers areinterested in installing distributed or off-grid gen-eration.

Discount rates that are too high may lead totwo effects. First, if firms investing in renewableenergy technologies have distorted discount rates,this could lead to underinvestment in renewableenergy resources relative to the economically effi-cient level. Second, if discount rates are too highfor firms extracting depletable resources, such asfossil fuels, then the fuels are extracted too rapidly,leading to prices that are lower than economicallyefficient. Because the depletable resource wouldbe depleted too rapidly, the transition to renew-able energy technologies may then be hastenedrelative to the efficient transition. However,investment in renewables may be second best, inthat it would still be optimal to invest more, con-ditional on the too-rapid extraction of depletableresources.

This phenomenon is applicable not only toenergy-related investments, but also to invest-ments throughout the economy. Thus this issueprovides reasons for changing incentives forinvestment throughout the economy, but it doesnot provide a particular reason for shifting invest-ments from other parts of the economy to renew-able energy, unless evidence suggested that highdiscount rates are particularly important forrenewable energy. However, we are aware of noevidence that could give a sense of the magnitudeof this distortion.



JOBNAME: Earthscan−renewenerg PAGE: 8 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 66F9EFCC/production/earthscan/books/renewenergy/chap05

Economies of Scale

Economies of scale, particularly increasing returnsto scale, refers to a situation where the averagecost of producing a unit decreases as the rate ofoutput at any given time increases, resulting froma nonconvexity in the production function for anynumber of reasons, including fixed costs. Thisissue may inefficiently result in a zero-outputequilibrium only when we have market-scaleincreasing returns, where the slope of the averagecost function is more negative than the slope ofthe demand function, and the firm cannot over-come the nonconvexity on its own.

Market-scale increasing returns refer to anonconvex production function at output levelscomparable with market demand. Figure 5.1graphically illustrates the second condition. If thequantity produced is small (e.g., quantity a), thenno profit-seeking firm would be willing to pro-duce the product, but if production could beincreased to some level above the crossing point(e.g., at the quantity b), then it would be profit-able for the firm to produce: price would exceedaverage cost.

Usually a firm could overcome the situationin Figure 5.1 on its own simply by selling at a lowprice. Even if this is a risky endeavor, it is notlikely that all firms would ignore this opportunity.However, firms may not be able to take advantage

of the opportunity because of capital constraintsor a simultaneous coordination problem.

Capital constraints may be a problem only ifthe aggregate investment required is extremelylarge; otherwise, it is likely that some firm couldbe expected to raise the necessary capital. Capitalconstraints facing an economy, as occurred in the2008–2009 recession, could limit such capitalinvestments for an entire economy. Because suchevents tend to be transient, however, these con-straints at most could be expected to delay theinvestments.

Often economies of scale are accompanied bya “chicken-and-egg” problem, wherein multipleactors must simultaneously invest and ramp upproduction in order to commercialize a new tech-nology.This may be most relevant in technologiesthat require a new infrastructure, such ashydrogen-fueled vehicles, which may or may notuse renewable energy depending on the hydrogengeneration source. Such possibilities requireinterindustry cooperation and thus may greatlydelay investments. Similar chicken-and-egg prob-lems have been overcome in the past, as with per-sonal computers, operating systems, and applica-tion software or automobiles, gasoline, service sta-tions, and roads, but these problems greatlycomplicate investments.

It should be noted that the equilibrium thatwould occur with market-scale increasing returnswould unlikely be a workable competitive equi-librium, but rather a single-firm monopolisticequilibrium. In fact, the situation of market-scaleincreasing returns is often referred to as a “naturalmonopoly.” This situation raises the possibility ofmarket power.

Market Power

Uncompetitive behavior may influence the adop-tion of renewable energy technologies in severalways. First, market power in substitutes forrenewable energy can influence the provision ofrenewable energy through two channels. Firmseffectively exercising market power in substitutesfor renewable energy (e.g., at times the OPECcartel) would raise the price of energy above theeconomically efficient level, making investment in

Price

Demand

Average cost

Quantityba

Figure 5.1. Economies of scale: slope of averagecost function is more negative than slope ofdemand function



JOBNAME: Earthscan−renewenerg PAGE: 9 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 793A0EFA/production/earthscan/books/renewenergy/chap05

renewable energy more profitable and leading toan overinvestment in renewable energy. On theother hand, firms that have market power in sub-stitutes for renewable energy may have an incen-tive to buy out fledgling renewable energy tech-nologies to reduce competitive pressures—leadingto a possible underprovision of renewable energyresources if that purchasing firm “buries” therenewable technology. However, the prospect ofbeing bought by a competitor could provide astrong incentive for a new firm to be created withthe explicit intention of selling itself to a largercompany. Which effect dominates and whetherthere is market power in substitutes for renewableenergy can be determined only empirically.

Market power may also influence the adop-tion of renewable energy resources by influencingthe rate and direction of technological change. Ifless competition exists in a market, firms are morelikely to be able to fully capture the benefits oftheir innovations, so incentives to innovate arehigher (e.g., see Blundell et al. 1999; and Nickell1996). Conversely, if more competition exists,firms may have an incentive to try to “escape”competition by investing in innovations that allowthem to differentiate their product or find a pat-entable product. Some evidence suggests that therelationship between competition and innovationmay be an inverted U-shaped curve, with a posi-tive relationship at low levels of competition and anegative relationship at higher levels of competi-tion (Aghion et al. 2005; Scherer 1967).This rela-tionship likely holds in all industries, not just therenewable energy industry.

Finally, in some cases, vertically integratedutilities may effectively exercise market power byfavoring their own electricity generation facilitiesover other small generation facilities, includingrenewable energy facilities.This was a concern forthe implementation of renewables when utilitiesinvested mostly in nonrenewable energy, but utili-ties now typically invest in renewable energyalong with conventional generation plants.8

Intertemporal (Stock-Based) Deviations

An important intertemporal deviation may occurwith the existence of stock-based environmental

externalities. A second intertemporal deviationmay occur if an imperfect capture of the stock ofknowledge is created as a result of current actions,leading to underinvestment or underproductionof those activities that lead to growth of theknowledge stock. These can occur with knowl-edge generation processes, such as learning bydoing or research and development; market diffu-sion of a new technology; or network externali-ties. Intuitively, when others can capture some ofthe benefits from the choice made by a firm orconsumer, the uncaptured benefits will be sociallyvaluable but will not be taken into account by thefirm or consumer.

Stock-Based Environmental Externalities

As discussed above, some environmental exter-nalities have consequences based on the stock ofthe pollutant, rather than the flow, and the stockadjusts only slowly over time. For such environ-mental externalities, the intertemporal nature ofthe damages from the stock imposes additionalstructure on the time pattern of deviations.

Particularly relevant to renewable energy sup-plies are carbon dioxide and other greenhousegases. For CO2, every additional metric ton emit-ted remains in the stock for more than a century.Thus emitting a ton today would have roughly thesame cumulative impacts as emitting a ton in 20years. This implies that, absent changes in theregulatory environment, the magnitude of thedeviation for emissions now will be the same asthe magnitude of the deviation for emissions 20years from now. Economic efficiency implies thata society should be almost indifferent betweenemitting a ton of CO2 now, 20 years from now, orany year in between .9 As will be discussed belowin the section titled “Policies for Stock-BasedEnvironmental Externalities: Carbon Dioxide”, itis this relationship that imposes a structure on thetime pattern of efficient policy responses.

Similar issues arise for toxic metals releasedinto the waterways, radioactive nuclear waste,mercury in waterways and oceans, sequestrationof carbon dioxide in the deep oceans, and rainfor-est land degradation.



JOBNAME: Earthscan−renewenerg PAGE: 10 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 7885BEB7/production/earthscan/books/renewenergy/chap05

Imperfect Capture of Future Payoffs fromCurrent Actions

R&DWhen firms invest in increasing the stock ofknowledge by spending funds on R&D, they maynot be able to perfectly capture all of the knowl-edge gained from their investment. For example,successful R&D (e.g., creating a new class of solarphotovoltaic cells) by a particular firm could beexpected to result in some of the new knowledgebeing broadly shared, through trade magazines,reverse engineering by its competitors, or techni-cal knowledge employees bring with them as theychange employment among competitive firms. Inaddition, patent protection for new inventionsand innovations has a limited time frame (20 yearsin the United States), so after the patent lapses,other firms may also benefit directly from theinvention or innovation.

Fundamentally, R&D spillovers can bethought of as an issue of imperfect property rightsin the stock of knowledge: other firms can sharethat stock without compensating the original firmthat enhanced the knowledge stock.To the extentthat those spillover benefits occur, the social rateof return from investment in R&D is greater thanthe firm’s private rate of return from investment inR&D. Indeed, although estimates differ by sector,there appears to be substantial empirical evidencethat the social rate of return is several times that ofthe private rate of return. For example, in theUnited States, the social rate of return is estimatedin the range of 30% to 70% per year, while theprivate rate of return is 6% to 15% per year(Nordhaus 2002). However, the magnitude of theR&D spillovers depend on the stage in the devel-opment of a new technology, with more funda-mental research having significantly greater R&Dspillovers than later-stage commercializationresearch (Nordhaus 2009).

Evidence of high social returns to R&D isfound not just in the renewable energy sector, butthroughout the economy.Thus, to the extent thatsome R&D in renewable energy technologiescomes at the expense of R&D in other sectorswith a high social rate of return, the opportunity

cost of renewable energy R&D may be quite high(Pizer and Popp 2008). Empirical work suggeststhat additional R&D investment in renewableenergy will at least partly displace R&D in othersectors. Popp (2006) finds that approximately halfof the energy R&D spending in the 1970s and1980s displaced, or crowded out, R&D in othersectors. Part of the rationale for this may be thatyears of training are required to become a compe-tent research scientist or engineer, and thereforethe supply of research scientists and engineers is,at least in the short term, relatively inelastic. In thelonger term, crowding out is less likely to be anissue, as universities train more scientists and engi-neers.

Learning by doingA similar intertemporal market imperfection dueto a knowledge stock spillover may also occur ifthere is a significant learning-by-doing (LBD)effect that cannot be captured by the firm. LBDhas a long history in economics, dating back toArrow (1962). The basic idea behind LBD is thatthe cost of producing a good declines with thecumulative production of the good, correspond-ing to the firm “learning” about how to producethe good better.10 One interpretation is that withLBD, the cost is dependent on the stock ofknowledge, which is proxied by the stock ofcumulative past production. In the standard modelof LBD, the firm today bears the up-front cost ofproducing an additional unit and thereby alsoincreasing the knowledge stock, while all firms inthe industry benefit from the increased stock ofknowledge, leading to reduced costs in the futurefor all firms—an intertemporal spillover.

Importantly, LBD alone does not necessarilyimply a market failure. In some situations, onecould imagine that all knowledge leading to costreductions could be used only by the single firmmaking the decision. In this special case, there areno spillovers, and the firm would have the incen-tive to produce optimally, weighing the up-frontcost of learning against the benefits of the costreductions in the future as it would any invest-ment decision.11

Outside of this special case, the existence ofLBD can represent an externality with the poten-



JOBNAME: Earthscan−renewenerg PAGE: 11 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 785F8C1B/production/earthscan/books/renewenergy/chap05

tial to be an important market imperfection inrenewable energy markets. There is little or noempirical evidence on the degree of spilloversfrom LBD, but ample evidence exists that the costof several important renewable energy technolo-gies tends to decline as cumulative productionincreases (Jamasb 2007). This evidence alone doesnot prove the existence of a market failure, forother factors may also be able to explain the costdecreases (e.g., R&D or even time-dependentautonomous cost decreases).

The magnitude of a LBD market failure isspecific to each technology, and each has to beassessed on a case-by-case basis. Moreover, muchlike R&D spillovers, LBD spillovers are notunique to renewable energy technologies, butmay also be present in any number of fledglingtechnologies as they diffuse into the market.Hence both R&D and LBD spillovers can beconsidered broader innovation market failures thatlead to underinvestment in or underproduction ofcertain renewable energy resources.

Network externalitiesNetwork externalities occur when the utility anindividual user derives from a product increaseswith the number of other users of that product.The externality stems from the spillover one user’sconsumption of the product has on others, so thatthe magnitude of the externality is a function ofthe total number of adoptions of the product.Often quoted examples of network externalitiesinclude the introduction of the “QWERTY”typewriter keyboard, telephone, and fax machine(David 1985).

An important caveat about network externali-ties is that the externality may already be internal-ized. For example, the owner of the network mayrecognize the network effects and take them intoaccount in his or her decisionmaking (Liebowitzand Margolis 1994). Alternatively, in some cases,the recipients of the network spillover may be ableto compensate the provider. For example, for net-work effects in home computer adoption, thenew adopter might take the previous adopter tolunch as thanks for teaching him or her how touse the computer (Goolsbee and Klenow 2002).When the externality is already internalized, net-

work externalities are more appropriately titled“network effects” or “peer effects” and do notlead to market failures (Liebowitz and Margolis1994).

In the context of renewable energy, networkexternalities may play a role in the adoption ofdistributed generation. This may come about ifconsumers believe that installing renewableenergy systems on their homes sends a message totheir neighbors that they are environmentallyconscious—and that more installations in theneighborhood will increase this “image motiva-tion” or “snob effect.” Evidence for this effect hasbeen shown in Sacramento for solar panels(Lessem and Vaughn 2009). Little evidence isavailable to indicate whether this is truly a net-work externality or just network effects in distrib-uted generation renewable energy.

Policy InstrumentsEach of the failures described above providesmotivation for policy to correct the failure, but itis not always a simple task to appropriately matchthe policy to the failure.Table 5.1 lists some of themore common classes of policy instruments avail-able to address failures relevant to renewableenergy. This table is meant to be illustrative, asthere exist an almost uncountable variety of dif-ferent policy instruments.

How do we choose among the policy instru-ments? Economic theory along with carefulanalysis can provide guidance. First, both theoryand evidence indicate that multiple market fail-ures will likely require multiple interventions, so asensible policy goal involves matching the mostappropriate intervention to the failure (Aldy et al.2009; Goulder and Schneider 1999). In somecases, several policy instruments can address, orpartly address, a given market failure. In thesecases, if economic efficiency is the goal, the com-bination of policy instruments that provides thegreatest net benefits should be chosen. In addi-tion, many of the market failures relevant torenewable energy are broader market failures thatmay apply to a wide range of markets or tech-nologies. Therefore, economic efficiency would



Kenneth Gillingham

Inserted Text

can

Kenneth Gillingham

Cross-Out

JOBNAME: Earthscan−renewenerg PAGE: 12 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: A034E63F/production/earthscan/books/renewenergy/chap05

be further enhanced if the interventions to addressthese market failures were not focused solely onrenewable energy.

Several concerns warrant careful attention inthis matching process. First, we care about howeffective the intervention will be at actually cor-recting the market failure. Second, the benefitsfrom the intervention must be weighed againstthe costs of implementing the policy, includingboth government administrative costs and indi-vidual compliance costs—taking into account therisk of poor policy design or implementation. Inaddition, careful consideration of any equity ordistributional consequences of the intervention isimportant, both for ethical reasons and for gainingthe political support for passage of the policy.

Uncertainty about the magnitude of the mar-ket failure and the effectiveness of the interven-tions is another important concern. In some cases,

potential damages from a market failure may belarge enough that the most sensible intervention isdirect regulation, so that we can be certain therisk is mitigated. For example, if a toxin is deemedto have sufficiently high damages with a highenough probability, it may be sensible for the gov-ernment to simply ban it. A comprehensive analy-sis of the costs and benefits of different policyoptions that explicitly includes uncertainty, in thiscase could be expected to reveal the need for sim-ply banning the toxin.

Finally, the temporal structure of the marketfailures may have a profound influence on thetemporal structure of optimal intervention. Eco-nomic theory suggests that not only should anintervention be matched to the failure, but alsothe temporal pattern of the intervention shouldbe matched to the temporal pattern of the failure.For example, the optimal correction for failures

Table 5.1. Some potential policy instruments

Direct regulation Command-and-control methods (e.g., requiring firms to generate electri-city from renewable energy resources)

Direct government-sponsored R&D Government funding for scientists and engineers working on improvingdifferent renewable energy technologies, support for national laboratories,funding research prizes such as “X prizes”

R&D tax incentives Subsidies for private renewable energy technology R&D

Instruments to correct market prices:excise taxes, cap-and-trade, subsidies

“Get prices right” by adding to the cost of goods (e.g., through a tax or apermit price) or reducing the cost of goods (e.g., through a subsidy)

Feed-in tariffs Require electric utilities to purchase electricity from other generators (oftensmall renewable energy generators) at a specified price

Information programs Education campaigns and required labels

Product standards Require firms to improve their product characteristics to meet a specifiedgoal (e.g., efficiency of solar PV cell or energy efficiency of lighting)

Marketable marketwide standards:renewable portfolio standards, low-carbon fuelstandards, corporate average fuel, economystandards

Require firms (e.g., utilities) to meet a specified standard (e.g., produce aspecified amount of electricity from renewables) or purchase permits orcertificates from other firms that overcomply with the standard

Transparency rules Require firms to provide more information about their current conditionsto investors

Macroeconomic policy Fiscal or monetary policies to stabilize the economy and provide liquidity tomarkets to reduce credit constraints

Corporate taxation reform Adjusting the corporate income tax to improve corporate incentives

Competition policy/laws Reduce the exercise of market power through antitrust action

Restructured regulation Reduce regulatory failures and loopholes in regulations that allow for mar-ket power

Intellectual property laws Laws to encourage innovation by allowing innovators to appropriate thebenefits of their work



JOBNAME: Earthscan−renewenerg PAGE: 13 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 304EE95E/production/earthscan/books/renewenergy/chap05

that decrease in magnitude and eventually vanishover time would be a transient intervention.

Table 5.2 summarizes the matching, with thevarious market failures listed as rows and policyinstruments from Table 5.1 as the columns. Forthose instruments that appear to be potentiallywell matched to the market failures, the letters Pand T indicate whether the instrument could beexpected to be permanent or transient. Of course,the particular circumstances of each market failureand the potential policy must be assessed. Somepotential policies may be useful only under lim-ited circumstances, and the evidence for somemarket failures in renewable energy is muchweaker than others. Moreover, some of the policy

options listed may be reasonably well matchedwith a market failure but may be second best toother policy options.

Policy Instruments for Atemporal(Flow-Based) Deviations

Atemporal deviations lend themselves to policyinterventions that vary, perhaps greatly, withchanging external conditions. If the underlyingmarket deviation is a continuing problem, thenthe policy interventions can be expected to have arelatively permanent nature. If the deviation istransient, the appropriate policy interventionwould likewise be transient.

Dire

ct re

gula

�on

Dire

ct g

over

nmen

t-sp

onso

red

R&D

Com

pe��

ons,

suc

h as

X p

rize

R&D

tax

ince

n�ve

s

Exci

se ta

xes

Prod

uc�o

n su

bsid

ies

Feed

-in ta

riffs

Info

rma�

on p

rogr

ams

Prod

uct s

tand

ards

Cap-

and-

trad

e

Mar

keta

ble

mar

ketw

ide

stan

dard

s

Tran

spar

ency

rule

s

Mac

roec

onom

ic p

olic

y

Corp

orat

e ta

xa�o

n re

form

Com

pe��

on p

olic

y/la

ws

Rest

ruct

ured

regu

la�o

n

Inte

llect

ual p

rope

rty

law

Labor supply/demand imbalances T TEnvironmental externali�es P P P P P P P P P P P PNa�onal security externali�es P P P P P P P P P P P PInforma�on market failures P P PRegulatory failures PToo-high discount rates P P PImperfect foresight P Economies of scale T T PMarket power P P PR&D spillovers P P P P PLearning-by-doing T TNetwork externali�es T P

Table 5.2. Sources of market failure and some illustrative potential policy instruments

Note: P indicates permanent change or instrument; T indicates transient instrument



JOBNAME: Earthscan−renewenerg PAGE: 14 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 7C862773/production/earthscan/books/renewenergy/chap05

Policies for Labor MarketSupply–Demand Imbalances

Labor market unemployment in well-functioningdeveloped economies can be expected to be atransient problem, associated with economicrecession. Typically, policies are crafted at thenational or international level and focus oneconomywide transient monetary or fiscal policiesthat are terminated when the economy returns tofull employment.12 However, it is often assertedthat subsidizing new renewable technologies isadvantageous because it would create jobs.

In theory, in order to align private incentiveswith socially optimal incentives, the labor cost ofproviding renewable energy could be subsidizedby the difference between the market price oflabor and the social cost of that labor. Thus inorder to improve economic efficiency, such alabor subsidy must vary sharply over the course ofthe business cycle; be zero during times of fullemployment; and differ across employees,depending on the options facing the unemployedperson. This set of conditions may make itextremely difficult to implement such a policy.

Moreover, and perhaps most important,unemployment is an economywide phenomenon,so an equally valid argument could be made forsubsidizing labor throughout the economy,including in the fossil fuels sector. Thus the “cre-ating jobs” argument does not clearly justify tar-geting the labor subsidies to the renewable energyindustry, unless a particularly large deviationexisted between the social cost of the labor andthe market price relative to the rest of theeconomy.

Policies for Environmental Externalities

Table 5.2 lists a wide variety of different policyinstruments to address environmental externali-ties.The most straightforward of these is to simplyprice the environmental externality, following thetheory first developed by Pigou (Baumol 1972).In doing so, firms and consumers will take intoaccount the externality in their decisions of howmuch to produce and consume. The price couldbe imposed directly as a pollution tax or pollution

fee, with the optimal tax set at the magnitude ofthe externality. Or a cap-and-trade system couldimpose a marketwide limit on the emissions, inwhich case trading of the allowances under a cap-and-trade system would lead to a market-clearingprice for the allowances. The cap should be set sothat the resulting permit price is equal to the mag-nitude of the externality.13 The magnitude of theexternality can be estimated based on damageestimates from scientific and economic literature.

In some cases, the risk from particularly severepollutants (e.g., possibly some criterion air pollut-ants) may be sufficiently high that the marginaldamage associated with the release of the pollut-ant would always exceed the economic costs ofreducing that pollution—implying that directregulation could be an economically efficientpolicy. Direct regulation would entail the govern-ment setting strict limits on the amount of thesevere pollutant emitted or, in some cases, possiblyeven banning emission of the pollutant entirely.

Environmental externalities from renewableenergy can be treated the same way as environ-mental externalities from fossil-fuel combustion.For most renewables, the environmental exter-nalities are small, so the appropriate tax would besmall. A few, such as corn-based ethanol and palmoil biodiesel, may have significant emissions ofsome pollutants, and the damages from theseshould be added to the price of the resource.

A second tax or subsidy approach to address-ing environmental externalities more closely fol-lows the policies in many countries. Rather thanputting a price on both fossil fuel and renewableenergy generation corresponding to the magni-tude of each externality, the same cost differentialcould be maintained by subsidizing low-emittingresources and not subsidizing (or taxing) high-emitting resources. However, this approach wouldhave the unintended consequence of makingenergy use less expensive than its actual socialcost, because the external costs would remainunpriced. An additional subsidy on energy effi-ciency investment can correct for the overuse ofenergy, removing the primary distortion inenergy markets. Unfortunately, this may still leadto a distortion through an overinvestment in thesubsidized energy-efficient technologies, because



JOBNAME: Earthscan−renewenerg PAGE: 15 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 7CC9382E/production/earthscan/books/renewenergy/chap05

the optimal choice may have involved moreenergy conservation and less investment inenergy-efficient technologies.14 In addition, sucha combination of subsidies may lead to furtherdistortions in factor markets, such as markets forthe inputs in the production of energy efficiencyequipment. Thus the economic theory suggeststhat the first-best approach to addressing environ-mental external damages is through taxes or per-mits, and the subsidy approach outlined above canbe considered a second-best approach to be pur-sued if the first-best approach is not politicallyfeasible.

Other approaches rely on the idea that if firmsmust clearly disclose their environmental impacts,they will be motivated to reduce those impacts,and consumers will be motivated to shift theirpurchases away from damaging products andtoward those that are environmentally benign.Information programs designed to publicize theenvironmentally damaging product or transpar-ency rules designed to document and communi-cate the environmental damages are prompted bythis idea. Enterprise software available from com-panies such as Hara Software15 have made it pos-sible to document and broadly communicate car-bon dioxide and other environmental impacts in atransparent manner.

Policies for National Security Externalities

Each of the policy instruments available forresponding to environmental externalities is alsoavailable for responding to national security exter-nalities. Again, the first-best policy interventionworks by getting prices right. By pricing theexternal costs imposed by the consumption of thefuel, firm and consumer decisions will take intoaccount the externality, resulting in an economi-cally efficient outcome. In this case, getting theprice right inherently involves taking into accountthe full external effect, such as the externality thatone country’s spending an extra dollar on defensecauses other countries to spend more on defense.With a correct price on the fuel, firms and con-sumers will substitute other energy resources thatdo not lead to national security risks, such as coalor renewable energy, or will find ways of reducingenergy use.

Just as for environmental externalities, a sec-ond approach would be based on maintaining aprice differential between fuels with high and lownational security external costs. This approachwould face the same issues: overuse of fuel withhigh external costs and overuse of energy in gen-eral. Policies to subsidize energy efficiency wouldhelp but may come at the cost of distortionsthrough overinvestment in energy efficiency oroverconsumption in some factor markets.

Other policy instruments may also improveeconomic efficiency by reducing consumption ofoil, such as product standards (e.g., for fueleconomy), but these approaches inherently lead toadditional economic distortions and thus are alsonot a first-best approach. For example, fueleconomy standards lower the effective cost permile of driving and thus induce more driving, areaction known as the rebound effect. The addi-tional driving may increase the use of oil, redu-cing the energy security (and environmental)benefits and at the same time increasing the exter-nal costs from accidents and congestion.

Policies for Information Market Failures

Information market failures stem from a variety ofsources, and some may be very difficult to address.Those that lead to an underinvestment in distrib-uted generation renewable energy by householdsmay be addressed through information programsto raise awareness. Similarly, consumers typicallycannot readily obtain information about theirinstantaneous use of electricity; they normallyreceive only a monthly bill for their total electri-city use. Programs to provide households withfeedback on the price and usage of electricity(e.g., “smart” meters or in-home dashboards todisplay instantaneous energy use) can help con-sumers make more informed choices relating touse of energy. Both feedback and informationprograms may also reduce behavioral failures, pos-sibly providing an additional benefit.

For interventions to address imperfect infor-mation, such as imperfect foresight for firms, theintervener—presumably a government agency—would have to possess better information. In situ-ations in which a government agency has superior



JOBNAME: Earthscan−renewenerg PAGE: 16 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 76EFBD0E/production/earthscan/books/renewenergy/chap05

knowledge, such as of future probable energyprices, an obvious intervention is for the agencyto share that information. The Energy Informa-tion Administration of the U.S. Department ofEnergy provides exactly such data and projectionsaccessible to anyone. In fact, given the ability for agovernment agency to share informationbroadly,16 and at low cost, it is very unlikely thatimperfect foresight about future energy condi-tions would provide a strong case for other gov-ernmental interventions.

In cases when information is particularly diffi-cult to process or a principal-agent issue exists,consumers may be unable to make informed deci-sions, suggesting that the government canimprove economic efficiency by using its superiorinformation-processing ability to make sensiblechoices.This reasoning underlies appliance energyefficiency standards and may perhaps pertain todistributed generation renewables in limited cases.

If managerial incentives are misalignedbecause of the imperfect knowledge of stock mar-ket investors, accounting and information rules topromote transparency and a clearer flow of infor-mation may be warranted. These accounting andinformation rules are not specific to renewableenergy and may also improve economic efficiencyin general.

In addition, if the managers of some firms takea short-term perspective and underinvest inrenewable energy, then other firms with a longer-term perspective may invest more to take advan-tage of the long-term profit opportunities. Ifother firms with a longer-term perspective do notstep in, then there may be motivation for publicsupport for R&D, either through public R&D orsubsidies for private R&D.This may not be a verylikely outcome, but it could occur in the presenceof behavioral failures on the part of stock marketinvestors that lead to a systematic bias towardrewarding short-term performance.

Policies for Regulatory Failures

Policy interventions to reduce regulatory failuresinvolve simply changing the regulatory structureto reduce perverse incentives. For example, toimprove on average cost pricing of electricity,

real-time pricing (RTP) of electricity at thewholesale level could be expanded to the retaillevel.17 However, the benefits of RTP or time-of-use (TOU) pricing would have to be weighedagainst the technology and implementation costs,however.

Policies for Too-High Discount Rates

If the discount rate is too high because of thecorporate income tax, then the failure here is aregulatory failure, and the appropriate responsewould be a tax reform. One tax reform thatwould alleviate this issue would be to allow for theexpensing of capital investments. Other optionsinclude accelerated depreciation for investments,tax credits for research and development, or theelimination of the corporate income taxentirely.18 These issues are not particular torenewable energy development, however, and adeeper examination of tax reform is beyond thescope of this chapter.

If the discount rates are too high because ofcredit limitations, then the appropriate govern-ment response involves macroeconomic policyactions, primarily by the central bank. Both taxreforms and macroeconomic policy actions areeconomywide policy actions that may affectrenewable energy, but they are unlikely to have adisproportionate effect on the renewable energysector in particular.

Policies for Imperfect Foresight

If the evidence is sufficiently strong that a system-atic bias exists as a result of imperfect foresight,this would imply a variety of government inter-ventions designed to provide information aboutpossible future states of the world in order toimprove long-term decisionmaking. Governmentinformation programs that involve data collectionand possibly forecasting reports may help alleviatethe systematic bias by improving firms’ ability topredict future conditions. Increasing regulatoryconsistency by governments implementing clear,predictable long-term renewable energy policiescould also help improve long-term decision-making by firms.



JOBNAME: Earthscan−renewenerg PAGE: 17 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 788E23BA/production/earthscan/books/renewenergy/chap05

Policies for Economies of Scale

Although economies of scale are not likely to playa significant role for renewable energy in general,it may play a role in specific areas. One approachto address economies of scale would be a tempo-rary direct subsidy sufficient to induce firms toproduce at the higher level. Once a sufficientlyhigh level of production is achieved such that thepositive competitive equilibrium can be reached,the subsidy can be removed. As indicated above,because in many cases firms can individually over-come problems of economies of scale, it isunlikely that such approaches are in fact needed.

Policies for Market Power

For market power relating to the possibility offirms buying out competing technologies, possi-bly including renewable energy technologies,enforcement of antitrust laws is likely to be themost effective intervention. In some cases, verticaldisintegration may be warranted to ensure a com-petitive market. Direct government subsidies forprivate R&D investment, coupled with limita-tions on the sale of the subsidized company, areanother possible alternative to address marketpower.

For market power motivating utilities to favortheir own generation over that from outside sup-pliers, a feed-in tariff or equivalent policy mayincrease economic efficiency if the price is setappropriately.The appropriate price would be thewholesale market price for electricity, adjusted forrisk and intermittency. Such a price would pre-vent utilities from favoring their own generation,but it also would prevent any distortions from aprice that does not correspond with the market.19

As an alternative to a feed-in tariff, regulatorscan restructure utilities to ensure that they do notfavor their own generation over that from outsidesuppliers. Alternatively, careful oversight of utili-ties by public utility commissions can also helpaddress market power.

Policies for Intertemporal(Stock-Based) Deviations

Intertemporal deviations are those in which theexternal costs are based primarily on a stock that

changes over time. Individuals influence thesestocks only indirectly, by altering the flows into orout of the stock. But once the flow is determined,those individuals have no further control of thestock. For that reason, policy instruments cannotbe directed toward the stock but must be directedtoward influencing the flow. For economic effi-ciency, the strength of the incentives must beguided by the intertemporal nature of the stockexternality and can be determined from the dis-counted net present value of the entire flow offuture impacts.

Policies for Stock-Based EnvironmentalExternalities: Carbon Dioxide

Carbon dioxide is perhaps the most importantstock-based environmental externality, and there-fore the following discussion focuses on this pol-lutant, but a similar result would hold for anystock-based pollutant.

In any given year, a firm can alter the amountof carbon dioxide it releases into the atmosphere,but once the carbon dioxide is released, the firmhas no further control. That additional carbondioxide remains in the atmosphere, increasing thestock of carbon dioxide for the next century. Theeconomically efficient carbon price in any givenyear (e.g., 2010) can be determined by taking thedamages each subsequent year and discountingthem back to the chosen year using the socialdiscount rate. In this sense, the optimal carbonprice is still the magnitude of the external cost,just as with atemporal environmental externali-ties.

The calculated optimal carbon price differs byyear (e.g., in 2010 compared with 2020) in threeways. First, and most important, the damages arediscounted back further at the earlier date, imply-ing that the carbon price is lower at the earliertime.20 Second, some damages occur during thetime between the two dates. Depending on thedamage function, this difference may be small(perhaps as it is between 2010 and 2020) and maynot change the increase over time of the optimalcarbon price.Third, a natural rate of decline of thestock occurs from dissipation of emissions in the



JOBNAME: Earthscan−renewenerg PAGE: 18 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 7599602B/production/earthscan/books/renewenergy/chap05

atmosphere. This fact would slightly increase therate of growth of the optimal carbon price.

The magnitude of the damages from carbondioxide is controversial, and estimates willimprove with increased scientific knowledge. Dif-ferent time patterns of damages would lead todifferent time patterns of the carbon price, basedon the three points discussed above. For example,if the damages from an additional metric ton ofcarbon dioxide grow (in real terms) at the socialdiscount rate, then the optimal carbon price willalso grow (in real terms) at approximately thesocial discount rate. Under the unlikely assump-tion that the incremental damages are constant inreal terms into the future, we could find a nearlyconstant (in real terms) economically efficientcarbon price.

Policies for Imperfect Capture of FuturePayoffs from Current Actions

R&DWhen a market failure occurs as a result of R&Dspillovers from imperfect property rights inknowledge generation, several possible govern-ment interventions might increase economic effi-ciency. The government could directly subsidizeprivate R&D to bring the private rate of returnfrom R&D closer to the social rate of return, anexample of getting the prices right. Such a subsidywould continue as long as a deviation existsbetween the private and social rate of return, andperhaps indefinitely. The economically efficientsubsidy would be set equal to the present dis-counted value of the spillovers from R&D.Importantly, R&D spillovers are likely to exist inmore than just the renewables sector, so an appro-priate policy would also provide the subsidy toprivate R&D in these other sectors.

The government could also directly fundR&D in sectors where spillovers are particularlyhigh. For example, the U.S. government directlyfunds research in renewable energy in nationallaboratories, universities, and some research insti-tutes. Theoretically, public R&D can improveeconomic efficiency if it is focused on researchareas where the social rate of return is sufficientlyhigh relative to the private rate of return. In these

cases, little R&D would have been undertaken byfirms relative to the economically efficientamount, so public R&D complements privateR&D. On the other hand, public R&D cancrowd out the private, depending on the nature ofthe R&D. For example, pure science public R&Dwould be much less likely to crowd out privateR&D than would demonstration projects. Theempirical evidence on public R&D is not clear-cut. David et al. (2000) review the empirical evi-dence on whether public R&D complements orcrowds out private R&D and find an ambiguousresult, suggesting that the result is situation-dependent and underscoring the importance ofthe nature of the R&D in the social rate of returnof public investment.

Intellectual property law plays a key role inhow well firms can capture the rents from theirinnovative activity. Determining the direction inwhich to change intellectual property law is notsimple. If intellectual property law is tightened(e.g., by increasing the length of time patents holdforce), then two opposing effects would exist.Firms could capture more of the benefits of R&Dand thus would have a greater incentive to investin it. But fewer spillover benefits may result fromthe R&D activity, so the social rate of return fromthe activity would be lower. Little empirical evi-dence is available to suggest that either a tighten-ing or loosening of intellectual property lawwould increase economic efficiency.

LBDIf learning by doing occurs in the production of anew technology (e.g., solar photovoltaic installa-tions), then the act of producing increases thefirm’s stock of cumulative experience and thusleads to declines in future costs. The stock ofcumulative experience grows when insights fromprevious production by that or another firmallows it to improve its production techniques.The stock also may decline if some of these tech-niques are forgotten. Theories of LBD oftenproxy all of these complex dynamics by postulat-ing that the cost of future production for all firmsat any time will be a function of the cumulativestock of experience from production in the mar-ket. But the market failure can be thought of in a



JOBNAME: Earthscan−renewenerg PAGE: 19 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 50137E67/production/earthscan/books/renewenergy/chap05

more general sense as a spillover from the stock ofany single firm’s cumulative experience from pro-duction to other firms.

Once a firm chooses how much to produce atany given time (i.e., the flow into the stock ofexperience), it subsequently has no further con-trol over the stock of experience and its impact onfuture costs. Thus economically efficient policiesfor LBD must focus on the quantity produced,while taking into account the fact that experienceis a stock. The most straightforward policy instru-ment to address LBD spillovers is a subsidy. Theeconomically efficient per-unit subsidy equals thediscounted present value of all future cost reduc-tions resulting from the additional production thatcannot be captured by the individual firm.

However, a second element to the economictheory behind LBD also may affect the economi-cally efficient policy: the spillovers from LBD arepostulated to decline along with the costs. Forexample, in the standard formulation, as illus-trated in Figure 5.2, the percentage cost decreasedepends on the percentage increase in the stock,so every additional unit has a progressivelydecreasing percentage impact on costs. Conse-quently, as cost decreases with a greater stock, agiven percentage cost decrease leads to a smallerabsolute cost decrease. These factors togetherimply that the LBD externality—and hence the

appropriate magnitude of the intervention—willbe declining over time. Acting in the oppositedirection, if the sales are growing rapidly, the costreduction is applied to a larger amount of produc-tion, reducing the rate of decline of the interven-tion.

Thus optimal subsidies for LBD will likely betransient and decline over time as LBD runs itscourse. The speed at which the subsidies arephased out will depend on the particular technol-ogy and may require adjustment if different con-ditions arise than were initially expected. In oneexample, the optimal solar PV subsidies for Cali-fornia calculated under the baseline assumptionsin van Benthem et al. (2008) follow a decliningpath and are phased out over 15 years.

Network externalitiesNetwork externalities may play a role in theadoption of distributed generation renewableenergy. If it can be demonstrated that a networkexternality truly exists, rather than networkeffects, then one approach to correct for thisexternality would be a temporary production sub-sidy (Goolsbee and Klenow 2002). Once a prod-uct has taken over (nearly) the entire market, therewould be no room for further spillovers and thusno need for the subsidy policy.

Installationsin 2010Discountfuturecumulativebenefitsto 2010 Installations

in 2015Discountfuturecumulativebenefitsto 2015

Installationsin 2020Discountfuturecumulativebenefitsto 2020

2010 2012 2014 2016 2018 2020Year of benefit

2022 2024 2026 2028 2030

Figure 5.2. Illustrative incremental benefits from additional cumulative installations: LBD



JOBNAME: Earthscan−renewenerg PAGE: 20 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 7A3D9453/production/earthscan/books/renewenergy/chap05

ConclusionsRenewable energy has an immense potential toserve our energy needs, and in the long run, atransition from depletable fossil fuel resources torenewable energy is inevitable. This chapter hasdelved into reasons why policymakers should beinterested in policies to promote renewableenergy, pointing to a variety of market failuresthat may lead to a divergence between the optimaltransition to renewables and the observed transi-tion. Economic theory suggests that we canimprove economic efficiency by matching thepolicy instrument to the market failure.

The structure and nature of each market fail-ure have important ramifications for the appropri-ate policy actions to correct for the market failureand move closer to an optimal transition torenewable energy. The discussion above distin-guished between atemporal market failures andintertemporal (i.e., stock-based) market failures.In either case, the economically efficient policyaction matches the temporal pattern of the marketfailure.This implies a temporary policy (e.g., LBDspillovers) in some cases and a permanent policy(e.g., R&D spillovers) in others.

Renewable energy policy is likely to requireseveral different policy instruments to address thevarious kinds of market failures. When the marketfailures are closely related, a single policy instru-ment can address, or partly address, more thanone market failure. For instance, provision ofinformation about low-cost or low-effort oppor-tunities to save energy and help preserve the envi-ronment may reduce the informational marketfailure and also influence consumers to partlyinternalize the environmental externalities(Bennear and Stavins 2007).

For renewable energy, the most importantmarket failures, with the strongest empirical evi-dence, appear to be environmental externalities,innovation market failures, national security mar-ket failures, and regulatory failures. Only a few ofthe market failures identified in this chapter areunique to renewable energy. Environmentalexternalities due to fossil-fuel use are the mostimportant of these, but if policy action is alreadyunder way to correct for externalities from fossil-

fuel emissions, then we must look to other marketfailures for motivation for renewable energypolicy. As these other market failures often applyto other parts of the economy, addressing themmay entail policy actions that extend muchbeyond renewable energy.

Political feasibility is a final consideration withimportant ramifications for renewable energypolicy. In some cases, the first-best policyapproach may not be politically feasible. Asecond-best approach may involve multipleinstruments, even in cases when the first-bestapproach involved only a single instrument. Forexample, rather than a single tax to internalizeenvironmental externalities, the same price differ-ential can be achieved by combining a smaller tax(or no tax) on fossil fuels with a subsidy forrenewable energy. Similarly, a cap-and-trade sys-tem may not be politically feasible because ofuncertainty about how high the costs of abate-ment might be, so a more viable option might beto use two instruments in a hybrid cap-and-tradeand tax system, commonly known as a cap-and-trade with a safety valve (Jacoby and Ellerman2004; McKibbin and Wilcoxen 2002; Pizer 2002).

In other cases, the only politically feasibleoptions for addressing the market failures relevantto renewable energy are not the first-choiceinstruments, but rather second-choice instru-ments that address the market failures indirectly.Renewable portfolio standards (RPSs) are one ofthe most prominent examples of a policy instru-ment that only indirectly addresses the marketfailures relevant to renewable energy. By setting arequirement on the amount of renewable energyin each utility’s electricity generation mix, anRPS adds an implicit subsidy on renewableenergy, with the magnitude of the subsidy directlyrelated to the stringency of the cap. If the RPS iscarefully set, this implicit subsidy could act justlike an appropriately set actual subsidy—leadingto the second-best outcome described above.However, finding this appropriate level for anRPS may be exceedingly difficult and subject tointense political disputes.

A sensible set of policies to address the marketfailures relevant to renewable energy has thepotential to greatly improve economic effi-



JOBNAME: Earthscan−renewenerg PAGE: 21 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 791DBC6F/production/earthscan/books/renewenergy/chap05

ciency—and at the same time would have otherbenefits, such as improving air quality and miti-gating the risk of catastrophic global climatechange. Much future work remains to betterquantify the most relevant market failures and fur-ther improve our understanding of how todevelop policies to best address these market fail-ures.

AcknowledgmentsThe authors thank Richard Schmalensee, BoazMoselle, Arthur van Benthem, Douglas Hannah,and Jonathan Leaver for very helpful comments.Any remaining errors are the sole responsibility ofthe authors.

Notes

1. The concept of behavioral failures stems frombehavioral economics and is quite new to environ-mental economics. See Shogren andTaylor (2008)and Gillingham et al. (2009) for recent reviewsdiscussing the concept in the context of environ-mental economics.

2. Economic theory defines “economically efficient”in technical terms as an allocation of resourceswhere no potential Pareto improvement exists,which refers to a reallocation of resources thatbenefits at least one individual and imposes nocosts on any others. Note that economic effi-ciency is a distinct concept from the equity orfairness of an allocation of resources.

3. It is still theoretically unclear how to disentanglesystematic biases in decisionmaking from inherentpreferences, but behavioral welfare analysis is anarea of active theoretical development and mayeventually shed light on this issue. See, e.g.,Bernheim and Rangel (2009).

4. Important equity or fairness concerns may also beinvolved. This chapter focuses on economic effi-ciency as a policy goal, while noting that equityconsiderations, in theory, can often be dealt withthrough lump-sum transfers of wealth that do notdistort incentives or through modifications of theincome tax rates. If the policy goal is reducing

global inequity, other distributional policies arelikely to be more effective than renewable energypolicy.

5. It is important to note that unless a behavioralfailure is a systematic, rather than random, depar-ture of observed choice from a theoretical opti-mum, it may be very difficult to formulate poli-cies. If the systematic departure is in a consistentdirection, the intervention can work in the oppo-site direction to correct this deviation. But ran-dom deviations would require an interventioncontingent on the deviation. For example, poorinformation about the operating characteristics ofdistributed photovoltaics could lead some peopleto install these devices even though they wouldultimately come to regret the decision and otherpeople not to install them even though the deviceswould have turned out to be beneficial. In suchcircumstances, development and dissemination ofinformation about photovoltaic operating charac-teristics for alternative locations could improvesuch decisions. For the most part, however, policyoptions designed to compensate for randomdeviations would be difficult to formulate andeffectively implement.

6. “Full employment” for a well-functioning devel-oped economy refers here to “at the natural rate ofunemployment.” Some unemployment will alwaysexist as a result of transitions between jobs andmismatches of available and needed skills.

7. Personal income taxes or labor taxes, such as theU.S. Social Security tax, provide incentives toreduce the supply of labor, so that the marginalsocial value of labor exceeds the value of that laborto the worker. However, issues of the distortionsassociated with and the reform of such tax systemsgo well beyond the scope of this chapter.

8. This remains a concern for the overall economicefficiency of investment, even when it does notdistort the mix of renewables versus nonrenewabletechnologies.

9. “Almost indifferent” because the cumulativeimpacts of emitting a ton now may be somewhatdifferent from the impacts of emitting a ton in 20years and because the regulatory environmentcould change in that period.

10. LBD is closely related to economies of scale,except that learning by doing has a distinctly dif-ferent intertemporal relationship, where costsdecline as a function of cumulative production,and increasing returns to scale implies that averagecosts decline with production at a given time.



JOBNAME: Earthscan−renewenerg PAGE: 22 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 896A624E/production/earthscan/books/renewenergy/chap05

11. If the knowledge leading to cost reductions by onefirm could have been used by other firms, but thatfirm somehow manages to keep all the knowledgeprivate, even if it does not use it (so it is effectively“wasted knowledge”), there would still be a mar-ket failure in that some of the potential benefits ofthe learning would not be captured by anyone.

12. As noted above, this does not deal with the labormarket problems associated with income taxes orlabor taxes that provide incentives to reduce thesupply of labor.

13. A substantial literature addresses the trade-offsbetween a tax and a cap-and-trade system, par-ticularly relating to policymaking under uncer-tainty. For a recent review discussing these issues,see Aldy et al. (2009).

14. If behavioral failures have caused an under-investment in energy efficiency, then there maynot be an overinvestment in energy efficiencyfrom the subsidy.

15. Information available at www.hara.com.16. In some cases, release of information is not possi-

ble or is undesirable, such as when it involvesweapons programs or other programs closelyrelated to national security.

17. In order to be effective, RTP would have to becomplemented with real-time feedback on theelectricity price at the current time.

18. Note that reducing taxes in some areas mayrequire increasing taxes in others, so a full analysisshould examine the relative distortions from eachof the different taxes.

19. For example, in the United States, the Public Util-ity Regulatory Policies Act (PURPA) of 1978required electric utilities to buy power from small-scale nonutility producers at the avoided cost rate,which is the cost the utility would incur were it toacquire the electricity from other sources. Choos-ing the appropriate price turned out to be remark-ably problematic.

20. For instance, the 2020 efficient tax would begreater than the 2010 efficient tax by a factor of (1+ r)10, where r is the annual social discount rate. Ata 5% discount rate, the 2020 carbon tax would be63% greater than the 2010 carbon tax.

ReferencesAghion, P., N. Bloom, R. Blundell, R. Griffith, and P.

Howitt. 2005. Competition and Innovation: AnInverted-U Relationship. Quarterly Journal of Eco-nomics 120: 701–728.

Aldy, J., A. Krupnick, R. Newell, I. Parry, and W. Pizer.2009. Designing Climate Mitigation Policy. Discus-sion paper 08–16. Washington, DC: Resources forthe Future.

Archer, C., and M. Jacobson. 2005. Evaluation of Glo-bal Wind Power. Journal of Geophysical Research D110: 1–20.

Arrow, K. 1962. The Economic Implications of Learn-ing by Doing. Review of Economic Studies 29: 155–173.

Baumol, W. 1972. On Taxation and the Control ofExternalities. American Economic Review 62: 307–322.

Bennear, L., and R. Stavins. 2007. Second-BestTheoryand the Use of Multiple Policy Instruments. Environ-mental and Resource Economics 37: 111–129.

Bernheim, D., and A. Rangel. 2009. Beyond RevealedPreference: Choice-Theoretic Foundations forBehavioral Welfare Economics. Quarterly Journal ofEconomics 124: 51–104.

Blundell, R., R. Griffith, and J. Van Reenen. 1999.Market Share, Market Value and Innovation in aPanel of British Manufacturing Firms. Review of Eco-nomic Studies 66: 529–554.

Bohi, D., and M. Toman. 1996. Economics of EnergySecurity. Norwell, MA: Kluwer Academic Publish-ers.

Borenstein, S. 2008. The Market Value and the Cost ofSolar Photovoltaic Electricity Production. CSEMWorking paper 176. Berkeley, CA: University ofCalifornia Energy Institute.

Cherni, J., and J. Kentish. 2007. Renewable EnergyPolicy and Electricity Market Reforms in China.Energy Policy 35: 3616–3629.

Da Rosa, A. 2005. Fundamentals of Renewable EnergyProcesses. Amsterdam; Elsevier Academic Press.

David, P. 1985. Clio and the Economics of QWERTY.American Economic Review 75: 332–337.

David, P., B. Hall, and A.Toole. 2000. Is Public R&D aComplement or Substitute for Private R&D?Research Policy 29: 497–529.

EIA (Energy Information Administration). 2008. Inter-national Energy Annual 2006. Washington, DC:Department of Energy, EIA.

Gillingham, K., R. Newell, and K. Palmer. 2009.Energy Efficiency Economics and Policy. AnnualReview of Resource Economics 1: 597–619.

Goolsbee, A., and P. Klenow. 2002. Evidence onLearning and Network Externalities in the Diffusionof Home Computers. Journal of Law and Economics35: 317–343.



JOBNAME: Earthscan−renewenerg PAGE: 23 SESS: 34 OUTPUT: Wed Jul 14 08:17:26 2010 SUM: 6D56AC65/production/earthscan/books/renewenergy/chap05

Goulder, L., and S. Schneider. 1999. Induced Techno-logical Change and the Attractiveness of CO2 Emis-sions Abatement Policies. Resource and Energy Eco-nomics 21: 211–253.

Heal, G. 1993. The Optimal Use of ExhaustibleResources. In Handbook of Natural Resource andEnergy Economics, Vol. 3, edited by A. Kneese and J.Sweeney. Amsterdam: Elsevier.

Hermann, W. 2006. Quantifying Global ExergyResources. Energy 31: 1685–1702.

Jacoby, H., and A. D. Ellerman. 2004. The Safety Valveand Climate Policy. Energy Policy 32: 481–491.

Jaffe, A., R. Newell, and R. Stavins. 2004. The Eco-nomics of Energy Efficiency. In Encyclopedia ofEnergy, edited by C. Cleveland. Amsterdam: Elsevier,79–90.

Jaffe, A., and R. Stavins. 1994.The Energy Paradox andthe Diffusion of Conservation Technology. Resourceand Energy Economics 16: 91–122.

Jamasb, T. 2007. Technical Change Theory and Learn-ing Curves: Patterns of Progress in Electricity Gen-eration Technologies. Energy Journal 28: 51–71.

Lessem, N., and R.Vaughn. 2009. Image Motivation inGreen Consumption. Working paper. Los Angeles:University of California–Los Angeles.

Levinson, A., and S. Niemann. 2004. Energy Use byApartmentTenants When Landlords Pay for Utilities.Resource and Energy Economics 26: 51–75.

Liebowitz, S., and S. Margolis. 1994. Network Exter-nality: An Uncommon Tragedy. Journal of EconomicPerspectives 8: 133–150.

McKibbin, W., and P. Wilcoxen. 2002. The Role ofEconomics in Climate Change Policy. Journal of Eco-nomic Perspectives 16: 107–129.

Murtishaw, S., and J. Sathaye. 2006. Quantifying theEffect of the Principal-Agent Problem on US Residential

Use. LBNL-59773. Berkeley, CA: LawrenceBerkeley National Laboratory.

Nickell, S. 1996. Competition and Corporate Perform-ance. Journal of Political Economy 104: 724–746.

Nordhaus, W. 2002. Modeling Induced Innovation inClimate Change Policy. In Technological Change andthe Environment, edited by A. Grubler, N.Nakicenovic, and W. Nordhaus. Washington, DC:Resources for the Future Press.

Nordhaus, W. 2009. Designing a Friendly Space forTechnological Change to Slow Global Warming.Snowmass Conference on Technologies to CombatGlobal Warming. August 2009, Snowmass, CO.

Pizer, W. 2002. Combining Price and Quantity Con-trols to Mitigate Global Climate Change. Journal ofPublic Economics 85: 409–434.

Pizer, W., and D. Popp. 2008. Endogenizing Techno-logical Change: Matching Empirical Evidence toModeling Needs. Energy Economics 30: 2754–2770.

Popp, D. 2006. R&D Subsidies and Climate Policy: IsThere a“Free Lunch?” Climatic Change 77: 311–341.

Rappaport, A. 1978. Executive Incentives vs. Corpo-rate Growth. Harvard Business Review 56: 81–88.

Scherer, F. 1967. Market Structure and the Employ-ment of Scientists and Engineers. American EconomicReview 58: 524–531.

Shogren, J., and L. Taylor. 2008. On Behavioral-Environmental Economics. Review of EnvironmentalEconomics and Policy 2: 26–44.

U.S. Congress. 2009. American Recovery and Reinvest-ment Act of 2009. 111th Cong., 1st sess.

van Benthem, A., K. Gillingham, and J., Sweeney.2008. Learning-by-Doing and the Optimal SolarPolicy in California. Energy Journal 29: 131–151.

Young, P. 2010. Innovation Diffusion in Heterogene-ous Populations: Contagion, Social Influence, andSocial Learning. American Economic Review 99: 1899–1924.



Market Failure and the Structure of Externalities · 2018-09-26 · Although markets are not...

Documents

Transcript of Market Failure and the Structure of Externalities · 2018-09-26 · Although markets are not...