balancing environmental protection and energy production

Balancing Environmental Protection and EnergyProduction in the Federal HydropowerLicensing ProcessLea Kosnik

ABSTRACT. The U.S. Federal Energy RegulatoryCommission (FERC) must balance environmentalprotection of riverine resources with the nation'sgrowing demand for power production every time iiissues a hydroelectric license. This paper models thebureaucratic agency's decision-making process inissuing these licenses. Data on nearly 500 hydro-power licenses issued from 1983 to 2005 are utilized.It is discovered that legislative ami institutionalconstraints are. by far, the largest influences onFERC's regulatory decisions, implying that if thecurrent allocation of surface water in the UnitedStates is considered inefficient, the most effectiveway to alter this allocation is by passing newlegislation, or by implementing institutional reformai FERC ( JEL'Q25. Q42)

I. INTRODUCTION

Energy needs and environmental con-cerns are two of the most pressing issuesfacing the world today. Frequently, theycome into conflict. We can increase domes-tic oil production, but at the cost of drillingin an arctic wildlife refuge. We can burnmore abundant coal stocks, but at the costof rising air pollution. And we can dammore rivers and generate more hydroelec-tric power, but only at the cost of inter-rupting natural river fiow and harmingaquatic resources. How are such importanttrade-offs to be made? The objective of thisresearch is to model and analyze how suchtrade-off choices between energy generationand environmental concerns are actuallybeing made, in particular, in the context ofhydroelectric power generation.

Land Economics • August 2010 • 86 (3): 444-466ISSN 0023-7639; E-ISSN 1543-8325© 2010 by the Board of Regents of theUniversity of Wisconsin System

Hydroelectric power, or electricity gener-ated by large turbines located at the bottomof reservoirs created from large dams, hasbeen a viable technology since the latenineteenth century (Edwards 2003). At thattime, when the industrial revolution was atits height and economic development of thenation was at center stage, the federalgovernment's official mandate was to en-courage hydroelectric production in theUnited States. Competing demands for theservices of the nation's rivers toward moreenvironmental ends did not begin to makeinroads into ofílcial legislative expressionuntil the late 1960s and early 1970s, afterpassage of legislation such as the NationalEnvironmental Policy Act (1969), the CleanWater Act (1972), and the EndangeredSpecies Act (1973). Today, the FederalEnergy Regulatory Commission (FERC),the agency in charge of hydropower regu-lation, has a clear mandate to balance bothpower interests and environmental con-cerns. Over the past 25 years, FERC hasbeen struggling to interpret just how, inpractice, to balance these competing inter-ests over one of the nation's most importantresources: its rivers.

FERC is able to affect this environmentalprotection/hydropower production trade-off when it issues licenses for hydropowerdams. All nonfederal hydropower dams inthe United States must be licensed every 30to 50 years in order to legally operate. This

The author is associate professor. Department ofEconomics. University of Missouri-St. Louis. The authorgratefully acknowledges the insightful comments and helpprovided by John Whitchead, John Tschirhart, LynneLewis, Trudy Ann Cameron, J. R. DeShazo. andnumerous conference participants. This research was alsogenerously funded by the University of Missouri-St.Louis internal grants program.

86(3) Kosnik: Federal Hydropower Licensing 445

licensing process is the only official oppor-tunity during the lifetime of a dam for acomplete review of its safety and opera-tions. It is also the only opportunity formandated revisions of its inherent energy/environmental operational trade-offs. Forexample, FERC could require dam ownersto limit the flow of water through the dam'sturbines. This would reduce electricityproduction, but it would also help eliminateirregular rushes of water that cause streambank erosion and damage to aquatic nestingsites. Other methods FERC could imposeto help the riverine environment (such asfish ladders, modified turbines, etc.) simi-larly tend to negatively affect electricityproduction. More environmental protec-tion implies less hydroelectric it y produc-tion. In the end, when FERC issues a licensemandating a certain level of environmentalprotection, it is implicitly choosing a certaintrade-off between environmental protectionand hydropower production. What influ-ences FERC's choice?

The significant influences on regulatoryagency decision making are still debated inthe literature. Four popular theories ofregulatory influence include regulatory cap-ture theory, congressional dominance theo-ry, discretionary theory, and theories ofinstitutional/historical constraints. I includ-ed variables on all four of these categoriesof influence in the regression analyses todetermine which, if any, have affectedFERC's environmental regulatory deci-sions when issuing hydropower licenses.The data spans 1983-2005 and includes theapproximately 500 hydropower licensesthat were issued over this period.

The findings are that, while there issupport for all four categories of influencein affecting FERC's behavior, by far thelargest influences on FERC's regulatorydecisions are Congress and institutionalhistory. Legislation passed by Congressand put into eíTect in 1990, for example,increased the level of environmental pro-tection mandated by 19%. Similarly, Senateoversight can affect the hydropower pro-duction/environmental proteetion balanceby as much as 17%. Historically, we find

that relicensing decisions are surprisinglypath dependent. Higher environmental pro-tection decisions issued by FERC in theyear before a particular license comes up forreview lead to significantly less environ-mental protection on wholly unrelatedlicenses in the following year. It is as ifFERC is trying to "balance" the environ-mental protection/hydropower productiontrade-off' year by year.

These results from past licensing deci-sions should help us predict how futurelicensing decisions will likely proceed. Evenmore importantly, these results imply that ifthe current allocation of surface water in theUnites States between its competing enduses is found to be inefficient, the mosteffective way to alter the current allocationis by passing new legislation or by imple-menting institutional reform at FERC. Thissuggests that, despite the increasing atten-tion given in the literature to interest groupeffects on regulatory behavior (Mitchener2007; Knittel 2006; Li, Qiang, and Xu 2005;Aidt 2002; Oates and Portney 2001),interest group lobbying is not as infiuentialin affecting regulatory behavior (at leastin this context) as direct congressionalaction.

. BACKGROUND

The History of

The federal govemment took its firstlegislative interest in monitoring the devel-opment of the nation's riverine resources in1899 when it passed the Rivers and HarborsAct, which recognized the interstate natureof river flows and the need for coordinatedeconomic development of these resources ata federal level. Energy generation was at thecore of the nation's burgeoning prosperity,and hydropower production was just oneimportant and growing subset of that eoreindustry. In the late 1800s permits forhydroelectric production were still beingissued by the Secretary of War, with the

' This historical summary draws on the one publishedby Kosnik (2006).

446 Land Economics August 20W

ultimate regulating authority over hydro-power development resting with the De-partment of War, but by 1920 hydropowergeneration had reached 20 bilhon kilowatthours of electric power per year (from anoperating capacity of 4,800 MW), and thenation's appetite for hydroelectric produc-tion was still far from satiated. There was aclear need for the creation of a morecoordinated and centralized federal regula-tory authority focused specifically on hy-dropower development issues.

In 1920 Congress passed the FederalWater Power Act, which was later replacedby the 1935 Federal Power Act, to stream-line the administrative processes involved inhydropower regulation and to create asingle regulatory agency focused entirelyon developing and encouraging hydropow-er production in the United States. Today,that agency is known as the Federal EnergyRegulatory Commission (FERC).

It was not until the late 1960s and early1970s, as society's preferences for protec-tion and enhancement of environmentalresources developed, that legislation waspassed that changed FERC's mandate froman exclusive focus on hydropower produc-tion, to an incorporation of protection for ariver's unique ecosystem. Prominent amongthis new legislation was the NationalEnvironmental Policy Act of 1969, whichidentified environmental protection as amajor policy objective for the federalgovernment. For FERC, this new act meantthat an envi ro timen tal assessment had to befiled for every hydropower permit request-ed, often leading to changes in the permitissued in order to protect environmentalinterests. In 1972, the Clean Water Actprohibited the discharge of pollutants intomost waterways of the United States. ForFERC. this legislation translated into arequirement for state certification of cleanwater at the project site before either aninitial license or a relicense could be issued.In 1973, the Endangered Species Actprovided a legislative foundation for pro-tecting and conserving endangered andthreatened species and their associatedhabitats. For FERC this translated into an

assessment requirement to determine whe-ther a hydroelectric project would affectany threatened or endangered species ortheir habitat, and often subsequent require-ments to alter hydroelectric production toprotect any identified species. Additionally,it required FERC to consult with the Fishand Wildlife Service or the National MarineFisheries Service in determining the impactof any hydropower project on any at-riskspecies. In 1986 Congress passed the ElectricConsumers Protection Act (ECPA), the firstpiece of legislation since 1935 to directlyamend the Federal Power Act. This impor-tant act required that "in addition to thepower and development purposes for whichlicenses are issued, [FERC] shall give equalconsideration to the purposes of enhance-ment of fish and wildlife (including relatedspawning grounds and habitat), the protec-tion of recreational opportunities, and thepreservation of other aspects of environ-mental quality."^

These acts, as well as others, forcedFERC to formally open up the regulatoryprocess and involve commentary fromoutside resource agencies and interestgroups. This interaction between FERCand outside agencies in the regulation of thehydroelectric industry led to new require-ments, now attached to the end of licensesissued, that limit water flow and hydroelec-tric production in return for environmentalenhancements such as improved waterquality, increased aquatic resources, andprotected wildlife and vegetation. FERCnow regularly seeks to balance power andenvironmental objectives in its hydro-power licensing process, the administrativedetails of which are explained in the nextsection.

Issuing a Hydroelectric Dam License

In the United States today, there are over79,000 working dams. Most of these aresmall structures with a crest elevation of lessthan 50 feet and little to no hydroelectric

^ Section 4(e) of the Federal Power Act, as amendedby the ECPA.

86(3) Ko.snik: Federal Hydropower Licensing 447

generation. Of the larger dams that remain,some are federally owned (the most familiarinclude the Hoover Dam, the Grand CouleeDam, and the Shasta Dam) and the rest arenonfederal or privately^ owned dams. Thenonfederal dams that generate hydropowercome under the regulating authority ofFERC. which since 1935 has issued thou-sands of licenses, all of which are for a termof 50 years or less. At the end of an initiallicense period, the dam owner must file for arelicense,, again not to exceed 50 years, inorder to legally continue operation."* As theapex of nonfederal dam building took placeafter World War II, and many of thesedams received initial 50 year license terms, itso happens that many dams are currentlyup for relicensing. For example, in 1983,fewer than 10 dams were up for relicensingbefore FERC, while in 1993, 157 damscame up for relicensing. Additionally,between 1998 and 2010, more than a thirdof all nonfederal hydropower dams, affect-ing over 100 different rivers, have come upfor relicensing.

Currently, there are three different pro-cedural methods for acquiring a hydroelec-tric license. The oldest and by far the mostcommon method for acquiring a newlicense is through the traditional licensingprocedure (TLP). This licensing processtakes place in four rough stages (Figure 1).The first step (Stage 1) is for the prospectivedam owner, or licensee, to file an officialnotice of intent to seek a license with

• While nonfederal dams are commoniy labeled as"private." ihis is somewhat misleading as many nonfed-eral dams are in fact owned by local or state governmentalenlilies. For example, the Azusa Dam on the San GabrielRiver in California is owned by the city of Pasadena. Thekey is ihat they are nonfcJeral and so come under thejurisdiction of FERC.

" Note that as a result of the 1984 Supreme Courtdecision in Yakima Indian Nation v. The Federal EtiergyRegulatory Commission (746 F. 2d. 466), "relicensing issubstantially equivalent to issuing an original license," Inother words, procedurally FERC is supposed to approachnew license applications and reiicense applications inessentially the same manner; we test for whether this is sowith a dummy variable on new license applications in ourempirical regressions.

' Details of this process have been published byFERC (2004),

Concurrent with the filing of thenotice of intent, technical informationregarding the project is required, includingall known and "reasonably" availableengineering, economic, and environmentalinformation. After submission of thesematerials, tentative impact studies arecarried out as the licensee begins to preparethe licensing package. Approximately threeyears later (Stage 2), the prospective damowner then submits an official licensingapplication, which FERC reviews, ap-proves, and then formalizes with a noticein the Federal Register. It is most commonlyat this point in the TLP (Stage 3) thatinterest groups concerned with a particularlicensing application become involved in theprocess by submitting comments, protests,and official requests for information and/orstudies. The licensing process is the onlyofficial opportunity during the lifetime of adam for a complete review of its safetyoperations atid environmental effects. It isalso the only opportunity for interestgroups and concerned individuals to influ-ence the operations of the dam throughFERC's formal regulatory process. Nego-tiations between the dam owner and theresource agencies should be worked outduring this phase, and either an environ-mental assessment or an environmentalimpact statement must also be completedat this juncture.

The final step (Stage 4) of the licensingprocess is scheduled to occur two years laterwhen FERC holds a hearing regarding thelicensing application and issues its regula-tory decision. FERC has three options, itcan (1) license a project for a term ranginganywhere from 30 to 50 years, (2) not issue alicense (if it is an original license applica-tion) or have the project decommissioned (if

^ Prefiling consultations (before Stage 1 begins) maytake place between ihe licensee, FERC, and publicresource agencies, in order to gel a feel for the necessarystudies and information that a license package willrequire, but this is not required, nor common, with theTLP. Part of the purpose of creating aliemaiive licensingprocedures, which are discussed below, was to formalizeand encourage this prefiling stage and make earlyconsultation and collaboration with outside agencies anintegrated part of the licensing process.

Land Economics August 2010

Stage 1 :-5 or 5'/2 yrs

Stage 2:- 2 yrs

Stage 3:-2 yrs to X yrs

Stage 4;Xyrs

FileNotice of

Intent

FileLicense

Application

ReviewProcess

RelicenseDecisionIssued

X+yrs

FIGURE 1ISSUING A HYDROELECTRIC DAM LICENSE

it is a relicense), or (3) recotntnend a federaltakeover of a project (with compensationgiven to the current owner) and have itoperated by the government.

An additional stage in the relieensingprocess can take place if one of the officialintervenors to the relicensing process choos-es to appeal FERC's relicensing decision. Insuch an instance, a review of the decision isconducted by the ftve FERC head commis-sioners. Barring finality ofthat decision, thelast word is reserved for the U.S. Court ofAppeals.

In 1998, due to increasing delays in theTLP process and contention among manyof the stakeholders involved in the licensingprocess. EERC instituted the alternativelicensing procedure (ALP). Procedurally,

^ As outlined in Figure I. the TLF is expected toaverage five years to completion, but in the 1990s thisprtKess began to get drawn out (Kosnik 2006), with somelicenses taking as long as 20 years to be issued.

the ALP involves the same steps as the TLP,but it is distinguished by at the same timeactively encouraging collaboration andprenegotiation among the stakeholders.Use of the ALP is specifically designed toincrease communication early on in theprocess (well before Stage 3) so thatultimate outcomes are less contentious,more politically legitimate, and issued withless delay. Settlement agreements amongthe interest groups and the dam owner, forexample, are heavily promoted. Between1998 and 2005 the ALP was chosen for useover the TLP 34 times.

While the ALP has, for the most part,been successful in its primary aims ofdecreasing licensing times and increasingstakeholder collaboration, use of the pro-cedure is limited. Acceptance of the ALPrequires special EERC approval, and this isnot granted if the parties to the processcannot agree ahead of time as to the successof the collaborative nature of the proce-


dure. Additionally, deadlines are not spe-cific in the ALP, but defined by the group,and this has the potemial for managerialand procedural difficulties. In the historicalprogression of the dam licensing process,the ALP is best viewed as a trial phase inreform of the TLP, as a side step along theway of reforming the TLP into what, since20()5, has become the new default licensingprocedure, the integrated licensing process(ILP).

While the ALP and the TLP can still beutilized upon FERC approval, the ILP hasbecome the default licensing procedure. Itmarries the structure of the TLP with thecollaborative negotiating encouragement ofthe ALP. Procedurally, it involves the samerough outline as illustrated in Figure 1,with official deadlines and publications inthe Federal Register, but now the precon-sultation phase is formally integrated intothe entire process. Stakeholder involvementis now required as early as Stage 1, insteadof expected only after Stage 2. Similarly, atan advisory level, FERC involvement be-gins in Stage 1, and the entire process isnavigated with a heavy emphasis on overallcollaboration, negotiation, and consulta-tion.

Reform of the licensing process from anexclusive utilization of the historical TLP,to development of the ALP and ILPalternatives, was a major procedural shiftfor FERC; as significant in its own way aspassage of the ECPA more than a decadeearlier. These two structural shifts were thetwo most fundamental reforms to thehydroelectric licensing regulatory processin its 75-year history.

The Hydroelectric Power Industry

The hydroelectric power industry ischaracterized by the fact that new construc-tion of hydropower dams, as well as finalretirement of older hydropower dams, hasessentially stalled in the United States. Sincethe 1980s, the stock of nonfederal hydro-power dams has been fundamentally static.

Figure 2 shows the annual figures fortotal dam storage capacity in the United

Year

FIGURE 2RESERVOIR STORAGE CAPACITY IN THE

UNITED STATES. 1880-2000SOURCES: ALLEY, REILLY, AND

FRANKE 1999; ANDERSON AND WOOSLEY 2005

States from 1880 until 2000; this graphmakes clear the history of dam constructionfrom its initial viability and take-oíT aroundthe turn of the twentieth century, until itsstagnation around the early 1980s. Sincethat time there have been very few newdams (ofany type) constructed anywhere inthe United States.** This has occurred fortwo primary reasons: first, there simply areno more viable sites containing the neces-sary technical requirements (groundstrength and narrow river bank, for exam-ple) for sound dam construction, andsecond, even in the few areas along ournation's navigable rivers where new damsmight conceivably be built, environmentalconstraints and regulations have limited thepolitical viability of any new construction.This is particularly true for hydropowerdams, which cause harm not just from theinterruption of river fiow, but from themechanized turbines that generate theelectricity and entrap and kill fish attempt-ing to pass through them. Even theNational Hydropower Association, theofficial trade group for the nonfederalhydropower industry, admits that any

" Only 4.1% of the observations in our 1983-2005dalaset, for example, are for new dam construction; thevasl majority are relicenses.

450 Land Economics August 2010

future increase in hydropower productionin the United States can only come fromtechnological improvements to existinghydropower stock, not from the construc-tion of any new large dams.^

Similarly, not many existing dams havebeen retired in recent decades. In FERC's75-year history, only once has it ordered thedecommissioning of a dam.'" While a fewdams have been voluntarily surrendered—instead of ordered for decommission byFERC—the number of such instances isastonishingly small. This is because hydro-power dams are, essentially, cash cows. Themain cost of hydroelectric power is theinitial construction costs of building a largedam and reservoir. Once a dam exists, itsmaintenance and operation costs are min-imal. Compared to other forms of generat-ing power, where conversion rates of energyinto electricity average 50% or less, hydro-electricity has an efficiency conversion rateof 90*/ .—better than any other form ofgeneration. Existing hydroelectric powerplants are therefore extremely cost-effec-tive. With rising energy prices in recentyears, the returns to the hydropowerproduced have only increased.

Therefore, when a hydropower projectcomes before FERC for a relicense, therequest is almost inevitably granted. Theregulatory issue is rarely whether or not toissue a license at all, it is, in issuing a license,how many environmental requirements tomandate alongside the license issuance.Because there are only so many end usesto which any single river can be allocated(endangered species protection, recreationaluse. aesthetic concerns, for example), it isFERC's job to balance the competingdemands when issuing a license and itsaccompanying requirements. This researchwill help to shed light on the mandates that

^ Although technical improvemenis to the existinghydropower stock could potentially yield as much as 50%more capacity (U.S. DOE 2006; Veltrop 1997).

'" In !997 FERC denied a license to the EdwardsDam. on the Kennebec River near Augusta. Maine, andordered that il instead be removed through federaltakeover.

accompany a hydropower license, includingwhat types of factors influenced how manywere ultimately chosen. This will help usunderstand how FERC has interpreted itsregulatory mandate in the past, whichshould help us anticipate how other reli-censing proposals will proceed in the future.

III. FRAMEWORK

The influences on regulatory agencydecision making are still debated in theliterature. An early assumption, which inrecent years has made something of acomeback, was that regulators sought tomaximize the public welfare, and in issuingregulatory decisions had the independenceand discretion to pursue a "public-interest-driven" policy. Today this assumption byitself appears naive, but as at least acomponent of regulatory agency decisionmaking (Crone and Tschirhart 1998; Now-ell and Tschirhart 1993). ideology anddiscretion have been found in many studiesto have empirical support. Olson (1995,1997. 2000) and Dranove and Meltzer(1994). for example, both found that theFood and Drug Administration (FDA)moved drugs along the regulatory pipelinein part based on a drug's perceived impor-tance. Ando (2001), Cropper et al. (1992),Magat. Krupnick. and Harrington (1986),and Yaffee (1982), in different contexts, allfound that Environmental ProtectionAgency (EPA) regulatory decisions havebeen influenced by independent bureaucrat-ic discretion. And Bartel and Thomas(1987) have found the same discretionaryeffect with regards to Occupational Safetyand Health Administration regulations.Discretion on the part of the regulator, itshould be noted, may not necessarily implybenevolence, but it does at a minimumimply the independence to pursue a possiblyideological agenda.

A significant challenge to the publicinterest view of regulatory agency decisionmaking came with the seminal works ofStigler(1971), Peltzman (1976), and Becker(1983). "Regulatory capture theory," as it is


known, implies that regulators are beholdento interest groups (either for politicalsupport, votes, funds, or future job oppor-tunities), and so they make decisions basedon the outcome of interest group lobbying,in a manner to maximize overall interestgroup support. Empirical evidence in sup-port of this theory, for agencies as diverse asthe FDA. the EPA, and the National ForestService, has been found in research byCarpenter (2002), Sigman (2001), Ando(1999, 2001), Crone and Tschirhart (1998),Sabatier, Loomis. and McCarthy (1995),Nowell and Tschirhart (1990), Cropper etal. (1992), and Yaffee{ 1982).

A third strand of the literature focuses onbureaucratic agency decision making asinfluenced by congressional control. "Con-gressional dominance theory." as it iscalled, suggests that agencies are notindependent but are reliant on congressio-nal support either because of budgetappropriations, oversight, or threatenedlegislation. An important influence onregulatory decision making then should bemembers of Congress and the actions theyimpose. Olson (1995), Anderson (1993),Heiland (1999), Gilligan, Marshall, andWeingast (1989), McCubbins, Noll, andWeingast (1987), and Weingast and Moran( 1983) have all found evidence for this effectat the FDA, the International- TradeCommission, the Interstate CommerceCommission, and other regulatory agen-cies."

Finally, most recently in the literature, afew voices have been heard suggesting thathistory, or the status quo starting point, isimportant for understanding bureaucraticagency decision making. As Beard, Kaser-man, and Mayo (2003, 594) note, "Thelocation of the unregulated equilibrium andother prior market parameters provide theprincipal constraint on the regulator'sability to deliver . . . through regulation."This argument echoes behavioral theories

about the importance of "endowment" or"starting point" for subsequent decisionmaking. Indeed, Sabatier. Loomis, andMcCarthy (1995) give evidence that ForestService planning decisions regarding outputtargets for timber and other resourceshave been heavily influenced by existing(or "historical") output levels.'" But this isnew in the literature. Empirical tests of thevalidity of such historical effects are com-mon in individual-level behavioral studies,but this paper appears to be the first toinclude such variables in regressions ofregulatory behavior.

In the following model we assume theregulator attempts to maximize a utility-based objective function in implementingregulatory policy, and that all of the abovefactors (discretion Ö. interest groups /,Congress C and history //) may beinfluential in affecting utility:

Discretion brings utility to the regulatorby allowing the regulator to follow throughon personally held beliefs, goals, or inter-ests, and by minimizing any cognitivedissonance on the job. Interest groups bringutility to the regulator through politicalsupport, votes, positive public-relationscampaigns, and other political activities.Congress brings utility by limiting itsoversight and by granting greater budgetappropriations. And history affects utilitythrough a starting point bias (Kahnemanand Tversky 1979), often related to institu-tional constraints (North 1990).

In the context of hydroelectric powerregulation, the regulator maximizes totalutility from these factors when achieving abalance between hydroelectric power pro-duction (P) and environmental riverineprotection (£*), per license:

U = U[D{E,P),I(E,PIC(E,P),HIE,P)].

' ' Spence (1997) and Wood ( 1988) arc two researcherswho provide contrary evidence that regulators arerelatively immune to input or oversight from Congress.

Peltzman (1976. 222-24) in his infiuentialarticle on regulatory behavior suggests defining hisequations "with respect to an anarchistic referencepoint."


Due to the physical nature of hydropow-er production, environmental protectiongenerally comes at the cost of reducedhydropower capacity (Kotchen et al. 2006:Edwards, Flaim, and Howitt 1999; FERC1991). For example, FERC could requiredam owners to limit the flow of waterthrough the dams' turbines. This wouldreduce electricity production, but it wouldalso help eliminate irregular rushes of waterthat cause stream bank erosion and damageto aquatic nesting sites. Other methodsFERC could impose to help the riverineenvironment (such as fish ladders, modifiedturbines, etc) similarly tend to negativelyaffect electricity production. In the end,when FERC issues a license mandating acertain number of environmental require-ments, it is implicitly choosing a certaintrade-off between environmental protectionand hydropower production. The issue forFERC is to decide just how many environ-mental mandates to issue per license, andhow to balance the trade-off. Therefore,either £ > 0 and i '< 0 or £'<0 and/ '>0 . Inother words.

U =s.t. E=fip)-and f<0.

FERC achieves the balance between Eand P in setting the specific number ofenvironmental requirements mandated perlicense. The number of mandated require-ments is analogous to setting the price,tariff, or subsidy level in other regulatorymodels (see, e.g., Becker 1983 or Beard,Kaserman, and Mayo 2003). Comparativestatics can then be derived on the relativeeffects of discretion, interest groups, con-gress, and history on the total number ofrequirements mandated per license.

IV. DATA

From 1983 to 2005 FERC issued 498 newhydroelectric licenses.'"^ Data were gatheredon these projects from a number of sources

'• A single license can authorize one, or more thanone, hydroelectric dam al a time.

including the licenses themselves, the licensedocket sheets (a formal recording of all thepaperwork submitted during the licensingprocess), an internal FERC database on thetechnical and physical characteristics ofeach of the projects, and public andgovernmental records on relevant politicaland river basin characteristics."* Summarystatistics are provided in Table 1.

The dependent variable in our analysis isthe number of environmental requirementsissued by FERC for each license applica-tion. It is through these requirements thatFERC carries out its contemporary man-date to balance both power and nonpowerobjectives when issuing a new hydropowerlicense. FERC has issued anywhere fromzero to over 30 such requirements perapplication (Figure 3), implying significantheterogeneity in ultimate outcomes. Arepresentative sample of these requirementscan be found in Table 2. Not all require-ments constrain hydropower production inthe same manner or to the same degree, butthey are all designed to limit hydropowercapacity somewhat. It is true that a morecalibrated model would weigh each require-ment in association with its degree of effecton hydropower capacity, but determiningsuch a weighting scale, given the ratherlimited technical information provided in alicense, is not at all straightforward. At-tempts to do so could be made on a varietyof levels, leading to a degree of subjectivityin the results that the author wishes toavoid. It is left to future research, therefore,to fmd a way to extend the current model toaccount for not just number, but type, ofenvironment requirements and its morespecific associated degree of constraint onhydropower capacity.

The explanatory variables are groupedinto four categories, following our fourfactors of influence on FERC's objectivefunction: D, /, C, H.

'** The final number of observations used in laterregressions is less than 498 because of random missingdata on ditTerenl observations from any one of the above-listed sources.

'• r


TABLE 1

SUMMARY STATISTICS

Variable

Dependent VariableTotal environmental requirements

Independent Variables

Dam height

Electricity priceEastWestRivermileEndangered riverPrivate ownerMultiple owner

Intervenors

SportsftshitigBoatingHistoricalEndangered speciesWater qualityRecreationFederalStateLocalPublicPrivateNGO

LCV scoreSenate committeeECPARelicenseHistoryALP

Sociodemographic Variables

PopulationllO^% age 0-24%age>75% bach, degreeMedian incomelW)^Vii below povertyHousing unitsiyO-Labor force! \0^UneniployednO^Total farmsManf. earningsllO^Gov. revenuel\d^

Mean

10.37

44.9224.157.710.780,20

233.280.090.394.52

10.62

4.880.390.715.484.771.074.634.074.267.121.234.7

0.560.33

650,938.910.06

140-1834.975.52

15.1328.1410.9054.8668.994.46

742.38260.84152.31

Std, Dev.

5.44

62,59112.27

2.050.420.40

1,183.070.280.495.07

7.55

3.330.910.953.643.341.393.22.89

21.764.381.8

22.03

0.290,470.480.263.100.24

423.649.0!2.186.079.183.75

148.26208.71

13-8766.19680.31366.37

Mm.

0

003.90000000

0

000000000000

000030

1,14005.4

10.972.40.930.550.02000.01

Max.

35

6322.000

13.4311

21.11011

18

39

1675

18188

161928271916

1I11

13.91

8,295.956.512.142.157.8125.6

2,855.584.095

2737,5906.798.763,859.49

Discretionary Explanatory Variables

We assume that if FERC wishes toexhibit discretion in a relicensing applica-tion, it does so according to the type ofproject up for relicensing. For example, ifFERC is predisposed to the hydropower

production potential of a dam (as opposedto protecting the riverbed it is built upon),then FERC might issue fewer requirementsfor larger dams, or those that produce morevaluable electricity, in more valuable re-gions of the country. Dam height, measuredin feet, is an indicator of the size of a


40-

30-

20-

10-

42

36

29

22

16

1212

40 40

32

2524

664 4

2| n~| 2 21 1

10 20 30

Total Environmental Requirements

F I G U R E 3ToTAi. ENVIRONMENTAL REQUIREMENTS

40

15project. Similarly, kW, which measuresthe available kilowatt production capacityat a project, is a second size/productionindicator. Electricity price, meanwhile, isthe average residential price of electricity (incents) per kilowatt-hour, for the state inwhich the project is located, in the year itapplies for a license.'^ And East and West,dummy variables indicating whether thepower produced at a dam is transmittedthrough the western U.S. power grid or the

' For projects with more than one dam. average damheight, maximum dam height, and lolal of all dams'height were tried as regressors, with similar ultimateresults. The regressions reported in the ibllowing tablesutilize tullesl dam height. In Table I. the minimum valuefor Dam fieighi is zero. This is not an error. Rarely, aproject can have a zero recorded dam height eitherbecause ( I ) the project is a low-impact run-of-riverhydroelectric facility that generates power as the riverflows, but without much ofan official diversion structure:or (2) because the nonfederal hydroelectric facility isutilizing a diversion structure owned by some other entity(for example, a municipal waier supply system): in thiscase because the dam structure is not owned by thelicensee, it is not officially a part of the hydro project aslicensed by FERC.

" As deflated by the Consumer Price Index, base year1970.

eastern U.S, power grid, are our geographicindicators.'^- '*

If. on the other hand. FERC's ideologicalbias is in the environmentally friendlydirection, then we would assume that largerdams might receive more environmentalrequirements, all else equal, and that damsbuilt on longer or more unhealthy riverswould also receive more requirements aftercontrolling for other factors. Rivertvilemeasures the total river miles of all thedams (creating their respective reservoirs) inthe project, and Endangered river is adummy variable used to indicate the baseenvironmental health of a river on which aproject is located. Endangered river indi-cates whether the river is listed as threat-ened/endangered in the year the projectapplied for a license.'^

'^The default altemative is an Independent powergrid in Texas. Alaska, or Hawaii.

'^ These power grids are distinct and have limitedinterconnections with each other, and while we have nopriors concerning the signs of the coefficients betweenthese two variables, it may be that electricity is morevaluable on a particular transmission grid.

** As assessed by the national organization AmericanRivers (www.amrivers.org).


II

4 0 -

3 0 -

20-

1 0 -

0 -

•

• • •• • •

ii::::i:.

•

• •• •• •II• •

••

ii• •i

•

• •

• • •• « • •

• • • • • • • • •

• •

10 20

Intervenors

30 40

FIGURE 4CORRELATION OF INTEREST GROUP INVOLVEMENT WITH TOTAL ENVIRONMENTAL

REOUIREMENTS ISSUED

Because FERC's initial 1935 mandatewas written specifically to encourage hy-dropower production, and because it wasonly altered to directly consider environ-mental objectives as late as 1986, many

TABLE 2

TYPES OF MANDATED

ENVIRONMENTAL REQUIREMENTS

License #

P-2520P-2436P-2512P-618P-663

P-1979P-2302P-1953P-2323P-1388

P-2496

P-1267

P-308

Abbreviated Description

Install upstream fish passage facilitiesProtect shoreline from erosionOperate project in a run-of-river modeMinimum flow releases of .v requiredProtect habitat aquatic species populations in

the bypassed reachDrawdown rates of reservoir limited to .vImplement a water quality planMonitor dissolved oxygen levelsLimit the maximum change in ramping ratesRevegetate disturbed areas with plant species

beneficial to wildlifeModify fish ladders so as to enhance

anadromous fish migrationManage and protect all historic properties on

siteMinimum flows required into the bypassed

reach

believe that FERC still acts with a latentbias — left over from this historical tradi-tion—in favor of hydropower production.If the coefficients on the seven discretionaryvariables are significant, they may confirmthat FERC does, or does not, hold such apro-power bias.

Interest Group Explanatory Variables

The hydropower licensing debate can becharacterized as a standotT between hydro-power interests and environmental inter-ests. The hydropower side is represented,for the most part, by the dam owner. Wehave two explanatory variables that give usan indication of the type of dam ownerinvolved in a project licensing. Privateowner is a dummy variable indicating thatthe owner is a private utility company, asopposed to the dam being municipallyowned, owned by a nonutility (also knownas a generation company), or owned by acooperative (for example, by an Indiantribe). The second firm ownership variable.Multiple owner, is a measurement of the sizeof the firm owner. Large firms generally


own more than one project and thus oftenhave multiple licensing experiences withFERC. Owners of large revenue-generatingprojects are also often heavily invested inhydroelectric generation and may be evenmore strongly opposed than other types ofowners to alternative allocative uses of theproject's waters. This variable is a count ofthe number of other projects simultaneous-ly owned by a particular licensee. If thecoefficients on both of these variables aresignificant and negative, it would indicatethat larger, privately owned utilities arebetter able to capture a less stringentrelicensing mandate from FERC.

Environmental interests are representedin the licensing process by advocacy groupswho formally lobby the regulator. All inter-ventions are posted on the license applica-tion's docket sheet, and from these docketsheets lists were tabulated, for each project,of all of the interest groups involved in therehcensing processes, as well as the numberof times each representative group chose tocomment (in order to capture the degree oftheir involvement, and not simply theirpresence). Intervenors is a count of the totalnumber of these interest groups involved ineach relicensing exercise."*' Figure 4 illus-trates the correlation of Intervenors with thetotal number of environmental require-ments mandated.

All kinds of interest groups tend to getinvolved in the process,^' from nationalto local level interest groups (AmericanRivers; Camp Grady Spruce), public andnot-for-profit interest groups (CaliforniaBoating and Waterways; Federation ofFly Fishers), and interest groups withadvocacy interests that range from sports-fishing to water quality (California Trout;Idaho Department of Water Resources). Inthe empirical section Intervenors is broken

"" Models were also estimated in Ihe empirical sectionwith count data on ihe total numbers oí commentssubmitted, but as the results were similar to those of themodels using ¡nierwnors. In the interest of brevity wereport only the latter results.

^' See Kosnik (2010) for a look into what motivates aninterest group to get involved in the process in the firstplace.

down by these additional category con-structions to try to make more detailedinferences regarding the effects of specifictypes of environmental interest group ac-tivity on the number of ultimate licensingrequirements mandated. Intervenors is bro-ken down first by advocacy type: Sports-fishing, Boating, Historical, Endangeredspecies. Water quality, and Recreation', thenby geographic range: Federal, State, andLocah and finally by market orientation:Public, Private, NGO. See Table 3 for a listof intervenor descriptions and examples. Bydifferentiating the intervenors by geograph-ic range and by market type, for example,we explore whether intervenors with alarger membership base have more cloutthan purely locally based intervenor groups,and whether the market type of an interestgroup has any significant effect on regula-tory clout.""

It is worthwhile to alert the reader eariyto the method by which the intervenor datawas broken down when disaggregated byadvocacy type. A few of the intervenors(though not enough to justify additionalcategories) were known to advocate onbehalf of multiple interests (AmericanRivers, for example, is concerned withendangered species, water supply, andsportsflshing issues). When the data weretabulated, these multiadvocacy intervenorswere counted in each of the categories towhich their interests applied. Due to thisconstruction, the reader should be fore-warned of a degree of multicollinearity builtin to the dataset, in particular, in theintervenor group data when it is disaggre-gated by advocacy type.

Congressional Explanatory Variables

The ability of the legislature to affectFERC's regulatory decisions was tested inthree ways. LCV score is the averageLeague of Conservation Voters score for

^ While it may be informative to differenliate theseintervenor variables by geographic level und by advocacytype simultaneously, unfortunately doing so results in toofew observations per calegory (of which in this instancethere would be 27) to run any meaningful regressions.

86(3) Kosnik: Eederal Hydropower Licensing 457

4-U

ozu>LU-

ooî.

c

<

c

U u.2 M u '5y E -3 5

as .°sCL -O

y Q

U

ç ïO

CL.

Ci

c o c2 i 2

o

2 c

^ i

•S .2m 13 1 1

« S

Í 3il

Sa:« 6 0.g iHO

C

g.2 Ü o s§ I

|

•co cTA O

CA

Boa

tu

or/l

arts

i

£.

neri

can

<

o

<

Can

oëat

io

u.•§Lu

ïorg

ia'O

Ut

S3

brni

u

c

Can

oë:

Ass

n.as

taP

addl

eun

dtt

eri

^^ Ú

3

iHfo

rnii

U

£•

ishe

u.o

(£

-oM

Boa

tin

ent

1

lana

S.

Wat

er*

:=c

de u

t5 U.

Í5.5

OÜ

.Ë-li. ¿ •§ <

II

u O


the two senators in Congress from the statewithin which a project is located, in the yearthat the project went up for licensing. It wasincluded as a proxy for the "congressionalmood" in Congress at the time of licensing.The idea is that, at different times or indifferent political environments, pressurefrom congressional representatives mayhave a varying effect on the regulatoryagency. Specifically, if the mood in Con-gress is heavily Democratic and pro-envi-ronmental, then congressional pressure insupport of environmental benefits, in such asetting, would lead to more environmentalrequirements mandated than if Congresswere at that time dominated by morehydropower-supporting Republicans. TheLeague of Conservation Voters score is arating between 0 and 1 for environmentalsensitivity (0 being none. 1 Implying ex-treme environmental sensitivity). T'his rat-ing is gleaned from the past voting records,on environmental issues, of each electedsenator. Traditionally, Republicans havehad lower League of Conservation Votersscores than Democrats, and eastern states

have had much higher scores than western•»Iones.

Senate cotwnittee is a dummy variableindicating whether at least one of thesenators in a project's state, at the time oflicensing, was a member of the oversightcommittee in charge of EERC. Accordingto congressional control theories of regula-tory capture, an individual senator who hasdirect congressional influence over FERCshould be able to particularly influence thecommission's regulatory decisions in his orher home state. The direction of thatcapture, however, is ambiguous. If damowners are supremely adept at influencingcongressional representatives, then wewould expect the coefficient on this variable

'' For example, tor the year 2000 scorecard. Repub-licans in the Senate averaged a score of 0.12, whileDemocrats averaged 0.79. Meanwhile, eastern statesposted the highest environmental scores—Rhode Islandat 1. Connecticut at 0.91. New Jersey at 0.91. Vermont at0.88. and Massachusetts at 0.88—and western states thelowest, with Idaho. Wyoming, Alaska, Utah, andOklahoma all posting scores of 0.

to be negative, indicating decreasing regu-latory stringency. If, on the other hand,pro-environmental pressures hold sway incongress, we would expect the coeiTtcient onthis variable to be positive, indicating anincreasingly stringent regulatory environ-ment.

Finally, we include a dummy variable,ECPA, which indicates whether a projectwas relicensed after the implementationdate of the ECPA amendments (January2, 1990). As the ECPA amendments weredesigned by Congress to increase FERC'sproclivity toward environmentally friendlyregulation, we would expect to see asignificant increase in environmental re-quirements issued after the passage of thisamendment. In other words, we wouldexpect the coefficient on this variable to bepositive.

Historical Explanatory Variables

The types of historical variables thattheoretically could affect FERC's decisionmaking do so either through a startingpoint effect (in other words, a tendency forFERC to favor, or unwarrantedly "grand-father."" historical positions), or throughinstitutional constraints (in other words, thedefined bureaucratic structure within whichFERC operates constrains decision mak-ing). The starting point effect is investigatedthrough the variables Relicense and Histo-ry, and the institutional constraint variableis ALP.

Relicense is a dummy variable indicatingwhether a particular application is for thelicensing of older hydroelectric projectworks. As stated in Section II, very fewnew hydropower dams are being builtanywhere in the United States, but as thereare some, with this variable we test to seewhether there is any significant difference inregulatory stringency between original li-censes and relicenses. Note that thereshould not be. In 1984 in Yakima IndianNation V. The Federal Energy RegulatoryCommission (746 F. 2d. 466), the SupremeCourt established the criteria under which ahydroelectric dam relicensing application


14-

12-

10^

6-

4-

2-

0 "1 1 1 1 1 1 ' r 1 1 1 1 T r -1 1 1 f—1 1 1

C\J - « t C O C O O C M t i C O O O O C J ^OO O O C O C O C 3 ) 0 : ) C Î O J O > 0 0 0o O J C D c j î o i o i o i o i c n o o oT- • • - • ^ • • - • i - i - - . - i - . . - c \ j c y c \ i

Year

FIGURE 5AVERAGE NUMBER OF ENVIRONMENTAL REQUIREMENTS

should be considered. It determined that arelicensing application should be given thesame scrutiny as an initial license applica-tion, with all the attendant requirementsand regulation. There should be no grand-fathering of older projects, in other words.•'Relicensing." the Supreme Court stated,"is substantially equivalent to issuing anoriginal license. . . . Relicensing, then, ismore akin to an irreversible and irretriev-able commitment of a public resource thana mere continuation of the status quo.Simply because the same resources hadbeen committed in the past does not makerelicensing a phase in a continuous activity.

TABLE 4

AVERAGE NUMBRR or ENVIRONMENTALREQUIREMRNTS. BY YEAR

Year

2O0520042003200220012000199919981997199619951994

Avg.

8.9611.1110.4210.9410.913.089.36

13.913.4512.6713.1411.68

Year

199319921991199019891988198719861985198419831982

Avg.

9.1110.239.727,544,275.29864.2534.83

Note: Electric Consumers Protection Act went into eíTect in1990.

Relicensing involves a new commitment ofthe resource."

The basis for the Supreme Court'sinterpretation stems from the initial legisla-tive decision, in the Federal Water PowerAct of 1920. to allow license terms forhydroelectric projects to last up to 50 years.During the debate at that time regardinghow best to undertake development of thenation's rivers, there was disagreement overwhether hydropower development shouldbe a private or public enterprise. Thecompromise was to allow project licenseterms of up to 50 years, which wasconsidered long enough time to allow aprivate developer to amortize the costs ofany hydroelectric project, but at the sametime would eventually return the waterresources to the public domain where itcould be determined anew how best toutilize them. In 1984 the Supreme Courtconcluded that when a dam then comes upfor a relicense. it must be treated withessentially a blank slate. This can work ineither the dam owners' or the environmen-tal interest groups' favor, because it impliesthat nothing is off the table for historicalreasons. We would expect the coefficient onRelicense, therefore, to be insignificant asthere is no a priori reason to expect olderdams to receive heavier or lighter regulatoryburdens. Newer dams are presumably inmore pristine areas, so a case could be made


TABLE 5DETERMINANTS OF THE NUMBER OF ENVIRONMENTAL

REQUIREMENTS PER LICENSE

TABLE 6

FIRST-STAGE OLS IV REGRESSION

Variable Model 1 Model IV

ConstantDiscretionat7

Dam heightkfVIWÊlectricity priceEastWestRivermileEndangered river

Interest groupPrivate ownerMuitipie ownerIntervenors

Congressional

LCV scoreSenate committeeECPA

HistoryReiicenseHistoryALP '

15.39***

0.01***0.01**

-0.47***-2.87-2.31

0.002.59***

1.69***-0.06

0.27***

2.59***-0.54

2.90***

-4.50***-0.22***

2.84***

(2.38)

(0.00)(0.00)(0.14)(1.73)(1.79)(0.00)(0.76)

(0.52)(0.05)(0.03)

(0.97)(0.56)(0.51)

(1.17)(0.08)(0.90)

R^ = 0.46

14.57*** (2.82)

0.01*** (0.00)0.00** (0.00)

-0.50*** (0.17)-2.03 (2.01)-1.53 (2.18)

0.00 (0.00)2.51*** (0.81)

1.63*** (0.54)-0.06 (0.05)

0.29»** (0.10)

2.68*** (0.98)-0.52 (0.56)

2.92*** (0.60)

-4.29*** (1.23)-0.25*** (0.09)

2.95*** (0.92)

R^ = 0.46

Note: Figures in parentheses are standard errors.• Significant at the IC i. level; •• significant at the S"A> level: • "

significant at the 1% level.

that they require greater environmentalprotection. Older dams, meanwhile, havepresumably been negatively affecting river-beds for longer., so a case could be madethat they require greater environmentalprotection. On balance and in the aggre-gate, the effect a priori is expected to beinsignificant.

History is a count variable indicating theaverage number of environmental licensingrequirements issued by FERC in the yearbefore a particular project applied for a newlicense. Figure 5 and Table 4 both indicatethat, for the most part, the average numberof environmental requirements issued byFERC, per license, has been increasing overtime. But there is no particular reason toassume that the average number of require-ments issued in the year directly before alicense application should have any effecton the number of requirements subsequent-ly issued in a wholly unrelated license. Asidefrom behavioral theories, there is no obvi-

Variable CoefTicient Std. Error

Constant

Sociodemographic

Popula tionl\0^% age 0-24% age >75"/o bach, degreeMedian inconieñO% below povertyHousing units!\0'Labor force! \0Ûneniployedl 10Tota! farmsManf EarningsliO'Gov. revenueHO'

Dam characteristics

Dam heightkiV!\OÊlectricity priceEastWestRivermileEndangered riverPrivate ownerMultiple ownerLCV scoreSenate committeeECPARelicenseHistoryALP

-18.37***

-0.12***0.010.120.090.16**0.23*0.18***0.09**0.45**

-0.00-0.00*

0.00

0.03***0.01***0.81***6.48*

10.03***0.00**2.41*

-0.650.17**0.110.321.633.55*

-0.02-0.95

4.78

0.030.050.230.090.070.130.060.040.200.000.000.00

0.010.000.253.363.370.001.280.870.081.620.940.991.950.171.52

R^ = 0.34

• Significant at the WA level;significant at the 1% level.

significant at the $% level;

ous reason why the coefficient on thisvariable should not be insignificant. If itturns out instead to have an effect, thiswould be an indication that FERC issubject to behavioral effects and thatlicensing decisions are unexpectedly (andperhaps unwarrantedly?) path dependent.

Finally, ALP is a dummy variableindicating that a project applied for a newlicense through the ALP. as opposed to theTLP. The most recent licensing procedurepossibility, the ILP, is not represented inour sample as no new licenses had as yetbeen fully processed under this procedure.The ALP was designed to encouragecooperation among all the stakeholders ina licensing process, which has tended toimply a rebalancing of the environmental/

86(3) Kosnik: Eederal Hydropower Licensing 461

TABLE 7DETERMINANTS OF THE NUMBER OF ENVIRONMENTAL REQUIREMENTS PER LICENSE {wrm ENVIRONMENTAL

INTEREST GROUPS DISAGGREGATED)

Variable Model 2 Mode! 3 Model 4

Constant

Discretionary

Dam height

Electricity priceEastWestRivermikEndangered river

Interest group

Private ownerM id lip le ownerSporlsfishingBoatingHistoricalEndangered speciesWater qualityRecreationFederalSlateLocalPublicPrivateNGO

CongressionalLCV scoreSenate committeeECPA

History

RelicenseHistoryALP

13.82***

0.02***0.00**

-0.32**-2.68-1.570.002.25***

1.69***-0.06-0.09-0.170.000.190.51**0.01

2.22**-0.402.86***

-4.65***-0.19*-2.41'**

(2.45)

(0.00)(0.00)(0.14)(1.74)(1.81)(0.00)(0.77)

(0.53)(0.05)(0.26)(0.25)(0.25)(0.32)(0.20)(0.20)

(0.99)(0.56)(0.52)

(1.18)(0.08)(0.91)

P^ = 0.47

15.45**" (2.34)

0.01*** (0.00)0.00** (0.00)

-0.49*** (0.13)-3.10* (1.70)-2.75 (1.76)

0.00 (0.00)2.40*** (0.75)

1.45"* (0.51)-0.06 (0.05)

0.33*** (0.09)O.39***(O.IO)0.02*** (0.01)

2 .29" (0.96)-0.83 (0.55)

2.76*** (0.50)

-4.42*** (1.15)-0.18** (0.08)

2.51*** (0.88)

R^ = 0.48

14.69***

0.01***0.00**

-0.44***-2.71-2.53

0.002.08***

1.42***-0.03

0.46*"-0.01

0.03"*

2.04*«-0.85

2.75"'

-4.80***-0.14*

2.55***

(2.32)

(0.00)(0.00)(013)(1.69)(1.75)(0.00)(0.75)

(0.52)(0.05)

(0.06)(0.15)(0.01)

(0.96)(0.55)(0.50)

(1.16)(0.08)(0.88)

R^ = 0.49

Note: Figures in parentheses are standard errors.• Significant at the 10% level; "* significant at the 5% level: •** significant at the 1% level.

hydropower trade-off toward more envi-ronmental concern. We would thereforeexpect the coefficient on this variable to bepositively significant.

V. EMPIRICAL APPROACH ANDINITIAL RESULTS

We analyze FERC's power/environmen-tal trade-off decision through regressionanalysis, with the number of environmentalrequirements as the dependent variable (R)and a set of variables corresponding to thecategories of influence in our theoretical

model (discretion D, interest groups /.Congress C, and history H) as the indepen-dent variables."'* Each observation y repre-sents a hydroelectric dam license:

Model 1 in Table 5 presents the initial

•' Ordinary least squares (OLS) and Poisson specifi-cations were both run, and as there were no changes in thesize or significance of any of the key regressors. for ease ofinterpretation we present the OLS results in the followingtables.

462 Land Economics August 20J0

results. We find evidenee in support of allfour of the categories of influence describedin the framework setup. We find, forexample, that Dam height and kW arc bothpositively significant, implying that thelarger the dam, the greater the regulatorystringency imposed on a license. Similarly,the more unhealthy the river, the greater theenvironmental regulatory burden thatFERC mandates. In a nod to the consumer,however, as electricity prices increase,FERC reduces the regulatory burden slight-ly. While geography, in the form of trans-mission grid utilized, is insignificant inFERC decision making, the significance ofthe other variables in the discretionarycategory implies that FERC is able andwilling to exhibit a measure of authority inissuing its regulatory decisions.

The interest group variables are alsosignificant, though not necessarily in thedirection expected. The coefficient on Pri-vate owner is large and positive, implyingthat private utility dam owners receive, onaverage, an 1 T/i (1.69/15.39) greater regu-latory burden than other types of damowners. We had hypothesized a priori thatprivate utility dam owners would havedeeper pockets, greater experience, and amore substantial interest in lobbying theregulator to decrease the regulatory burden,but we found instead that FERC tended toincrease the regulatory burden of privateutility dam ovraers relative to all otherownership types. This may be becauseprivate utility owners generally own pro-jects that require substantial environmentalattention,"" or because FERC is battling itsreputation as being biased toward suchowners due to their initial mandate toencourage hydroelectric production. Theenvironmental interest groups, Inlervenors,are also found to have a significantlypositive effect on the regulatory burden.The coefficient is small, however (althoughits standard deviation is large), so weexplore this variable in greater detail inlater regressions (Table 7).

Although interaction terms on this were insignifi-cant.

It is the congressional variables and thehistory variables that have high significanceand the largest coefficients in the table,implying that, overall, while all categories ofinfluence matter. Congress and historymatter the most to FERC decision making.The positively significant coefficient on LCVscore implies that liberal-leaning senatorsfrom a project's state have the power toincrease the regulatory burden by 11% onaverage. This is a substantial effect. Evengreater is the effect of passage of the ECPAlegislation. Projects that were licensed afterthis legislation went into effect received aregulatory burden that was \9% higher onaverage. These results lend strong supportto the congressional dominance theory ofregulatory agency decision making.

Relicense turns out to be the regressorwith the largest coefficient, by far. Itsnegative significance implies that olderprojects receive a less stringent regulatoryburden than brand new hydroelectric facil-ities. Exactly why this would be, is unclear.Newer projects could be located on morepristine land, implying that in order toprotect them FERC grants their licenseonly with high environmental barriers topass. A detailed look at each of theseobservations, however, fails to imply this.An alternative explanation is that FERC isexhibiting a form of de facto grandfalheringon older licenses, even though such behav-ior is not a part of their mandate and is,indeed, forbidden by judicial decree.

History, too, is negatively significant,though its coefficient is small. This impliesthat the stringency of FERC's regulatorydecisions are path dependent; higher aver-age regulatory burdens from the year beforelead FERC to cool off slightly the followingyear, on wholly unrelated licenses (thiseffect can also be seen in Figure 5). This isan entirely surprising result. It implies, inessence, that a dam relicensed in, say,California in 1993 affected the relicensingprocedure of a wholly unrelated dam inMaine in 1994. Historical biases have beenstudied extensively in the psychology andeconomics literatures at the individualdecision-making level, but extending such


effects to federal regulatory decision mak-ing is still quite novel.

Finally. ALP is significant in the positivedirection, implying that FERC is actingwithin the institutional constraints of itssystem. ALP is a form of regulatoryprocedure (the alternative in this timeperiod being the TLP) designed to increaseearly consideration of environmental issuesand goals in the licensing process. When it isused, it indeed tends to correlate withgreater environmental stringency on alicense, by 18% on average.

Endogeneity

A central concern in the estimationstrategy is the potential simultaneity ofenvironmental interest group involvementand environmental requirements issued. Itis conceivable that levels of intervenorinvolvement reflect knowledge of impend-ing environmental requirements, ratherthan acting solely as a stimulus for them.If this is the case, intervenors who actuallyhave zero infiuence on the outcome of thelicensing process could appear instead tohave a statistically significant positiveeffect. Any significant coefficients reportedfor such an intervenor in a non-endogene-ity-corrected model would then be spurious.

The existence of factors causing this typeof bias would be quite a policy revelation,implying enormous wasted efforts by inter-venors, who regularly spend thousands ofdollars mobilizing individuals to commentand protest directly to FERC. It would alsoimply enormous wasted social costs by theregulatory staff, who must by law spendtime and effort processing intervenor com-mentary. Intuitively, such a conclusionseems implausible, especially when oneconsiders the numerous legislative andjudicial rulings designed to explicitly em-phasize and encourage intervenor effects onthe relicensing process."^ Official interve-

• The Administrative Procedures Acl of 1946 estab-lished the legal mandate for public participation inregulcitory decision making. The courts also solidifiedthe role of interest groups, in the relicensing process inparticular, as early as 1967 in Udall v. Federal PowerCormiission (387 US 428).

nors in the relicensing process carry the veryreal power to appeal any relieense issued byFERC to the U.S. Court of Appeals, withthe potential for the entire relicense to beoverturned and renegotiated. With suchpower, it would be foolhardy for FERCnot to incorporate intervenor commentaryinto their objective function, in order toavoid the costly and politically damningfate of the U.S. appeals process.

Nevertheless, strongly intuitive judgmentcalls have been proven false by thoroughresearch before. For this reason, a trulycomplete exposition of the model shouldinclude some econometric attention topotential simultaneity bias. We have chosento address this concern by instrumentingfor intervenor involvement with a set ofexogenous sociodemographic and damcharacteristic variables.' ^ The rationalefor these instruments stems from environ-mental economic theory. Kolstad andothers (Braden and Kolstad 1991) haveshown that many environmental goodssuch as clean air and clean water are normalgoods, demand for which generally increas-es as income increases. Studies have addi-tionally shown that more educated, youngerpopulations demand greater environmentalamenities than less educated, poor, agingpopulations. As well, populations with ahigh unemployment rate or a large percent-age of the workforce employed in manu-facturing will desire fewer environmentalamenities if they come at the expense ofworking conditions or employment pros-pects. Therefore, there is a sound theoreticalbasis for using instruments for intervenorinvolvement such as median income, unem-ployment rate, manufacturing output, andeducation level.

We additionally include the dam charac-teristic variables as instruments, under theassumption that intervenors. after theirinitial proclivity to act is determined bysociodemographic factors, ultimately get

The complete list of instruments used in our initialspecification can be found in Table I. The results of theinstrumental variables regression can be found in Table 6.


involved in a relicensing proposal due to thespecific characteristics of a project.

The results for our first specificationappear almost identical whether correctedfor potential endogeneity bias or not(Model I vs. Model IV, Table 5). In orderto test exogeneity explicitly, we conduct aDurbin-Wu-Hausman test and find that theendogeneity of intervenor participation isnot significant (Prob > F = 0.7082). Forthis reason, as well as for the practicalconsideration of the near impossibility offinding unique instruments for all of thefurther specifications presented, we do notattempt to correct for potential endogeneitybias in the rest of the specifications present-

VI. THE SPECIAL ROLE OFINTEREST GROUPS

Interest group effects are increasingly ofinterest in the literature (Kosnik 2010;Mitchener 2007; Knittel 2006; Li et al.2005; Aidt 2002; Oates and Portney 2001).For this reason, in Table 7, we explore theeffect of Intervenors in greater detail,disaggregating this variable first by itsadvocacy type, then by its membershipbase, and finally by its market type. Wefind that not much of significance can bedetected when the intervenors are brokendown by advocacy type. This implies thatthe influence interest groups gamer fromFERC does not appear to be related to whatthey are arguing for, instead, what appearsto matter more in gaining inlluence over

^ Further support for the decision not to be overlyconcerned with polential endogeneity of inlervenorcomments comes from recent research that bolsters suchan assumption: In a study of intervenor group commentsreceived by the EPA during consideration of additions tothe LI.S. endangered species list. Ando (2001) specificallyasks the question whether interest group comments aresubmitted endogenously, whereby the pressure from oneinterest group increases or decreases with the pressureintensity coming from an opposing group, or whetherinterest group comments are instead motivated bybenefit/cost considerations that can accrue to the groupsas a result of the final regulatory decision. Her resultsconfirm that interest groups do not appear to actendogenously but instead respond to internal benefit/costcalculations.

FERC decision making is the geographicrange of the interest groups, and an interestgroup's market type. Federal and state-based interest groups, for example, havelarge influences as measured by coefficientsize, implying that broad-based groups havemore clout than smaller, local interestgroups. The significance of Public andNGO market types imply that interestgroup structure matters too, with intrago-vernment lobbying showing by far thebiggest effect on FERC decision making.This is not surprising to those withingovemment, at organizations such as theFish and Wildlife Service and the NationalMarine Fisheries Service, who note that theyoften have a statutorial obligation to involvethemselves in hydropower relicensing pro-ceedings. These intriguing results call formore research in the future, as they hint at adegree of complexity behind interest groupinfluence on the regulatory process that hasnot been fully explored in the literature todate.

VI!. CONCLUSION

The environmental protection/energyproduction trade-ofT that FERC is makingevery time it issues a new hydroelectriclicense has complex roots. T'his papersought to uncover some of those roots andfound that many categories of influenceexist on the trade-off decision, includingregulatory discretion, regulatory capture,congressional dominance, and historicaland institutional factors. Of these, thegreatest influences appear to come fromCongress and history. Additional researchinto these factors would be worthwhile. Forexample, we know that direct legislation onFERC's mandate, such as the ECPA, madea difference in shifting the trade-off towardenvironmental protection, but does indirectlegislation have similar eflfects? How aboutamendments thrown into broader bills, likethe 2005 Energy Bill? Also, incorporatinghistory and behavioral theories, such asstarting point bias, in regulatory models isstill a novel undertaking. There needs to bemore research into decision making at other


regulatory institutions, such as the FDA orthe EPA, to see if these types of behavioralvariables are effectual there as well.

Now that we have a better idea how toaffect FERC decision making, future re-search should also investigate the level ofenergy production versus environmentalprotection we seek to achieve. If it turnsout that we are currently not balancingthese twin goals efficiently, then the resultsfrom this research indicate thai we shouldpass new legislation or implement reform atFERC in an attempt to get there. Actionslike these can help us achieve the difficultbalance of efficient levels of energy produc-tion attä environmental protection.

References

Aidt. Toke S. 2002. "Strategic Political Participa-tion and Redistribution." Economics and Poli-tics 14(1): 19^0.

Alley, WilliatTi M., Thomas E. Reilly, and O. LehnFranke. 1999. Sustainability of Ground-WaterResources. U.S. Geological Survey. Circular1186. Denver. CO: U.S. Geological Survey.

Anderson. Keith B. 1993. "Agency Discrelion orStatutory Direction: Decision Making at theU.S. International Trade Commission." Journalof Law atid Economics 36 (2): 915-35.

Anderson. Mark T.. and Lloyd H. Woosley, Jr.2005. Water Avaihibility for the Western UnitedStates—Key Scientific Challenges. Circular1261. Reston, VA: U.S. Geological Survey.

Ando, Amy Whritenour. 1999. "Waiting to BeProtected under the Endangered Species Act:The Political Economy of Regulatory Delay."Journal of Law and Economics 42 ( 1 ): 29-60.

. 2001. "Economies of Scope in Endan-gered-Species Protection: Evidence from Inter-est-Group Behavior." Jottrnal of EtivironntentalEcortomics and Management 41 (93): 312-32.

Bartel. Ann P., and Lacy Glenn Thomas. 1987.•"Prédation through Regulation: The Wage andProfit Effects of the Occupational Safety andHealth Administration and the EnvironmentalProtection Agency." Jottrnal of Law and Eco-nomics 30 (2): 239-64.

Beard. T. Randolph. David L. Kaserman, andJohn W. Mayo. 2003. "A Graphical Expositionof the Economic Theory of Regulation."Economic Inquiry 41 (4): 592-606.

Becker. Gary S. 1983. "The Theory of Competitionamong Pressure Groups for Political Influ-

ence," Quarterly Journal of Economics 98 (3):371^00.

Braden, John B.. and Charles D. Kolstad. 1991.Measuring the Demand for Environmental Qual-ity. New York: Elsevier Science,

Carpenter. Daniel P. 2002. "Groups, the Media,Agency Waiting Costs, and FDA Drug Ap-proval." American Jotirnal of Political Science46 (3): 490-505.

Crone. Lisa, and John Tschirhart. 1998. "Separat-ing Economic from Political Influences onGovernment Decisions." Jotirnal of EconomicBehayior and Organization 35 (4): 405-25.

Cropper. Maureen L.. William N. Evans. StephenJ. Berardi. Maria M. Ducla-Soares, and Paul R.Portney. 1992. "The Determinants of PesticideRegulation: A Statistical Analysis of EPADecision Making." Journal of Political Econo-my 100(1): 175 97.

Dranove. David, and David Meltzer. 1994. "DoImportant Drugs Reach the Market Sooner?"RAND Journal of Economics 25 (3): 402-23.

Edwards, Brian K. 2003. The Economics ofHydroelectric Power. Northampton, MA: Ed-ward Elgar Publishing.

Edwards. Brian K.. Silvio J. Flaim, and Richard E.Howitt. 1999. "Optimal Provision of Hydro-electric Power under Environmental and Reg-ulatory Constraints." Land Economics 75 (2):267-83.

Federal Energy Regulatory Commission. Office ofHydropower Licensing (FERC). 1991. Evaluat-ing Relicense Proposals at the Federal EnergyRegulatory Commission. Washington, DC:FERC.

. 2004. Handbook for Hydroelectric ProjectLicensing and 5 M W Exemptions from Licensing.Washington. DC: FERC.

Gilligan. Thomas W., William J. Marshall, andBarry R. Wcingast. 1989. "Regulation and theTheory of Legislative Choice: The InterstateCommerce Act of 1887." Journal of Law andEconomics 32 ( 1 ): 35-61.

Heiland. Eric. 1999. "The Waiver Pork Barrel:Committee Membership and the ApprovalTime of Medicaid Waivers." ContemporaryEconomic Policy 17 (3): 401 -11.

Kahneman, Daniel, and Tversky Amos. 1979."Prospect Theory: An Analysis of Decisionunder Risk.'" Econometrica 47 (2): 263-91.

Knittel, Christopher R. 2006. 'The Adoption ofState Electricity Regulation: The Role ofInterest Groups." Journal of Industrial Econom-ics $^(2): 2Q\-22.

Kosnik, Lea-Rachel. 2006. "Sources of Bureau-cratic Delay: A Case Study of FERC Dam


Relicensing." The Journal of Law, Economics,and Organization 22 (1): 258-88.

-. 2010. "Time to Pick a Fight? InterestGroup Decision Making to Enter the Hydro-power Regulatory Process." Eastern EconomicJuurnan6(\): 11-32.

Kotchen, Matthew J., Michael R. Moore. FrankLupi. and Edward S. Rutherford. 2006. "Envi-ronmental Constraints on Hydropower: AnEx-Post Benefit-Cost Analysis of Dam Reli-censing in Michigan." L(md Economics 82 (3):384-403.

Li. Wei, Christine Qiang, and Lixin Xu. 2005.••Regulatory Reforms in the Telecommunica-tions Sector in Developing Countries: The Roleof Democracy and Private Interests." WorldDevelopment 33 (S): 1307-24.

Magat, Wesley A.. Alan J. Krupnick, and WinstonHarrington. 1986. Rules in the Making: AStatistical Analysis of Regulatory Agency Be-havior. Washington. DC: Resources for theFuture.

McCubbins, Matthew D., Roger G. Noll, andBarry R. Weingast. 1987. "AdministrativeProcedures as Instruments of Political Con-trol."' Journal of Law, Economics, and Organi-zation 3 (2): 243-77.

Mitchener. Kris James. 2007. "Are PrudentialSupervision and Regulation Pillars of FinancialStability? Evidence from the Great Depres-sion." Journal of Law and Economics 50 (2):273-302.

North, Douglass C. 1990. Institutions. InstitutionalChange, and Economic Performance. New York:Cambridge University Press.

Nowell, Clifford, and John Tschirhart. 1990. "ThePublic Utility Regulatory Policy Act andRegulatory Behavior." Journal of RegulatoryEconomics 2(1): 21-36.

. 1993. "Testing Theories of RegulatoryBehavior." Review of Industrial Organization 8(6): 653-668.

Oates. Wallace E., and Paul R. Portney. 2001."The Political Economy of EnvironmentalPolicy."' RFF Discussion Paper 01-55. Wash-ington, DC: Resources for the Future.

Olson, Mary K. 1995. "Regulatory Agency Dis-cretion among Competing Industries: Inside theFDA." Journal of Law, Economics, and Orga-nization 11 (2): 378^05.

• . 1997. "Firm Characteristics and the Speedof FDA Approval." Journal of Economics andManagement Strategy 6 (2): 377^01.

-. 2000. "Regulatory Reform and Bureau-cratic Responsiveness to Firms: The Impact ofUser Fees in the FDA." Journal of Economicsand Management Strategy 9 (3): 363-95.

Peltzman, Sam. 1976. "Toward a More GeneralTheory of Regulation." Journal of Law amiEconomics 19 (2): 21 I^W.

Sabatier. Paul A., John Loomis, and CatherineMcCarthy. 1995. "Hierarchical Controls. Pro-fessional Norms, Local Constituencies, andBudget Maximization: An Analysis of U.S.Forest Service Planning Decisions." AmericanJournal of Political Science 39 (1); 204-42.

Sigman, H. 2001. "The Pace of Progress atSuperfund Sites: Policy Goals and InterestGroup influence." Journal of Law Economics44 (I): 315^14.

Spence, David B. 1997. "Agency Policy Makingand Political Control: Modeling Away theDelegation Problem." Journal of PublicAdministration Research and Theory 1 (2):199-219.

Stigler. George J. 1971. "The Theory of EconomicRegulation." Bell Journal of Economics andManagement Science 2(1): 3-21.

U.S. Department of Energy (U.S. DOE). 2006."Feasibility Assessment of the Water EnergyResources of the United States for New LowPower and Small Hydro Classes of Hydroelec-tric Plants." DOE-ID-II263. Idaho Falls, ID:Idaho National Laboratory, U.S. DOE.

Veltrop, Jan. 1997. "Sustainable Hydropower atHome and Abroad.'" In Waterpower '97:Proceedings of the International Conference onHydropower. ed. Daniel J. Mahoney, 1257-66.Reston. VA: American Society of Civil Engineers.

Weingast, Barry R., and Mark J. Moran. 1983."Bureaucratic Discretion or CongressionalControl? Regulatory Poiicymaking by theFederal Trade Commission."' Journal of Polit-ical Economy 9\ (5): 765-800.

Wood, B. Dan. 1988. "Principles. Bureaucrats, andResponsiveness in Clean Air Enforcements."American Political Science Review 82 (1):213-34.

Yaffee. Steven Lewis. 1982. Prohibitive Policy:Implementing the Federal Endangered SpeciesAct, Cambridge, MA: MIT Press.

Copyright of Land Economics is the property of University of Wisconsin Press and its content may not be

copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.

balancing environmental protection and energy production

Documents