Tài liệu Studies in Public Regulation doc

Analysts customarily measure totalbenefits and a limited set of aggregate total costs estimated by comparingthe equilibrium prices and outputs predicted with the static economicmodel.2 O

Volume Title: Studies in Public Regulation

Volume Author/Editor: Gary Fromm, ed.

Volume Publisher: The MIT Press

Chapter Author: Robert A Leone, John E Jackson

Chapter URL: http://www.nber.org/chapters/c11433

Chapter pages in book: (p 231 - 276)

5 The Political Economy of Federal Regulatory Activity:

The Case of Water-Pollution Controls

Robert A Leone John E Jackson

Increasingly, policymakers have resorted to regulation of private corporateactivity as a means of achieving socially desirable ends (Schultze 1977;Leone 1977) Despite this growth in regulation, there has been littleinvestigation of the dynamic political process by which regulations areformulated and implemented There has also been little systematic analysis

of the dynamic economic process of regulatory compliance

Several aspects of the regulatory approach to public problem solvingmerit investigation Perhaps the most obvious question is whetherregulations achieve desired ends It is certainly possible for Congress tomandate certain goals and to establish bureaucratic machinery to promul-gate the necessary rules; yet these two acts alone do not guaranteeattainment of the stated objectives We leave examination of this mostbasic question to others who are already so engaged.1

To address these questions we present a model of policy-developmentprocesses and industry-response processes At the core of this modelare the economic costs and benefits created by regulation and their distri-bution among firms and regions Analysts customarily measure totalbenefits and a limited set of aggregate total costs estimated by comparingthe equilibrium prices and outputs predicted with the static economicmodel.2 Our model provides for the important role that distributionaleffects and industry dynamics play in determining regulatory impacts.Distributional effects are required because, among other things, theycreate many pressures on the political organizations that develop andadminister policies The dynamic analysis is motivated by the hypothesisthat the constraints and difficulties firms encounter in the short run inattempting to adjust to specific regulations have important aggregate anddistributional implications These short-run effects relate to the availability

of capital and the ease with which different firms can adjust their capitalstock and their manufacturing and marketing strategies to new conditions

Economic Impacts of Regulation

Our analysis begins with costs The costs of regulation are more difficult

to define and more uncertain than those of other public activities Inpublic-works projects, for example, the principal cost uncertainties areorganizational (such as unforeseen delays and unanticipated obstacles toconstruction) and economic (inflation), and are largely exogenous todecisions about the project itself Costs of most business regulations alsoare subject to organizational and external economic uncertainties But, inaddition, these costs depend on factors internal to the regulated industry(such as the rate and direction of technological change), on the existence

of capacity pressures within an industry, and on differences in costsbetween new and existing facilities Costs also are sensitive to uncertaintiescreated by the regulatory process itself: How much time will be left forcompliance? Are standards likely to change? Will enforcement be uniformand equitable? Stated differently, the definition of costs for a public-worksproject is basically an engineering and managerial exercise; the identifica-tion of costs associated with a business regulation is primarily an exercise

in dynamic economic and political analysis, with all the attendant culties and uncertainties this implies This distinction is intended to stressthe variety of methodological approaches that may be required to analyzeregulatory policies

diffi-When not seen from an engineering perspective, costs usually areviewed from the standpoint of the competitive-market model and asso-ciated static equilibrium Viewed this way, regulations prohibit certainproduction processes, require additional capital and operating expendi-tures, and increase some factor prices, thus shifting the long-run supplycurve within an industry upward This method is deficient in two impor-tant ways: It only estimates aggregate costs, and it ignores all short-runand dynamic adjustment problems

Distribution of Economic Impacts

The focus on total costs obscures some very important characteristics ofregulatory costs From the standpoint of aggregate efficiency, compari-sons of total costs and estimated benefits may be an appropriate decisioncriterion However, decisionmakers' objectives are not solely focused onthis criterion Hidden within any specific set of aggregate costs are highlyvariable consequences for different plants within a firm, for firms within

an industry (and among industries, for that matter), and for regions of thecountry These distributional effects may run counter to other policy

Cumulative Percentage of Industry Capacity

Figure 5.1 Compliance-cost curve.

objectives, such as antitrust goals or regional development concerns Atthe same time, the firms and regions most affected presumably will work

to influence policy choice Thus, any eventual policies will not be basedsimply on aggregate efficiency effects, but will reflect accommodation topolitical pressures created by distributional effects.3

Estimating the distributional impacts of national policies is particularlydifficult For a variety of reasons, the incidence of compliance costswithin an industry and among regions need not be uniform Plant-to-plantdifferences in costs may be quite large, depending on the nature of regula-tions, the ages and values of existing capital stocks, and the constraintsplaced on manufacturing processes by regulations

However, if the cost structure of each plant in an industry were known,

we could measure disaggregated effects of regulatory policies from shifts

in the cost curve of each plant brought about by new regulations; and ifindividual plants were then arrayed in descending order according to theiraverage unit costs, an industry cost curve could be generated

Figure 5.1 shows the distribution of industry costs due to a hypotheticalregulation that results when the compliance costs of individual plants arearrayed in descending order The vertical axis represents the unit cost ofcompliance, the horizontal industry capacity For this hypothetical regula-tion, 60 percent of the industry (in terms of capacity) can comply with

a unit cost increase of $1 or less, while for 10 percent of capacity pliance increases costs by more than $5 per unit of output

com-The importance of figure 5.1 is that it can be used to determine whichplants will be hardest hit by proposed regulations and which actuallymay benefit If an aggregate economic analysis at the industry levelyields an overall price increase of $2, close to 75 percent of industrycapacity will receive revenue increases that exceed their compliance costs.For the remaining proportion of the industry capacity, represented bythe rightmost portion of the curve, the regulations entail an economicloss.

Identifying where individual plants fall on the cost curve is an importantcomponent of our impact analysis Even if the aggregate costs are less thanthe aggregate benefits, the incidence of costs on certain plants may conflictwith other policy objectives, raising questions of whether a program should

be implemented For example, various national policies encourage petition in manufacturing industries and try to prevent the economicdecline of various regions Yet if heavily impacted plants are those ofsmaller producers or are concentrated in a specific region, the effects ofproposed regulations may be to increase concentration in an industry or

com-to exacerbate regional economic disparities These possible deleteriouseffects ought to be identified and considered prior to the promulgation

of regulations Other governmental tools may make it possible to come such unwanted side effects, but they are seldom used when regu-lations are taking effect and are altering the structures of industries andthe regional distribution of jobs and income

over-Short-Run Economic Impacts

Accurate assessment of intraindustry and interregional impacts requiresunderstanding how individual plants may respond to proposed standards,

in both the short and the long run Thus, we try to model how regulatoryconstraints affect representative plants and how effects on individualplants are distributed among firms and regions This requires identifi-cation of the age and other attributes of the capital stock of variousplants, their specific production processes, and their existing effluentcontrol measures It becomes critical to specify required productionchanges and capital investments for various plants These conditionsdetermine both short- and long-run consequences for the industry Newplants may have compliance costs substantially different from those ofolder plants

Identification of short-run distributional effects becomes very tant when considering alternative regulatory policies Firms and regionsthat see themselves suffering from promulgated regulations—even if only

SLR

QLR QSR Quantity

Figure 5.2 Short-run and long-run equilibrium D: demand SSR: supply, short-run

(based on ascending average variable costs) SLR: supply, long-run (based on total economic costs, including return on investment) QLR, QSR: equilibrium quantity, long-run and short-run respectively PLR, PSR: equilibrium price, long-run and short-run respectively.

temporarily—are likely to mount campaigns aimed at altering theregulations, weakening their enforcement, or simply preventing furtherregulation Conversely, firms and regions benefiting from regulations,either because they receive benefits at relatively low cost or because theygain financially, can be expected to oppose changes in regulations.Our model is expanded to consider short-run effects as well as distri-butional impacts In figure 5.2 we depict a demand curve (D), a short-runsupply curve (SSR), and a long-run supply curve (SLR) For the moment,assume that D does not change over time SSR is an upward-slopingcurve which, as described, arrays individual plants in an industryaccording to ascending average variable costs

SLR is drawn as a horizontal line, representing the underlyingassumption that in the long run an effectively unlimited supply of newcapacity can be brought on line at the average total cost (including areturn to capital) of the lowest-cost source of new capacity

Understanding the relationship between SSR and SLR is critical todetermining the impact on industry of a government regulation As drawn

in figure 5.2, new capacity is relatively costly We could just as easily

i i i i

i

+ VC SSR + TEC

/ SSR

/ / / /

Figure 5.3 Short-run and long-run equilibrium with added costs of regulation (case 1;

PSR' < PLR') TEC: total economic costs of regulation (including return on investment) VC: variable costs of regulation All variables with prime are the after-regulation equivalents of the unprimed variables See figure 5.2 for definitions of other abbreviations.

have depicted new capacity as relatively inexpensive—owing, perhaps,

to scale advantages, technical change, or factor substitution The point,

of course, is that the actual relationship between SLR and SSR is anempirical question

Furthermore, the path the industry follows from SSR to SLR depends

on a number of factors, perhaps the most significant of which is theeconomic longevity of existing facilities Thus, as drawn in figure 5.2,the long-run equilibrium quantity QLR is less than the short-run equili-brium quantity QSR The time path of adjustment from QSR to QLRdepends on how rapidly the existing capacity is retired

If the highest variable cost capacity is retired first, then at someintermediate point in time the supply curve will be marked by the pointsABC and the long-run price and output levels will have been reached

As more old facilities are retired the supply curve will continue to shiftfrom ABC to SLR, but in so doing it will merely dissipate the quasirents

of existing facilities without influencing price and output levels.4

In figure 5.3 government regulation is imposed on this situation, andits costs shift both the short- and long-run supply curves upward Considerfirst the impact on the short-run curve The curve labeled SSR + TECreflects the old short-run supply curve plus the total economic costs

of compliance This cost recovery need not imply price increases equal

to cost increases, however As part of this recovery, the profit-maximizingfirm will count contributions to sunk costs of all prior investments,which it would have to forgo if it did not comply This new curve is acombination of variable and total economic costs The maximumcapacity a rational firm will bring into compliance will be QSR

Once compliance investment decisions have been made, short-runbehavior will be predicated on an industry's variable cost structure Thecurve marked SSR + VC in figure 5.3 is such a curve, for it reflects theold short-run supply curve, the variable costs of government regulation,and the capacity constraint implied by QSR This short-run curve isnow vertical at the desired quantity level, because government regulationsforce firms to rationalize the industry's capital stock

A similar vertical shift in SLR is due to added costs of regulation fornew facilities Again, depending on the rate of retirement of old capacity,there is a time path of adjustment from PSR', OSR' to PLR', OLR'.Nothing in this logic just described requires PLR' to be greater thanPSR' Indeed, if the costs of retrofitting existing capacity are high andthe incremental costs of compliance in new facilities are low, it is quitepossible for PSR' to exceed PLR' (see figure 5.4.)

Obviously, government regulation, by shifting long-run supply costs,will influence the total size of an industry; as depicted, however, it willnot influence profit margins of new plants However, returns (or quasi-rents) to existing plants will be materially affected by both the height ofany vertical shift in the supply curve and the time path of adjustment

to the new long-run equilibrium Furthermore, the economically criticalfactor is not the average vertical shift, but the shift that occurs at therelevant margin This may be more or less than the average For example,

if any industry's high-cost producers also have relatively high compliancecosts, then the marginal cost of compliance will exceed average costsand new quasirents will be created by regulation If compliance costs ofmarginal facilities are low, then regulation will dissipate some quasirents.(In both cases, some quasirents are terminated as the industry contracts.)Whether the net impact on rents is positive or negative is an empiricalquestion we will address below in the context of tissue manufacturing

It should be clear that to calculate impacts of regulations we mustconsider short-run effects, and not merely show differences in long-runcosts, prices, and outputs To obtain a better estimate of regulatorycosts, the observed time stream of these cost and price increases needs

to be discounted

The above discussion was predicated on the assumption of constantdemand More realistically, demand is likely to shift over time In someinstances, this shift will be upward (as income grows, for example); inother instances, the shift will be to downward (as when lower-cost foreignsupplies become available) The addition of a dynamic element to demandonly reinforces our conclusion on the importance to impact analysis ofthe time path an industry follows in adjusting to government regulation

In this same vein, any movement from a short-run to a long-runequilibrium requires investment, the timing of which can significantlyinfluence the costs of compliance For example, Leone et al (1975) con-cluded that the annualized price per ton customers would pay as a result

of the Federal Water Pollution Control Act Amendments of 1972 could

be as low as $2.45 per ton if expenditures for pollution-control devicesdid not necessitate deferral of investments in production capacity Incontrast, if investments in new capacity were deferred on a dollar-for-dollar basis to allow for financing of pollution control devices, theannualized price per ton could be as high as $14.20 Under both as-sumptions, the estimated long-run (15-20 years) price per ton wasvirtually the same, as would be expected.5

Furthermore, short-run costs may have substantial impacts in the longrun because they may seriously affect the competitive structure of anindustry If high short-run costs fall disproportionately on the smaller,more marginal plants and firms within an industry, these firms may not

be able to stay in business The result would be a more oligopolisticindustry by the time the "long run" was attained If this is the case,then long-run market conditions will be different from those described

by the competitive-market equilibrium discussed above

There is one further important reason for considering these short-rundistributional effects: Members of the public at large, as well as variousinterest groups, may substantially alter their support for intended policyobjectives if they do not perceive the costs initially or if they feel theyare bearing a disproportionate share of the costs People may favorimproving water quality and support legislation promising to do sowhen costs of the program are as vague and hidden as they are in theenvironmental area; they may also change their positions radically oncecosts are perceived

In her insightful article, Dorfman (1975) estimates the total economiccosts of air and water pollution-control programs passed in the 1970sand shows that the near-term costs are substantially greater than thelong-run costs She further suggests that the vast majority of these costsare in the form of higher industry costs, which will be passed on toconsumers We think it is fair to speculate that these costs may nothave been accurately perceived and fully discussed when the legislation waspassed and may have turned out to be far greater than what the public

is willing to pay for improved environmental quality Once these costsbegin to be perceived, which will occur as regulations are written andenforced, political support for these programs may erode Our model ofpolicy development must take such changing forces into account

We have briefly outlined the model necessary to define and measureeconomic impacts of proposed regulations on an industry The remainder

of this paper applies the model to development of the 1972 amendments

to the Federal Water Pollution Control Act and subsequent rulemaking

by the EPA for the pulp and paper industry We hope to demonstratehow the model is estimated in practice, to indicate the magnitudes ofquasirents being created and dissipated by the EPA, and to show whereand how costs and rents influence regulatory policy

Economic Impacts and Congressional Consideration of Water-Pollution Regulations

The 1972 amendments to the Federal Water Pollution Control Actrepresent a major turning point in the public approach to water-pollutioncontrol Previous legislation required state enforcement of policies based

on local water quality There was a general feeling that this approachhad not been effective because each state had a considerable incentive

to protect its economic base and thus not to set or enforce strong standards.The 1972 amendments decreed that policy would now be based onuniform effluent standards defined and enforced by the federal Environ-mental Protection Agency, regardless of the quality of the receivingwater or local economic impacts The act, as passed over a presidentialveto, established 1985 as the date for "ending all discharge of pollutants"(EDOP) and set interim standards for 1977 based on "best practicabletechnology" (BPT) and for 1983 based on "best available technology"(BAT) Not only are the costs of attaining these standards large; theyare anything but uniform in their economic impact on different regionsand on competitors within an industry

The passage of this legislation provides important clues about possiblereasons for continued use of the regulatory approach to solve importantsocial problems Prior to passage, there had been growing concern amongthe American people about environmental problems Opinion pollsrecord that between 1965 and 1970 the proportion of persons who saidwater pollution was a "serious" or "very serious" problem rose from

35 percent to 74 percent (Erskine 1972)

Numerous politicians tried to capitalize on this concern by proposingnew legislation The most notable was Senator Edmund Muskie (D,Maine), who was preparing for his expected presidential campaign of 1972(Ingram 1979) Muskie was chairman of the Air and Water Subcommittee

of the Senate Public Works Committee, and thus had a major role indrafting the original version of the 1972 amendments Muskie had beenstung by accusations from Ralph Nader's organization that he was weak

on water-quality policy He responded by promoting a bill embodyingthe regulatory approach with strong effluent standards The full SenatePublic Works Committee unanimously (16-0) reported the Muskie billwithout questioning the regulatory approach or the likely economicconsequences The Senate passed the bill unanimously (86-0) five dayslater, essentially unchanged

The Senate's biggest concern appeared to be the expenditures the

federal government ultimately might incur The committee supplementalreport (filed by five Republicans) and two of the three floor roll-callvotes concerned provisions in the bill providing federal grants to localgovernments for constructing water-treatment facilities (The other votewas on a proposal to allow the Small Business Administration to establish

a subsidized loan program for small businesses finding it hard to raisenecessary capital to install pollution-control devices.) The only alternative

to the regulatory scheme was a proposal from Senator William Proxmire(D, Wisconsin) to include an effluent-tax system This proposal wasdebated briefly and then rejected on a voice vote

In the House of Representatives, the Public Works Committee sidered the bill in early 1972 and reported its version in March 1972 Therewere several marked differences between the Senate and the HouseCommittee versions The most important was that the House wouldestablish as goals rather than as mandatory the zero discharge standardfor 1985 and the BAT requirement for 1981 (later amended in conference to1983) The House version also required a study by the National Academy

con-of Sciences on the environmental, technological, economic, and socialfeasibility of meeting the 1981 and 1985 goals prior to the mandating oftheir implementation.6 Thus, the House Public Works Committee ex-pressed concern over the uncertainty of economic impacts and sought todelay implementation until further study However, it did not questionthe regulatory approach or the ultimate objectives In dissenting reports,Democrats Bella Abzug and Charles Rangel of New York supported aset of modifications, known as the Reuss-Dingell Clean Water Package,and proposed several amendments of their own These proposals wouldhave reinstituted the Senate language mandating the 1981 and 1985policies and added other strengthening requirements A minority reportfiled by Republicans Roger Zion (Indiana) and John Terry (New York)expressed concern about the long-run economic impacts of the regulationsand requested more information on potential price increases, employ-ment impacts, balance-of-trade effects, and budget commitments Theydid not propose any specific alterations to the committee version, however.House floor debate focused largely on the stronger Reuss-Dingell amend-ments, which were defeated, and the municipal-treatment grant program.The final version, passed by the House on a 380-14 vote, was very close tothe version reported by the House committee The final version of the bill,

as reported by the conference committee and passed over PresidentRichard M Nixon's veto, kept the EDOP goals, but mandatedimplementation of BAT standards in 1983

What role did potential costs and economic impacts play in the passage

of these 1972 amendments? There was, of course, the expected conflictbetween industry, labor, and environmentalist groups over the proposedregulations Presumably those individuals who would bear a smallproportion of the costs and who value environmental quality would bemore likely to support the legislation and to expect their representatives

to do likewise Beyond that, our model suggests that some firms andregions will be more disadvantaged than others; some may end up betteroff, depending upon their location on the compliance-cost curve infigure 5.1 We would then expect to find differences in positions taken

by different firms in their testimony before congressional committees,

in the behavior of various representatives, in industry and regionalpresentations and comments to the regulatory agency, and in firms'participation in lawsuits filed over the promulgated regulations, de-pending upon relative economic impacts Our concern is how influentialthese various political activities may be in policy processes

Perhaps the most conspicuous arena for these political considerations

is the House of Representatives, which is organized on a regional basisand comprises legislators closely tied to district concerns It would besurprising if the expected or most obviously perceived distributionalimpacts of proposed policies were not influential in the congressionaldecision process Thus, by examining regional distributions of costs andbenefits one ought to be able to explain, at least in part, the behavior ofthe House in establishing the regulatory policy

We have already commented that total costs of the 1972 Water PollutionControl Amendments apparently were not considered in the developmentand passage of the legislation However, some of the direct cost andemployment effects were sufficiently large and concentrated to haveinfluenced congressional decisions Other things being equal, such aslocal pressures for environmental programs, it would be expected thatrepresentatives from districts with large concentrations of potentiallyimpacted businesses would be more likely to oppose stronger versions

of the bill than representatives from districts likely to suffer little adverseeconomic impact

The amendments offered and voted on in the House of Representativesprovide an opportunity to examine how these direct regional consequencesinfluenced the legislation Table 5.1 shows the six (of nine) recorded floorvotes taken in the House, which we will analyze.7 The amendments areordered in terms of increasing regulatory strength, and presumably inseverity of impacts on industry For example, passage of the Abzug and

Table 5.1 House votes on the 1972 Water Pollution Control Amendments.

Location of Amendment Included in Analysis Relative to Vote for Passage a

Passage of the bill Passed 380-14 0.00

McDonald (R, Michigan) Amendment to exempt

industries from paying capital costs on federally funded

municipal waste treatment plants which they use, in

addition to paying user charges for maintenance,

operation, and expansion Rejected 66-337 A Nay 0.56

vote was considered to be in support of the act (0.12)

W Ford (D, Michigan) Amendment to guarantee

public hearings in EPA investigations of employee

firings or layoffs resulting from effluent limitations or 1.52

orders under the act Adopted 275-117 (0.15)

Reuss (D, Wisconsin) Amendment to require adoption

of toxic-pollutant standards and effluent limitations

before EPA could transfer permit programs over to

the states, and to give EPA permit-by-permit veto

power over state programs; amendment would also

have eliminated a provision in the bill giving immunity

until 1976 to polluters who applied for discharge 2.72

permits Rejected 154-251 (0.17)

Reuss (D, Wisconsin) Amendment to require industries

to use the "best available" water pollution control 3.04

technology by 1981 Rejected 140-249 (0.17)

Abzug (D, New York) Amendment to require impact

statements under the National Environmental Policy

Act of 1969 for all activities covered by the bill 3.50

Rejected 125-268 (0.18)

a For an explanation of the entries in this column see the discussion in the text.

Reuss amendments would have produced a much stronger bill with moresevere impacts on manufacturing firms Conversely, the McDonaldamendment was designed to limit impacts on firms by exempting themfrom capital costs of new municipal waste-treatment plants

We have categorized legislators who voted according to which ments they supported or opposed to obtain a measure of support for thelegislation For example, someone voting against all amendments andthe bill itself was placed in the first category, while someone who justvoted for the bill is put in the second category, and so on Representativeswho voted for all amendments fall into the seventh, or highest, category.8

amend-The distribution of representatives by respective categories is given intable 5.2 Fifteen representatives did not cast enough votes to be located

on the scale and are omitted from further analysis The scale has a fairlybroad distribution, with only some bunching at the upper end Thus,

Table 5.2 Distribution of representatives voting,

repre-Table 5.3 Model of House voting on the 1972 Water Pollution Control Act Amendments.

Dependent variable: Scale category from table 5.2.

Population density (log)

Median house value x % owner

Median rent x % renter

Median age in district

(population over 25)

% suburban

Northern Democrats

Southern Democrats

Age of congressman (log)

Cost 11 Primary metals

- 0 4 6 0.30

- 0 3 1 0.10 0.01 0.62

0.06 0.38 1.08 0.66

- 1 4 8 0.24

- 0 0 2

- 0 5 3

- 0 1 8 -0.003

Standard Error 1.64

0.22 0.28 0.23 0.21 0.35 0.22 0.04 0.21 0.54

0.02 0.26 0.15 0.28 0.32 0.18 0.32 0.24 0.19 0.03

a These coefficients and the coefficients reported in table 5.1 were estimated using

n-chotomous multivariate probit techniques.

b Maryland, West Virginia, Kentucky, Tennessee, Missouri, Oklahoma.

c Virginia, North Carolina, South Carolina, Georgia, Florida, Alabama, Mississippi, Louisiana, Arkansas, Texas.

d Michigan, Wisconsin, Minnesota, North Dakota, South Dakota.

e Indiana, Illinois, Iowa, Nebraska, Kansas.

f Montana, Idaho, Oregon, Washington, Alaska, Wyoming.

g Utah, Colorado, New Mexico, Arizona, Nevada, California, Hawaii.

h The primary metals, mining and petroleum extraction variables are dummy variables equal to 1 if the industry is important to the district's economy (as reported by the Almanac

of American Politics) Paper production is tons of capacity from Lockwood's Directory of

the Pulp and Paper Industry The cost/ton data are derived from Leone et al 1975.

it should provide a good measure of the positions of most legislators onthe issue of water-pollution-control regulations.

The next step is to relate these positions to demands for environmentalquality and to the economic impacts in each district We focus on thepulp and paper industry, and we want to know specifically if the likelihood

of a representative's supporting strong environmental legislation decreases

if there are pulp or paper mills in the district, and if this support is furtherreduced if likely compliance costs for district mills would increase.The model and the statistical method used to explain House voting

on the 1972 amendments are described elsewhere (Jackson and Leone1978), but we can easily summarize its contents and results The majordeterminants of a representative's position on the legislation are hypothe-sized to be the district's demand for environmental cleanup, the likelyeconomic consequences for the district, and the representative's partyaffiliation and personal preferences The precise variables used in theanalysis, their estimated effects on legislators' positions, and the standarderrors of the coefficients are shown in table 5.3.9

The statistical method provides an estimate of relative location, orspacing, of votes used to constitute our scale of support for strongerwater pollution regulation The numbers in table 5.1 are these estimatedlocations, with the vote for passage arbitrarily defined as zero to locatethe scale (The parenthetical numbers are standard errors of estimatedlocations.) The three amendments in the Clean Water Package are locatedclose together, while the largest distances are between these amendmentsand the Ford amendment, and between the Ford amendment and thetwo weakest votes The magnitudes of distances between votes aid ininterpreting coefficients in the underlying model For example, it wouldtake a difference of at least 1.2 in estimated positions of two representatives

to bridge the gap between the Ford amendment and the first Reussamendment

The specification of demand variables is based on a model of people'swillingness to pay (in dollars) for environmental cleanup, as estimatedwith data from a 1969 Harris survey (Jackson 1979) This model showsthat willingness to pay is strongly related to a person's region and place

of residence, age, family size, education, and income.10 Unfortunately,data on income and education levels of congressional-district residents

is not available for 1972.11 We hope that inclusion of housing-stockvariables serves as a proxy for these characteristics The population-density variable is included to represent estimated willingness-to-paydifferences between rural and metropolitan areas Coefficients on these

variables are consistent with the demand model Only a district's medianvoter age did not perform as expected This may be explained by thesmall variation in this variable and by the fact that what variation exists

is largely attributable to the location of large military bases, with theiryounger nonresident populations We thus conclude that representatives'positions were related to the environmental demands of their constituents.The party-affiliation variables are self-explanatory Northern Demo-crats exhibited much stronger support for the legislation than did Re-publicans, with the support of southern Democrats somewhere in between.The results are consistent with the pressure President Nixon put on theHouse to weaken the bill The age of legislators is used to proxy theirown preferences, and this has the expected sign: Younger representativesshow preferences for more stringent regulations

In addition to economic impacts associated with the pulp and paperindustry, we make a crude attempt to include the presence in districts ofother industries likely to be adversely affected by the regulations The

Almanac of American Politics (Barone et al 1972) gives a brief summary

of the industrical base of each district, compiled from the 1970 census.The variables, primary metal, mining, and petroleum extraction, aresimply dummy variables based on whether the almanac mentioned theappropriate economic activity As such, they may not be completelyreliable, particularly for metropolitan areas where census data are notavailable on a congressional-district level The authors of the almanacadmitted that for these districts the description is based largely oncharacteristics of the entire metropolitan area This difficulty may helpexplain the wrong sign on the primary metal variable, since many steelmills are located in metropolitan areas It may also be the case thatwater quality is lower in such districts and the legislator was voting forcleanup in spite of economic effects

Local compliance costs for the pulp and paper industry are our ticular concern here The importance of this industry to a district ismeasured by production capacity in thousands of tons per day of mills

par-in the district Cost data, par-in dollars per ton to meet 1983 BAT standards,are from the National Commission on Water Quality (NCWQ) study.The NCWQ model used for estimating these costs is the predecessor tothe industry model described elsewhere in this paper Our hypothesis isthat the greater the estimated cost of complying with the 1983 standards,the greater the adverse economic impact on a district—in terms of bothpersonal income and employment—and the less likely a representativewas to support stronger versions of the 1972 amendments.12

There are two alternative hypotheses about the expected signs on thetwo pulp and paper variables A naive hypothesis is that paper companies(and, presumably, representatives from districts with economies based

on pulp and paper) are sensitive to any regulation that may raise theircosts and reduce output and profits Particularly, given uncertainty abouthow regulations might be written and enforced, firms may simply bewary of the unknown and oppose any government regulation If this isthe case, we would expect congressional voting to be sensitive to theamount of pulp and paper production in a district, but relatively insensitive

to costs, which were only determined after the EPA began to defineindustry categories and establish effluent standards The hypothesis based

on more sophisticated behavior is that firms are aware not just of theirown expected pollution-control costs, but how these costs compare withthose of competing firms With this knowledge, firms estimate theirexpected change in net worth, not just cost increases This calculation isbased on the shape of the compliance-cost curve of figure 5.1 According tothis analysis, firms in the left-hand portion of the curve potentially stand tohave increases in net worth because expected price increases resultingfrom a shift in the aggregate supply curve may exceed their compliancecosts, resulting in increased profit margins Sophisticated firms in theleft portion of the curve may actually gain economically from imposition

of pollution standards and might be expected to support the legislation

A large negative coefficient on the cost multiplied by the productionvariable and a positive (or at least a zero) coefficient on the productionvariable would support this hypothesis

The estimates do not support the more sophisticated model of firmbehavior, but indicate that representatives from districts with pulp andpaper mills were less likely to support the legislation Although thestandard error of each coefficient is large, the null hypothesis that pulpand paper presence and costs have no effect (that is, that both coefficients

equal zero) can be rejected at the 0.01 level The x 2 value for this pothesis is 10.0 with 2 degrees of freedom (If the production capacitymultiplied by cost per ton variable is deleted, the coefficient of productioncapacity is —0.20, with a standard error of 0.06.)

hy-Overall, voting was not sensitive to estimated compliance costs For

a district with the maximum capacity (6,500 tons per day), the predictedeffect on congressional voting of a $10 per ton cost difference (themaximum difference among districts) is only — 0.20.13 For smaller costdifferences or for districts with less capacity, expected voting differenceswill be even smaller

In contrast, the mere presence within a district of 6,500 tons of dailypulp and paper production capacity at the average compliance cost of

$6 per ton affects a representative's vote by —1.30.14 Each 1,000-tondecrease in capacity increases the expected support for pollution-controlregulations by 0.20 at the average compliance cost of $6 per ton Thesignificance of these magnitudes can be ascertained by consulting table5.1 for differences in locations of votes

We conclude from these results that congressional voting on the 1972amendments was sensitive to potential direct economic effects on adistrict, but in a rather unsophisticated manner Representatives did notseem to be pressing a possible economic advantage to their region bysupporting legislation that would give local mills a competitive advantage;they simply opposed regulations affecting local industry

An obvious rationale for the relative unimportance of variations incompliance costs in explaining congressional voting is that intraindustryimpacts were not known at the time the legislation was considered Thedevelopment of a regulatory policy is largely defined by the stream ofadministrative decisions made by the regulatory agency once the legisla-tion is passed In the case of the 1972 Water Pollution Control ActAmendments, the bill simply specified that industry had to satisfy effluentstandards consistent with the BPT by 1977 and the BAT by 1983 forgiven industries and industry subcategories It was left to EPA to definethe subcategories, to specify what constituted BPT and BAT standardsfor various industries and subcategories, and to establish the norms Onlywhen the EPA begins this process are firms able to predict how they will

be affected Without these predictions, one might expect a strong, generalopposition to the concept of being regulated (possibly in anticipation ofadverse economic consequences), with no variation in opposition inresponse to variations in economic impacts

Administrative Rulemaking and Regulatory Impacts

The fact that distributional consequences within an industry (and thusbetween regions) are not defined until rulemaking regulatory processesbegin has strong implications for the effect of distributional impacts onpolicy processes The regulating agency (in this case the EPA) determinesthese impacts, which implies that the agency and not the Congress becomesthe focus for the political forces they generate The question now to beasked is: How, and to what extent, are an administrative or regulatoryagency's rulemaking and enforcing decisions affected by pressures related

to the economic consequences of their decisions? A brief glimpse at thisprocess, and at possible influences of economic effects, is obtained bynoting that the EPA published three different versions of BPT and BATstandards for a large section of the pulp and paper industry Thesedifferent versions and solicitation of industry and public comments arepart of the required rulemaking procedure Subsequent alterations ofthe standards and industry categorizations provide clear evidence ofaccommodation to various pressures from segments of the industry andfrom environmentalists.

Investigation of the question posed in the preceding paragraph andanalysis of different standards proposed by the EPA require that weestimate for a segment of the pulp and paper industry the detailed modeldescribed in the first section of this paper This is done for the tissueportion of the industry

In estimating impacts of regulations, first it is necessary to identifythe distribution of costs the tissue industry will confront in meeting theregulations That is, we must estimate the compliance-cost curve (figure5.1) Second, these costs must be translated into price effects and microlevelimpacts The specific procedures used to calculate the costs the tissueindustry will incur in complying with mandated water-pollution reduc-tions and associated total costs of manufacture are described elsewhere(Leone 1980) Here, we merely note that these costs are estimated on aplant-by-plant basis This is done by taking cost levels estimated for

"representative" or hypothetical mills and regressing them on variousmill characteristics The resulting equations permitted estimation of costsfor sixty-four existing mills in the industry.15

In calculating pollution-control costs, we assume that each plant willminimize the discounted present value (with an interest rate of 15 percent)

of compliance costs for the anticipated sequence of 1977, 1983 and 1985standards, and that all in-process or end-of-pipe changes that reducepollution loads and yield a 15 percent return will be made Occasionally,this assumption may not be valid For example, the 1985 standards aremerely a "goal" in the 1972 act If they are not enforced, then on a present-value basis a company may take a different course of action to meet the

1983 standards Furthermore, we ignore issues of regulatory uncertainly;that is, we assume that standards are known and will be strictly enforced.The principal benefit of this microlevel cost orientation is that it permitssimulating the distributional consequences of regulation which we arguedare so important to understanding the political economics of businessregulation

Capacity (Percent of Industry Total)

Figure 5.5 Estimated 1977 BPT compliance-cost curve (economic-cost basis).

The estimated compliance-cost curve for the tissue portion of the pulpand paper industry for 1977 BPT effluent control levels is shown in figure5.5 The BPT costs range from $1.85 to $82.82 per ton.16 About 80 percent

of capacity has unit costs of $12.34 per ton or less The average economiccost for the tissue industry is $9.41 per ton

Figure 5.6 is identical in construction to figure 5.5, but shows variousdefinitions of total manufacturing costs estimated in 1974 prices.17 Themiddle curve reports total costs of production on an accounting costbasis before the act; thus, it excludes all water-pollution-control costsassociated with the act except those with a rate of return of 15 percent

on the required investment The range of production costs is substantial:from $556 to $693 per ton The weighted average cost of $613 per ton

is exceeded by thirty-five of the sixty-four mills in our sample, whichindicates that the lower-cost mills are predominantly the larger ones.Although it cannot directly be discerned by comparing figures 5.5 and5.6, it is worth noting that there is no obvious correlation between amill's BPT compliance costs and its total manufacturing costs It is notthe case, for example, that mills with high production costs necessarilyhave high BPT compliance costs Accordingly, to understand the eco-nomic consequences of BPT regulations it is not sufficient to examine

$638 per ton (Arthur D Little, Inc., 1974) In view of the extensivepossibilities for error in our cost-estimation procedures, these observedprice and utilization levels are quite consistent with the numbers shown

in figure 5.6 Indeed, at a 94 percent utilization rate, figure 5.6 implies

a price of $627 per ton; at a price of $638, figure 5.6 forecasts a utilizationrate of 97 percent These predictions are very close to the observed values

Table 5.4 The tissue industry's costs of production with BPT controls in place (1974 prices,

accounting-cost basis), by mill.

10 12 17 18 21 22 31 32 33 35 37 39 39 41 43 43 47 48 49 49 50 59 60 60 61 65 66 66 67 67 68 69 69 70 71 71 72 72 72 73 73 73 73 75

Mill Capacity (Tons per Day) 648

204 208 453 104 272 68 816 29 116 181 133 172 37 140 181 36 362 45 90 18 72 816 54 45 36 408 72 45 40 31 54 58 58 108 27 36 72 7 22 22 22 31 18 145

Identification

No of Mill 2 4 42 50 57 49 16 1 11 38 55 44 20 30 10 43 12 53 8 63 58 7 47 27 3 22 51 61 33 31 46 9 15 32 5 36 37 59 41 48 62 34 21 26 52

Percentile Ranking of Mill with

No Controls 7 10 13 18 11 21 22 31 32 35 34 41 37 39 39 43 47 47 48 49 48 60 58 59 68 66 64 66 65 67 69 67 68 70 74 69 71 73 47 69 70 74 71 71 77