Does International Harmonization of EnvironmentalPolicy Instruments Make Economic Sense?The Case of Paper Recycling in Europe
ANNI HUHTALA1 and EVA SAMAKOVLIS2
1National Institute of Economic Research, Box 3116, 103 62 Stockholm, Sweden(E-mail: [email protected]); 2Department of Economics, Umeå University, Sweden
Accepted 11 July 2001
Abstract. Harmonization of the instruments used in environmental policy has been considerednecessary to guarantee “fair” competition in international markets. We examine the economic costsof harmonizing paper recycling standards in countries where the urgency of the waste disposal prob-lems differ. Using data of seven European countries we estimate the technologically feasible inputcombinations of pulp and waste paper for paper production. Short-term effects of two environmentalpolicy measures, minimum content requirement and utilization rate target, are analyzed. By translat-ing the two administrative instruments into taxes and subsidies, we show that the shadow costs of theharmonization vary considerably between countries. The difference in the domestic availability ofwaste may explain the variation, and a modification of the policy measures to incorporate this aspectis suggested.
Paper recycling entails two key environmental concerns: the conservation of rawmaterials (energy, forests) and the alleviation of waste disposal problems. Due tothese socially attractive features, promoting recycling has become one of the polit-ically most popular environmental objectives. One approach has been to encouragethe increased use of waste paper in the manufacture of newsprint and paperboard.In the United States, for example, local and national policies have made a certainrecycled content mandatory, and similar policy proposals appear from time to timeon the environmental policy agendas of the European Union member states andinternational organizations.1 The reasoning is that even if sorting and collection ofpost-consumer waste are well organized by public authorities, these measures donot necessarily make firms utilize extensively the post-consumer waste collected.Minimum recycled content scheme2 seem to address exactly the right problem. Ifsuch an environmental standard is to be introduced in domestic markets, it is feltthat foreign competitors should also meet that same standard, because production
262 ANNI HUHTALA AND EVA SAMAKOVLIS
costs may be higher when recycling technology is used as opposed to conventionaltechnology with virgin, or primary, raw material. International harmonization ofenvironmental standards has therefore been considered necessary.
However, a recycled content scheme is not an unproblematic policy instrumentin an international context (e.g., Grossman 1981 and Beghin and Sumner 1992).The availability of waste, or secondary, material depends on domestic consump-tion, i.e., the proportion of post-consumer waste which is recyclable. The higherthe proportion of production that is exported, the more difficult it is to meet theminimum standards of secondary material use, by domestic recycled material.For example, since Scandinavia exports about 90 percent of its paper production,complying with a certain internationally given content scheme of recycled materialcould necessitate a relatively high domestic recovery rate and import of wastepaper. This is exactly what happened in Canada when the US legislation intro-duced minimum levels for recycled fiber content (Roberts and Johnstone 1996).The economic availability of secondary material is also affected by collection costs,which differ from country to country according to population density and transportdistances. The urgency of the waste disposal problems differs. Due to a lack oflandfill space, densely populated areas in Central Europe have had more urgentwaste management problems than the relatively sparsely populated, forest-richScandinavian countries.
Harmonization of policy instruments to promote the use of waste paper maythus have unintended effects on forest management, affect trade patterns signifi-cantly, and even change the location of industries. The importance of studyingpolicies that affect input uses is accentuated by the projections which indicate thatthe world consumption of fiber furnish in paper and paperboard manufacture willincrease from 250 million metric tons in 1990 to 400 million metric tons by theyear 2010 (FAO 1997).
We examine here the economic costs of an internationally harmonized contentscheme for recycled paper in Europe. To illustrate the distributional effects ofharmonization, we show how the common standard for waste paper input could bereached using input taxes and subsidies. We suggest that if the reasons behind needsfor a policy intervention are purely environmental, a socially optimal recovery rateand the domestic availability of waste paper in different paper producer countriesshould be investigated in order to determine an international waste paper utilizationstrategy.
The literature on recycling and “green design” policies is extensive. Forexample, Calcott and Walls (2000), Choe and Fraser (1999), Conrad (1999),Eichner and Pethig (2001), Fullerton and Wu (1998) and Palmer and Walls(1997) present theoretical models to study different policy instruments to promoterecycling. However, they do not consider input endowments, or the extent/severityof the environmental problems related to the amount of secondary material, orwaste paper in different countries. Grace et al. (1978) is one of the earliest studiesto focus on trade in waste paper. More recently, empirical studies by Weaver et al.
ENVIRONMENTAL POLICY 263
(1997) and van Beukering and Duraiappah (1998) seek to minimize the environ-mental impacts of paper product life cycles and illustrate the trade implications forEurope and India respectively. They use operational research techniques and donot aim at maximizing social welfare or consider different policy instruments foroptimal recycling. Copeland (1991) presents a theoretical model of trade in wastedisposal services and studies the welfare effects of restricting such trade.
Analytically, the spirit of our theoretical model on which the content schemeanalysis is based comes closest to that of Bhagwati and Srinivasan’s (1969) paperon non-economic objectives of trade policy. Given the recycling target as a non-economic objective, we compare empirically different policies for reaching thecommon recycling goal to evaluate the effects of harmonization. We are inter-ested in the shadow prices of common regulation, or constraints that harmonizedstandards impose on paper producers in different countries.
The paper is organized as follows. First, we discuss the current state ofwaste paper recovery and raw material use in Europe. Section 3 presents theanalytical framework for our policy comparisons and shows that if the use ofrecycled material in paper production is to be encouraged for environmentalreasons, the policy instruments for this purpose need to be improved to takeinto account country-specific features. In section 4, the empirical analysis iscarried out. Using aggregate production data for seven European countries, weestimate the technologically feasible input combinations of pulp and waste paperfor paper production. Given that currently the input choices within individualindustries are freely optimized, we impose a common standard for waste paperinput to see the extent of relative changes in input uses which this policy measurewould imply for each country. The standard is then “translated” into market-basedinstruments to illustrate how a common recycling goal could be implemented indifferent countries by taxes or subsidies. Finally, we contrast the trade and distri-butional effects of a harmonized policy with our alternative policy design, whichacknowledges country-specific differences in paper trade in order to promote paperrecycling.
2. Current Input Choices and a Need for Environmental Policy Intervention
The bulk of the fiber furnish used for manufacturing paper and paperboard consistsof waste paper and wood pulp.3 The use of waste paper, in particular, has increasedglobally for four main reasons: good price competitiveness of recycled fiber,technological progress, regulations influencing demand for recovered paper, andthe environmental concern of waste disposal affecting the paper recovery sector.Despite the increased use of and trade in waste paper, the composition of fiberfurnish still reflects to a major extent the domestic supply of these inputs in eachcountry. This is shown in Figure 1, which illustrates the use of wood pulp and wastepaper in the production of paper and paperboard in Europe. At the one extreme arethe large producers, such as Finland and Sweden, which use mainly virgin fiber; at
264 ANNI HUHTALA AND EVA SAMAKOVLIS
Figure 1. Share of pulp and waste paper in production of paper and board.
Table I. Paper, pulp, waste paper consumption, production and recovery 1996 (kilotons).
Countries Paper Paper Pulp Pulp Waste paper Waste paper
consumption production consumption production consumption recovery
Austria 1446 3653 1832 1550 1537 1054
Belgium 2633 1328 653 383 361 1020
Denmark 1141 322 61 71 395 615
Finland 1634 10442 8184 9676 575 563
France 9382 8531 4100 2517 4192 3857
Germany 15471 14733 5105 1816 8888 10912
Greece 912 352 120 20 307 300
Italy 8250 6954 3309 540 3515 2531
Netherlands 3166 2988 584 125 2106 2056
Portugal 836 1026 679 1594 315 329
Spain 5171 3684 1282 1461 2774 2125
Sweden 1748 9018 7321 9779 1502 1158
UK 11443 6188 2168 575 4323 4551
Source: Pulp and Paper International (1997).
the other are the small producers, such as Denmark, Greece and the Netherlands,which rely extensively on recycled fiber. Table I displays consumption and produc-tion data for paper, pulp and waste paper in Europe. As will become evident below,the share of domestic consumption of paper in paper production plays an importantrole when considering recycling policies in different European countries.
ENVIRONMENTAL POLICY 265
The extent of waste paper recycling in each country is generally described usingtwo indicators: the waste paper recovery rate and the waste paper utilization rate.4
It should be noted that waste paper consumption refers to the volumes used in theproduction of new paper and board, whereas waste paper recovery equals wastepaper consumption minus imports plus exports of waste paper. The waste paperrecovery rate (to be denoted by α) is defined as the ratio of waste paper recoveryto total domestic paper and board consumption. The utilization rate (µ) is definedas the ratio of waste paper consumption to total paper and board producton. InEurope, the recovery rate increased from 40 percent in 1989 to 49 percent in 1996and is expected to reach 55 percent in 2005. The utilization rate increased from 36percent in 1989 to 44 percent in 1996.
National variations in these rates are considerable, however, as can be seen inTable II. For our analysis, it is important to recognize that low utilization rates donot necessarily mean that the country is not recycling a large share of its paperconsumption. For example, at 17 percent, Sweden’s utilization rate is among thelowest, while its recovery rate – 66 percent – is among the highest. If the policygoal is a high utilization rate, this indicates the importance of waste paper importsfor high-volume paper-exporting countries. The opposite is true for countries suchas Greece, Italy, and Spain, which import a large share of the paper they use: theutilization rates are high even though the recovery rates are low.
The other two ratios in Table II indicate whether a country is a net exporter(or importer) in its trade of paper and waste paper. The parameter γ capturesthe ratio of paper consumption to paper production, and ω is the ratio of wastepaper consumption to waste paper recovery. If γ (ω) exceeds 1, the country is anet importer of paper (waste paper). In 1996, for example, Austria, Finland andSweden were net exporters of paper and net importers of waste paper, whereasBelgium, Germany and the United Kingdom were net importers of paper and netexporters of waste paper.
Given the current production and consumption structure of paper and paper-board in Europe, the harmonization of policy instruments to promote utilizationof waste paper is not necessarily a straightforward task. It is crucial to take intoaccount country-specific differences. If a harmonized standard initially imple-mented for environmental reasons should not principally aim at increasing tradeof waste, then the domestic availability of waste paper would be a measurableindicator of the seriousness of the waste disposal problems where waste paper isconcerned. Utilization rate µ (the ratio of waste paper input to total paper andboard production, R/Y) and proportion of recycled input β (ratio of waste paperinput to virgin material, R/V) are not necessarily the most adequate measures, sincethey conceal the domestic availability of waste paper. Instead, the utilization ratecan be expressed as µ = R/Y = (αγ Y)/Y = αγ , or as a product of α, the recoveryrate, and γ , an adjustment parameter capturing the ratio of paper consumption topaper production in a country. Consequently, an environmental objective seeking
266 ANNI HUHTALA AND EVA SAMAKOVLIS
Table II. Waste paper recovery and utilization rates 1996.
Countries Recovery Utilization Paper consumption/ Waste paper consumption/
rate (α) rate (µ) paper production (γ ) waste paper recovery (ω)
Austria 0.73 0.42 0.40 1.46
Belgium 0.39 0.27 1.98 0.35
Denmark 0.54 1.23 3.54 0.64
Finland 0.44 0.06 0.16 1.02
France 0.41 0.49 1.10 1.09
Germany 0.71 0.60 1.05 0.81
Greece 0.33 0.87 2.59 1.02
Italy 0.31 0.51 1.19 1.39
Netherlands 0.65 0.71 1.06 1.02
Portugal 0.39 0.31 0.81 0.96
Spain 0.41 0.74 1.37 1.31
Sweden 0.66 0.17 0.19 1.30
UK 0.40 0.70 1.85 0.95
Source: Own construction based on figures from Pulp and Paper International (1997).
to promote the input use of recovered waste paper (WP) in paper production isexpressed as a restriction:
R ≥ µ ∗ Y = α ∗ γ ∗ Y = (WP recovery/paper cons.)∗(paper cons./
paper prod.)∗paper prod.
= WP recovery (a)
or the goal for increased input use of waste paper is related to the amount ofdomestically recoverable material, which potentially would end up in landfills, ifthe utilization of waste paper were not actively promoted.5 The decomposition ofutilization rate µ to α and γ reveals that µ would not necessarily be commonto all countries. For a given geographical distribution of recycled raw material asreflected by γ , the waste paper recovery rate α should capture the environmentaloptimality, and in each country, the waste paper recovery rate α should be at a levelwhere the social marginal net benefit from recycling is zero. Next we present ananalytical framework to elaborate the social optimality conditions.
3. Input use Optimization under Alternative Policy Goals
The paper producing industry can use both recycled material (waste paper), R, andvirgin resources (wood pulp), V, as raw material. We assume that these are the onlyvariable inputs used and that the industry production function, for a fixed level oflabor L and capital K, can be represented by f(V, R; L, K).6 Taking the production
ENVIRONMENTAL POLICY 267
level, Y, as given, an objective of the representative industry will be to minimizecosts. The Lagrangian then becomes L = pV V + pRR + λ(Y − f(·)) where pV is theprice of virgin material, pR is the price of recycled material and λ is the Lagrangianmultiplier. The necessary conditions are
pV
pR
= fV
fR
and Y = f (V, R; L, K) (1)
or the optimal amount of virgin and recycled material used in production is deter-mined by the relative input prices. Note that in a competitive market economyinput prices reflect the costs of felling (cV ) and the costs of waste paper recovery(cR), or forest owners maximize �V = pV V − cV V and waste paper recovery firmsmaximize �R = pRR − cRR.
The environmental goal of promoting recycling is often motivated by thesocial costs (benefits) associated with the use of virgin (recycled) material. Letus denote these social costs of wood material by ϕV (fellings leading to loss ofbiodiversity/absorption of carbon dioxide etc.) and the social benefits of recycledmaterial by ϕR (saved landfill space etc.). An objective of a social planner is tominimize production costs and net costs of waste management and forestry. Takingthe production level, Y , and the corresponding amount of waste paper generateddomestically, γ Y , as given, the Lagrangian then becomes
L = pV V + (cV + ϕV − pV )V + pRR + (cR − ϕR − pR)αγ Y
+(cE − pE)(1 − α)γ Y + λ(Y − f (·))where pE is the price of recycled material used as raw material in alternative wastemanagement (e.g. energy value), cR is the waste paper recovery cost (sorting,collection, transportation), cE is the cost of alternative waste management method(landfilling/incineration/export of waste), and cV is the cost of “producing” virginmaterial (felling, transportation).7 To emphasize the importance of the optimalrecovery rate, the necessary condition for an optimal α is written separtely
∂L/∂R = ∂L/∂V = ∂L/∂λ = 0 ⇔ cV + ϕV
pR
= fV
fR
and Y = f (V, R; L, K) (2a)
∂L/∂α = 0 ⇔ pR = cR − ϕR − cE + pE (2b)
Equation (2b) is an optimal condition to guarantee the social marginal net benefit ofrecovery to be zero, or pR − cR + ϕR + cE − pE = 0. The cost of an alternative wastemanagement method (incineration/landfilling/export of waste) cE may vary fromcountry to country. For example, if cE increases (due to, e.g., diminishing landfillspace), then the social marginal benefit of recovery increases, and there shouldbe more utilization. In a similar way, if it is relatively inexpensive to export/burnwaste, the optimization leads to a decreased marginal benefit of recovery. Bycontrasting equations (1) and (2a and 2b) we can see that the optimum is reached
268 ANNI HUHTALA AND EVA SAMAKOVLIS
when relative input prices reflect the production and waste recovery and disposalcosts, or pV /pR = (cV + ϕV )/(cR − ϕR − cE + pE). The objective of a policymarker is to affect input prices as presented in equation (1) in such a way that theyreflect all costs derived in equations (2a) and (2b). A socially optimal recycling andwaste paper recovery policy could be based on α, which is determined by takinginto account environmental considerations, given the availability of waste paperas captured by γ .8 We consider now two intervention cases where environmentalobjectives for the use of recycled material are initially imposed as strict harmonizedstandards instead of using market-based instruments.
3.1. MINIMUM CONTENT REQUIREMENT
To promote the use of recycled paper, a minimum content requirement could beimplemented for the industry such that the use of recycled inputs should be at leasta certain minimum proportion β of the use of virgin material, (R/V) ≥ β. Given thatthe purpose would be to increase the utilization of recycled waste from the initiallevel, the constraint would be binding with a strict equality. The Lagrangian wouldhave an additional constraint, or L = pV V + pRR + λ(Y − f(·)) + δ1(βV − R), whereδ1 is a multiplier, or a shadow price reflecting the impact of the minimum contentrequirement. The necessary conditions would read
pV + δ1β
pR − δ1= fV
fR
and Y = f (V, R; L, K). (3)
Compared to the industry optimization conditions presented in equation (1), theseimply that the additional environmental objective has an effect on the relativeshadow prices that favors the use of secondary material by increasing the cost ofvirgin material and decreasing the cost of recycled material. If this policy is aimedat being socially optimal, δ1β should reflect the social costs of virgin materialto be taxed (ϕV ) and δ1 should reflect social benefits (ϕR) and opportunity costs(−cE + pE) of the use of recycled material to be subsidized. Our concern is that arigid harmonized standard, as determined by the environmental constraint βV = R,cannot capture the differences in the benefits and costs in different countries.
3.2. UTILIZATION RATE TARGET
An alternatively policy measure to promote the use of recycled material wouldbe a target for utilization of waste, whereby recycled input use should be a certainfraction of the production of final goods, (R/Y) ≥ µ. The amount of recovered wastethat would potentially go to landfills (if not for input use) is R = αγ Y, where γ isthe share of paper consumption of domestic production and α the waste recoveryrate. We decompose the utilization rate into µ = αγ in order to take into accountcountry-specific differences in γ .9 Consequently, the Lagrangian L = pV V + pRR +λ(Y − f(·)) + δ2(Y − (R/αγ )) would include a multiplier δ2 capturing the shadow
ENVIRONMENTAL POLICY 269
price of the utilization rate target for recycled material. The necessary conditionswould read
pV
pR − δ2αγ
= fV
fR
and Y = f (V, R; L, K). (4)
Compared to Equation (3), the utilization rate policy would not involve the priceof virgin material; only the price of recycled inputs. The interpretation is that whenthe utilization rate target is used as policy instrument, the social costs of the useof virgin material as indicated by ϕV in Equation (2a) are not the policymaker’sexplicit concern. The stringency of the policy target would correspond to a subsidyof δ2/αγ per unit of recycled input. The problem is that if αγ is replaced bya common standard µ, the social optimality of the standard can be questioned.There is simply no possibility to take into account the variation in γ and thecorresponding optimal level of α which depends on the social benefits and costsassociated with recycling in each country.
3.3. IMPLICATIONS OF POLICIES BASED ON HARMONIZED STANDARDS
In the long run, economic agents adjust to price changes, which leads to realloca-tion of resources. The purpose of our static model is to analyze short run effects ofpolicy measures in order to reveal the shadow prices of environmental constraints.The immediate effects are illustrated in Figure 2. An industry production isoquantis Y = f(V, R; L, K). The cost-minimizing input combination is determined by theprice ratio of the inputs (pV /pR), initially at point A for a given producer country.Point A lies on the line βA, along which the input ratio is constant when the outputlevel is changed.
To underline the importance of γ in choosing appropriate policy instruments,let us consider the maximum physical amount of R in each country, i (i = d, for domestic and foreign, respectively). The domestic availability of waste paperdepends on the share of domestic consumption of production, γi , the wasterecovery rate, α (where 0 < α < 1), and production, Y; the maximum amountof secondary material available in each country is then αγiY at a given productionlevel Y = YR. If a country exports more (consumes less) of its own production thana foreign competitor does, less secondary material is available in domestic markets.Even if two different countries produced at the same production level and they hada certain common waste recovery rate, α, which would reflect the environmentalgoal of recycling, the maximum amounts of recycled material available for eachcountry would differ as depicted on the vertical axis of Figure 2, i.e., Rc = αγf Y forthe foreign competitor and RB = αγdY for the domestic producer when γf > γd .
Consider now a case where a minimum content requirement (R = βV) forsecondary material use is introduced. Let us assume that the stringency of thecommon standard is motivated by the seriousness of waste disposal problemsand, hence, the abundance of waste paper in the domestic market of the foreign
270 ANNI HUHTALA AND EVA SAMAKOVLIS
Figure 2. The effects of environmental policy objectives imposed.
competitor. The new standard can be depicted by input ratio line βc which corre-sponds to waste paper input use R = Rc = αγf Y at the given production level. Inthe short term, the industries in both countries would be expected to move on theisoquant from the initial equilibrium A to point C, if the same harmonized standardwere applied in both countries. However, even if the waste paper recovery rate α
were the same in both countries, there would be less secondary material availabledomestically and hence the domestic production could only reach point B, with R= RB = αγdY, which lies on input ratio line βB . In order to reach C, an amount Rc −RB of recyclable waste would have to be imported. In other words, a policy goal thatdetermines input ratios should not be the same in the two countries with differentγ even if the waste recovery rate of α were identical (i.e., βC > βB). If market-based instruments were used instead, the intervention input combinations could bereached by promoting the use of recycled input by a subsidy, δ1, and by taxingthe primary input by δ1β, as equation (3) suggests. An alternative policy measurewhich corresponds to R = βcV, at a given production level, is a utilization rate targetR = RC . Recall that the input use goal is determined according to the environmentalconcerns in the country with γf (>γd), i.e., RC = µY = αγf Y. As described byequation (3) the target suggests a subsidy for recycled input only, but the outcomeof the policy intervention would be the same as in the case of a minimum contentrequirement. Again, if differences in γ were not acknowledged, the country withless waste paper could only meet immediately the utilization target by importingwaste paper. However, given that an optimal α were to capture the environmentalgoal for recycling, different countries would use different proportions of recycledmaterial in production, since the subsidy, δ2/(αγi), would depend also on γi .
ENVIRONMENTAL POLICY 271
We have emphasized that standards β and µ conceal the environmental originof the need for policy intervention: the abundance of waste paper that creates wastemanagement problems. We point out that the utilization rate, µ, can be defined intwo alternative ways: 1) as the ratio of waste paper consumption to total paper andpaperboard production (µ = R/Y), or 2) as the product of the waste paper recoveryrate and the share of domestic paper and paperboard consumption in production(i.e., µ = R/Y = αγ , because R = αγ Y). The latter acknowledges country-specificdifferences in paper trade, or the domestic availability of waste paper. Furthermore,a socially optimal level of α can be determined. It is likely that an optimal µ wouldonly by coincidence be the same for all countries. This distinction is important torecognize to analyze further the effects a policy intervention may have on wastepaper trade.
Another issue is that, if the availability of waste paper is a constraining factor,it is possible to adjust the production level instead of importing waste paper.In Figure 2 this can be seen as a shift from isoquant Y to YR and to point D,given that the constraint R = RB = αγdY would be binding. It is even likely thatstrict standards lead to relocation of industries. The question is whether increasingtrade in waste paper or relocation of industries should be the outcome when theinitial purpose was to increase the utilization of waste paper to alleviate environ-mental/disposal problems. In the next section, we analyze empirically the impactson different countries when industries move from the initial equilibrium A to policy(or relocation) equilibrium B or C (D) as depicted in Figure 2.
4. Empirical Analysis
4.1. DATA, MODEL SPECIFICATION AND ESTIMATION
To study empirically the effects of policy instruments on input choices in theEuropean paper and paperboard branch, we estimate a production function for theindustry; production is represented by a family of isoquants in input space, whereeach isoquant corresponds to a country-specific level of output. The data sampleused in this study is based on an unbalanced panel containing annual data from 7European countries over the period 1989–1996, comprising 52 observations.10 Thecountries included are Austria, Finland, France, Germany, Italy, the Netherlands,and Sweden – the producers for which consistent data series on all the explanatoryvariables needed were available. Since the paper and board production of thesecountries amounts to 81 percent of Europe’s total production11, the data set shouldgive a fair representation of the production technology in Europe.
The data consist of observations on paper and paperboard (Y) produced, fibers(virgin wood, V, recycled paper, R), capital (K), and number of employees (L).For wood input we use as a proxy consumption of pulp, and for recycled paperinput we use consumption of waste paper. Data units are in 1000 tons for paperand paperboard production and fibers, million ECU for the capital stock and 100
272 ANNI HUHTALA AND EVA SAMAKOVLIS
employees for the labor input. The capital stock is calculated using the perpetualinventory formula:
Kt = (1 − d)Kt−1 + It−1 (5)
where subscript t indicates time period, d a constant rate of depreciation, and K andI capital stock and investment at a given time, respectively.12
We estimate a representative industry production function using a flexibletranslog specification, which is a local second order approximation of any arbitraryfunction. To reduce the number of parameters to be estimated, each variable wasmultiplied by (1/L), the transformed variables being denoted by y, k, r and v. Inessence this imposes a restriction of constant returns to scale, or homogeneity ofdegree one, on the production function (see, e.g., Berndt 1991, Chapter 9).
The model was specified as (ignoring time and country subscripts forsimplicity):
where dummy variables for the four most significant producer countries, Finland(denoted by DF I ), France (DF ), Germany (DG) and Sweden (DS) are included.13
These countries have high waste paper recovery rates (from 41 to 71 percent), buttheir waste paper utilization at current production differs substantially (from 6 to60 percent). The following restrictions on the estimated parameters follow fromthe assumption of linear homogeneity of the production function:
αV + αR + αK + αL = 1 βV V + βV R + βV K + βV L = 0
The translog is symmetric, meaning that βij = βji .A series of model diagnostics and specification tests were applied to check
the econometric reliability of the estimation results; test results are reported inTable III. As regards the functional form of the production, an F-test favored thetranslog specification compared to the Cobb-Douglas, which would have restrictedthe cross-product terms between the inputs to zero (Functional form test A). AnF-test indicated that the use of four dummy variables is justified to capture vari-ations in the intercepts between the selected countries (Heterogeneity test B).Since we use pooled cross-section and time series data, there is reason to expectheteroscedasticity. White’s test (Heteroscedasticity test C) indicated that hetero-scedasticity is not a problem when transformed variables are used (Model 2; y, k,r, v) instead of using labor, L, as a separate variable (Model 1; Y, K, L, R, V). Thetransformation could, however, generate endogeneity problems, since the multi-plicative variable (1/L) now appears on both sides of the estimated equation. To test
ENVIRONMENTAL POLICY 273
Table III. Tests of selected hypotheses.
Test Restrictions Degrees of Test Critical value
freedom statistic 95th percentile
F or χ2
A. F-test: Functional form
H0:βkk=βkv=βvv=βvr=βrr =0 6 42 39.69 2.33
B. F-test: Heterogeneity
H0:γF I =γF =γG=γS=0 4 38 5.17 2.62
C. White’ heteroscedasticity test
Model 1
H0: Homoscedasticity – 17 34.31 27.59
Model 2 CRS
H0: Homoscedasticity – 12 16.53 21.03
D. Hausman’s endogeneity test
H0: Exogeneity – 14 11.36 23.69
E. F-test: CRS
H0: αV + αR + αK + αL = 1
βV V + βV R + βV K + βV L = 0
βRR + βRV + βRK + βRL = 0
βKK + βKV + βKR + αKL = 0
βLL + βLV + βLK + βLR = 0 5 33 0.58 2.51
for endogeneity, the Hausman test (test D) was applied; this indicated that the nullof exogeneity can not be rejected, but the test (test D) was applied; this indicatedthat the null of exogeneity can not be rejected, but the test seems to be very sensitiveto the specification of the instrument used, as is usual. The assumption of CRS wastested using an F-test (test E). The null hypothesis of CRS could not be rejected.
4.2. RESULTS
The OLS coefficient estimates for the model in equation (6) are reported in TableIV. The t-values indicate statistically significant relationships between output andfiber uses. In particular, the coefficients for v and r as well as the cross-term vr andthe second-order term rr are significant at the 5 percent level, but the coefficientsincluding the capital stock variable are not.14 The results show that the conventionalgoodness-of-fit statistic R2 is high, 0.99.
Using the estimated coefficients, the isoquants for the benchmark countries canbe depicted in pulp/recycled fiber space for given country-specific average capitalstocks. As a visual check Figure 3 suggests that the production functions are wellbehaved. The translog function is, however, only a local approximation, which
274 ANNI HUHTALA AND EVA SAMAKOVLIS
Table IV. Parameter estimates of translog production function.
Parameter OLS
Estimate t-value
α0 0.8521 0.3859
αv 1.0385∗ 2.8110
αr 0.8128∗ 3.4548
αk −0.8795 −0.5843
βvv 0.0929 1.1258
βvr −0.1524∗ −2.3629
βvk −0.1035 −0.6365
βrr 0.1932∗ 2.6480
βrk −0.1374 −1.0075
βkk 0.4896 0.8530
γF I 0.0969 0.8094
γF −0.0179 −1.4041
γG 0.0350∗ 3.3519
γS 0.0440 0.8118
R2 = 0.9963.
means that it does not necessarily satisfy the restrictions for production functionsglobally. Therefore, we need to examine monotonicity and convexity, i.e., thatoutput increases monotonically with all inputs and that the isoquants are convex. Asregards convexity, the bordered Hessian matrix of first and second partial deriva-tives needs to be negative definite for the isoquants to be strictly convex. If at leastone βij is not equal to zero, there exist combinations of inputs where neither mono-tonicity nor convexity is satisfied. However, there can be well-behaved regionsthat are large enough so that the translog function is a good representation. Fora presentation of how to check these criteria, see Berndt and Christensen (1973,pp. 84–85).
The monotonicity condition was verified for all the existing combinations ofinputs in all the countries, with the exception of waste paper in Finland. For all thecountries the bordered Hessian analysis rejected the strict convexity requirementdue to the fact that the third determinant is approximately zero. In other words,the isoquants proved to be convex, instead. Again, the only exception was Finland,for which the convexity condition was not fulfilled. A scatter plot of observationsrevealed that the estimated translog function is not a representative approximationin the input region where only a small amount of recycled fiber is used. Thisexplains why the estimated translog function does not seem to be well behaved inthe case of Finland, where the level of recycled paper input use is low. Therefore,
ENVIRONMENTAL POLICY 275
Figure 3. Isoquants for the benchmark countries with production volumes (fitted values,kilotons).
we cannot derive reliable estimates for Finland in illustrating the price effects andpolicy implications in the following discussion.
4.3. ESTIMATED ECONOMIC EFFECTS OF DIFFERENT POLICIES
The first-order conditions are used to calculate the magnitude of the effects ofimposing standards for use of waste paper in the pulp and paper industries in theselected European countries. Recall that a cost-minimizing optimum for input usesis
pv
pr
= ∂y/∂v
∂y/∂r(8)
which corresponds to the following technical rate of substitution between r and vin the case of the translog production function
T RS = ∂y/∂v
∂y/∂r= r
v∗ αv + βvr(lnr) + βvv(lnv) + βvk(lnk)
αr + βvr(lnv) + βrr(lnr) + βrk(lnk)(9)
where the capital stock variable k is fixed at the country-specific mean valuelevel. To illustrate the outcomes of alternative policies to promote the utilizationof waste paper, we will use the estimated isoquants of two of the most signifi-cant producer countries representing different trade patterns: Germany (importspaper and exports waste paper) and Sweden (exports paper and imports wastepaper). For each country we will compare the initial, currently observed inputcombination (denoted by point A) to those input combinations which are imposed
276 ANNI HUHTALA AND EVA SAMAKOVLIS
by a common harmonized European standard (point B; a utilization rate targetand the corresponding minimum content requirement) or by a country-specificmaximum potential recycling target (point C) (See Figures 4a-b for illustrations).Point C represents a decomposed utilization rate, µ = α ∗ γ , which acknowledgescountry-specific differences in paper trade. As shown in section 3, the waste paperrecovery rate, α, should in each country be at a level where the social marginal netbenefit from recycling is zero. Since determining the optimal recovery rate for eachcountry would be a subject of a separate study we will here show that even if α werecommon, the standard for recycled content (as determined by µ or correspondingβ) would differ from country to country due to γ .
Recall that the utilization rate target was defined as a certain share of wastepaperin production, or µ = R/Y = αγ , and the corresponding minimum content require-ment is determined by calculating the ratio for recycled and virgin inputs, β =R/V at the given output level (isoquant). First, we need to determine a relevantpolicy point of reference B. Since Germany is the largest exporter of waste paperin our data set, it is reasonable to expect that it faces the strongest pressure topromote domestic utilization of waste paper. During the data period consideredhere, the mean utilization rate in Germany was µ = 0.53. We have therefore chosena slightly higher common utilization rate target of µ = 0.60 and correspondingminimum content requirements to illustrate the effects for each country.15 Next, itis interesting to compare how these “administrative” targets for the use of recycledmaterial could be reached with market-based instruments, i.e., by affecting theinput prices instead of quantities. Our comparison also clarifies why the govern-ments of the countries concerned may have different interests and strategies forpromoting recycling.
Estimated country specific effects are reported in Tables V and VI. (SeeAppendix 2 for a description of the calculations.) We start with Sweden, thecountry for which the proposed recycling standards would have the most dramaticconsequences.16 The initial Swedish fiber use is 1248 kilotons of waste paper and7526 kilotons of pulp.17 The technical rate of substitution at point A is calculatedas TRS = pV /pR = 0.88, which indicates that the use of waste paper as a rawmaterial is a more expensive option. If a utilization rate target of µ = 0.60 wereimplemented, the input combination would correspond to point B at the currentSweden output level indicated in Figure 4a. Consequently, the technical rate ofsubstitution at the new equilibrium would equal 1.30 and fiber uses should beadjusted to 5367 kilotons of waste paper and 3121 kilotons of pulp. In other words,the use of waste paper would increase more than fourfold and the use of virgin fiberwould be reduced to less than half of the initial amount.
Instead of imposing standards, taxes and subsidies could be used as economicinstruments, as derived in section 3. Sweden could reach the utilization rate target,or point B, by subsidizing waste paper such that the subsidized price (pR − δ2/αγ
in equation (4)) would be 30 percent lower than the initial price. Alternatively, acorresponding minimum content requirement could be translated to a simultaneous
ENVIRONMENTAL POLICY 277
Figure IVa. Comparing policy instruments for Sweden (production, Y = 8946 kilotons).
Figure IVb. Comparing policy instruments for Germany (production, Y = 13664 kilotons).
278 ANNI HUHTALA AND EVA SAMAKOVLIS
Table V. Estimated country specific effects of different policies.
The values of R, V and Y are in kilotons Finland France Germany Sweden
Mean R (recycled paper) 504 3685 7028 1231
Mean V (virgin wood) 7904 4017 5550 7472
Mean K (capital, million ECU) 7288 7225 14863 5755
Mean L (labor, 100 employees) 207 206 443 166
Yf it 9910 7961 13664 8946
A. Initial position (given Yf it and mean βAi
)
βAi 0.06 0.92 1.27 0.17
RAi 509 3699 7090 1248
VAi
7959 4025 5583 7526
µAi
0.05 0.46 0.52 0.14
PAV /PA
R – 1.09 1.30 0.88
B. Policy I: Common standards∗Utilization rate target (common α = 0.6 and γ = 1)
µB = αγ = R/Y = 0.60
RBi 5946 4777 8198 5367
VBi 3009 3123 4778 3121
∗Minimum content requirement (corresp. µB )
βBi
1.98 1.53 1.72 1.72
PBV /PB
R 1.43 1.32 1.46 1.30
C. Policy II: Country-specific Ri,max
(Given a common recovery rate α = 0.6, increase the use of recycled paper to country-specific Ri,max)
Ri,max = 0.6∗γi,max∗Yi,max 1050 6456 11031 1291
VCi
7863 2027 3114 7477
µCi 0.11 0.81 0.81 0.14
βCi
0.13 3.19 3.54 0.17
PCV
/PCR
1.21 1.79 2.00 0.88
D. Policy III: Country-specific Ri,max 1050 6456 11031 1291
(And requirement βBi
simultaneously fulfilled) 1.98 1.53 1.72 1.72
VDi 531 4219 6413 728
YDi 1779 10642 18277 2143
use of taxes and subsidies as shown in equation (3). Sweden could reach B bytaxing virgin fiber such that the price of pulp (pv + δ1β) would ultimately increase27 percent and by subsidizing waste paper such that the final price (pR − δ1) wouldbe 14 percent lower.
The problem for Sweden is that the above policies would require such a largeamount of waste paper that extensive imports of recycled fiber would be necessary.
ENVIRONMENTAL POLICY 279
Table VI. Estimated country specific distributional effects.
Finland France Germany Sweden
Use economic instruments to move from
A to B∗Utilization rate target
subsidize recycled paper (R) – 17% 11% 30%∗Minimum content requirement
subsidize recycled paper (R) – 8% 5% 14%
and tax virgin wood (V) – 11% 7% 27%
Corresponding changes in raw material
uses when moving from A to B
0Ri (change in recycled fibers) +5437 +1078 +1108 +4119
0Vi (change in virgin wood) −4950 −902 −806 −4405
The values of 0R and 0V are in kilotons.
The potential maximum amount of waste paper available in Sweden can be esti-mated by increasing the recovery rate to 60 percent (compared to the mean rate of53 percent). Using the country-specific maxima of γmax and Ymax (0.23 and 9354respectively for Sweden during the data period), the potential maximum recycledfiber amount would be R = αγmax Ymax = 1291 kilotons. If this maximum amount ofwaste paper were used, Sweden would end up on the isoquant at C, which is onlyslightly above A. However, even if Sweden used all of its potentially recyclablewaste paper, it could not meet a common utilization rate target µ = 0.60 at itscurrent production level. In order to fulfill the policy requirement (corresponding tothe line depicted by βc = R/V = 1.72). Sweden would need to import approximately4000 kilotons of waste paper to produce at B or, alternatively, to cut its produc-tion substantially. At point D, where the domestic waste paper input constraint isbinding, Sweden could produce only one fourth of its current output level. Thisis a purely hypothetical outcome, but it illustrates in a striking way that a well-intentioned harmonization policy may result in wholly unanticipated outcomes ifcountry-specific differences are not taken into account.
Germany represents the other extreme compared to Sweden: large domesticconsumption with respect to domestic paper production enables a much largersupply of waste paper (Figure 4b). Initially, Germany produces at point A using7090 kilotons of waste paper and 5583 kilotons of pulp with a technical rate ofsubstitution of TRS = pV /pR = 1.27. In other words, waste paper is a less expensiveraw material than virgin fibers. When the utilization rate target of µ = 0.60 corre-sponding to a minimum content requirement of βc = 1.72 is introduced, the inputbundle becomes 8198 kilotons of waste paper and 4778 kilotons of pulp with TRS= pV /pR = 1.46. In terms of taxes and subsidies, this point could also be reached
280 ANNI HUHTALA AND EVA SAMAKOVLIS
if waste paper were subsidized by 11 percent to meet the utilization rate target, orif wood pulp were taxed by 7 percent and waste paper subsidized by 5 percent tomeet minimum content requirement. In physical terms, Germany would not haveproblems in reaching B, since its hypothetical maximum amount of waste paperavailable is about 11000 kilotons. (See Table V, panel C.)
In addition, the relative price changes summarized in Table V illustrate thedistributional effects of harmonized policy. By increasing the use of recycled paperto point C (Policy II in Table V), the country-specific maximum recycling targetwhich acknowledges differences in paper trade, France, Germany and Swedenwould altogether use approximately as much recycled material as in point B(Policy I), the minimum recycled content scheme. However, relative price/quantitychanges required to move from initial position A to C would be substantiallylarger in France and Germany compared to Sweden which in fact would nothave any need to intervene to affect prices. This suggests that the availability ofwaste paper reflecting the seriousness of this particular environmental problemshould also be acknowledged when the international standard aims at promotingthe use of recycled material. When striving for an internationally optimal policy, theenvironmental policy goal (expressed as a standard) and the corresponding pricechanges (the shadow prices of the standard) should reflect the social net benefitsof recycling. Therefore, it is important that the recovery rate α captures the socialoptimality of the standard in terms of environmental efficiency compared to otherdisposal options.
In sum, a common recycling standard, justified by a need to harmonize interna-tional environmental policy, would have widely varying impacts on input combina-tions in different producer countries, as illustrated in Figures 4 a–b. When standardsare translated into monetary terms, e.g., taxes and subsidies, the effects becomemost evident. Table VI summarizes the relative price changes needed for eachcountry to move from the initial input combination A to an input combination Bwhich satisfies the common policy goal of a utilization rate target of µ = 0.60. Twodifferent policies are a tax solely on virgin wood pulp (corresponding to a commonutilization rate target) or, alternatively, a combination of a tax on wood pulp and asubsidy on recycled pulp (corresponding to a common minimum content). As canbe seen from Table VI, the policies affecting relative prices would require substan-tial taxes and subsidies in Sweden, whereas the relative price changes in Germanywould not be as dramatic. Also, the changes in volumes of use of wood pulp andrecycled pulp reflect the magnitude and direction of price changes. This wouldindicate that to justify the use of harmonized standards, Sweden should have thehighest social benefits/costs to be internalized. We claim that this is not the case. Inaddition to the fact that Sweden would have to increase heavily its import of wastepaper, there would be significant distributional effects; the principal losers wouldbe the Swedish forest owners who supply wood to the pulp and paper industry.By decomposing the waste paper utilization rate µ to α∗γ we have shown with our
ENVIRONMENTAL POLICY 281
calculations that even if the environmental goal α were common, µ as a policy goaland as a standard would differ from country to country due to γ .
Our calculations are based on estimations for Europe, but the policy relevance ofcomparing instruments to promote recycling is confirmed by the impact on Canadaof the minimum content legislation already adopted in the US. The heated debateon “the garbage crisis” made the US to introduce recycled content legislationwhich resulted in a “crisis” for the Canadian newsprint industry in the early 1990s(Roberts and Johnstone 1996). Even with 100% recovery the Canadian consumerswould not be able to generate sufficient waste paper to supply industry require-ments at existing levels of exports to the US. The shortfall has been made up byimporting waste paper from the US putting firms with high transportation costs ata disadvantage. Michael (1998) claims that due to the larger factor endowment ofwaste paper, the US has reduced imports of paper and shifted domestic produc-tion from higher grade paper products toward waste paper intensive outputs. Heconcludes that recycling may be more beneficial for the US industry than for thedomestic environment.
5. Conclusions
In recent years, there has been pressure on international policy arenas to encouragepaper recycling largely because of environmental concerns. At the same time, inter-national harmonization of environmental standards has been deemed necessary. Weargue here that this poses a challenge calling for an appropriate policy measure:both the international and national economic viewpoints should be acknowledgedwithout compromising the initial environmental viewpoints.
The contribution of this paper is to show both analytically and empiricallythat economic costs of internationally harmonized environmental standards (oradministrative instruments, as they are often referred to vary considerably betweencountries. The analysis is based on a theoretical model, which describes howadministrative measures impose shadow costs that affect producers’ input choices.A production function for the European paper and board industry is empiricallyestimated and the effects of different recycled content schemes on the input choicesof wood pulp and recycled paper are studied. The schemes analyzed are a minimumcontent requirement and a utilization rate target. We translate these standards intomarket-based, or economic, instruments to show the short term marginal costs ofpromoting the use of waste paper in paper and board production. The results showthat a given common standard corresponds to a wide range of marginal (shadow)costs, or country-specific subsidies by our empirical results is that a commonstandard may lead to heavy subsidies on recycling in areas where waste manage-ment problems are not the most urgent environmental problems. As a naturalexplanation for why the harmonized standards may fail to hit the environmentaltarget we suggest the domestic availability of waste paper, which reflects the extentof the environmental problem in each country. This indicates that implementing a
282 ANNI HUHTALA AND EVA SAMAKOVLIS
common international waste paper utilization goal is not a social cost minimizingpolicy measure for promoting waste paper use. In the short term, strict standardsmay only lead to trade effects rather than environmental improvements.
Using a common standard as a policy instrument to increase utilization ofrecovered waste paper is frequently justified by arguments of “fairness”, i.e., bothdomestic and foreign competitors should meet the same environmental standard.However, our estimations predict significant distributional effects that would resultfrom a well-intentioned common policy. There are two lessons to be learnt fromour calculations. First, harmonized policy does not guarantee “fairness” as such(neither “harmonized” environmental benefits nor equal marginal costs) if country-specific features are not taken into account. Second, translating the correspondingshadow prices for standards into terms of market-based instruments shows thatusing these administrative instruments for environmental policy may lead tonon-optimal recycling policy.
In an international context, a harmonized policy instrument should impose anenvironmentally justified, recycling goal (such as recovery rate α), but shouldacknowledge potentially drastic trade impacts by taking into account the geograph-ical distribution of recycled material (as reflected by paper consumption/productionratio γ ). One way to proceed is to refine the standard such that r = α ∗ γ ∗Y, whereα captures the environmental optimality in each country (i.e. the social marginalnet benefit from recycling should equal zero) and γ then captures the geograph-ical distribution of waste paper. An optimal α for each country could be derivedfrom country specific cost-benefit studies of the different waste managementalternatives. If policy instruments to promote recycling result only in increasedimport/export of waste paper, the green labels used to inform consumers about therecycled material content of paper products may also be telling only half the truth.
Acknowledgements
Earlier versions of this paper have been presented in seminars at the Departmentof Economics, University of Umeå, and at the National Institute of EconomicResearch, Stockholm, at the Bromarv workshop on forestry economics at theFinnish Forest Research Institute, and at the World Congress of Environmental andResource Economists in Venice. The authors would like to thank Runar Brännlund,Lauri Hetemäki, Karl-Gustaf Löfgren, Per-Olov Marklund, Anne Toppinen, Lars-Erik Öller, and two anonymous referees for their comments. We also appreciateSara Kristenson’s help with the graphics. The usual caveat applies. Financialsupport from the Academy of Finland and the Swedish Council for Forestry andAgricultural Research is gratefully acknowledged.
ENVIRONMENTAL POLICY 283
Notes
1. Ecolabeling is already used in some European countries to discriminate in favor of recycledpaper products; for example, the Dutch Stichting Milieukeur and the German Blaue Engelaward their environmental labels to recycled paper products made of 100 percent recycledpaper. Some international organizations also recommend that public agencies purchase environ-mentally friendly goods: e.g., the United Nations Development Programme (UNDP, 1995) hasspecified standards for their offices worldwide including, among others, a 50 percent recycledpaper minimum-content requirement for paper. For a further discussion of waste paper cyclemanagement incentives, see, e.g., Bertolini (1994).
2. Here, the term “(minimum) recycled content scheme” refers generally to content standards,including a minimum content requirement and a utilization rate target which will be specifiedexplicitly in section 3.
3. In 1994, the composition of fiber furnish in Europe was (figures for North and Central Americain parentheses): wood pulp 58.3% (66.9%), other fiber pulp 0.3% (0.4%), and waste paper 41.4%(32.8%) (FAO 1997, p. 45).
4. A list of variables and measures used is presented in Appendix 1.5. An alternative policy goal could be to increase the input use of waste paper, no matter where the
waste paper is generated:R ≥ µ ∗ Y = α ∗ γ ∗ ω ∗ Y = (WP recovery/paper cons.)∗(paper cons./paper prod.)∗
(WP cons./WP recovery)∗ paper prod.
= WP consumptionTechnically, restriction (a) from section 2 and restriction (b) coincide when ω = 1, or when allrecovered waste paper is consumed domestically. Restriction (a) is in the line with the initialenvironmental goal, since the domestic availability of waste paper determines the seriousnessof the environmental problem. However, to meet constraint (b), some countries would have toincrease imports of waste paper. In other words, the main problem would no longer be the amountof waste paper which would go to domestic landfills, or the environmental problem, but thechallenge would be to meet the recycling standard which must be done by importing waste.Therefore, we focus on the policy goal of restriction (a).
6. The motivation for not treating separately imported and domestic waste paper/wood pulp inthe production function is that imported and domestic inputs are perfect substitutes in thepapermaking process. In addition, the most relevant feature of the production function is thesubstitutability between R and V given the combination of L and K. The theoretical productionfunction is consistent with our empirical estimations in section 4, where the linearly homogenousfunctional form could be rejected.
7. Note that waste management sector maximizes �E = pER − cER.8. To determine empirically the optimal recovery rate for each country could be a subject of a
separate study; see, e.g., Huhtala (1997). The point we want to make here is that α plays animportant role, if standards for waste paper use are to be promoted for environmental reasons.
9. Recall that the utilization rate is currently measured by µ (Table II), which conceals the effectsof trade flows of paper industry products.
10. The panel is unbalanced since data for the Netherlands were only available for four years.11. This figure is for 1996. All of the aggregate level data for the pulp and paper industry are taken
from different issues of Pulp and Paper International.12. The depreciation rate used is 3 percent. Due to lack of data we could not empirically estimate the
depreciation rate. A study of the Finnish paper industry by Hetemäki (1990) used the procedurepresented in Kuh and Schmalensee (1973) to calculate depreciation rates. Hetemäki arrived ata depreciation rate of 3.5 percent for building structures and 6.9 percent for equipment andmachinery. We believe that even though we had to choose a depreciation rate without an empir-ical estimation, another rate in this range would not change the point that we want to make. in
284 ANNI HUHTALA AND EVA SAMAKOVLIS
order to obtain the initial values for the capital stock, we assume that the capital per ton of paperproduced is equal for all the countries. This ratio is calculated from Swedish data on the pulpand paper sector. Since the data on investment and labor are only available for the pulp and papersector as a whole, we assume that these variables are proportional to the production of paper asa share of the productionn of pulp and paper, and investment and labor are multiplied by thisshare. This simplifying assumption is made only because of lack of data.
13. Other specifications were tested before selecting upon a model, e.g., input variables only, fixedeffects using period dummies, and both period and country dummies. After applying a battery ofstatistical tests and careful model diagnostics, the statistical performance of the model presentedproved to outperform the other specifications in adequacy.
14. Indirect parameter estimates of labor could be obtained by rearranging the homogeneity restric-tions in (7) in terms of the directly estimated parameters. Variances of the indirectly estimatedparameters could then be calculated as a linear combination of the directly estimated variancesand covariances (Berndt 1991).
15. It was difficult to determine the most adequate parameter value for illustrations of a potentialutilization rate target, because minutes of the meetings of the Eco-Label Competent Bodies(where all European member countries are represented) are confidential during the planningprocess. In the US, the secondary content requirements vary from 40% to 80% (Ruston andDesser 1988).
16. This is likely to be true also for Finland, which is as large a producer and exporter, but which infact currently uses relatively less waste paper as raw material.
17. The average values of the data period are used in the calculations.
References
Beghin, J. C. and D. A. Sumner (1992), ‘Domestic Content Requirements with Bilateral Monopoly’,Oxford Economic Papers 44, 306–316.
Berndt, E. R. (1991), The Practice of Econometrics: Classic and Contemporary. Addison-WesleyPublishing Company.
Berndt, E. R. and L. R. Christensen (1973), ‘The Translog Function and the Substitution of Equip-ment, Structures, and Labor in U.S. Manufacturing 1929–1968’, Journal of Econometrics 1,81–114.
Bertolini, G. (1994), ‘Wastepaper Cycle Management: Incentives and Product Chain Pressure Pointor ‘Leverage Point’ Analysis’, in H. Opschoor and K. Turner, eds., Economic Incentives andEnvironmental Policies, Kluwer Academic Publishers.
Beukering, P. van and A. Duraiappah (1998), ‘The Economic and Environmental Impact ofWastepaper Trade and Recycling in India: A Material Balance Approach’, Journal of IndustrialEcology 2(2), 23–42.
Bhagwati, J. N. and T. N. Srinivasan (1969), ‘Optimal Intervention to Achieve Non-EconomicObjectives’, The Review of Economic Studies 36, 27–38.
Calcott, K. and M. Walls (2000), ‘Can Downstream Waste Disposal Policies Encourage Upstream“Design for Environment”?’, American Economic Review 90, 233–237.
Choe, C. and I. Fraser (1999), ‘An Economic Analysis of Household Waste Management’, Journalof Environmental Economics and Management 38: 234–246.
Conrad, K. (1999), ‘Resource and Waste Taxation in the Theory of the Firm with RecyclingActivities’, Environmental and Resource Economics 14, 217–242.
Copeland, B. R. (1991), ‘International Trade in Waste Products in the Presence of Illegal Disposal’,Journal of Environmental Economics and Management 20, 143–162.
Eichner, T. and R. Pethig (2001), ‘Product Design and Efficient Management of Recycling and WasteManagement’, Journal of Environmental Economics and Management 41, 109–134.
ENVIRONMENTAL POLICY 285
FAO (1997), FAO Provisional Outlook for Global Forest Products Consumption, Productionand Trade to 2010, Forestry Policy and Planning Division, Forestry Department. Food andAgriculture Organization of the United Nations, Rome.
Fullerton, D. and W. Wu (1998), ‘Policies for Green Design’, Journal of Environmental Economicsand Management 36, 131–148.
Grace, R., R. K. Turner and I. Walter (1978), ‘Secondary Materials and International Trade’, Journalof Environmental Economics and Management 5, 172–186.
Grossman, G. M. (1981), ‘The Theory of Domestic Content Protection and Content Preference’, TheQuarterly Journal of Economics 96, 583–603.
Hetermäki, L. (1990), ‘Factor Substitution in the Finnish Pulp and Paper Industry’, Acta ForestaliaFennica, 211.
Huhtala, A. (1997), ‘A Post-Consumer Waste Management Model for Determining Optimal Levelsof Recycling and Landfilling’, Environmental and Resource Economics 10, 301–314.
Kuh, E. and R. Schmalensee (1973), An Introduction to Applied Macroeconomics, North-Holland.Michael, J. A. (1998), ‘Recycling, International Trade, and Distribution of Pollution: the Effect
of Increased U.S. Import Demand for Canadian Paper’, Journal of Agricultural and AppliedEconomics 30, 217–223.
Palmer, K. and M. Walls (1997), ‘Optimal Policies for Solid Waste Disposal – Taxes, Subsidies, andStandards’, Journal of Public Economics 65, 193–205.
Roberts, S. and N. Johnstone (1996), ‘Transport in the Paper Cycle, Towards a Sustainable PaperCycle’, Sub-Study Series No. 12, International Institute for Environmental and Development,United Kingdom.
Ruston, J. and S. Desser (1988), ‘Policy Options for Developing Secondary Materials Markets’,Journal of Resource Management and Technology 16(2), 52–64.
UNDP (1995), ‘Introducing The “Green” Office Programme, Sustainable Development and theEnvironment’, United Nations Development Programme, Inter-Agency Procurement ServicesOffice.
Weaver, P. M., H. L. Gabel, J. M. Bloemhof-Ruwaard and L. N. van Wassenhove (1997), ‘OptimizingEnvironmental Product Life Cycles’, Environmental and Resource Economics 9, 199–224.
Appendix 1
LIST OF VARIABLES AND MEASURES USED
R: waste paper consumption (volumes used in the production of paper and board)V: wood pulp consumption (volumes used in the production of paper and board)Y: paper and board productionK: capitalL: laborI: InvestmentpV : price of wood pulppR: price of waste paperpE: price of recycled material used as raw material in alternative waste managementcV : cost of producing virgin materialcR: cost of waste paper recoverycE : cost of alternative waste management methodϕV : social costs (benefits) associated with the use of wood materialϕR: social costs (benefits) associated with the use of recycled materialλ: Lagrangian multiplier
286 ANNI HUHTALA AND EVA SAMAKOVLIS
δi : shadow pricesDi : country dummiesd: depreciation ratewaste paper recovery = waste paper consumption-waste paper imports + waste paperexportsα: waste paper recovery rate = waste paper recovery/paper and board consumptionγ : paper and board consumption/paper and board productionω: waste paper consumption/waste paper productionµ = R/Y = (αγ Y)/Y: utilization rate = waste paper consumption/paper and board produc-tionβ = R/V: minimum content requirement = waste paper consumption/wood pulpconsumption
Appendix 2
CALCULATIONS FOR TABLES V AND VI AND FIGURES 4A AND 4B
Initial position, A
Use mean R, V, K and L to get Yf it , and calculate β = (mean R/mean V). Given Yf it andβ, derive Vf it from the translog function, then Rf it = βVf it . Substitute Vf it and Rf it intoexpression (6) to calculate the technical rate of substitution.
Common standards: Utilization rate target and minimum content requirement, B
To recaluate the first-order conditions, solve for v from the translog. Then v becomes asolution to a second-order polynominal where A = 1/2 βvv , B = αv + βrv lnr + βvk lnk, andC = α0 + αr lnr + αk lnk + βrr1/2(lnr)2 + βrk lnr lnk + βkk1/2(lnk)2 − lny. Calculate r asr = αγ Y, where αγ equals 0.6. Substitute v and r are into expresion (6). Use mean inputvalues to get the technical rate of substitution. To see which minimum content requirementcorresponds to this utilization rate target, calculate r/v.
If we instead use the minimum content requirement as the reference policy: solve for vfrom the translog function by replacing r with βv. Then v becomes a solution to a second
order polynomial vm = −B±√
B2−4AC2A
where A = 1/2βvv + βvr + 1/2βrr , B = αv + αr + αrv
