Resources, Conservation and Recycling

34 (2002) 175–191

A note on inter-country differences in wastepaper recovery and utilization

Christer Berglund a, Patrik Soderholm a,*, Mats Nilsson b,1

a Di�ision of Economics, Lulea Uni�ersity of Technology, SE-971 87 Lulea, Swedenb Swedish Competition Authority, SE-103 85 Stockholm, Sweden

Received 28 February 2001; accepted 29 August 2001

Abstract

Countries worldwide express waste paper recycling targets in terms of recovery andutilization rates. The main purpose of this paper is to identify and analyze the mostimportant determinants of inter-country differences in these waste paper rates. By employingtwo regression models and using data for 89 and 81 countries, respectively, the paperconcludes that relative waste paper recovery and use are largely market-determined, anddepend thus on long-standing economic factors such as population intensity and competitive-ness in the world market for paper and board products. We also find evidence that supportsthe conjecture that rich countries tend to recover relatively more waste paper than is the casein low-income countries, reflecting the higher demand for waste management and environ-mental policies in more developed economies. As recovery and utilization rates are deter-mined largely by long-standing economic and demographic characteristics the degree ofpolicy flexibility in affecting these rates may be limited. In particular, an ambitious utilizationrate target may be very costly to enforce as it can conflict with existing trade patterns ofpaper and board products as well as with other environmental goals. Additional policytargets may, therefore, be desirable, especially since paper recycling is motivated primarily byenvironmental concerns and seldom is a benign activity in itself. © 2002 Elsevier ScienceB.V. All rights reserved.

Keywords: Waste paper; Recovery rate; Utilization rate; Inter-country differences; Recycling policy

www.elsevier.com/locate/resconrec

* Corresponding author. Tel.: +46-920-72078; fax: +46-920-72035.E-mail addresses: patrik.soderholm@ies.luth.se (P. Soderholm), mats.nilsson@kkv.se (M. Nilsson).1 Fax: +46-8-245543.

0921-3449/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved.

PII: S0921 -3449 (01 )00101 -X

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191176

1. Introduction

Throughout the world we witness an increased interest in policies aimed atencouraging waste paper recycling. Concerns about the natural environment andsolid waste management are generally the prime motivations for these policyefforts. As a part of attempts to measure ‘recycling success’, countries worldwidetend to formulate recycling goals in terms of reco�ery and utilization rates for paperand board products (e.g. Buclet and Godard, 2000; Van Beukering and Sharma,1996). However, if these recycling rates are to serve as major policy variables, policymakers need to understand what determines the magnitude of waste paper recoveryand utilization. The purpose of this paper is to identify and analyze the mostimportant determinants of inter-country differences in waste paper recovery andutilization rates.

The main hypothesis of the investigation is that recovery and utilization rates arelargely market-driven, i.e. they are determined by long-run supply and demandfactors in the waste paper market. Thus, in addition to policy differences in termsof the number and the quality of municipal recycling schemes, countries areendowed with different sets of demographic and geographical features (e.g. popula-tion intensity, virgin forest supplies), which in turn affect the economics of paperrecycling. OECD (1976, p. 29) provides an early support for these arguments andconcludes:

[T]here exists very significant differences between countries both in rates ofrecovery and utilization. In part these differences stem from variations betweencountries in economic characteristics about which little can, or should, be done.In part, however, they reflect attitudes, institutions and policies which areamenable to change.

An analysis of what are the main determinants of recovery and utilization ratesis thus important for understanding the degree of policy flexibility in affecting theserates. For example, if the relative availability of waste paper is an importantdeterminant of relative waste paper utilization, a strict utilization rate goal may bedifficult to enforce as it conflicts with prevailing paper and board trade patterns(Huhtala and Samakovlis, 1999). Also, the recovery rate may be determined byimportant cost elements (e.g. population density etc.), which make it costly topursue very ambitious recovery targets.

Previous research efforts on waste paper economics have focused strongly onshort-run supply and demand behavior in waste paper markets (e.g. Anderson,1977; Blaut and Steiker, 1978; Edwards, 1979; Gill and Lahiri, 1980; Edgren andMoreland, 1989; Nestor, 1992). In general these studies conclude that in theshort-run both the supply and the demand of waste paper tend to be own-priceinelastic2. Furthermore, as the prevailing level of economic activity heavily influ-

2 This implies that a 1% increase in the price of waste paper will lead to a less than 1% change in thewaste paper quantities supplied and demanded.

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191 177

ences demand for paper and board products, this price insensitivity makes the wastepaper market very volatile in terms of price fluctuations. These earlier studies havebeen motivated largely by argument that policy makers need a more detailedunderstanding of waste paper supply and demand in order to have recycling policiesaffect production and consumption levels. The question of whether recyclingpolicies are in any sense economical from society’s point of view is thus not theissue under scrutiny in this earlier work. A second set of studies does, however, takethis latter matter in hand and focuses on the efficiency of recycling policies. Inparticular, it analyzes whether collected waste paper should be incinerated, land-filled or recycled in paper and board mills (e.g. Bystrom and Lonnstedt, 1997;Bruvoll, 1998; Radetzki, 2000).

This paper differs from earlier research efforts in the following aspects. Very fewstudies recognize explicitly that different countries (and regions) possess differenteconomic prerequisites in terms of waste paper recovery and utilization potential.In contrast to the earlier market studies, this paper attempts to understand thelong-run factors that drive recovery and utilization rates. As was noted above, thisshould be of particular interest as these rates in themselves represent importantpolicy targets and since the consequences of their enforcement need to be under-stood. Such an understanding is not only important for policy makers; it alsofacilitates aggregate projections of waste paper usage and recovery worldwide. Byconcentrating on inter-country differences, we also broaden the focus of wastepaper studies, which in the past have been heavily biased towards the US andWestern European markets.

The paper proceeds as follows. Section 2 presents waste paper recovery andutilization rate measures, and discusses some potentially important determinants ofthese rates. Thus, in this section we suggest why these rates may (and perhapsshould) differ between countries. In Section 3 we present two simple regressionmodels, which are used to analyze the impact of a set of independent variables onthe recovery and utilization rates respectively. The empirical results from theseestimations are presented and discussed in Section 4. Finally, Section 5 providessome concluding remarks.

2. Recovery and utilization rates: definitions and determinants

In general there is no mystery to the basis for waste paper recycling. First, paperis a major component of the overall municipal solid waste stream, and waste paperrecycling is seen as one important waste management strategy. It reduces the needfor disposal capacity, which in turn leads to lowered emissions from landfills.Second, recycled paper is often a relatively inexpensive input factor in the produc-tion of new paper and board products, not the least because its energy requirementsare low. Extended use of waste paper in paper production may also be motivatedby the long-term value of forest conservation. Thus, the growing waste paper stockmay displace some of the virgin raw material, i.e. wood pulp used in the productionprocess. In other words, paper and board recycling is both supply- and demand-

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191178

driven, and it is shaped by both economics and politics. This has not alwaysbeen recognized in the recycling debate, which in recent years often has tendedto focus on (a) the municipal solid waste (or supply) view of paper recycling and(b) on the policy aspects of the issue (Smith, 1997). Other important issues, suchas the economic determinants of waste paper use, are however often neglected(e.g. Nestor, 1992).

The two most common indicators of recycling progress, the recovery andutilization rates, differ in one very important respect. The recovery rate (RR) isbasically grounded in the waste management view of recycling, as it measuresthe success with which one is able to recover waste paper from the waste stream.It is normally expressed as:

RR=WPRE

PBCONS

(1)

where WPRE denotes the amount of waste paper recovered and PBCONS is paperand board consumption. The utilization rate (UR) reveals the extent to whichthe recovered paper actually is being used in paper and board production. Themost commonly used measure of the utilization rate is defined as waste paperconsumption in domestic paper and board production, WPCONS, divided by pa-per and board production, PBPROD, so that:

UR=WPCONS

PBPROD

(2)

Thus, the UR measure is primarily directed toward the demand side of thewaste paper market. It should be clear that the above recycling rates both havepolicy relevance. The recovery rate provides an indication of to what extent litterand improper disposal has been reduced. If, on the other hand, policy makersare interested in the forest protection or energy conservation advantages of recy-cling, they also need to know more about the utilization of waste paper.

In 1997, the average global waste paper recovery and utilization rates wereboth around 40% (Pulp and Paper International, 1999). Table 1 summarizesrecovery and utilization rates in some selected countries worldwide. Substantialinter-country differences exist, especially with respect to utilization rates. Theserange from 6% in Finland to 100% in Hong Kong. The recovery rates vary lessbetween countries but the differences are still substantial. In 1996, Israel recov-ered 24% of its waste paper, while the corresponding figure for Germany was ashigh as 71%. It is also the case that high recovery rates do not necessarily implyhigh utilization rates. In Sweden, for example, the recovery rate of 52% is fairlyhigh by international standards, but the utilization rate is ‘only’ 17%. Sincewaste paper recovery and utilization rates have become important policy targets,it is important to understand the main determinants of these rates. A number ofplausible determinants are discussed below, beginning with those that are likelyto affect recovery rates.

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191 179

Table 1Waste paper utilization and recovery rates in selected countries (1996)

RR (%)UR (%)Country

7160Germany100Hong Kong 61

52Sweden 1739 45USA24 43Canada

4069UK346Finland

Argentina 314437 28China78Israel 24

Source: Pulp and Paper International (1998).

2.1. Reco�ery rates

Fig. 1 shows the historical patterns of waste paper recovery rates in Sweden andthe USA. Together with Table 1, this figure provides an appropriate starting pointfor discussing the factors that drive these rates. Both Sweden and the USA haveexperienced more or less consistent increases in recovery rates over the years. In thispaper, we hypothesize that growths in per capita income are likely to explain partsof this pattern. As incomes increase so does the value that people assign to theenvironment (e.g. Wiseman, 1990). In economic terms, the income elasticities ofenvironmental goods will normally be greater than one. This is consistent withthe environmental Kuznets curve hypothesis, which predicts a correlation be-

Fig. 1. Waste Paper Recovery Rates in Sweden and the USA, 1940–1997 Sources: Swedish Federationof Forest Industries (Skogsindustrierna), American Forest and Paper Association (2000), and Smith(1997).

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191180

tween incomes and the extent to which environmental protection measures areadopted3.

The increase in environmental awareness is largely reflected in governmental lawsand regulations. For example, recycling obligations of packaging material andmunicipal collection schemes clearly enhance the recovery of used materials.Moreover, many governments try to inform and educate consumers with the purposeof encouraging them to undertake recycling activities. Previous research resultsconfirm that people tend to engage in recycling activities largely because they believethat recycling is necessary for achieving environmental goals, and they feel committedto these goals (e.g. Hornik et al., 1995). Environmental preferences in favor ofincreased recycling, whether expressed by policies or by consumer volunteerism, arelikely to be important determinants of the increase in recovery rates in Sweden andthe USA since the early 1970s.

The above policies, we hypothesize, are especially prevalent in relatively richcountries, and for this reason we would expect waste paper recovery rates to bepositively correlated with per capita income. Table 1 also provides some preliminaryempirical support for this hypothesis. Argentina and China, for example, are countrieswith relatively low per capita income levels, and their waste paper recovery rates arelow as well.

However, if one accepts our hypothesis that waste paper recovery is affected alsoby fundamental economic factors, we have to consider the costs of collection andrecovery. One important cost component in waste paper collection, especially in thepast, has been labor. ‘Although a substantial investment in physical capital is requiredto produce recycled paper, the sorting, collecting, packing, and even transportationinvolved in the recycling process tend to be less capital- and more labor-intensive thanproduction methods utilizing non-recycled materials inputs’, (Wiseman, 1990, p. 42).Furthermore, as per capita income grows so does the real wage, and hence theopportunity cost of labor-intensive activities. In other words, with growing per capitaincome levels, time-consuming recovery and recycling activities become less attractive.Hence, when considering both the demand for and the cost of waste paper recovery,it becomes clear that the net effect of per capita income changes on the recovery rateis ambiguous and, therefore, remains an empirical question.

Since waste paper recovery ultimately stems from recent paper and boardconsumption, the cost of waste paper recovery will also be affected by demographiccircumstances. Other things equal, actions to separate and collect waste paper willbe more viable in regions which are densely populated and/or in which people liveclustered in highly urbanized areas. Moreover, in such regions the likelihood for highlandfill taxes/prices will be high, and so will the cost of disposal. Thus, since themarginal cost of recovery will depend on the size of the waste stream, we hypothesizethat recovery rates will be particularly high in small but densely populated regions.Hong Kong, with a recovery rate of 61%, provides an illustrating example of the latter(Table 1).

3 See, for example, World Bank (1992) and Beckerman (1992). For an influential critique of theenvironmental Kuznets curve hypothesis, see Arrow et al. (1995).

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191 181

Fig. 2. Waste Paper Utilization Rates in Sweden and the USA, 1940–1997 Sources: Swedish ForestIndustries Federation (Skogsindustrierna), American Forest and Paper Association (2000), and Smith(1997).

It should be noted, of course, that these collection costs are not purely market-driven; they are also affected indirectly by government policies. In the USA, forinstance, the alternative cost of landfill disposal (‘tipping fees’) rose dramaticallyduring the end of the 1980s, in part as a way of encouraging recycling activities andin part because of stricter regulation of landfill systems to control leachate etc. So,in effect, the cost of waste paper collection in the USA, and indeed throughout mostdeveloped countries, were being driven at least in part by these contemporaneouspolicy developments4.

2.2. Utilization rates

Fig. 2 shows the waste paper utilization rates for Sweden and the USA since 1940.Both countries have experienced significant increases in utilization rates since the1970s. However, before this increase, we have witnessed an almost consistent decline,at least in the USA. The above suggests that long-term forces have been at work.

Since the utilization rate measures to what extent waste paper is actually used inproducing new paper and board products, we have to consider both the input andthe output side of the paper and board industry in order to understand what drivesthis rate5.

The decision to use waste paper as input for the production of paper is primarilydetermined by the rate of return of different production methods. Thus, undertraditional profit maximizing conditions, waste paper consumption will depend on

4 We are indebted to one of our reviewers for this remark.5 There exist technical limitations with respect to how much waste paper can be used in the production

process. A common assumption is that a fiber can circulate 5–6 times before it is exhausted, i.e. whenthe quality of the fiber no longer is able to meet the demands of the final product due to lack of strengthand brightness (Antti, 2000). However, these technical constraints are unlikely to be importantdeterminants of inter-country differences in utilization rates, and will, therefore, not be dealt with inmore detail in this paper.

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191182

the price of waste paper and the price of its substitutes, which in most cases is woodpulp. Hence, the availabilities of waste paper and wood pulp (or virgin fiber) arelikely to be important determinants of waste paper utilization rates. This impliesthat countries with relatively large forestry resources, utilized by modern pulp mills,will have access to inexpensive virgin fiber relative to secondary fiber. Under suchcircumstances waste paper is obviously less economically attractive as a rawmaterial, and the utilization rate will be comparatively low (e.g. Ince, 1995).

The potential availability of waste paper supplies, on the other hand, is largelyreflected in the extent of a country’s consumption of paper and board products. Forexample, in countries where consumption is significantly greater than production(implying net imports of paper and board) waste paper will be relatively abundantand we would expect the utilization rate to be high. This is sometimes known as thestructural effect or the trade effect, as it ultimately indicates the way in whichcomparative advantage of the domestic paper and board industry affects wastepaper use (OECD, 1976). Furthermore, the intensity with which this waste paperpotential is exploited is also an important determinant of utilization rates. Thus, areasonable hypothesis is that the higher the actual recovery rate, the greater thelikelihood for significant use of secondary fiber6.

This relative abundance of waste paper is likely to explain some of the differencesin utilization rates displayed in Fig. 2 and Table 1. Sweden possesses largeinventories of forests and is a major exporter of paper and board products. As aconsequence its utilization rate is still comparatively low, even though it hasincreased substantially since 1940. Finland, also a country blessed with largevolumes of virgin forests, has a very low utilization rate compared with the worldaverage (Table 1). Major net importers of paper and board, such as Hong Kongand Israel, display much higher utilization rates. Fig. 2 also indicates someinteresting developments over time. We note in particular that both in Sweden andthe USA, the 1960s witnessed stagnant and even decreasing utilization rates. Partsof these declines are probably explained by the changing relative economics ofvirgin versus secondary fiber during this period. Specifically, advances in pulp andpaper technology and the introduction of semi-chemical pulp led to a lower relativecost of primary fiber use (Wiseman, 1990). Nevertheless, in order to gain a morecomplete picture of the determinants of waste paper utilization, we need to considerthe output side of the paper and board industry as well.

Differences in waste paper utilization rates may reflect variations in the paperand board product mix. Due to differences in quality demands, some productssimply have higher waste paper coefficients than others. The highest utilizationrates are generally found for newsprint and some paperboard qualities, while thelowest rates normally are in printing and writing papers7. Thus, countries in which

6 It should be clear that this relationship may not necessarily exist as much paper and board products(including waste paper) are imported and exported. The actual impact of changes in recovery rates onthe size of utilization rates, therefore, remains an empirical question.

7 For example, in Sweden, roughly 85% of paper industry demand for secondary fiber comes fromnewsprint and liner and fluting production (Swedish Forest Industries Federation, 2000).

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191 183

overall paper and board production is dominated by waste paper intensive prod-ucts, will, ceteris paribus, experience a higher utilization rate than otherwise. Thisfactor can contribute to the depressed utilization rates in the USA during the1960–1990 period. For instance, during this period, the printing and writing papershare of paper supply increased from 46% in 1960 to almost 60% in 1990 (AmericanPaper Institute, 1983, 1991).

As waste paper recycling tends to be heavily connected to environmental aware-ness, consumer preferences may not only affect the output mix but presumably alsothe firms’ input choices (Van Beukering and Sharma, 1996). Consumers may bewilling to pay a premium for recycled paper and in this way ‘convince’ producersto increase their share of waste paper in the fiber mix. The paper and boardproducers are keen to pursue this as it opens up new markets and/or increasescompetitiveness.

Environmental preferences in favor of increased use of recycled paper productsmay also be reflected in government policies, for example, through the use ofrecycled content standards. Such environmental motives for purchasing recycledpaper are likely to be especially prevalent in rich countries, and we would expectalso the utilization rate to be influenced by per capita income levels.

Thus, in this paper it is hypothesized that increasing environmental awareness,expressed here as higher GDP per capita levels, will influence the utilization rate intwo distinctly different ways. First, there is what might be referred to as the indirectimpact. Higher GDP is expected to lead to more intense waste collection andsorting activities undertaken. This gives a higher recovery rate, which in turnstimulates increased use of waste paper. Second, higher GDP also increases thelikelihood that peoples’ willingness to pay for recycled paper is high, and there willbe a direct positive impact on waste paper use.

3. The econometric models

Two simple regression models, which attempt to explain inter-country differencesin waste paper recovery and utilization rates respectively, are developed in thissection. The first model, the waste paper recovery rate equation, is written inlog-linear form and can be formulated as:

ln RRi=�0+�1 ln GDPi+�2 ln URBi+�3 ln POPi (3)

where RRi is the recovery rate in country i defined as in Eq. (1), and GDPi measuresthe gross domestic product per capita (in US$) for the same country. URBi is theurbanization rate for country i expressed as the percentage share of the totalpopulation living in urban areas, while POPi denotes the population density, i.e. thetotal population divided by the country i ’s land area in square kilometers. Thelog-linear form allows us to interpret the � coefficients as elasticities.

Following the discussion in Section 2, it should be clear that it is hard to formany a priori expectations about the sign of the �1 coefficient. GDP per capita, wehave suggested, reflects two opposing determinants of recovery rates. As income is

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191184

high so is the labor cost of waste paper collection activities, but the same may holdtrue for the value that people place on environmental amenities and hence on wastemanagement policies. The former should have a negative influence on recoveryrates, while the impact of the latter factor ought to be positive. The size of the neteffect thus remains an empirical question. Since both the urbanization rate and thepopulation density measure the ‘cheapness’ of separating and collecting wastepaper, we would expect �2 and �3 to be positive.

The second regression model, the waste paper utilization rate equation, is also inlog-linear form and it is expressed as:

ln URi=�0+�1 ln GDPi+�2 ln FRi+�3 ln PMi+�4 ln SEi+�5 ln RRi (4)

where URi is the utilization rate in country i defined as in Eq. (2). Also in thisequation GDPi is assumed to reflect environmental preferences towards recycling.There is one important difference, however. The environmental preferences affect-ing the utilization rate are assumed to be reflected in policies that aim directly at thedemand side of the waste paper market (e.g. material contents standards) and inconsumer demand for recycled paper and board products. We would thus expect tofind a positive sign for the �1 coefficient.

FRi denotes the growing stock of forests divided by total land area, a variableintended to measure the relative availability of virgin fibers. Since virgin fibers aresubstitutes to waste paper in paper and board production we would expect thecoefficient �2 to be negative.

Inter-country differences in the composition of paper and board production aremeasured by the share of total paper and board production (in tons) that consti-tutes newsprint and liner and fluting board so that:

PMi=(Newsprint production + liner and fluting board production)

Total paper and board production(5)

Increases in PMi should lead to increases in the utilization rate, since intensiveuse of waste paper is particularly high in the production of these two paperqualities.

SEi and RRi are both waste paper supply indicators, where SEi (the structuraleffect) is defined as the share of production to consumption of paper and board incountry i. If, for example, consumption is significantly greater than production thisis a reflection of the fact that potential availability of waste paper is relatively high,while the demand for waste paper (in production) is relatively low. In this case theratio SEi will be low and the utilization rate will be high, i.e. the �4 coefficient isexpected to have a negative sign. RRi, on the other hand, measures instead theintensity with which the above potential is exploited. Naturally, if the recovery rateincreases, it would be relatively easier to obtain waste paper for use in paper andboard production. Hence, �5 should have a positive sign.

In order to employ the regression models in Eq. (3) and Eq. (4) empirically weadd additive stochastic disturbance terms, �i and �i respectively, to each equation.Eq. (3) will be estimated using ordinary least squares (OLS) methods, while

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191 185

equation Eq. (4) is estimated using instrumental variable techniques. The latterapproach is taken since the endogenous variable, RRi, is included in the utilizationrate equation and it is likely to be contemporaneously correlated with the errorterm �i. In such a case, OLS estimates would be biased and inconsistent. Bothmodels were also tested for the possible existence of heteroskedasticity using White(1980) test, but in both cases the null hypothesis of homoskedasticity could not berejected.

Data for 89 (recovery rate equation) and 81 (utilization rate equation) countriesworldwide have been collected. The use of cross-country data can be assumed toreflect long-run responses since the differences between variables in this case exhibita wide range of variation and the result of long-standing political and economicfactors (e.g. Atkinson and Manning, 1995). Data sources and definitions aresummarized in Table 2.

4. Empirical results

Table 3 presents the parameter estimates for the coefficients in both regressionmodels, together with adjusted R-square measures. Starting with the recovery rate

Table 2Data sources and definitions

Definitions SourcesVariables

Recovery rate Pulp and Paper InternationalWaste paper recovery divided by paper and(1998)board consumption (%)(RRi)

Waste paper consumption divided by paper Pulp and Paper InternationalUtilization rate(URi) (1998)and board production (%)

World Bank (1998) and CIAPurchasing power parity (PPP) estimates inGDP per capita(1999)(GDPi) international dollarsa

Population Number of inhabitants per square kilometer World Bank (1998)density(POPi)

Urbanization Percent of total population living in urban World Bank (1998)areas (%)rate (URBi)

Structural effect Paper and board production as a share of Pulp and Paper International(SEi) (1998)paper and board consumption (%)

Pulp and Paper InternationalPaper product The sum of newsprint and liner and flutingboard production as a share of total papermix (PMi) (1998)and board production (%)

FAO (1996), World Bank (1998)Virgin forest Growing stock of forest in million cubicsupply (FRi) meters divided by square kilometer and CIA (1999)

The data for the growing stock of forest are for the year 1990. All other figures are for 1996, except inthose (rare) cases when 1996 data were missing. 1995 or 1997 estimates were then used instead.

a The international dollar, developed by the World Bank, is the unit of account that equalizes pricelevels in all participating countries. It has the same purchasing power over total GNP as the US dollarin a given year, but purchasing power over subaggregates is determined by average international pricesat that level rather than by US relative prices.

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191186

Table 3Parameter estimates for the recovery rate and utilization rate equations

Recovery rate (RRi) Utilization rate (URi)

Constant (�o)Constant (�0) 0.284 (0.220)−3.926 (−6.887)−0.089 (−0.831)GDP (�1)0.214 (2.912)GDP (�1)

FR (�2)URB (�2) 0.107 (0.624) −0.035 (−0.977)PM (�3)POP (�3) 0.091 (2.111) −0.095 (−0.645)SE (�4) −0.611 (−5.666)RR (�5) 0.636 (1.587)

0.560.24R-square (adj) R-square (adj)

t-statistics are given in parentheses. The instruments used in the utilization rate equation include thenatural logarithms of GDPi, FRi, PMi, SEi, and POPi.

regression model we note that two factors in particular tend to be importantdeterminants of inter-country differences in recovery rates. First, the coefficientrepresenting the impact of GDP per capita (�1) has the expected sign, and it isstatistically significant at the 1% level. A 1% increase in GDP per capita will, ceterisparibus, yield a 0.21% increase in the recovery rate. This result suggests that wastepaper recovery and in particular the environmental benefits that are expected tofollow from it are luxury goods. As countries grow richer, the value that people puton environmental issues will increase and there is likely to be a greater demand forwaste reduction policies. We also hypothesized that growing GDP per capitaimplies a higher opportunity cost of labor, and since paper recovery and collectionin many cases has been a relatively labor-intensive activity growth in per capitaincome would lead to decreases in the recovery rate. We are unable to separate thedifferent impacts of these two effects, but we can nevertheless conclude that the neteffect of increasing GDP per capita on the recovery rate appears to be positive.

Second, the coefficients for the demographic variables (�2 and �3) both have theexpected positive signs, but only the population density coefficient is statisticallysignificant (at the 5% level)8. This indicates that the lower the cost of waste papercollection and recovery, in terms of transport etc. the higher the recovery rate.Hence, in countries with high population densities recovery rates will tend to behigher, this since waste paper collection will be cheaper and disposal will tend to becostly. This result is in some contrast with earlier studies, which primarily stress theimportance of urbanization on waste paper recovery (without, however, presentingany solid empirical evidence), especially in developing countries (e.g. Van Beukeringand Sharma, 1996). Projections show that the share of total population living incities will grow at a fast rate in the future (World Bank, 1995). However, our

8 It should be noted that the urbanization rate and the population density are not highly correlated.In our sample, the correlation coefficient between the two is only −0.18. Thus, it seems not to be thecase that a large part of the variation in URBi is explicable by the variation of POPi.

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191 187

empirical investigation does not suggest that this development will have a majorimpact on waste paper recovery9.

Thus, in general we would expect increases in recovery rates to be highest incountries with high population and per capita income growth rates. In most cases,this involves developing countries. However, our goodness-of-fit measure, R-square(adj), is only 0.24. In other words, as much as 76% of the variation in recovery ratesis left unexplained and is thus due to variations in the error term (�i) or tovariations in other variables that implicitly form part of the error term10. Oneimportant variable that we have not included explicitly in the recovery rateequation is the impact of waste management policies on recovered paper supply. Asnoted above, parts of this impact are likely to be reflected in the GDP measure.However, what is probably not reflected in the GDP measure is, for example, theefficiency with which these policies are implemented. Future research efforts should,therefore, explore in more detail such policy differences between countries.

If we proceed to analyze the parameter estimates for the utilization equation, wefirst note that the R-square measure is comparatively high (0.56). We have thusbeen more successful in explaining the inter-country variation in utilization ratesthan was the case with recovery rates. All coefficients, except �1 and �3 representingthe impacts of GDP per capita and the paper product mix respectively, have theexpected signs (for the GDP per capita effect, see below). Thus, our results implythat in countries in which the share of ‘intensively-recyclable’ paper products ishigh, utilization rates tend to be low. This effect, though, is insignificant from bothan economic and statistical point of view. Moreover, in countries blessed withsubstantial forestry resources waste paper utilization is (as expected) relatively low,but also this effect is statistically insignificant and fairly small.

The �1 coefficient indicates that a 1% increase in GDP per capita leads to adecrease in the utilization rate by 0.09%. Still, also this effect is highly insignificantfrom a statistical point of view. Thus, we do not gain support for our GDP-relatedhypothesis. It is interesting to note that while GDP per capita was a majordeterminant of waste paper recovery rates, this tends not to be the case forutilization rates. This suggests that richer countries do place a higher value onenvironmental goods, and thus on recycling activities. However, these preferencesmainly manifest themselves in recovery and collection activities, and not so much inconsumer demand for paper and board products. Of course, further investigationsinto this issue are needed before more definite conclusions can be drawn.

Our two indicators of waste paper availability are both major determinants ofinter-country differences in utilization rates, but only the structural effect (SEi) isstatistically significant. Furthermore, the two measures are correlated (with a

9 One important caveat is, however, in place. GDP per capita and the urbanization rate are quitehighly correlated (the correlation coefficient is 0.62), and this may explain the statistical insignificance ofthe latter variable.

10 One should note, however, that low R-squares are very common for cross-section samples.R-squares of 0.5 is generally believed to be ‘high’, and even estimates as ‘low’ as ‘0.2’ are sometimesnoteworthy (Greene, 1993).

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191188

correlation coefficient of 0.56), and it is thus hard to sort out in detail the exactcontribution of each of these variables. What is clear, however, is that countriesendowed with abundant waste paper supplies (whatever the measure) overall havehigh utilization rates. This suggests that waste paper utilization largely is market-driven. In other words, paper and board mills do not use waste paper primarilybecause public policy or preferences ‘mandate’ them to do so; they use it becauseit is widely available and thus cheap11. Of course, public policy will affect wastepaper utilization indirectly through the effect it has on the recovery rate, and henceon the relative cheapness of waste paper.

As was noted in Section 3, our reliance on inter-country variations primarilyreflects long-run responses. In the short-run, the link between increased recoveryand utilization are not always apparent in waste paper markets. For example, inmany cases collection schemes have led to abundant waste paper supplies, but thishas not led directly to increased use but instead to price slumps12.

Finally, it needs to be stressed that a major part of the supply potential is notpolicy determined. It is instead largely an effect of national comparative advantagein foreign trade of paper and board products. Waste paper supply ultimately stemsfrom recent paper and board consumption. As a consequence, in countries with acomparative advantage in producing and exporting paper products, paper produc-tion will be significantly greater than consumption implying net exports. In suchcases, the availability of waste paper will necessarily be comparatively low, notbecause of a lack of efficient waste management policies but as a consequence ofthe relative competitiveness in the world market for new paper and board products.It is not clear why policy makers should try to alter this situation only to achievea higher waste paper utilization rate (Huhtala and Samakovlis, 1999).

5. Concluding discussion

By using a data set of 89 and 81 countries, respectively, this paper has attemptedto identify and analyze the main determinants of inter-country differences in wastepaper recovery and utilization rates. The following general conclusions can bedrawn from the analysis. Both economic and political factors affect recovery andutilization rates. Rich countries with high per capita income levels will experiencehigher recovery rates. The maintained hypothesis in this paper is that this relation-ship is explained by the fact that in richer countries, the demand for wastemanagement and environmental policies tend to be high. As higher recovery ratesimply higher waste paper availability, this will in turn positively affect utilizationrates. However, we also gain support for our hypothesis that the two recycling ratesalso are market- (and not policy-) determined. For example, to some extentdemographic features such as the population intensity, which determines the cost of

11 In some regions (e.g. a number of states in the USA), there do of course exist recycled contentstandards, but such policies do not appear to explain inter-country differences to any great extent.

12 See, for example, Browne (1996) for an analysis of such behavior in the US newsprint market.

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191 189

collection and recovery, influence recovery rates. Furthermore, waste paperavailability is the most important determinant (both statistically and economically)of waste paper utilization rates. Waste paper availability is in turn affected largelyby the so-called structural (or trade) effect, which ultimately reflects the countries’comparative advantage in world paper markets. In other words, countries, whichare net exporters (importers) of paper and board products will in general experiencea low (high) utilization rate.

In sum, differences in waste paper recovery and utilization rates are to a greatextent due to economic characteristics about which very little can, or even perhapsshould, be done. This indicates that the two recycling rates may be of limited useas policy measures. Policy efforts intended to affect the rates may turn out to bevery costly or difficult to enforce as they run counter to other environmental oreconomic goals.

For example, attempts to mandate a certain utilization rate (in terms of recycledcontent standards for paper and board products) may lead to unintended behavior.If a country that is a major net exporter of paper and board faces a recycledcontent goal, it may be forced to import waste paper in order to comply since itsdomestic supply will be too scarce. Canada is one example where this has occurred.Also Swedish paper and board producers tend to import cheap waste paper (partlydue to subsidized collection) from Germany to meet requirements and customerdemand (Bystrom and Lonnstedt, 1995). As transport over long distances arerequired it is not clear that this behavior favors the environment, which the policiesprimarily were put in place to do. This suggests that policy makers need torecognize that waste paper recycling seldom is a desirable goal in itself. We areprimarily interested in the environmental and economic benefits of recycling, andour policy variables need to reflect this. Recovery and utilization rates may be onlyof limited use in this sense.

Finally, we wish to emphasize that this paper represents only a first attempt totrace the main determinants of waste paper recovery and utilization rates. Futureresearch efforts should be welcomed and these, we argue, ought to focus in moredetail on public policy differences between countries. In addition, it would be usefulto extend the data set by employing panel data. This enables one to divide the dataset into developed and developing countries. Some authors suggest that the impactsof some determinants on recycling rates are likely to be more significant indeveloped countries compared to developing ones and vice versa (Van Beukeringand Sharma, 1996). For example, environmental awareness is probably of moreimportance in rich countries. Our inability to fully resolve the above issues in thispaper should open the field for further research and improve our understanding ofwaste paper recovery and use.

Acknowledgements

The research undertaken in preparation for this paper was initiated while one ofthe authors, Dr Soderholm, was a Wallenberg Fellow at the Center for Energy and

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191190

Environmental Policy Research, Massachusetts Institute of Technology, Cam-bridge, USA. Financial support from the Jan Wallander and Tom HedeliusFoundation, the Knut and Alice Wallenberg Foundation, the Marcus and AmaliaWallenberg Foundation and the Kempe Foundations is gratefully acknowledged, asare valuable comments from Mikael Bask, Stefan Hellmer, Marian Radetzki andtwo anonymous reviewers. Any remaining errors, however, reside solely with theauthors. Finally, the views expressed in the paper do not necessarily reflect those ofthe Swedish Competition Authority.

References

American Forest and Paper Association. Paper Recycling: Statistical Highlights, available at http://www.afandpa.org/recycling/recycling.html, 2000.

American Paper Institute. 1983 Statistics of Paper, Paperboard and Wood Pulp, New York, 1983.American Paper Institute. 1991 Statistics of Paper, Paperboard and Wood Pulp, New York, 1991.Anderson RC. Public policies toward the use of scrap materials. American Economic Review

1977;67:355–8.Antti BM. Utvecklingschef vid ASSI Doman Kraftliner, Pitea (Head of R&D at AssiDoman Kraftliner,

Pitea, Sweden), personal communication, 17 February, 2000.Arrow K, Bolin B, Costanza R, Dasgupta P, Folke C, Holling CS, Jansson B-O, Levin S, Maler K-G,

Perrings C, Pimental D. Economic growth, carrying capacity and the environment. Science1995;268:520–1.

Atkinson J, Manning N. A survey of International energy elasticities. In: Askin AB, Kraft J, editors.Econometric Dimensions of Energy Demand and Supply. Lexington: Lexington Books, 1995.

Beckerman W. Economic growth and the environment. Whose growth? Whose environment? WorldDevelopment 1992;20:481–96.

Blaut T, Steiker G. Characteristics of Waste paper Markets and Trends in Scrap Paper Recycling, Prices,Demand and Availability: A National and Regional Overview, RSRI Discussion Paper Series, No.103, Regional Science Research Institute, Philadelphia, 1978.

Browne AG. Essays on Recycling Economics and Policy, Ph.D. dissertation, Graduate School, BostonUniversity, 1996.

Bruvoll A. The Costs of Alternative Policies for Paper and Plastic Waste, Report 98/2, StatisticsNorway, Oslo, 1998.

Buclet N, Godard O. Municipal Waste Management in Europe—a Comparative Study in BuildingRegimes. The Netherlands: Kluwer Academic Publishers, 2000.

Bystrom S, Lonnstedt L. Waste paper usage and fiber flow in Western Europe. Resources Conservationand Recycling 1995;15:111–21.

Bystrom S, Lonnstedt L. Paper recycling: environmental and economic impact. Resources Conservationand Recycling 1997;21:109–27.

CIA. The World Factbook 1999, available at http://www.odci.gov/cia/publications/factbook/-index.html, 1999.

Edgren JA, Moreland KW. An econometric analysis of paper and waste paper markets. Resources andEnergy 1989;11:299–319.

Edwards R. Price expectations and the supply of waste paper. Journal of Environmental Economics andManagement 1979;6:332–40.

FAO. European Timber Trends and Prospects: Into the 21st Century, Geneva Timber and Forest StudyPapers, No. 11, United Nations Publication, New York, 1996.

Gill G, Lahiri K. An econometric model of waste paper recycling in the USA. Resources Policy1980;6:434–43.

Greene WH. Econometric Analysis, Second ed. New York: Macmillan, 1993.

C. Berglund et al. / Resources, Conser�ation and Recycling 34 (2002) 175–191 191

Hornik J, Cherian J, Madansky M, Narayana C. Determinants of recycling behavior: a synthesis ofresearch results. The Journal of Socio-Economics 1995;24(1):105–27.

Huhtala A, Samakovlis E. Does International Harmonization of Environmental Policy InstrumentsMake Economic Sense? The Case of Paper Recycling in Europe, Working paper No. 65, NationalInstitute of Economic Research, Stockholm, Sweden, 1999.

Ince P. What won’t get harvested where and when: the effects of increased paper recycling on timberharvest, Working Paper c3, School of Forestry and Environmental Studies, Yale University, NewHaven, USA, 1995.

Nestor DV. Partial static equilibrium model of newsprint recycling. Applied Economics 1992;24:411–7.OECD. Prospects and Policies for Waste Paper Recycling in the Pulp and Paper Industry, OECD, Paris,

1976.Pulp and Paper International. International Fact & Price Book 1998. San Francisco, Calif. Miller

Freeman Publications, 1998.Pulp and Paper International. International Fact & Price Book 1999. San Francisco, Calif. Miller

Freeman Publications, 1999.Radetzki M. Fashions in the Treatment of Packaging Waste: An Economic Analysis of the Swedish

Producer Responsibility Legislation, Multi-Science Publishing Company Ltd., Brentwood, 2000.Smith M. The US Paper Industry and Sustainable Production. Cambridge, MA: MIT Press, 1997.Swedish Forest Industries Federation. Waste Paper Statistics, available at http://www.forestindustries.se/

oh/index.cfm, 2000.White H. A heteroskedasticity-consistent covariance matrix and a direct test for heteroskedasticity.

Econometrica 1980;48:817–38.Wiseman AC. US Waste Paper Recycling Policies: Issues and Effects, ENR 90-14, Resources for the

Future, Washington, DC, 1990.Van Beukering P, Sharma VK. International Trade and Recycling in Developing Countries: The Case

of Waste Paper Trade in India, IVM Report No. W96/29, Institute for Environmental Studies, VrijeUniversity, Amsterdam, 1996.

World Bank. World Development Report 1992: Development and the Environment, Oxford UniversityPress, New York, 1992.

World Bank. World Development Report 1995: Workers in an Integrated World, Oxford UniversityPress, New York, 1995.

World Bank. World Development Indicators, CD-ROM, Washington, DC, 1998.