at SciVerse ScienceDirect

Journal of Cleaner Production 37 (2012) 229e242

Contents lists available

Journal of Cleaner Production

journal homepage: www.elsevier .com/locate/ jc lepro

An investigation into e-waste flows in India

Maheshwar Dwivedy a,*, R.K. Mittal b

aMechanical Engineering Department, BITS, Vidya Vihar, Pilani 333 031, Rajasthan, IndiabBITS Pilani e Dubai, Dubai International Academic City, P. O. Box No. 345055, Dubai, United Arab Emirates

a r t i c l e i n f o

Article history:Received 7 November 2010Received in revised form25 May 2012Accepted 10 July 2012Available online 31 July 2012

Keywords:WEEEe-wasteEOLReuseRecycleMarkov chain

* Corresponding author. Tel.: þ91 1596 245073x220E-mail addresses: dwivedy_m@bits-pilani.ac

(M. Dwivedy).

0959-6526/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.jclepro.2012.07.017

a b s t r a c t

Reverse supply chains that characterize reuse and recycling remains the primary focus of large busi-nesses in a globalized economy. This article critically examines the environmental and social benefits ofreuse that would result through systematic interventions in the existing WEEE trade chain in India. Thereexists an increasing body of scientific evidence documenting the deleterious effects of informal recyclingin India. Though formal recycling remains the focus of existing e-waste management systems indeveloped nations, we argue that alternative systems should be explored by making reuse as a policyinstrument through appropriate interventions in the existing disposal practices found in developingnations. We show that prompt reselling of WEEE to other users can potentially go a long way inincreasing their lifespan. The study uses a Markov chain model to analyze the underlying relationshipthat exist within the reverse supply chain partners by quantitatively evaluating the performance measureof different policy scenarios. Finally we discuss the critical factors affecting the reuse business in thecontext of Extended Producer Responsibility (EPR).

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Globally, e-waste is identified as a popular, informal name forelectronic products after the end of their “useful life”. ‘e-waste’comprises of waste from electronic products such as personalcomputers (PC) mobiles and household appliances. Acceleratingtechnological changes in tandem with rapid obsolescence ratesmakes electronic products obsolete very quickly. Electronic productdiscards are one of the fastest growing segments of the currentwaste stream. According to the European Union’s Waste Electricaland Electronics Equipment (WEEE) directive (Directive, 2002/96/EC, 2003) electrical and electronic equipment is the fastest-growing waste stream in the EU. The directive in general definesthe requirements necessary to comply with the mandatorycollection and recycling objectives of themember states. E-waste asopposed to other municipal waste is more hazardous as it containsthousands of toxic ingredients (BAN and SVTC, 2002) includingheavy metals and harmful chemicals such as lead, cadmium,mercury, arsenic etc., with the potential to pollute the environmentand damage the human health if treated inappropriately usingprimitive recycling methods or when disposed to unsecured

; fax: þ91 1596 244183..in, dwivedym@gmail.com

All rights reserved.

landfills. It is estimated that about 70 per cent of heavy metalsfound in US landfills come from electronic discards (Toxics Link,2003). However, WEEE also has a high residual value. It is anabundant source of metals that can be recovered and brought backinto the production cycle.

In India, e-waste is becoming an important waste stream interms of both quantity and toxicity, as typical of any developingeconomy in transit. The Indian electronics industry has emerged asa fast growing sector in terms of production, internal consumptionand export (Dimitrakakis et al., 2006). According to the latest studyby Dwivedy and Mittal (2010a), the total WEEE estimates during2007-11 will be around 2.49 million MT. During the same period,the number of obsolete PC’s (both desktop and notebook) and colortelevision was estimated at 31 million units and account forapproximately 30 per cent of the total units of WEEE during thesame period. In the span of 5 years until 2011, the projected totalobsolete electronic items will be expected to reach upwards of 100million units. Roughly, around 1.95 million MT of e-waste areestimated to be generated from PC and television accounting for 78per cent of the total e-waste weight. The study reports the currentannual growth rate of e-waste in India to be within 7e10%.According to the Greenpeace assessment report (Greenpeace,2008), India in 2007 generated 0.38 million MT of e-waste fromdiscarded computers, television and mobile phone. The reportfurther states that this figure is projected to grow to more thandouble by 2012, to 0.8 million MT per annumwith a growth rate of

Fig. 1. WEEE management impact hierarchy.

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242230

15 per cent. Yet another study by MAIT-GTZ (2007) reports that e-waste in India is expected to touch 0.47 million MT by 2011.

According to Electronic Industries Association of India (ELCINA),the service/commercial sector accounts for 80 per cent of the totalmarket penetration rate of computer and IT hardware in India. Themajor e-waste inventory that will drive the development of e-wastemanagement system in India will be personal computers (Jain,2010a). The economic performance of an e-waste recovery systemdepends significantly on the return product mix (Boma et al., 2010).Though laptops have a higher concentration of valuable commod-ities, it is their low mass limit that renders them unfavorable forrecovery in comparison to desktop PCs. Contrary to the worldaverage of 27 computers per 1000 people and over 500 computersper 1000 people in the US, India in the year 2004 had one of thelowest PC penetration rate at just 9 computers per 1000 people(Moskalyuk, 2004). However, the size of India’s market in absoluteterms is larger than most of the high income countries (Sinha-Ketriwal et al., 2005). For instance, Dwivedy and Mittal (2010b)report that India could overtake the PC penetration of the U.S. (1computer per capita) sometime by the year 2046. The authorsestimated that the number of obsolete computer inventory in Indiawill double that of the US by the year 2022 while the cross-overtaking place sometime between the year 2017e2018. They furtherstate that the generation of obsolete PCs in India, after the crossoverwill dramatically rise. A similar study by Yu et al. (2010) report thatthe generation of obsolete PCs in the developing regions (within400e700 million units) will double that of the developed regions(within 200e300 million units) by the year 2030, with the cross-over expected to occur between 2016 and 2018.

The existing system of e-waste processing in India is mostlyhandled by a very well-networked informal sector (Sinha andMahesh, 2007) involving key players like the vendors, scrapdealers, dismantlers and the recyclers. However, the disposal andrecycling of computer specific e-waste in the informal sector arevery rudimentary as far as the recycling techniques employed andsafe recycling practices are concerned resulting in low recovery ofmaterials. The process followed by these recyclers is product reuse,refurbish, conventional disposal in landfills, open burning andbackyard recycling (Dixit, 2007). Most often, the discarded elec-tronic goods finally end-up in landfills along with other municipalwaste or are openly burnt releasing toxic and carcinogenicsubstances into the air. There is a huge potential for organizedrecycling industry in India given the fact that about 95 per cent of e-waste is processed by the informal sector (MAIT-GTZ, 2007a).Presently, around 23 recycling facilities in varying levels of infancyhave come up in the organized or the formal sector to address thisproblem, which when fully operational could recycle 60 per cent ofthe estimated annual e-waste inventory (Jain, 2010b). Thoughmostof the hazardous material found in e-waste are covered under thepurview of “The Hazardous and Waste Management Rules, 2008”,the current Indian legislation on classification of e-waste ashazardous is ambiguous with none of the laws directly referring toe-waste or it’s handling (Dutta et al., 2006). However, the Ministryof Environment and Forests (MoEF) have recently proposed a set ofdraft rules called the “e-waste (Management and Handling) Rules,2011”which comes into force fromMay 1st, 2012. The new rule putthe onus of e-waste management on the manufacturers in the linesof the principle of EPR and also restricts the use of hazardoussubstances in e-products. Through this enactment, companies nowhave to design their own take-back system. However, the rulesremain silent on collection, recycling and reuse targets as well asthe role of secondary reuse market.

The other major source of e-waste in India apart from thehousehold and the business sector are the illegal imports fromdeveloped countries where it is expensive to recycle the discarded

electronics. Despite India being a signatory of the Basel Conventionfor Transboundary Movement of Hazardous Substances, there hasbeen a spurt in such imports in the absence of proper importregulations. This notwithstanding, the presence of cheap labor hascontributed to a flourishing trade where the products are repairedand reused to extend their useful life. The major driver for theplanning, design and implementation of sustainable e-wastemanagement system in India will be personal computers(Jain, 2010a). The compounded annual growth rate of PC sales inIndia during 2004e2010 is expected to reach 16 per cent (MAIT,2010). It is the business sector in India which alone accounts forabout 80 per cent of overall annual PC sales (MAIT-GTZ, 2007b).Regulation of disposal practices in the business sector will signifi-cantly help in tackling the e-waste problem in India.

Literatures on e-waste in India have been limited to reviews onthe current status of e-waste recycling/disposal (Mundada et al.,2004), estimation of generation quantities (Streicher-Porte et al.,2005; Jain and Sareen, 2006; Streicher-Porte and Yang, 2007; Babuet al., 2007; Dwivedy and Mittal, 2010a, 2010b) and benchmarkingof e-waste management systems that exist in the EU with focus onIndia (Khetriwala et al., 2005; Widmer et al., 2005; Bandyopadhyay,2008; Khetriwal et al., 2009; Wath et al., 2010). Though there havebeen several studies documenting the major stakeholders in the e-waste trade value chain including the assessment of environmental,social and health impacts from their disposal, there are no empiricalresearch investigating the repercussion of the disposal behavioralpatterns of the concerned stakeholders in the e-waste trade chainthat exist in India. Therefore an attempt is made here to simulatealternate e-waste disposal policy scenarios from the business sectorin India and their long term ramifications towards reuse and lifetimeextension of e-products.

1.1. Reuse/lifetime extension of WEEE

Fiksel (2006) points out that industrial ecology lays the foun-dation for rethinking current product and process technologies anddiscovering innovative pathways for recovery and reuse of wastestreams in place of virgin resources. Extension of this concept in thecontext of WEEE, are primarily focused on minimizing environ-mental harm through lifetime extension and reuse. Though reuse isat the top of the hierarchy (see Fig. 1) for any waste management

Fig. 3. WEEE refurbish categories (Kimura et al., 1998).

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242 231

policy, the business sectors have focused more towards the recy-cling compatibility of products.

A closed loop material reuse cycle typically involves repairing,reconditioning and remanufacturing. According to King et al.(2006), repairing involves correction of specified fault ina product, while remanufacturing is a systematic process wherereturned products are brought back at least to original equipmentmanufacturer (OEM) performance specification. From thecustomer’s perspective, remanufacturing entails giving warrantiesthat are equivalent to those of new products. In between these twoextremities, is reconditioning which involves less work contentthan remanufacturing, but more than repairing. Comparisons of thethree reuse chains vis-à-vis their different performance measuresare shown in Fig. 2.

Subramoniam et al. (2010) argue that at the strategic level,companies have till recently focused more on innovative productdesign approaches giving much less thinking towards reuse/remanufacture. Tasaki et al. (2006) report the only study quanti-fying the level of WEEE reuse in society. The authors’ show that thelevels of virgin material consumption in society could be signifi-cantly reduced to two-third when the demand for new products isreadily met by the second hand market. Williams et al. (2008)report that the total energy required in the manufacture ofa desktop computer could be as high as 4 times greater than theelectricity consumed by the computer while in use. Griese et al.(2004) argue that from the environmental perspective, reuse andrecycling have quite different environmental impact. Therefore,extension of lifespan by making reuse as a national strategy can goa long way in mitigating the lifecycle impacts of WEEE. Thisrequires making appropriate interventions in the existing practicesfollowed in developing economies. Reuse of computers being a partof the reverse supply chain, mitigation of lifecycle impacts couldpartly be explored by studying the current WEEE reverse supplychain network that exist in India.

Product reuse through refurbishing includes process not limitedto inspection, cleaning, storage and testing but also may includedisassembly, reprocessing and reassembly (Matsumoto, 2010;Subramoniam et al., 2009; Pagell et al., 2007). Reuse throughrefurbishing can again be classified according to the intendedquality level required as can be seen in Fig. 3. Matsumoto (2010)argue that the success of reuse business depends largely onconsumer preferences for reuse products. The increased consumer

Fig. 2. Performance measures of reuse/refurbish chain (King et al., 2006).

exposure to WEEE reuse business in India is the result of wideningeconomic disparity resulting in digital divide, penetration of IT inthe rural educational sector, affordability and rising demand forreusable e-products.

1.2. Reuse/recycling system

The aspect which is often neglected while designing a reversesupply chain is the economics of reuse/recycling system. Nowonderthat such systems could be a significant employment generator indeveloping economies. Delhi alone employs 25,000 workers ininformal scrap yards handling about 10,000e20,000 tons of e-wasteannually (MAIT-GTZ, 2007b). While the European models havea well-developed system of visible and invisible fees for the recy-cling of WEEE, in India the consumers in contrast are paid anamount of money for their resource. Recently, on account of societalpressures and anticipated government legislation towards safedisposition of WEEE, formal recycling practices are increasinglybeing advocated. Additionally, the presence of limited landfill spacecombined with an increased cost of disposal in landfill has renewedfocus on recycling by all original equipment manufacturers.However, it is only economics of scale which can justify thecommitment towards reuse/recycling. Williams et al. (2008) reportthat informal reuse/recycling practiced in developing countries areeconomically driven because it runs on a net profit basis as opposedto the net cost for recycling in the U.S. Electronic scrap typicallycomprisesmetals, plastics and refractory oxides (Sodhi and Reimer)approximately in the ratio 40:30:30. The precious metals reclaimedfromelectronic recyclablesmake recycling a potential profitmakingbusiness. The recovery of precious metals is the major driver of allthe e-waste recycling activities in India. While bulk recycling indeveloped nations have reached a certain level of automation tofulfill and sustain mandatory recycling targets, manual disassemblyand segregation remains the preferred method for developingnations like India. We argue here that pure recycling towardsprecious metal reclamation though seems economically feasible,alternative approaches like reuse/recycling systems should also beexplored. Truttmann and Rechberger (2006) clearly document thefact that resource consumption is more sensitive to changes inrecycling efficiency compared to changes in product life. From Fig. 4,it is clearly evident that a 25% reduction in resource consumption

Fig. 4. Strategy to reducing resource consumption (Truttmann and Rechberger, 2006).

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242232

can be achieved either through extending the product life by 34% orfrom increase in recycling efficiency by 5%.

However, Williams et al. (2008) report that increase in reusesignificantly lowers the net environmental impacts but are poorlyunderstood. Product take back in recycling/reuse systemwhere thereverse supply chain partners make concerted efforts to prioritizereuse over recycling by diverting a proportion of the collected unitsback to the consumers needs to be properly investigated beforesuch systems could be put into place. Kronenberg (2007) arguesthat “reasonable” consumption is necessary for long-term survivalof ecosystem where economic reasoning should complementenvironmentally “reasonable” behavior. Tasaki et al. (2006)emphasize that the so-called rebound effects where while envi-ronmental efficiency inevitably increases from recycling, theabsolute level of total consumption increases even faster. Williamset al. (2008) believe that WEEE recycling by definition mobilizesmaterial and depending on the level of process control can emitlead, mercury and other hazardous substances. They further arguethat just because formal recycling lowers the lifecycle emission ofheavy metals, does not mandate it as the default environmentallypreferable alternative.

While we attempt to emphasize the benefits of domestic reuse/recycling in developing countries, there is an increasing body ofliterature contending the efficacy of international trade-in qualityof used equipments to poor countries, something which Williamset al. (2008) espouse citing mitigation of digital divide, low recy-cling cost and employment generation as positives from such flowsin these countries. The authors state that even if heavy metals andbrominated flame retardants are removed from future e-products,there would still be considerable generation of toxins from theinformal recycling process. As Carter (2003) contends that ecolog-ical modernization through clean technologies takes place in thedeveloped countries while outsourcing polluting activities to thepoor countries where both regulation and implementation are non-existent. Presumably, what has not been factored in this interna-tional trade concept is the exponential spurt in domestic WEEEgeneration, the increasing numbers of formal recycling entrepre-neurs and the growing awareness of the environmental impact ofsuch products in India.

Although system level mappings of reuse business have beenstudied in detail (Matsumoto, 2009, 2010), we attempt to extendthis concept by analyzing the operational framework of e-wastetrade chain for the case of India in the context of reuse and lifetimeextension. Several scenarios are studied for investigating thepossibility of reuse business by focusing on how the transitions take

place in the e-waste trade chain. The micro level interventionswithin the system boundary are modeled and evaluated.

2. WEEE flow modeling background

The e-waste trade flow in the business sector is modeled asa Markov chain. Such a model involves a sequence of stochasticevents completely characterized by a set of states and the transitionprobabilities associated with various state changes. While a simpleunivariate analysis is feasible for examining current or past rela-tionship, a Markov chain model is more effective since it provideslikelihood estimations for future outcomes. Also, it offers theadvantage of statistically examining multiple transitions simulta-neously. In doing so, we take recourse to one such study whichattempts to map the e-waste flow for the case of China (Veenstraet al., 2010). Much of their work analyzes the perturbations of thematrix of transition probabilities to simulate the WEEE flowpatterns in China. Typically, some part of the approach outlined inthis paper are in line with previous work illustrating the transit ofwaste in a remanufacturing environment (Yang and Gao, 2009) andforecasting the average number of times a material is used in thesociety right from cradle to grave (Matsuno et al., 2007).Throughout this article, we adopt the statistical notation given inVeenstra et al. (2010).

The major stakeholders involved in the WEEE flow are theprimary consumers depicting the business sector, the consumersfrom the household sector, collectors, the second hand marketvendors, dealers and the recycler. The scope of this study is limitedto the primary consumers i.e., the business sector which accountsfor about 80 per cent of the e-waste generated in India. Further-more, we restrict our study to only PCs given the fact that this traceranalysis will have a greater impact on any future e-wastemanagement models. The primary input to theWEEE flowmodel isthe data representing the disposal behavior of the consumer inquestion. The only available study is the survey conducted byMAIT-GTZ (2007) for the state boundary of Delhi. The survey instrumentsused in the study were in-depth interviews with the major stake-holders supported by secondary information. A total of 203respondents which included 98, 59 and 46 respondents respec-tively from small, medium and large business organizations,participated in the stated survey. The study also reports unstruc-tured interviews with the unorganized key channel members likecollectors and recyclers in and around Delhi. Additionally, formalrecyclers from Bangalore, Chennai and Mumbai were reportedlyinterviewed.

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242 233

The survey summarizes that at the corporate/business level, 11per cent of the replaced computers are sold to scrap collector, 21per cent of the replaced computers enter the second hand market,48 per cent are returned to dealers through exchange and buy backscheme, 7 per cent are donated and 11 per cent are sold toemployees while the rest 2 per cent are recycled. Apparently, 70percent of the collected e-waste are refurbished and sold at thesecond hand market while the rest 30 per cent are sold to recyclers.At the dealer end, 50 per cent of the computers replaced throughbuy back and exchange schemes, end up in the formal sector at theretail outlet, the rest being refurbished (Jain, 2010a). The survey issilent onwhat transpires to the 11 per cent e-waste disposed to thecollector. Much of the trade in this channel is clandestine andillegal. Therefore, an attempt is made here tomodel the uncertaintyin the disposal behavior of the collectors through a scenario anal-ysis, the objective of which is to capture the range of achievablepossibilities. We define scenario-1 as the case of scrap collectorsrefurbishing the major chunk of returned products (70%) with fewbeing sold directly at the second market (5%). Likewise, scenario-2addresses the case of the majority returned products sold at thesecond hand market (70%) by the scrap collector and a smallproportion getting refurbished (5%). In either of the cases, 20 percent of returned products are recycled.

3. Basic Markov chain model

A Markov chain is a stochastic process which is completelyspecified by the state definition or the set of states (sici); thetransition probabilities (pijci,j) which represents the probabilityassociated with going from state i to state j; and the set of uncon-ditional probabilities for the initial states. This knowledge assists usto determine the probability of being in any particular state at anyfuture point in time. With this information, the state transitionnetwork diagram of a Markov chain could be constructed usinga set of arcs passing from state i to state j given the associatedtransition probability are positive (i.e., pij > 0). We make thefollowing assumptions much of which are similar to those adoptedby Veenstra et al. (2010) for constructing the WEEE Markov chainnetwork (see Fig. 5):

i. The system describes the process of disposing computerspecific e-waste from the business or corporate sector. Thebusiness sector is the primary consumer.

ii. The Markov chain state space includes the consumer, scrapcollector, second hand vendor, dealer (or retailer), refurbish-ment and recycler. Intermediate traders are not modeled asa separate state.

Fig. 5. WEEE trade

iii. The stage donee (or recipient) and employees represent thesame state which is the consumers. In other words, thestatement representing the probability that computersdisposed off as donation is therefore equivalent to the prob-ability that the computers stay in the same state ‘consumer’.

iv. Refurbished products from the state; ‘Refurbishment’ areentirely sold forthwith.

v. Recycling is an absorbing state.

The one step stationary transition probability matrix forscenario-1 (P1) and scenario-2 (P2) are shown in Eq. (1) and (2)respectively.

P1 ¼

consumerscrap

2ndhanddealer

refurbishrecycling

0BBBBBB@

0:18 0:11 0:21 0:48 0 0:020:01 0 0:05 0:04 0:70 0:20:01 0 0 0 0:69 0:30:5 0 0 0 0:5 01 0 0 0 0 00 0 0 0 0 1

1CCCCCCA

(1)

P2 ¼

consumerscrap

2ndhanddealer

refurbishrecycling

0BBBBBB@

0:18 0:11 0:21 0:48 0 0:020:01 0 0:70 0:04 0:05 0:20:01 0 0 0 0:69 0:30:5 0 0 0 0:5 01 0 0 0 0 00 0 0 0 0 1

1CCCCCCA

(2)

Since in Eqs. (1) and (2), p66 ¼ 1, both scenario-1 and scenario-2are absorbing Markov chain consisting of both transient states andan absorbing recycling state. The transition matrix (P) of a genericabsorbing Markov chain having ‘t’ transient states and ‘r’ absorbingstates can be expressed in the canonical form as:

P ¼ TRAB

�Q R0 I

�(3)

In Eq. (3), the first (t) e states are transient followed by the next‘r’ absorbing states. Here, the matrix ‘Q’ is a t-by-t matrix, ‘R’ is a t-by-r matrix and ‘I’ is a r-by-r Identity matrix. Given that the startingstate is si, the probability of being in state sj after n e steps is givenby,

Pn ¼�Qn f ðQ ; RÞ0 I

�(4)

As n approaches infinity, the probability of being in the transientstates (see Eq. (4)) must approach zero (Qn / 0). The fundamentalmatrix (N) of P can be computed from the expression, N ¼ (I � Q)�1

where the entry nij of N represents the expected number of times

chain in India.

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242234

that the process is in the transient state sj given that it starts in thetransient state si.

N1 ¼

consumerscrap

2nd handdealer

refurbish

0BBBB@

9:38 1:03 2:02 4:54 4:397:36 1:81 1:64 3:61 4:206:56 0:72 2:41 3:18 3:769:38 1:03 2:02 5:54 4:899:38 1:03 2:02 4:54 5:39

1CCCCA (5)

N2 ¼

consumerscrap

2nd handdealer

refurbish

0BBBB@

7:81 0:86 2:24 3:78 3:484:61 1:51 2:02 2:27 2:615:46 0:60 2:57 2:65 3:137:81 0:86 2:24 4:78 3:987:81 0:86 2:24 3:78 4:48

1CCCCA (6)

The fundamental matrix for scenario-1 and scenario-2 areshown in Eqs. (5) and (6) respectively. For instance, if we start fromstate-1 which is the consumer, the expected number of times thereplaced computer remains in state-1, 2, 3, 4 and 5 before gettingfinally recycled is 10, 2, 3, 5 and 5 for scenario-1 and 8, 1, 3, 4 and 4for scenario-2, respectively. Noteworthy outcome from the simu-lation is the entry N11 which explicitly states the number of timesthe computer will remain in the consumer state given that it startsat the same state. In other words, a computer disposed by thebusiness sector has 10 lives in scenario-1 and 8 lives in scenario-2.The model does not take into account the end-of-life store phase,which explains the high computed value of N11. But certainly,a computer will remain in use for a longer period when a scrapcollector additionally gets involved in refurbishing instead of takingthe route of disposing the same directly to the second handmarket.

An equally important index of a Markov chain is the averagetotal time to absorption. It represents the expected number of stepsbefore a transition state is absorbed and is given by t ¼ Nc, where cis the unit column vector. Therefore, the average time to absorptionin both the scenarios are calculated as,

t1 ¼

consumerscrap

2nd handdealer

refurbish

0BBBB@

21:3618:6116:6422:8622:36

1CCCCAand t2 ¼

0BBBB@

18:1713:0114:4119:6719:17

1CCCCA (7)

Though a Markov chain does not allow for an analysis of time,what is obvious from Eq. (7) is that the case of scrap collectorsinvolved in the refurbishing of collected computers, eventuallydelay the process of computers disposed to recycling and thereforecontribute towards reuse as compared to the case when they aredirectly sold in the second hand market. Another interestinginference is that for scenario-1, the shortest route to recycling is thesecond hand market while in scenario-2; computers at the scrapcollectors’ end take the least number of steps before they could beultimately recycled. Therefore, the operational efficiency ofscenario-1 as far as lifetime extension of WEEE is concerned aremore palpable than scenario-2.

4. Investigation into recycler’s strategy

In this scenario, some proportion of the WEEE sourced from thecorporate sector could be potentially refurbished while the restrecycled for material recovery. This is true particularly withcomputer specific waste given that most of the organizations of thissector replace annually at least 10 to 40 per cent of their PC installedbase. Incompatibility with latest technology amongst others is thesingle most important factor driving the replacement decision. TheMAIT-GTZ (2007) survey reports that 60 per cent of the organiza-tions look for best monetary offers while disposing off their

computers. With the current Indian structure bereft of any legis-lative mandate, the recycling activity cannot be environmentfriendly. The major concern of the formal recycler’s in India is thecollection rate given that the compensation paid by the formalrecyclers for the resources are lower compared to the informalsector, where the labor and processing costs are much lower.

4.1. Case 4a: recycler refurbishing a part of the proceeds

Recycling companies have to pay a price affront for sourcinga regular supply from the corporate sector. Therefore, it makes realsense to refurbish some of the quality replaced computers andresold at higher revenue as compared to limited profits made fromrecycling. In retrospect, through refurbishing, we delay the envi-ronmental burden through lifetime extension of the computers. Forexample, a formal recycling company in Chennai, Ash recyclersrefurbishes old monitors into TV entertainment system which aresold at the second hand market. To realize this case, we makechanges in at least two entries of transition matrix P i.e.,

p65 ¼ p and p66 ¼ 1� p; given 0 < p < 1 (8)

This (see Eq. (8)) reflects the case where certain proportion ofcomputers collected for recycling (p66) are diverted to refurbishing(p65). By making these changes, the new Markov chain havingtransition probability ðPÞ is no more absorbing, but ergodic innature. The characteristic equilibrium probability of these ergodicMarkov chains represents the limiting probability (W) that thesystemwill be in each state j after a large number of transitions, andthis probability is independent of the initial state given asW ¼ lim

n/NPn. Here, W is the matrix all of whose rows are the fixed

probability row vector w and wP ¼ w. Much as with an absorbingchain, an ergodic Markov chain can best be analyzed through itsfundamental matrix given by, Z ¼ ðI � P þWÞ$�1 Given thefundamental matrix of the ergodicMarkov chainwhose ijth entry iszij, the mean first passage time matrix (M) can be computed from:

mij of M ¼ zjj � zijwj

(9)

where,wj is the jth state entryof thefixedprobability vectorwandmij

is the mean first passage time or the expected number of steps toreach state sj for the first time given that the ergodic Markov chainstarts in state si for is j. Anotherparameterof importance is themeanrecurrence time ri which essentially gives the expected number ofsteps to return to state si for the first time, assuming that the ergodicMarkov chain starts in state si for i ¼ jwhich is calculated from:

ri ¼ 1=wj (10)

where, wj has its usual meaning.Veenstra et al. (2010) report that the ideal way to analyze the

impact of different policy scenarios is through perturbation of thetransition matrix. For a stochastic matrix such as P, the associatedeigenvalue is 1 of multiplicity one, and all other eigenvalues li willhave magnitude strictly less than 1 (jlij < 1). Additionally, theauthors state that if the second largest eigenvalue is also close to 1,then the Markov chain is very sensitive to small perturbations. Forour case, the second largest eigenvalue of matrix P is 0.82 and 0.85for scenario-1 and scenario-2 respectively, underlying the fact thatthe base case model will behave unpredictably for small changes inany entries of the matrix P. Given that the n e state ergodic Markovchain P is perturbed toP, then the perturbationmatrix (E) is given asE ¼ P � P with respective equilibrium probability row vector wjandwj. In studying the sensitive measure through perturbation, theobjective is to estimate the changes in wj �wj in terms of the

Table 2Sensitivity analysis for recycle-refurbish (Scenario-2).

p ¼ 0.01 p ¼ 0.05 p ¼ 0.1 p ¼ 0.2 p ¼ 0.3

Consumer 6.5508e-002 1.9931e-001 2.6765e-001 3.2303e-001 3.4695e-001Scrap 7.2059e-003 2.1924e-002 2.9441e-002 3.5533e-002 3.8165e-0022nd hand 1.8801e-002 5.7203e-002 7.6815e-002 9.2708e-002 9.9576e-002Dealer 3.1732e-002 9.6547e-002 1.2965e-001 1.5647e-001 1.6806e-001Refurbish 3.7591e-002 1.1437e-001 1.5359e-001 1.8536e-001 1.9909e-001Recycle 8.3916e-001 5.1064e-001 3.4286e-001 2.0690e-001 1.4815e-001Second EV 9.3363e-001 8.9062e-001 8.3648e-001 7.2657e-001 6.1377e-001(r1, r6) (15.27, 1.19) (5.02, 1.96) (3.74, 2.92) (3.10, 4.83) (2.88, 6.75)Hunter’s 18.243 12.321 9.2969 6.8461 5.787

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242 235

changes in E using the normwise expression,kw�wk � kkEk or thecomponent wise expression

��wj �wj�� � kj

��E�� for suitable values ofcondition numbers k (and kj). We use the condition numbersexpressed in terms of mean first passage times proposed by Choand Meyer (2001) for the calculation of component wise pertur-bation bound as follows:

����wj �wj

���� � 12

"maxisjmij

mjj

#kEkN (11)

where, kEkN is the maximum row sum norm. A bound in terms ofcomponent wise perturbation gives us detailed information aboutthe sensitivity of the chain to perturbations. Yet another measure ofsensitivity for a s e state Markov chain is the Hunter’s statistic(Hunter, 2003) defined by vi ¼

Psj¼1 mijwj;, having lower bound

given by v ¼ s.For our analysis, we use different values of p65 ¼ p and through

adjusting the probabilities of p66, we study the relative impact ofsuch changes on the performance of the e-waste managementsystem.

Tables 1 and 2 show the equilibrium probabilities for the ergodicMarkov chain representing scenario-1 and scenario-2, respectivelyand the associated steady state characteristics. The decrease in thesecond highest eigenvalue for both the scenarios (0.94 to 0.63 inScenario-1 and 0.93 to 0.61 in Scenario-2), clearly suggest thatthe Markov chain equilibrium probabilities are less susceptible toperturbations with increase in proportion of e-waste diverted fromrecycling to refurbish. This is confirmed by the decrease in Hunter’sstatistic from 20 to 5.9 for Scenario-1 (Table 1) and 18 to 5.7 inScenario-2 (Table 2). It is also observed that the upper bound statedin Eq. (11) are high for the recycling state, specifically at lowervalues ofp35, but steadily decrease for higher values of p35. Forinstance in scenario-1, the highest upper bound for the recyclingstate related to p35 ¼ 0.2 and p35 ¼ 0.3 is 0.04177 which graduallyincreases to 0.1868 during p35 ¼ 0.01 and p35 ¼ 0.02. Apparently,the mean time for a computer to return back again to the consumerstate given that it currently exists with the consumer (r1) expect-edly decreases with increase in p for either of the scenarios. Thisnotwithstanding, for both the ergodic cases, the mean first passagetime to reach the consumer state from the ith state (m41) are bothidentical and smallest for the dealer state (results not shown due tolack of space). It is also observed from Table 3, thatm46 which is thetime for e-waste to reach the recycler from the dealer state arehighest amongst the remaining starting states. This arguablydemonstrates that consumers prefer the dealers closely followedby the second hand market for purchasing second hand computers.

4.2. Case 4b: recycler sells some refurbished products to secondhand market

We further extend this analysis to the case of recyclers refur-bishing part of their collection and selling the refurbished

Table 1Sensitivity analysis for recycle-refurbish (Scenario-1).

p ¼ 0.01 p ¼ 0.05 p ¼ 0.1 p ¼ 0.2 p ¼ 0.3

Consumer 7.6632e-002 2.2136e-001 2.8978e-001 3.4274e-001 3.6497e-001Scrap 8.4295e-003 2.4350e-002 3.1875e-002 3.7701e-002 4.0147e-0022nd hand 1.6514e-002 4.7704e-002 6.2447e-002 7.3860e-002 7.8651e-002Dealer 3.7120e-002 1.0723e-001 1.4037e-001 1.6602e-001 1.7679e-001Refurbish 4.4028e-002 1.2718e-001 1.6649e-001 1.9692e-001 2.0969e-001Recycle 8.1728e-001 4.7217e-001 3.0905e-001 1.8276e-001 1.2975e-001Second EV 9.4256e-001 9.0042e-001 8.4750e-001 7.4079e-001 6.3261e-001(r1, r6) (13.05, 1.22) (4.52, 2.12) (3.45, 3.24) (2.92, 5.47) (2.74, 7.71)Hunter’s 20.629 13.260 9.7777 7.0815 5.9496

computers to the second hand market. We also assume that thesecond hand market sell all the refurbished shipment directly tothe consumers. Nischalke (2007) report in their study, that thiskind of disposition by recyclers is not quite uncommon in India. Theproposed changes necessitate a revaluation of some of the entriesof the matrix P. We incorporate this by increasing the proportionp31 followed by a corresponding decrease in p36. Due to reasons ofspace, we only show the results of Scenario-2 in Table 4.

From Table 4, it is evident that when the recycler increases thequantity refurbished (p65) from 20% to 30%, the instability in theequilibrium probabilities reduces. This is evident from the secondhighest eigenvalues and also the Hunter’s statistics given that thelower bound of this statistic for six states is 6. The analysis ofequilibrium probabilities which gives the proportion of times theprocess is in each of the states in the long run, for both the cases(Sec 4.1 and 4.2) show some interesting outcomes. For instance, themaximumproportion of time a computer remains in the first case isthe recycling state and by increasing refurbishing even by a smallamount (case 4a), the equilibrium probability of the recycle statesubstantially decreases (see Tables 1 and 2) which means in thelong run there is a potential of reducing the recycling volumes byfocusing on refurbishing and reselling. On the other hand, when therefurbished computers are sold to the second hand market who inturn sells them directly to consumers (case 4b, Table 4), a morepowerful impact can be attained as can be seen from the increasingequilibrium probabilities of the consumer state. This is confirmedfrom the fact that the eventual time to return to the recycle state(r6) is substantially delayed when recycler sells the refurbishedcomputers to consumer through the second handmarket. Similarly,the case 4a showmarginally lower values of indicator r1 over that ofcase 4b, given that this indicator invariably decreases with increasein p for both the cases. Apparently, the behavior of the scrapcollectors have the least influence on the outcome of the case 4b asobserved from their impact which are not entirely different.

4.3. Case 4c: recycler sells small proportion to scrap collectorswithout refurbishing

Nischalke (2007) state that the recyclers in India not only haveaccess to the second hand market, but are also in liaison with thescrap collectors too. Therefore an attempt was made to assess the

Table 3Time to reach recycle state from ith state.

mi6 of M Scenario-1 Scenario-2

Recycle Recycle

Consumer 2.1358eþ001 1.8167eþ001Scrap 1.8610eþ001 1.3011eþ0012nd hand 1.6640eþ001 1.4407eþ001Dealer 2.2858eþ001 1.9667eþ001Refurbish 2.2358eþ001 1.9167eþ001

Table 4Stationary probabilities for scenario-2 (case 4b).

ap31 ¼ 0.11,p36 ¼ 0.2

ap31 ¼ 0.21,p36 ¼ 0.1

bp31 ¼ 0.11,p36 ¼ 0.2

bp31 ¼ 0.21,p36 ¼ 0.1

Consumer 3.4205e-001 3.6346e-001 3.6260e-001 3.7972e-001Scrap 3.7626e-002 3.9981e-002 3.9886e-002 4.1770e-0022nd hand 9.8169e-002 1.0431e-001 1.0407e-001 1.0898e-001Dealer 1.6569e-001 1.7606e-001 1.7564e-001 1.8394e-001Refurbish 1.8646e-001 1.8770e-001 1.9766e-001 1.9610e-001Recycle 1.7000e-001 1.2848e-001 1.2014e-001 8.9488e-002Second EV 0.7445 0.7621 0.6358 0.6571MRT (r1, r6) (2.92, 5.88) (2.75, 7.78) (2.76, 8.32) (2.63, 11.17)Hunter’s 7.0983 7.3820 5.9353 6.0975

Note:a p65 ¼ 0.2, p66 ¼ 0.8.b p65 ¼ 0.3, p66 ¼ 0.7.

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242236

third case,where a proportion of the recycler’s collection is sold backto the scrap dealers without refurbishing. Needless to say that thiscase situation does not reveal any significant impact as far as lifetimeextension of computers is concerned. The proposed study wasconducted by increasing the entry p62 with a matching decrease inp66. It was observed from the Hunter’s statistic that the equilibriumprobabilities are highly unstable for small values of p62, but convergeto stability with increase in units sold to the scrap collectors.

5. Case 5: investigation of the second hand market

The aim of this exercise is to investigate the case of refurbishing(or reselling) more at the second hand market over disposal torecycling. We realize this by gradually reducing the entry p36 whilemaking a corresponding increase in p35. Such an exercise is fairlyplausible for a country like India where e-waste is the main sourceof livelihood of the second hand market driven by the economics ofscale that justify refurbishing over selling the same at a reducedprice to the recyclers. In fact, in smaller cities of India such asChandigarh, Jaipur and Bangalore, around 65 per cent of the pop-ulation prefers second hand goods (Raghavan, 2010). Though, thesize of the organized and unorganized second hand market in Indiais difficult to estimate, there is no ambiguity that it will makeeconomic sense if these markets are taken into deliberation whenthe business sector dispense off their stockpile ofWEEE. The resultsof this case are shown in Table 5.

The parameter of significance is the number of times theproduct returns to the consumer state (N11) and the time toabsorption (tj). The entry N11 from our analysis gradually increasesfrom 9.38 in the base case to 15.74 in scenario-1 and from 7.81 to14.14 in scenario-2 (Table 5). Similarly, a steady increase in time toabsorption is found for either of the scenario’s indicating thestrategy of selling more through refurbishing by the second handmarket has a positive effect in extending the lifespan of computersin use, more so when fewer than 10 per cent are sold to recyclers.Invariably in both the scenarios, the shortest route to recycling isthe scrap collector while products that are with the dealer end uplast at the recycler.

Table 5Time to absorption for case-5.

State Scenario-1

p36 ¼ 0.3, p35 ¼ 0.69 p36 ¼ 0.2, p35 ¼ 0.79 p36 ¼ 0.1, p35 ¼ 0

Consumer 21.36 27.02 36.52Scrap 18.61 23.20 30.882nd hand 16.64 23.41 34.76Dealer 22.86 28.52 38.02Refurbish 22.36 28.02 37.52N11 9.38 11.75 15.74

6. Investigation into producer responsibility for producttake-back

With growing consumer awareness on corporate practices,businesses are forced to reciprocate with greener practices such as“Extended Producer Responsibility” (EPR) a notion or concept thatmake producers take responsible for their activities not only inproduction and distribution of goods but also towards the safedisposal of such products. The underlying assumption here is thatthe stakeholders along the product chain are expected to shareresponsibility for lifecycle impacts of products. Kautto andMelanen(2004) emphasize that the primary pressure for the business sectorto upgrade their environmental performance come from theircustomers. In a product take back program, the company’sresponsibility is not only limited to managing its own waste, butalso working with distributors and customers to take back productsfor potential reuse (French, 2008). WEEE take back or exchangethough not mandated in India upto now, there was evidence ofselect companies like Nokia, HCL, DELL, Panasonic and Wipro toname a few who have institutionalized this program obliging totheir moral and ethical responsibility; brand protection and IPRissues, in addition to the economic gains made from reuse business.However, this scenario will likely change in the coming days, moreso after the e-waste handling rules comes into force after May,2012. Matsumoto (2009) emphasize that original equipmentmanufacturers (OEMs) may favor reuse over recycling so long asreuse business do not impact their new product sales.

6.1. Case 6: product returns through dealers

Product returns through dealers replicating the corporate take-back are modeled assuming that any increase in collection by thischannel will result in a proportional decrease in collection by thescrap collectors. In other words, we assign higher probability to p14and correspondingly adjust the probability p12. We also assumethat dealers on an average refurbish 80 per cent of their productsselling the rest directly to the customers (p45 ¼ 0.8 and p41 ¼ 0.2).Additionally, we investigate three different possibilities: all refur-bished products from the dealers are directly sold to customers(p51 ¼1), all refurbished products from the dealers are directly soldto second hand market (p53 ¼ 1), and exactly half the refurbishedproducts are sold directly to the customers while the rest half soldat the second hand market (p51 ¼ 0.5 and p53 ¼ 0.5). Further, if thetotal collected and refurbished computers from the dealers are tobe resold to the consumer through the second hand route, then theproportion sold to consumers (p31) shall in all likelihood increasenecessitating a reduction in entry p35.

6.1.1. Case 6.1a: dealers refurbish and sell to consumersFor the case where all refurbished computers at the dealer end

are earmarked for resale to the consumers, the parameter ofinterest is the time to absorption especially of the consumer stateand N11. As observed from Table 6, the number of transitions in

Scenario-2

.89 p36 ¼ 0.3, p35 ¼ 0.69 p36 ¼ 0.2, p35 ¼ 0.79 p36 ¼ 0.1, p35 ¼ 0.89

18.17 23.70 33.7313.01 18.00 27.0514.41 20.75 32.2419.67 25.20 35.2319.17 24.70 34.737.81 10.06 14.14

Table 6Time to absorption for case 6.1a.

p14 ¼ 0.48,p12 ¼ 0.11

p14 ¼ 0.5,p12 ¼ 0.09

p14 ¼ 0.55,p12 ¼ 0.04

Consumer 22.72 23.66 26.41Scrap 19.69 20.43 22.592nd hand 17.59 18.25 20.18Dealer 24.52 25.46 28.21Refurbish 23.72 24.66 27.41N11 9.3765 9.770 10.917

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242 237

steps before a computer is eventually recycled gradually increaseswith increase in collection of EOL computers from the consumers.From Table 6 for instance, with increase in EOL product collectionby the dealers (p14) from 48% to 55%, the average time thata product takes to reach the recycling state increases from 22.72 to26.41. Additionally, this trend also persists for N11 too, reinforcingthe fact that the number of potential lives of a product will increasegradually when producers take responsibility in the collection oftheir products. Another perspective that emerges from the analysisis that in all possible eventualities, it is the second handmarket thatshows the shortest route to recycling.

6.1.2. Case 6.1b: dealers refurbish and sell to second hand marketHere we investigate the case of dealers collecting and refur-

bishing WEEE and selling the same to the thriving second handmarket. To make the analysis tractable, we assume that all refur-bished products at the dealer are sold to the second hand market(p53 ¼ 1). Further, if the total collected and refurbished computersfrom the dealers are to be resold to the consumer through thesecond hand route, then the proportion sold to consumers (p31)shall in all likelihood increase necessitating a reduction in entry p35.To better understand the implications of this all important e-wastetrade chain, a complete investigation is required so as to capturethe influence of the different parameters on the model’s behavior.

A Markov chain simulation of this case reiterate the limitedbenefit from increased refurbishing vis-à-vis the base casescenarios. We analyze the case by selectively varying the propor-tion sold by the second handmarket to consumers (p31) followed bythe proportion collected by the dealers, one factor at a time whileassuming that the dealers on an average refurbish 80 per cent oftheir products. So far as the case of scrap dealers refurbishingamajor portion of collected units (Scenario-1) is concerned, there ispractically very little evidence of improvement in overall reusebusiness efficiency as the time to absorption decreases withincrease in proportion collected by dealers. On the contrary,extended producer responsibility through increased involvementof dealers has a positive effect on the entire reuse business for thecase where scrap dealers sell a major portion of collected units tothe second hand market (Scenario-2) as observed from Table 7. It isobvious from Table 7 that with increase in dealer collection by 5%,the time to absorption to the recycling state increases from 10.17 to10.46. A similar increasing trend in number of lives of computer(N11) is also observed.

6.1.3. Case 6.1c: dealers only collect but do not refurbish and sell toscrap and second hand market

We now construct the case where the dealer under the purviewof extended producer responsibility limit their role only to collectthe returns from consumers and sell the same to scrap collectorsand second handmarket where they are refurbished and sold to theconsumers. This case reflects the case of OEMs selling new prod-ucts, and is negatively disposed towards the reuse business. TheOEMs invariably are uniquely positioned for undertaking reusebusiness, but they shy away from this prospect given their

disproportionate profit structures (Matsumoto, 2010). To investi-gate this case, we make changes in scenario-1 (p41 ¼ 0, p42 ¼ 0.4,p43 ¼ 0.4, p46 ¼ 0.2) reflecting the case where dealers sell allcollected units to scrap collectors, second hand market and recy-clers. This model can further be extended to analyze the case wherethe dealers make a conscious choice of not selling the returnedproducts to recyclers, rather preferring to sell them to the scrapcollectors and the second handmarket (p42¼ 0.5, p43 ¼ 0.5, p46¼ 0,p21 ¼ 0.05 and p24 ¼ 0). Such a case stems from the fact that thereturns are fairly new, technologically not obsolete and havinga market value as is generally the case of returns from the businesssector. In other words, the case situation mimics the case whereproducers are absolved from collecting orphaned or historicalWEEE, a term referenced for old obsolete items that have beenupgraded several times during their lifespan leaving no trace of theoriginal equipment manufacturers. This is always the case whena nation have just established and legalized extended producerresponsibility concept. Due to lack of space, we report only therelevant indicators of interest. The (N11,t1) values are found to be(3.26, 8.9) and (4.34, 12.44) respectively, for the case where dealersdispose and do not dispose to recyclers. For either of the cases,there is a marginal increase in both N11 and time to absorption,when the dealers sell more to the scrap collectors over the secondhand market. Additionally, there is no significant effect of increasein dealer collection on the performance parameters, whichevidently are much worse than the base case. This signifies that anyfuture take back policy where retailers and dealers resell via thesecond handmarket will not be successful. The implication of this isthat reselling directly to consumers (Case 6.1a) or directly moni-toring the second hand sales would be better alternatives.

7. Empirical findings

7.1. Environmental and social implication

A preliminary insight into the relevant implications of theillustrated case scenarios could be constructed by analyzing theirenvironmental impact from recycling and associated social impactfrom increased reuse. The relevant statistic used for analyzingenvironmental impact are the proportion waste handled by therecyclers as observed from the equilibrium probabilities in additionto time to absorption to the recycling state. Similarly, the socialimpact from increased reuse is observed by analyzing the propor-tion of disposed products at the consumer state, increased use ofsecond handmarket alongwith the number of extended lives of theproduct (N11). It is observed that scenario-1 where scrap dealersmaximize their revenue through refurbishing as against selling tothe second hand market, objectively outperforms as far as lifetimeextension of products is concerned. According to the recent reportby Toxics Link (2012), a leading environmental NGO, about 1 tonneof e-waste is daily passed through the hands of about 300dismantling units alone at Seelampur in Delhi. It is obvious thatconsumers are conscious about the quality of used products (Case4a) since they prefer the dealers over the second hand market inpurchasing used e-products. From the perspective of scenario-2where the scrap collectors sell a major portion of returned prod-ucts to the second hand market, the case 4b (see Table 5) showssimilar and identical performance. Much of the assumptions forboth the scenarios remain the same except for the dispositionbehavior of the scrap collectors. Therefore, the scrap collectors havethe least influence on the lifetime extension ofWEEE. However, it isthe second hand market that is expected to have a major stake inany policy decision towards reducing the environmental burdenoriginating from the recent spurt in domestic generation ofe-waste. Assuming that the dealers take back their EEE products at

Table 7Time to absorption for case 6.1b (Scenario-2).

p31 ¼ 0.21, p35 ¼ 0.49,p12 ¼ 0.04, p14 ¼ 0.55

p31 ¼ 0.11, p35 ¼ 0.59,p12 ¼ 0.04, p14 ¼ 0.55

p31 ¼ 0.21, p35 ¼ 0.49,p12 ¼ 0.09, p14 ¼ 0.50

p31 ¼ 0.11, p35 ¼ 0.59,p12 ¼ 0.09, p14 ¼ 0.50

Consumer 10.46 10.76 10.17 10.45Scrap 8.24 8.48 8.07 8.312nd hand 8.81 9.12 8.61 8.91Dealer 9.81 10.12 9.61 9.91Refurbish 11.46 11.76 11.17 11.45N11 3.78 3.78 3.69 3.69

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242238

the end of their useful life, appropriate mechanism for integratingthe recyclers with the second handmarket have to be pursued. Alsothe producers have to be proactive in implementing buy-back andresell strategy by mandating the dealers to refurbish and sell theproducts that are returned. The analysis of reuse and lifetimeextension in conjunction with dealer take back, purportedlyoutlines that positive impact from dealers refurbishing and sellingthe proceed directly to consumers can be achieved as against thecase of selling with or without refurbishing to the second handmarket. Clearly, the implementation of even limited extendedproducer responsibility through voluntary collection by theproducers, results in lower environmental impact. Even in theevent of mandated EPR regime, the current network of scrapcollectors needs to be kept insulated from hazardous dismantling/disassembly activities for refurbish/reuse while consolidating theireffort towards improving the collection efficiency only. Table 8shows the evaluation of the different cases by benchmarkingtheir performance with the base case scenarios.

The positive environmental impact of Case 4a and 4b is theoutcome of decrease in the volume of recycling, while the positivesocial impact for Case 4b is more pronounced on account of theincreased use of second hand market in addition to the observedincrease in the time to absorption to the recycling state. For Case 4c,the impacts though positive are not significant on account of moreproportion of products at the recycler end. Similarly, the positiveeffect of impacts from Case 5 and 6.1a are spectacular resultingfrom increase in time to absorption, number of lives and theincreased usage of the second hand market.

The resulting cases of interest are shown in Fig. 6. The rest of thecases (Case 6.1b and 6.1c) show negative impact given that theyperformmuchworse than the base case. However, any future policyresulting from extended role of dealers in a take back regime willin all likelihood translate to neutral profits as a consequence ofthe additional role of dealers (�) and the revenues from refurbish/resell (þ).

7.2. Survey result

The study focused on sampling the perceptions and reuseexperience of consumers by conducting an online survey througha well structured questionnaire in a university set-up. Therespondents included university students, teachers, the non-

Table 8Benchmarking with base case.

Case Environmental impact Social impact Profit

4a þ þþ4b þþ þþ4c þ þ5 þþ þþ6.1a þþ þþ �6.1b e e �6.1c e e �

teaching community and their families. A request for taking thesurvey was made through root mail to all and a total of 80responses were received. The sampled population comes fromdifferent ethnic and religious background, is educated, representsboth high and low-income groups and is predominantly from semi-urban and urban areas. The survey instruments focused onrespondent’s general awareness about the secondary marketstructure and their environmental attitudes towards reuse chains.About 64%, 81%, 98%, 85%, 69%, 38% of the respondents owneda desktop PC, laptop PC, mobile phone, refrigerator, washingmachine and air conditioners respectively. While 16% of therespondents were highly aware about the environmental andhuman health concerns emanating from WEEE, 73% of therespondents were limitedly aware, where much of this awarenesscame from electronic and the print media. Data from the surveyindicate that respondents preferred to exchange their old electronicgoods with the dealer/retailer followed by the alternative ‘selling tosecond hand shops’ (see Fig. 7).

To the question “When you will dispose e off your used elec-tronic equipment?”, a majority of the respondents (about 65%)stated functional obsolescence, while 20% of them choose thealternative ‘better and improved features in newly launchedproduct’ as a triggering point for disposing off their e-waste. Afurther breakdown of survey responses for the purchasing decisionof the respondents indicate that only 14% have actually purchaseda reusable product and about 45% of the respondents are willing topurchase with an additional 5% strongly willing to purchase

Fig. 6. Cases of high impact.

Fig. 7. Consumer WEEE selling strategy.

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242 239

a reusable electronic product. The results compare well with 42%willing to buy second hand electronics in Ireland, 46% in the UK and31% in Belgium (O’Connell et al., 2011). As observed from Fig. 8, therespondents prefer to purchase remanufactured or new-likeproducts specifically if the manufacturer through his dealer/retailer gets involved in the remanufacturing business. Thisconfirms the finding of our simulation effort. Furthermore, as isevident from Fig. 9, majority of the respondents have indicatedquality and cost as the deciding factors in the purchase of a rema-nufactured product followed by warranty and label.

Surprisingly, in contrast to remanufactured products, the attri-butes the respondents look for in a generic reusable product are inthe decreasing order of quality, warranty and cost. So, reusablee-products in the secondary market like second-hand shops can becompetitively positioned by offering warranty to the consumers.

8. Discussion

The implementation of a reuse framework through the principleof EPR in India is a challenge considering the number of stake-holders involved. The primary concern is the integration of scrapdealers and scrap collectors into any formalized take-back schemeand consumer awareness of the benefits of reused products. Themajor operational factor for the success of reuse business in India iscustomer service requirement. Here the consumer as observedfrom the study is principally concerned about the quality standardsof the product. Therefore, the demand for remanufactured productswhich guarantees repeatable and consistent quality levels equiva-lent to virgin products is evidently high. For instance, the Japanesemanufacturer Fujitsu has been successful in creating a demand forremanufactured PCs in India. With 20 major Japanese reusecompanies operating, the flow of reused computers from 2001 to2004 increased dramatically by a whopping 2.59 million units,majority of which exported to India and China (Yoshida et al.,

Fig. 8. Consumer preference for remanufactured products.

2007). Yet another estimate (Akenji et al., 2011) states that Japanin 2006 exported 30% of all home appliances to developing coun-tries. In US alone, remanufacturing business in the year 2001peaked to an astonishing $100 billion (Loomba and Nakashima,2012). Similarly appropriate systems for effective supply manage-ment along with the different EOL processing activities have to bedeveloped. High level recovery for reuse (Zoeteman et al., 2010)which is possible through remanufacturing in the original supplychain, not in another alternate supply chain is yet another factorwhich affects consumer preferences for reused e-products.

The decision to refurbish or to recycle will depend on the cost/benefit ratio of refurbishing. This decision requires knowledge ofincoming EOL product quality parameters such as its age, functionalcondition, physical condition, functional age, remaining useful lifeetc which can be estimated if there is adequate product information(Parlikad and McFarlane, 2005). The current legislation of e-wastehandling rules in India enforces manufacturers to adopt EPR witheffect from May, 2012. This regulatory pressure will motivatecompanies to implement projects faster (Subramoniam et al.,2010). The proposed EPR system should take financial measuresfor collecting and treating “orphans”, including products assembledby small-scale industries, imitation products, and smuggled prod-ucts since they are considered to be the most important precon-dition for applying EPR policies in developing countries (Kojimaet al., 2009). In other words, identification of the producer is themajor detriment in the implementation of EPR in India. This isa burden which has to be equally shouldered by all the identifiedproducers.

Recycling and reuse targets, though not part of the currentIndian legislation, are vitally essential for any EPR scheme tosucceed. Currently, India has a license scheme for formal recyclers,and for it to be effective, the disposers have a responsibility to sellto license holders (Shinkuma and Managi, 2010). This meansrecycling as well as reuse targets are vitally essential in the nearfuture. Hammond and Beullens (2007) prove that even a tokenlegislation that all new products must be recoverable at someminimum target can be sufficient in creating reuse and recyclingactivities within closed loop supply chains. They also furtheremphasize that legislating collection targets instead of recoverabletargets are detrimental since manufacturers might choose theirown devices like preferring recycling over reuse. Some authors(Manomaivibool, 2009) argue that a national mandatory pro-gramme that at least puts collection responsibility on the producerswill scale up the incentive for formalization in India. This beingalready there in the current legislation, what is lacking is someform of recoverable target.

With licenses and targets comes the responsibility of compli-ance monitoring and the capacity to enforce regulations at alllevels. Formal reuse activity necessitates accreditation of refur-bishers and incentives in the form of subsidy. Kojima et al. (2009)cautions that, any future subsidy to formal collectors and recy-clers might create incentives to over-report the amount of collectede-waste. Additionally, incentives need to be disbursed for sellingthese collected units to the second hand market. So, appropriatemonitoring mechanisms are to be in place for proper checks andbalances.

Amongst the strategic factors (see Fig. 8), commitment ofmanagement resource is critical to the success of reuse businessfollowed by the choice of take-back scheme. The total collectionand logistics expenses incurred in the collection of WEEE werealmost 22.5% of the total Advance Recycling Fund (ARF) received inSwitzerland (Bandyopadhyay, 2010). So, by having commoncollection points, the Producer Responsibility Organizations (PROs)are better able to manage logistics, benefit from economies of scaleand provide a consumer friendly, all-inclusive solution instead of

Fig. 9. Ranking of attributes for a remanufactured product.

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242240

a prohibitively expensive brand specific take-back scheme. Someargue that in a collective take back scheme (Webster and Mitra,2007), products from different manufacturers gets co-mingled,the outcome of which is recycling tends to dominate instead ofreuse. However, the current EU WEEE legislation following thelobbying of several leading manufacturers (McKerlie et al., 2006),stipulates individual responsibility for all new products launchedfrom August 2005 onwards. The legislation allows no visible fees tofund the management of waste resulting from new electrical andelectronic products. The reasoning was that individual responsi-bility offers the best feedback for design change and provides thehighest incentive for Design for Environment (DfE). Further, it alsoallows the companies a competitive choice in the marketplace forrecyclers and the freedom to form a collective group outside thenational PROs. Forslind (2009) reports that EPR finances are moreefficiently managedwhen the producers develop an insurance fundto finance future waste generation quantities as against creatingfunds to manage the immediate waste generation quantiles.

In a survey carried out by Subramoniam et al. (2010), majority ofthe respondents supported positively the need to protect the

Fig. 10. Critical factors a

intellectual property (IP) rights of their product thereby influencingtheir decisions to remanufacture. The same authors further arguethat design for remanufacturing is a strategic factor affectingremanufacturing which ensures the need for better product designguidelines and engagement of product designers as well as thecompany leaders in the product design process. Our study clearlyproves the positive impact of retailer involvement in the reusechain. This is corroborated by Hong and Yeh (2012) who analyticallyproved that a retailer collection model which captures the realitythat electronics manufacturers of consumer products typically arenot remanufacturers outperforms the non-retailer collectionmodel(where the retailer may sub-contract the EOL processing activity toa PRO) for the return rate, manufacturer’s profits and total profits ofthe system (Fig. 10).

Amongst the economic factors, refurbish/remanufacturing costis the major factor affecting the decision making process (Chen andChang, 2012) for firms trying to competewith secondarymarkets inselling refurbished/remanufactured products, in addition tomanufacturing new products. They further state that so long as themanufacturer is able to remanufacture at lower rates than the

ffecting reuse chain.

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242 241

secondary market, remanufacturing will not affect sales of newproducts. For the second scenario, it is profitable to focus only onmanufacturing new products as competition to remanufacturedproducts from the secondary market. However, our study reportsthat in the EPR regime, products that are with the second handdealers are first to be absorbed to the recycling state. Experiencesfrom Taiwan (Lua et al., 2006) clearly indicate that that the pricesoffered by the second hand market are much higher than thecollection subsidy offered by the Taiwan WEEE system. This meansthat any form of recycling fund needs to dynamically evolve so as tocreate a level playing field and subsidize the formalized processes.Therefore the second hand market necessitates monitoring andcontrol from the manufacturer. Additionally, the pollution controlboards have to strengthen pollution control measures against thedownstream informal processes that exist in India.

With EPR becoming a policy instrument in India, the possibilityof increased returns through the dealer/retailer channel is forth-coming. Michaud and Llerena (2006) and Subramoniam et al.(2010) emphasize that when an OEM collects and successfullyremanufactures its own products, a high willingness to return theproducts by the consumer can be expected. The manufacturer in anEPR regime (with recoverable targets) and having absolutepossession over the cores can control the profits of the remanu-facturer through competitive pricing so that the sales of newproducts are not hampered (Webster andMitra, 2007). The lifetimeof products used by first users primarily governs the level ofmaterial use of a system (Tasaki et al., 2006) in case that second-hand product demand is “suppliable” and constant. Hence, incen-tives for early returns are vital for the success of reuse business. Therespective stakeholder’s co-ordination is critical to reuse business.Currently, the business sector is the major source of returnedproducts to the existing formal recyclers in India. For instance,Nokia India has recently set-up about 1400 collection points for itscustomers. The formal recycling sector in India should explorewaysand means to integrate with the other two stakeholders: scrapdealers and second hand market for an uninterrupted supply ofWEEE. The recyclers as our study suggest should additionally alsoinvolve in refurbishing. In contrast, formal scrap dealers should bebanned from refurbishing, engaging them to collection alone. Allrefurbished units from recycler are best served by selling to secondhand market while all units collected by scrap dealer must route tothe recycler. These supply dynamics are in addition to the dealertake-back. Furthermore, the business sector in India should begiven economic incentives through appropriate disposal subsidyfor prematurely reselling their quality WEEE inventory for refur-bish/reuse. The results from our simulation effort interestinglyshow high values of the number of lives of a computer (8e10) in thereuse chain that currently exists in India. For instance, the ‘MeanTime Between Failures’ (MTBF) for an IBM z-series server isdesigned to be around 50 years (Parlikad and McFarlane, 2005),whereas on average it is returned by the customer in 5e10 years.Evidently, e-waste reuse business in India is considered to bea multiple revenue stream generation model. This assessmentneeds further validation through appropriate lifecycle tracking ofproducts from cradle-to-grave. RFID (Radio Frequency Identifica-tion Devices) tags could be assigned to new shipments which cantrack the PCs over its entire lifecycle. Alternatively, extensive fieldwork and consumer survey could be pursued.

9. Conclusion

While we emphasize that domestic generation of WEEE indeveloping country is rising rapidly, corresponding growth inrecycling infrastructure whether formal or informal are far fromevident. We argue that the secondary EEE market in the developing

world is robust and growing. In India, recovering and refurbishingEEE products are a source of income and livelihood to thousands ofpeople. The primary concern among policymakers is the fear thatthe current EPR system would lead to collapse of the secondarymarket. AWEEE program devised just to collect and recycle will beof limited objective. On the economic side, the high lifecycle cost ofrecycling WEEE and the ensuing environmental pressure from therecycling residue justifies reuse prior to recycling. Sustained massflow of used e-products to refurbish/reuse business sector is centralfor competing with new products. Apart from the requirement ofa well organized take-back strategy, appropriate measures forrefurbishment as well as quality assurance are other concerns ofthis business. The entire reverse logistic focus for collecting WEEEshould be reorganized so that not only the defective units reach therecycler, but also quality used products which could be furtherrefurbished. Illegal imports of WEEE have the potential to under-mine the current Indian legislation. Illegal imports together withthe informal downstream processes are existing market anomaliesthat needs to be addressed since they represent major dysfunctionsof the current EPR legislation.

References

Akenji, L., Hotta, Y., Bengtsson, M., Hayashi, S., 2011. EPR policies for electronics indeveloping Asia: an adapted phase-in approach. Waste Management &Research 29 (9), 919e930.

Babu, B.R., Parande, A.K., Basha, C.A., 2007. Electrical and electronic waste: a globalenvironmental problem. Waste Management & Research 25, 307e318.

Bandyopadhyay, A., 2008. A regulatory approach for e-waste management: a cross-national review of current practice and policy with an assessment and policyrecommendation for the Indian perspective. International Journal of Environ-ment and Waste Management 2 (1e2), 139e186.

Bandyopadhyay, A., 2010. Electronics waste management: Indian practices andguidelines. International Journal of Energy & Environment 1 (5), 793e804.

Boma, M.B.-W., Jeremy, R.G., Randolph, E.K., 2010. Modeling Electronic WasteRecovery Systems Under Uncertainty. IEEE International Symposium onSustainable Systems and Technology (ISSST), pp. 1e6.

Carter, N., 2003. The Politics of the Environment: Ideas, Activism, Policy. CambridgeUniversity Press, Cambridge.

Chen, J.-M., Chang, C.-I., 2012. The co-opetitive strategy of a closed-loop supplychain with remanufacturing. Transportation Research Part E 48, 387e400.

Cho, G., Meyer, C., 2001. Comparison of perturbation bounds for a stationarydistribution of a Markov chain. Linear Algebra and its Applications 335,137e150.

Dimitrakakis, E., Gidarakos, E., Basu, S., Rajeshwari, K.V., Johri, R., Bilitewski, B.,Schirmer, M., 2006. Creation of optimum knowledge bank on e-wastemanagement in India. In: ISWA Annual Conference. http://www.iswa2006.org/papersalpha.htm.

Directive 2002/96/EC of the European Union Parliament and of the Council of 27January 2003 on Waste electrical and electronic equipment (WEEE). OfficialJournal of European Union, 24e38.

Dixit, N., 2007. E-waste: a Disaster in the Making, vol. 7(2). CHANGE e The GoorejHouse Magazine.

Dutta, S., Upadhyay, V.P., Sridharan, U., 2006. Environmental management ofindustrial hazardous wastes in India. Journal of Environmental Science &Engineering 48 (2), 143e150.

Dwivedy, M., Mittal, R.K., 2010a. Estimation of future outflows in India. WasteManagement 30 (3), 483e491.

Dwivedy, M., Mittal, R.K., 2010b. Future trends in computer waste generation inIndia. Waste Management 30 (11), 2265e2277.

Fiksel, J., 2006. Sustainability and resilience: toward a system approach. SustainableSolutions 2 (2). http://ejournal.nbii.org/progress/2006fall/0608-028.fiksel.html.

Forslind, K.H., 2009. Does the financing of extended producer responsibility influ-ence economic growth? Journal of Cleaner Production 17, 297e302.

French, M.L., 2008. Improving sustainability through effective reuse of productreturns: minimizing waste in a batch blending process environment. Journal ofCleaner Production 16 (15), 1679e1687.

Greenpeace, 2008. Take Back Blues e An assessment of e-waste take back inIndia. http://www.greenpeace.org/raw/content/india/press/reports/take-back-blues.pdf.

Griese, H., Poetter, H., Schischke, K., Ness, O., Reichl, H., 2004. Reuse and LifetimeExtension Strategies in the Context of Technology Innovations, Global Marketsand Environmental Legislation. IEEE.

Hammond, D., Beullens, P., 2007. Closed-loop supply chain network equilibriumunder legislation. European Journal of Operational Research 183, 895e908.

Hong, I.-H., Yeh, J.-S., 2012. Modeling closed-loop supply chains in the electronicsindustry: a retailer collection application. Transportation Research Part E 48,817e829.

M. Dwivedy, R.K. Mittal / Journal of Cleaner Production 37 (2012) 229e242242

Hunter, J., 2003. Mixing times and applications to perturbed Markov chains.Research Letters in the Information and Mathematical Sciences 4, 35e49.

Jain, A., Sareen, R., 2006. E-waste assessment methodology and validation in India.Journal of Material Cycles Waste Management 8, 40e45.

Jain, A., 2010a. E-waste business models, policies and regulations in India. In:Proceedings of IEEE International Symposium on Sustainable Systems andTechnology (ISSST), Arlington, Virginia, USA, pp. 1e4.

Jain, A., 2010b. E-waste management in India: current status, emerging drivers andchallenges. In: Regional Workshop on e-waste/WEEE management, Osaka,Japan. gec.jp/gec/jp/Activities/ietc/fy2010/e-waste/ew_1-2.pdf.

Kautto, P., Melanen, M., 2004. How does industry respond to waste policy instru-ments e Finnish experiences. Journal of Cleaner Production 12 (1), 1e11.

Khetriwala, D.S., Kraeuchib, P., Schwaninger, M., 2005. A comparison of electronicwaste recycling in Switzerland and in India. Environmental Impact AssessmentReview 25 (5), 492e504.

Khetriwal, D.S., Kraeuchi, P., Widmer, R., 2009. Producer responsibility for e-wastemanagement: key issues for consideration e learning from the Swiss experi-ence. Journal of Environmental Management 90 (1), 153e165.

Kimura, F., Hata, T., Suzuki, H., 1998. Product quality evaluation based on behavioursimulation of used products. Annals of the CIRP 47, 119e122.

King, A.M., Burgess, S.C., Ijomah, W., McMahon, C.A., 2006. Reducing waste: repair,recondition, remanufacture or recycle? Sustainable Development 14, 257e267.

Kojima, M., Yoshida, A., Sasaki, S., 2009. Difficulties in applying extended producerresponsibility policies in developing countries: case studies in e-waste recyclingin China and Thailand. Journal of Material Cycles Waste Management 11,263e269.

Kronenberg, J., 2007. Making consumption reasonable. Journal of Cleaner Produc-tion 15 (6), 557e566.

Loomba, A.P.S., Nakashima, K., 2012. Enhancing value in reverse supply chains bysorting before product recovery. Production Planning & Control: The Manage-ment of Operations 23 (2e3), 205e215.

Lua, L.-T., Wernick, I.K., Hsiao, T.-Y., Yu, Y.-H., Yang, Y.-M., Maa, H.-W., 2006.Balancing the life cycle impacts of notebook computers: Taiwan’s experience.Resources, Conservation and Recycling 48, 13e25.

MAIT-GTZ, 2007. First MAIT-GTZ Study Reveals Extent of e-waste Challenge. PressRelease, New Delhi. http://www.mait.com/admin/press_images/press77-try.htm.

MAIT-GTZ, 2007a. E-waste Assessment in India e a Quantitative Understanding ofGeneration, Disposal & Recycling of Electronic Waste in India.

MAIT-GTZ, 2007b. E-Waste Assessment in Delhi e a Quantitative Understanding ofGeneration, Disposal & Recycling of Electronic Waste in Delhi.

MAIT, 2010. India IT Industry Performance Annual Reviews: 2009e2010.http://www.mait.com/industry-statistics.php.

Manomaivibool, P., 2009. Extended producer responsibility in a non-OECD context:the management of waste electrical and electronic equipment in India.Resources, Conservation and Recycling 53, 136e144.

Matsumoto, M., 2009. Business framework for sustainable society: a case study onreuse industries in Japan. Journal of Cleaner Production 17 (17), 1547e1555.

Matsumoto, M., 2010. Development of a simulation model for reuse business andcase studies in Japan. Journal of Cleaner Production 18, 1284e1299.

Matsuno, Y., Daigo, I., Adachi, Y., 2007. Application of Markov chain model tocalculate the average number of times of use of a material in society e part 2.International Journal of Life Cycle Assessment 12 (1), 34e39.

McKerlie, K., Knight, N., Thorpe, B., 2006. Advancing extended producer responsi-bility in Canada. Journal of Cleaner Production 14, 616e628.

Michaud, M., Llerena, D., 2006. An economic perspective on remanufacturedproducts: industrial and consumption challenges for life cycle engineering. In:13th CIRP International Conference on Life Cycle Engineering e Leuven,pp. 543e548.

Moskalyuk, A., 2004. India PC Sales on the Rise. http://blogs.zdnet.com/ITFacts/?p¼5151.

Mundada, M.N., Kumar, S., Shekdar, A.V., 2004. E-waste: a new challenge for wastemanagement in India. International Journal of Environment Studies 61 (3),265e279.

Nischalke, S.M., 2007. Sustainable e-waste Legislation and Social Responsibility inIndia: Opportunities and Limitations. Master’s thesis, Albert-Ludwigs-Uni-versität Freiburg i.Br. (Germany) and University of KwaZulu-Natal, Durban(South Africa).

O’Connell, M.W., Hickey, S.W., Fitzpatrick, C., 2011. Evaluating the SustainabilityPotential of a White Goods Refurbishment Program in Ireland. www.reevaluate.ie/wp-content/uploads/EST_paper_submission_August_11.pdf.

Pagell, M., Wu, Z., Murthy, N.N., 2007. The supply chain implications of recycling.Business Horizons 50, 133e143.

Parlikad, A., McFarlane, D., 2005. Recovering Value from End-of-Life Equipment.Technical Report No. CUED/E-MANUF/TR. 29, Centre for Distributed Automationand Control Institute for Manufacturing, Department of Engineering, Universityof Cambridge.

Raghavan, S., 2010. Don’t throw it away: the corporate role in product disposition.Journal of Business Strategy 31 (3), 50e55.

Shinkuma, T., Managi, S., 2010. On the effectiveness of a license scheme for e-wasterecycling: the challenge of China and India. Environmental Impact AssessmentReview 30, 262e267.

Sinha-Ketriwal, D., Kraeuchi, P., Schwaninger, M., 2005. A comparison of electronicwaste recycling in Switzerland and in India. Environmental Impact AssessmentReview 25, 492e504.

Sinha, S., Mahesh, P., 2007. Into the Future: Managing e-waste for protecting livesand livelihoods. http://www.toxicslink.org/pub-view.php?pubnum¼171.

Streicher-Porte, M., Widmer, R., Jain, A., Bader, H.P., Scheidegger, R., Kytzia, S., 2005.Key drivers of the e-waste recycling system: assessing and modeling e-wasteprocessing in the informal sector in Delhi. Environmental Impact AssessmentReview 25, 472e491.

Streicher-Porte, M., Yang, J., 2007. WEEE recycling in China: present situation andmain obstacles for improvement. In: Proceedings of IEEE Symposium on Elec-tronics and the Environment.

Subramoniam, R., Huisingh, D., Chinnam, R.B., 2009. Remanufacturing for theautomotive market e strategic factors: literature review and future researchneeds. Journal of Cleaner Production 17 (13), 1163e1174.

Subramoniam, R., Huisingh, D., Chinnam, R.B., 2010. Aftermarket remanufacturingstrategic planning decision-making framework: theory and practice. Journal ofCleaner Production 18, 1575e1586.

Tasaki, T., Hashimoto, S., Moriguchi, Y., 2006. A quantitative method to evaluate thelevel of material use in lease/reuse systems of electrical and electronic equip-ment. Journal of Cleaner Production 14, 1519e1528.

Toxics Link, 2003. A Report on Scrapping the Hi-Tech Myth: Computer Waste inIndia. http://www.toxicslink.org/pub-view.php?pubnum¼37.

Toxics Link, 2012. New e-waste Rules Risk Falling Flat Without Specific Guidelines:Experts. http://www.toxicslink.org/?q¼media/media-coverage/new-e-waste-rules-risk-falling-flat-withoutspecific-guidelines-experts.

Truttmann, N., Rechberger, H., 2006. Contribution to resource conservation by reuseof electrical and electronic household appliances. Resources, Conservation andRecycling 48, 249e262.

Veenstra, A., Wang, C., Fan, W., Ru, Y., 2010. An analysis of e-waste flows in China.The International Journal of Advanced Manufacturing Technology 47 (5e8),449e459.

Wath, S.B., Dutt, P.S., Chakrabarti, T., 2010. E-waste scenario in India, itsmanagement and implications. Environmental Monitoring Assessment.http://dx.doi.org/10.1007/s10661-010-1331-9.

Webster, S., Mitra, S., 2007. Competitive strategy in remanufacturing and the impactof take-back laws. Journal of Operations Management 25, 1123e1140.

Widmer, R., Oswald-Krapf, H., Sinha-Khetriwal, D., Schnellmann, M., Boni, H., 2005.Global perspectives on e-waste. Environmental Impact Assessment Review 25(5), 436e458.

Williams, E., Kahhat, R., Allenby, B., Kavazanjian, E., Kim, J., Xu, M., 2008. Environ-mental, social and economic implications of global reuse and recycling ofpersonal computers. Environmental Science & Technology 42 (17), 6446e6454.

Yang, X., Gao, F., 2009. Controlling remanufacturing time based on Markov process.In: Proceedings of IEEE International Conference on Information Management,Innovation Management and Industrial Engineering, pp. 176e178.

Yu, J., Williams, E., Ju, M., Yang, Y., 2010. Forecasting global generation of obsoletepersonal computers. Environmental Science & Technology 44 (9), 3232e3237.

Yoshida, A., Murakami-Suzuki, R., Terazono, A., 2007. Present Status of Reuse/recycling of WEEE in Japan. Thailand’s Electrical and Electronic Green SocietyInternational Conference.

Zoeteman, B.C.J., Krikke, H.R., Venselaar, J., 2010. Handling WEEE waste flows: onthe effectiveness of producer responsibility in a globalizing world. InternationalJournal of Advanced Manufacturing Technology 47, 415e436.