Territorial Agglomeration and IndustrialSymbiosis: Sitakunda-Bhatiary, Bangladesh,as a Secondary Processing Complex

Nicky GregsonDepartment of GeographyDurham UniversityScience LaboratoriesDurham DH1 3LEUKnicky.gregson@durham.ac.uk

Mike CrangDepartment of GeographyDurham Universitym.a.crang@durham.ac.uk

Farid Uddin AhamedDepartment of SociologySouth Asian Universityfuahamed@soc.sau.ac.in

Nasreen AkterDepartment of AnthropologyUniversity of Chittagongnasreen_shilpi@yahoo.com

Raihana FerdousDepartment of GeographyDurham UniversityRaihana.ferdous@durham.ac.uk

Sadat FoisalDepartment of AnthropologyUniversity of ChittagongSadatahmedfoisal@gmail.com

Ray HudsonDepartment of GeographyDurham Universityray.hudson@durham.ac.uk

Key words:wasteindustrial symbiosisagglomerative tendenciessustainable economicdevelopmentBangladesh

abst

ract

[Correction added after online publication March 16,2011: The contact information for two authors waslisted incorrectly. The email addresses for FaridUddin Ahamed and Nasreen Akter have been cor-rected in this version.]

This article both joins with recent arguments ineconomic geography that have made connectionsbetween work on industrial symbiosis and agglom-erative tendencies and recasts this work. Drawing onthe case of Sitakunda-Bhatiary, Bangladesh, it showsthat symbiosis is intricately bound up in the globalcirculation of wastes and their recovery through sec-ondary processing. It draws attention to the impor-tance of key places as conduits in the transformationof materials and secondary processing; emphasizestheir importance as sites of symbiotic activity; andshows how such places exemplify economies of recy-cling, reuse, and remanufacturing, but in conditionsof minimal environmental regulation. It thereforeshows that contemporary symbiosis is not necessarilyclean and green and may be very messy; that it canbe generative of agglomerations, not just dependentupon prior agglomerations; that such agglomerationsmay be cross sectoral, not just interplant; and thatsymbiosis needs to be thought of not just throughgeographic proximity, but through the spatialities ofglobalization.ecge_1138 37..58

37

ECO

NO

MIC

GEO

GR

APH

Y88(1):37–58.©

2011C

larkU

niversity.w

ww

.economicgeography.org

Acknowledgments

This article draws onresearch funded by GrantRES 000230007 from theEconomic and SocialResearch Council under TheWaste of the Worldprogram.Thanks to Ash Aminfor comments on an earlierdraft, to three anonymousreferees, and to Henry Yeungfor their suggestions forimprovements.The usualdisclaimers apply.

In this article, we use empirical research conducted inBangladesh to recast the economic geography literaturethat has drawn connections among agglomerative ten-dencies, industrial ecology, and industrial symbiosis(Gibbs, Deutz, and Proctor 2005; Deutz and Gibbs2008; Gibbs 2006, 2008; Deutz and Lyons 2008).Drawing on work in Sitakunda-Bhatiary, near Chit-tagong, the article makes three contributions to thedebate. First, it shows how looking to the South cancontest prevailing accounts of industrial symbiosis, spe-cifically its assumed relation to firms and sectors,markets, and regulation, and its portrayal as a clean,green form of organizing production. Second, by payingattention to global flows in wastes, particularly whatsubsequently happens to them, the article complicatesthe relationship between industrial symbiosis and terri-torial agglomeration, as it has been understood to date. Itshows that symbiosis can be generative of agglomera-tions, not just dependent on their prior existence, andthat it is multiscalar, rather than simply local or regional.Third, and more broadly, in so doing, the article bothchallenges the prevailing orthodoxies’ positioning ofthe economies of the South in globalization as sitesof low-cost manufacturing for consumer goods inglobal markets (Gibbon 2001; Henderson et al. 2002;Kaplinsky Morris, and Readman 2002; Coe et al. 2004;Coe et al. 2008) and suggests that it is possible to learnfrom the economies of the South (c.f. Murphy 2008;Pollard, McEwan, Laurie, and Stenning 2009).

The literature on industrial symbiosis focusesalmost exclusively on interplant flows of energy andmaterials in primary processing. Its emphasis is onhow physical, material wastes and the by-products ofspecific production processes are exchanged betweendiscrete economic entities, typically manufacturingplants, which may be, but are not necessarily, firms.This thinking comes with a strong normative agenda,grounded in industrial ecology, which seeks to drawaffinities between natural ecosystems and economies.The argument, made first in the seminal article byFrosch and Gallopoulos (1989), is that in the interestsof sustainable development, economies should movefrom being linear to circular and that the way to do sois to close material loops through the exchange ofby-products and wastes.A corollary of the argument isthat collective and/or cooperative interfirm actiongenerates environmental benefits by moving towardminimizing wastes. Industrial symbiosis, therefore, isseen by its proponents as a means to cleaner, greenerproduction.

ECONOMIC GEOGRAPHY

38

A striking feature of work on industrial symbiosis is its tendency to generalize fromparticular European circumstances. In this article we consider a different case of indus-trial symbiosis: the secondary processing complex of Sitakunda-Bhatiary, Bangladesh.We chose this agglomeration for two reasons. First, there is no reason a priori to confinestudies of symbiosis to primary processing, as is recognized in the limited literature thathas drawn parallels with scavengers, fungi, and bacteria in natural ecosystems to acknowl-edge the niche role of “scavengers” and “decomposers” in industrial systems (Geng andCôte 2002). Indeed, with the growing importance of the recovery (for recycling), reuse,and remanufacturing of secondary materials within understandings of waste minimiza-tion and sustainable production, it can be argued that it is important to turn to theconsideration of secondary processing activities in studies of industrial symbiosis.Second, there has been a dearth of work on symbiosis in the South. In this regard, thework of Chertow and her colleagues in Puerto Rico is the one sustained example(Deschenes and Chertow 2004; Chertow 2007; Chertow, Ashton, and Espinosa 2008) toset alongside Erkman and Ramaswamy’s (2006) demonstration of the potential ofresource flow analysis and studies on industrial metabolism for developmental planningin India.

The limited attention paid to symbiosis in the South may be explained, in part, by theemergence of the concept in those parts of the world where moves toward sustainableproduction systems have gathered pace, subtended by arguments about ecomoderniza-tion. To see symbiosis as an exclusively modern concept, however, is to neglect that muchof the development of industrial capitalism in the 19th century, particularly in Europe butalso in North America, was predicated upon the symbiotic exchange of by-products andwastes. The close connection between the wastes of the chemical industry and the alliedgrowth of the textile dye industry is a case in point, as is the further development of thechemical industry through its by-products (Bensaude-Vincent and Stengers 1996).Acknowledging this connection opens the door to admitting that there is a messier, dirtierform of symbiosis, or, in other words, that symbiosis is not necessarily clean and/or green.As this article shows, it is the messier form of symbiosis that has emerged in con-temporary economies in the South, in market and regulatory contexts that havestrong resonances with those that pertained in 19th-century Europe and NorthAmerica.

Unlike their northern predecessors, however, which developed within the largelyregionally and nationally bound spatial divisions of labor of the 19th century, contempo-rary instances of messy symbiosis in the South are as likely produced through globalizedflows of wastes. Trading wastes and the creation of and trade in by-products have a longhistory. Alfred Marshall, best known as the originator of arguments on industrial districts,noted that “many of the most important advances of recent years have been due to theutilising of what had been a waste product; but this has generally been due to a distinctinvention, either chemical or mechanical” (quoted in Desrochers 2004, 1100). However,the trade in wastes and the creation of by-products is not just about innovation. Geographymatters, too, in that the differences between waste regimes (Gille 2007) enable boththe movement and recategorizations of one person’s wastes as another’s resource.That geography matters was recognized shortly after Marshall, by the chemicalengineer John Kershaw who wrote in his 1928 book The Recovery and Use of Indu-strial and Other Wastes: “Dirt, from the philosopher’s standpoint, is simply matter inthe wrong place, and industrial waste may be regarded similarly as useful materialproduced or dumped in places where it is not required. When transported to theright spot an industrial waste will often form the raw material for some secondaryindustry or manufacture” [Emphasis added] (quoted in Desrochers and Lam 2007,

Vol. 88 No. 1 2012

39

TER

RITO

RIA

LA

GG

LOM

ERAT

ION

INB

AN

GLA

DESH

38).1 The global trade in wastes and the use of wastes as feedstock for secondaryindustries show that it is places in the global South that are now the right spots. Typically,these flows of wastes are portrayed as the dumping of the developed world’s industrialwastes on the countries of the South. Such dumping has long been a concern of those withinterests in environmental (in)justice (Clapp 2001; Basel Action Network 2002, 2005),with some contrarian accounts arguing for it as an efficient market solution to the cost ofwastes in the developed world (e.g., Johnson, Pecquet, and Taylor 2007). Less consideredby either side in this debate is the connection between global flows of wastes and theeconomic activities of secondary processing. Recognizing this connection opens up thequestion of how geographic scale is related to understanding closed-loop processes—whether these processes now have to be seen as international (Lyons, Rice, and Wachal2009) but simultaneously intersecting with local agglomerations. Using the instance ofSitakunda-Bhatiary in Bangladesh, the article shows that the current “right spot” for aparticular industrial waste (end-of-life ships) has created a local agglomeration foundedon a network of messy symbiotic exchanges that are based in secondary processing of aglobally supplied waste.

The article begins with a summary of the literature on industrial symbiosis and theconnection drawn in this work to research on agglomerative tendencies. Typically, thatlink is made through a focus on industrial districts and ecoindustrial parks, or EIPs. Next,it provides a brief overview of the secondary processing complex of Sitakunda-Bhatiarybefore focusing on three key sectors within this complex: secondary steel production,based on recovered recycled ferrous scrap from ship breaking, reconditioned engineeringproducts, and remanufactured furniture. It then tracks back to the literature on industrialsymbiosis to address what turning to the messy symbiotic agglomerations of theSouth offers to the academic debate about symbiosis. The article concludes by reflectingmore broadly on how messy symbiotic agglomerations that are founded on secondaryprocessing are related to wider issues of economic development in the South.

Industrial Symbiosis and Agglomerative TendenciesIn the literature on industrial ecology and symbiosis, one example is cited as the

paradigmatic case of symbiosis in practice: Kalundborg, Denmark. A useful descriptionis provided by Ehrenfeld and Gertler (1997) and has been repeated on many subsequentoccasions. The core features of Kalundborg are as follows: the complex evolved overabout 25 years, seemingly as a means of minimizing the cost of compliance with newenvironmental regulations, a point that is overlooked in all subsequent commentaries. Atthe heart of the complex are four plants—a coal-fired power station, an oil refinery, apharmaceutical plant that produces enzymes, and a plasterboard manufacturing plant.Waste heat is exchanged from the power plant to other plants, notably the pharmaceuticalsplant, and to fish farms and a CHP (combined heat and power) system. Deschenes andChertow (2004, 206) described Kalundborg thus:

There are some 20 resource exchanges involving an oil refinery, a power station, a gypsumboard facility, a pharmaceutical plant, the municipality of Kalundborg and other actors. Theparticipants literally share ground water, surface water and wastewater, steam and fuel. Theyalso exchange a variety of residues, such as ash and scrubber sludge from the power plant andsulphur from the oil refinery, that become feedstocks in other processes.

1 The philosopher in question was William James, from whom Mary Douglas (1966)—in Purity andDanger—borrowed the phrase for her summary of dirt as matter out of place.

ECONOMIC GEOGRAPHY

40

Each link is an independently negotiated business deal, in which the principles ofindustrial ecology have been applied only if they are seen to be economically beneficial.Effectively, this means using a by-product or feedstock at less than the cost of the primaryraw materials or a saving on costs incurred for “disposal.” Subsequent work has identifieda handful of further instances of symbiosis in the developed world, mostly in Scandinaviaand focused mainly on energy recovery—Jyväskylä, Finland (Korhonen 2001, 2002), andLandskrona, Sweden (Mirata and Emtairah 2005). Deutz and Lyons (2008) added Styria,Austria (c.f. Schwartz and Steininger 1997), whereas Chertow (2007) extended the list indeveloped economies to include Kwinana and Gladstone, Australia (see van Beers,Corder, Bossilkov, and Van Berkel 2007); Triangle J, North Carolina, United States; andBurnside Industrial Park, Nova Scotia, Canada (see Geng and Côte 2002). The far-from-extensive list of empirical cases, suggests that notwithstanding a lot of enthusiasm for theprinciples of symbiosis, there is little evidence of success.

Together, the sparsity of instances of symbiosis on the ground and the increasing swayof environmental issues in framing economic development have meant that Kalundborghas been held up as a model for sustainable economic development. Kalundborg isregarded in European academic and policy circles as emblematic of how symbiosis resultsin cleaner, greener production. Consequently, there have been attempts to distill a set ofprinciples from this paradigmatic case. These principles have been assembled into thepolicy concept of EIPs. EIPs seek to enhance the competitiveness of businesses by usingone or more of three agglomerative possibilities: a shared infrastructure, the joint provi-sion of services; or, most important, the creation of exchanges of by-products, wastes, andenergy among geographically proximate firms. The latter moves beyond “passive” to“dynamic” agglomeration and involves matching and developing interests (Chertow2007; Chertow et al. 2008). The trouble is that, notwithstanding their promotion (Hewesand Lyons 2008), EIPs remain more concept and rhetoric than empirically developed.Gibbs and Deutz’s (2005) work shows that although symbiosis is the ideal, EIPs in Europeand North America struggle to get beyond the aesthetic and the cosmetic, that is, beyondlandscaping and infrastructure management. Gibbs et al. (2005, 179) remarked: “Theconstant reference to Kalundborg as a working model and the limited range of subsequentexamples examined in the literature suggests that initiatives based upon the interchangeof wastes and cascading of energy are few in number and perhaps difficult to organise.”In response, a consensus seems to be emerging that normative attempts to apply theprinciples of industrial ecology in Europe and North America either will not or do notwork and that what needs to be taken from the Kalundborg paradigmatic case are qualitiesthat start from its economic sociology. Consequently, attention is beginning to turn in theEuropean- and North American—focused literature to issues that were first highlighted ineconomic geography in relation to the study of industrial districts and agglomerativetendencies. This is what Ehrenfeld and Gertler (1997, 74) stated in relation to theiroriginal study of Kalundborg, but was lost in the attention paid to material and energysynergies, namely, the “short mental distance between firms.”

The potential significance of social and cultural ties to the development of industrialecology and symbiosis on the ground in Europe was confirmed by a study of Murano,Veneto, that showed how an Italian industrial district based on glass manufacturing hasbeen able to develop symbiotic exchanges (of wastewater and oxygen) by drawing on priorties of social connection and shared culture (D’Amico, Buleandra, Veland, and Tanase2007). More generally, and in a number of review articles, Gibbs and Deutz and theircoworkers have argued for a greater connection between work in the industrial ecology/symbiosis tradition and the literatures on agglomerative tendencies and regional develop-ment (Gibbs et al. 2005; Deutz and Gibbs 2008; Deutz and Lyons 2008; Gibbs 2008).

Vol. 88 No. 1 2012

41

TER

RITO

RIA

LA

GG

LOM

ERAT

ION

INB

AN

GLA

DESH

If the examples in the West are sparse but often cited, then alternative studies havepointed to symbiosis in the developing world—notably the celebrated cases of Barcelon-eta and Guayama in Puerto Rico (Deschenes and Chertow 2004; Chertow and Lombardi2005; Chertow et al. 2008). Another strand of work has pointed to the East Asianeconomies as especially suited to adopting the principles of industrial ecology. One caseof the firm-based route is Zhu, Lowe, Wei, and Barnes’s (2007) work on the Guitanggroup of sugar refiners. That study documented the group’s development of a strategy ofinternal and external symbiosis over four decades. Internal symbiotic linkages character-ize the production of sugar, alcohol, cement, compound fertilizers, and paper, whileexternal relationships with other producers, which have developed organically but withinthe same region, focus on two product chains: paper and sugar. Zhu et al. also pointed toa potential downside to symbiosis: its capacity to make innovation harder through lock-inbased on material exchanges. Although recognizing their differences from the Westerndefinition of EIPs, other research has pointed to the success of ecoparks (e.g., Ulsan,Korea) and ecocities (e.g., Guiyang, Southwest China) and has highlighted the impor-tance of closed-loop thinking in economic planning, for example, in relation to Shanghai(Rock, Angel, and Feridhanuesetyawan 1999; Shi, Moriguichi, and Yang 2002; Park andWon 2007). A case in point here is Shi, Chertow, and Song’s (2010) study of the TianjinEconomic Development Area. In addition to documenting a network of 81 interfirmsymbiotic exchanges of materials and energy, grounded in the electronics, automobilesand machinery, biotechnology and pharmaceuticals, and food and beverages industries,Shi et al. showed that symbiotic exchanges can be discontinued, thus countering argu-ments regarding the propensity for symbiotic arrangements toward lock-in. At the sametime, their work recognized the potential conflict between the interfirm exchange ofmaterials and moves toward cleaner production, with the latter potentially working toreduce, or even eradicate, the former.

The body of empirical work in East Asia builds on a point recognized by Hamner(1997), that the regulatory context in East Asia provides industrial ecology with a uniqueopportunity. Shi et al. (2010) supported these arguments, showing that although theearliest symbiotic exchanges were formed spontaneously in response to savings oneconomic costs, public policies—particularly subsidies but also information sharing—have played a key role in the development of both symbiotic principles and cleanerproduction.

In sharp distinction, the existence of symbiotic principles in South Asia appears to fitthe industrial ecology tradition less well. Thus, the one instance of a South Asianapplication of industrial ecology methodologies to show their potential for developmentalplanning was Erkman and Ramaswamy’s (2006) resource flow analysis of Tiruppur’stextile industry, Haora’s foundries, and the leather tanneries of Tamil Nadu and theDamodar Valley. This work highlighted an alternative development platform to pollutionabatement, based on maximizing the productivity of the limited resources available and afull assessment of their various possible uses.

Staying with India, but very much within the tradition of research on clusters in theSouth, research on the leather-producing cluster in the Palar Valley (Tamil Nadu) showedhow the hundreds of tanneries in the area worked to install a collective effluent treatmentplant (Kennedy 1999), which involved mobilizing to create a shared infrastructure but notan economy of materials exchange. The critical point here is that those actions both drewon social ties, of community and shared local identities, and were a response to threats toclose the tanneries for failure to comply with regulations on the control of effluents. In thisrespect, the development of symbiotic links in the Palar Valley has affinities with theMurano case, and the initial, but largely neglected, impetus behind Kalundborg. As

ECONOMIC GEOGRAPHY

42

Kennedy stated, “Here, acting under constraint, firms have collectively undertaken thetask of internalising costs which previously constituted negative externalities, and whichwere borne by the local population” (1674). Although these studies all stuck to industrialecology’s classic and normative assumption in favor of geographic proximity, vanBeukering and Duralappah (1998) extended the boundary of the system to the global scaleto study the global trade in wastepaper and its recycling in India. The argument here is thattrade liberalization, at least in relation to nontoxic wastes, brings about an efficient globalallocation of resources and has the capacity both to engender economic development andto reduce environmental degradation. Moreover, the example poses the questions ofwhether studies have looked in the wrong places or at the wrong scales for exchangesof materials and how global flows of materials are related to local agglomerations(Lyons et al. 2009). Industrial ecology celebrates the transformation of one firm’s wasteinto another’s feedstock. But that transformation is often enabled by global flowsmoving materials from developed to less developed worlds, since in the former they areclassified as wastes, and thus problems, whereas in the latter, they are classified asresources.

Look beyond Europe, North America, and Australia, then, and there are examples ofthe principles of industrial ecology and industrial symbiosis on the ground, especially inEast Asia. In South Asia, the interconnections are not as dense and do not reach Chertow’s(2007) minimal definition of symbiosis as involving a 3-2 materials exchange betweenindependent plants, but material linkages do exist. All these cases of material symbiosisdepend on the conjuncture of markets, their regulation and the social economy, and theuniqueness of that interplay within and between economies.

Establishing that there are examples of industrial symbiosis beyond Europe leaves anexplanatory conundrum. To ignore the existence of these cases or to exclude them as onlypartial instances of symbiosis is clearly not good enough. This may be yet anotherinstance of the Euro-American centrism of economic geography (Yeung & Lin 2003:Murphy 2008; Pollard et al. 2009). Another reason why these instances of symbiosis donot appear in the literature on economic geography and industrial symbiosis may be that,as examples, they are perhaps not quite the right ones to highlight, especially for a modeof thinking that normatively supports clean production. These cases of symbiosis containmore contradictions and paradoxes and offer more ambiguous messages. As Desrochers(2000, 2002) recognized in his work on the uses of by-products in 19th-century Britain,symbiosis is not always clean. It can instead be messy, often very messy.2 And symbiosismay actually be easier to achieve in conditions of mess and environmental degradationthan in conditions of more stringent environmental regulation. Put simply, reducingwastes limits the scope for reuse. Desrochers’s work documented how, in the 19th-centurynational-scale economy, trade liberalization, and no environmental regulation, external-ized wastes and industrial residues were exchanged, reworked to become by-products,and then became the basis for whole new industries. Although this was a form ofsymbiosis par excellence, it certainly was not a clean form of production. The lack ofcontemporary Western examples of symbiosis, then, results from the interplayof markets and the governance of waste materials, along with globalized exchanges.Indeed, there are multiple permutations of regulation and market that can promote or

2 That symbiosis is not necessarily a form of clean production is also signaled by the reliance of the keyEuropean exemplars on coal-fired power plants as cornerstones of symbiotic exchanges (Kalundborg) or onunsustainable resources, such as peat (Jyväskylä). In this respect, lock-in becomes not just economiclock-in but also lock-in to carbon economies.

Vol. 88 No. 1 2012

43

TER

RITO

RIA

LA

GG

LOM

ERAT

ION

INB

AN

GLA

DESH

inhibit symbiosis; for example, some regulations that limit dumping can lead to increasedrecycling, but the designation of something as “waste,” or worse, as “toxic waste,” canform a barrier to reuse, even if it is chemically identical to a commercially produced“virgin” product (Desrochers 2002).

If we turn our attention to symbiosis in secondary processing, however, it is impossibleto overlook that notwithstanding recent investments in recycling and materials-recoverytechnologies in Europe, particularly in Scandinavia, the Netherlands, and Germany, thevast majority of this activity globally occurs neither in Europe nor in North America, butin India, China, and other countries in South and East Asia and Africa. That this is so isa matter of the flows of wastes in globalization, the importance of secondary materialsderived from wastes as feedstock for recycling and remanufacturing within the economiesof South and East Asia and Africa, and the regulatory conditions that shape theseeconomies. It is to an examination of one such complex of activities in Bangladesh thatwe turn in the following section—not, we emphasize, to generalize from the particularcase to the whole of the South but precisely to counter the tendency in the literature onindustrial symbiosis to generalize from the European case to the rest of the world, byshowing the emergence of a symbiotic agglomeration in SouthAsia that is based on globalflows that involve toxic wastes.

An important point to note is that the research on which this article was based did notset out to find industrial symbiosis or agglomeration per se. Instead, its genesis lay in aprogram of ethnographic research conducted around the global ship-breaking industryfrom 2008 to 2010 (Gregson et al. 2010). This research had begun by following flows ofmaterials across the globe, but initial fieldwork revealed the dense agglomeration ofsecondary processing activities around the Bangladeshi breaking yards. Consequently, afield survey was conducted in 2008 by research assistants who visited all secondhandretail and wholesale outlets in Bhatiary, near Chittagong, documenting their locations andproducts (see Figure 1) and using structured interviews to address their ownershipstructures, alliances, and employment patterns. The majority of these outlets manufacturethe products they sell. Where this was the case, data were collected on supply chains andsources of raw materials. The vast majority of these raw materials were wastes andby-products of the ship-breaking industry. Thus, the connection to symbiotic materialexchanges was established. This finding indicated that Sitakunda-Bhatiary should beexamined as a case of symbiosis, precisely because that is what occurs on the ground, ingeographic proximity, around ship breaking, at least in Bangladesh.

Sitakunda-Bhatiary:An Agglomeration Based on“Decomposition” and “Scavenging”

Together, India, Bangladesh, and Pakistan account for more than 75 percent by tonnageof the global trade in the ultimately footloose industrial waste: end-of-life commercialships. Bangladeshi imports of scrap ships rose in value from about U.S.$25 million in 1989to U.S.$400 million by 2007 (United Nations Statistics Division, Commodities and TradeDatabase), and industry insiders suggest the current value to be U.S.$1 billion. Industryreports give Bangladesh approximately 25 percent of the global market share by numberof ships broken and as much as 38 percent of the total tonnage in 2009.Yet these figures donot fully depict the spatial concentration of this industry. The ship-breaking industry inSouth Asia is based on just three locations—Alang (India), Chittagong (Bangladesh), andGaddani (Pakistan). Each site consists of a few kilometers of foreshore. The foreshore inthe Bangladeshi case stretches some 20 kilometers, from Fouzderhat to Kumira, north ofChittagong (see Figure 1). The attention of nongovernmental organizations (NGOs),

ECONOMIC GEOGRAPHY

44

Figure 1. The secondary processing complex of Sitakunda-Bhatiary, Chittagong, Bangladesh.

Vol. 88 No. 1 2012

45

TER

RITO

RIA

LA

GG

LOM

ERAT

ION

INB

AN

GLA

DESH

filmmakers, lawyers, and politicians has understandably been drawn to the dissolution ofships on the foreshore (Greenpeace/FIDH/YPSA 2005;Young People SocialAction 2005;International Federation for Human Rights 2008), the abysmal labor conditions, and theresultant environmental degradation. This situation is depicted as a classic example of the“pollution haven” hypothesis—that the developed world’s wastes are dumped on a poorcountry. But beyond the foreshore, and largely invisible to the scrutiny of the world, lies theengine for Bangladeshi ship breaking, a secondary steel sector that is based on recyclingscrap ship plate. The foreshore, then, is just one component—albeit the critical one—in anagglomeration complex founded on decomposition and scavenging activities.Things havenot always been so in Sitakunda-Bhatiary. Although relatively near the second city ofChittagong, with its export processing zones, export-oriented garment factories, and portfacilities, 20 years ago the area was heavily reliant on indigenous agriculture, fishing, andforestry. Figure 2 presents a schematic of the main industries in Sitakunda-Bhatiary nowand the flows of materials between them. Four points stand out about the materialexchanges depicted in Figure 2, working to differentiate Sitakunda-Bhatiary from theexemplars in the literature on industrial symbiosis as was discussed earlier. In what followswe highlight these points before drawing on empirical work to elaborate on the salientdetails.

First, Sitakunda-Bhatiary far exceeds the 3-2 exchange heuristic, involving two mate-rials exchanges among three independent economic entities (usually plants), that Chertow(2007) regarded as a minimal condition for symbiosis. Sitakunda-Bhatiary is character-ized not by a small number of interplant exchanges, but by a network of exchanges thatlink distinctive sectors and a myriad of firms. Thus, exchanges of materials in Sitakunda-Bhatiary work to constitute a wider agglomeration, grounded in the productionof a variety of secondary products from reconditioning, reusing, and remanufacturingthe wastes and by-products of ship breaking and the allied recycling of ferrous and

SHIP-BREAKINGYARDS

Power-related

equipment

Reconditioning

firms/

engineered

products/

repair firms

Engines, crankshafts,

gearboxes, brakes, axles,

boilers, compressors

Furniture

remanufacturers

Wholesale

Furnitureremanufacturers

Furniture Rods Rods

ManualRRMs

AutomatedRRMs

SECONDARY STEEL PROCESSINGScrap metal

merchants

Retail/

wholesale

Retail/

wholesale

Retail/

wholesale

Boards

Particleplywoodpartex

asbestos

Furniture

(as is)

Spot market Tied market

Process facing

ferrous scrap

Nonferrous

scrap

Electrical

cable

Kitchen equipment,

cutlery

Destruction andmaterials recovery

Recyclingactivities

Reconditioning andremanufacturing

Reuse throughresale

Materials/goods for secondarymanufacturing/ reconditioning

Inshore

craft

Products soldwholesale or retail

Lifeboats, chandlery,

marine equipment

Construction

Industry

Domestic

consumers

SMEs

SMEs and

domestic consumers

and some export

Commodities’

Markets

Wiring andconstruction industry,

SMEs includinggarment industry

Restaurants/

domestic

consumers

Coastal

shipping

fleet

Figure 2. Schematic of material flows and exchanges: Sitakunda-Bhatiary, Chittagong, Bangladesh.

ECONOMIC GEOGRAPHY

46

nonferrous metals recovered as scrap. In this respect, the agglomeration of Sitakunda-Bhatiary can be considered an exemplar case of a secondary processing cluster.

Second, as Erkman and Ramaswamy (2006) showed, the flows of materials in this(Global South) case link a low-technology infrastructure, not high-technology plants, andthey are not focused on power plants as many European cases are. Exchanges among thevarious economic entities here are almost exclusively one way, rather than potentially twoway. They are from the low-technology (and highly labor intensive) decomposing activi-ties that occur on the foreshore to a plethora of low-technology scavenging activities,located in Bhatiary, which produce products for both domestic consumers and for smalland medium-sized enterprises (SMEs) in the manufacturing sector, particularly but notexclusively in Chittagong. The resultant dependence of scavenging activities on decom-posing activities provides a different type of symbiotic lock-in to that identified in relationto high-technology/interplant exchange forms of symbiosis. This is not a dependence ofmaterially interdependent plants, but a dependence of sectors on one decomposingactivity, namely, ship breaking and the allied processing and recovery of scrap metal.

Third and consequently, the decomposing activity of ship breaking and the recovery ofscrap metal is the clear “kernel” activity (Chertow 2007) within this secondary processingcluster. That kernel activity, as we show later, emerged organically, responding to a nichethat was opened up by the limitations of primary steel production in Bangladesh that led,in turn, to the rise of a secondary steel-producing sector based on rerolling ferrous scrapderived from ship breaking. Further organic evolution over the course of the next 20-oddyears has seen the emergence of various sectors of scavenging activity, all of them basedon feedstock derived from the wastes and by-products of ship breaking.

In both the importance of a kernel to its emergence and the organic nature of itsemergence, Sitakunda-Bhatiary echoes Kalundborg or any other high-profile successstory of symbiosis. Where it differs, however, is in the nature of the business transactionsthat underpin the exchanges of materials. With the exception of some breaking yardsco-owned with rerolling mills (RRMs), each exchange of materials is a separatelynegotiated and unique business deal, based on the auction of objects and materials fromeach vessel beached on the foreshore. It has to be so because although the supply of scrapships is broadly predictable, each ship that arrives on the foreshore is an independentsale—of a bespoke object purchased from the world market. The nature of the supply,consisting of large discrete units, means that key scavenging firms in different materialsstreams work as intermediaries, breaking the object (and materials) down into smaller lotsfor purchase by those without the working capital to bid for whole consignments.Financial transactions therefore work to cascade recovered materials (and things) throughthe local economy of Bhatiary and Chittagong.

Fourth, what marks Sitakunda-Bhatiary’s difference from the extant cases in theliterature is that as a secondary processing cluster, decomposition and scavenging activi-ties are the primary activities within this local economy. Local businesses have estimatedthat such activities constitute 80 percent to 90 percent of the total economic activity in theimmediate area. As such, decomposition and scavenging activities not only characterizethis agglomeration, but are its motor. Rather than simply existing alongside and in nichescreated by firms that are engaged in primary processing (Geng and Côte 2002), inSitakunda-Bhatiary the original decomposing and scavenging niches now drive the localeconomy.

Such are the bare points that characterize the agglomeration’s relationship to keysymbiotic principles. In what follows, we elaborate on the emergence of these decom-posing and scavenging activities and their key characteristics, focusing particularly on thekernel activity and two key scavenging sectors: reconditioned engineering products and

Vol. 88 No. 1 2012

47

TER

RITO

RIA

LA

GG

LOM

ERAT

ION

INB

AN

GLA

DESH

remanufactured furniture.3 The selection of these two sectors, rather than others (suchas electronic waste), is not just to show that scavenging activities can produce goodswith high and low unit prices for both domestically oriented manufacturing industriesand consumer-facing markets, but to demonstrate the distributed vulnerability of thissecondary processing agglomeration.

Ship Breaking and Secondary Steel ProductionThe kernel of the Sitakunda-Bhatiary agglomeration is the decomposing activity of

ship breaking and the flow of ferrous scrap from the foreshore to the RRMs of Bhatiary,Chittagong, and Dhaka. Process requirements shape even the types of vessels that appearon the foreshore, for it is oil tankers and container ships or, in other words, ships with lotsof space, that are quicker and easier to cut and are made from particular qualities andthicknesses of steel that are sought in the global ship-demolition market by Bangladeshiship breakers. In 2010 there were some 119 ship-breaking yards, whereas in 2008 therewere only 57—a sign of the upturn in the global ship-demolition market after the globaldownturn and its consequences for the shipping industry.4 Field data from 2009–10, basedon observed truck movements at a medium-sized yard at regular intervals over a one-yearperiod, suggest a flow of scrap from the beaches to the RRMs of about 3.05 million tonnes.This figure, which is congruent with figures from a maritime database on the amount oftonnage scrapped, suggests that 80 percent of Bangladesh’s national demand for steel issourced from the ship-breaking industry. From the foreshore, ferrous scrap is trucked toproximate RRMs, of which more than 70 are active in the greater Chittagong area, as wellas to the RRMs of the greater Dhaka area.5 The six largest ship-breaking firms all displayvertical integration with rerolling operations,6 while the remaining RRMs purchase scrapon a local spot market, primarily from breakers without formal links to RRMs. The mills’output is rebar, or rod; the primary market is the construction industry. The demand forrod in Bangladesh is high and expanding, with major infrastructure projects and thehigh-rise building-construction industry being major markets.

Unlike the cases that stand as exemplars in the literature on symbiosis, however, thekernel decomposing activity of ship breaking and metals recovery in Sitakunda-Bhatiaryrests not on a niche opened up by the wastes of primary production but, rather, on anopportunity afforded by the problems of primary steel production. To explain the currentsymbiotic exchanges between the foreshore and the RRMs of Sitakunda-Bhatiaryand Chittagong, then, it is necessary to track back to the foundation of Bangladesh,specifically to the difficulties associated with its lack of a steel industry.

3 Twenty-one interviews were conducted with key informants and key players in these two sectors: with ninefurniture firms, including family-run businesses and partnerships; the pioneer and leading-edge firms; andthose at the entry level to the sector. In the reconditioned goods sector, a total of 12 interviews wereconducted with firms selling reconditioned boilers, generators, and electrical cable. Although we were ableto obtain entry to one steel rerolling mill in 2009, sustained access to the rerolling mills (RRM) sectorproved impossible to negotiate. We therefore rely here on secondary source material and informationsupplied by those in other sectors. Interviews covered: business ownership structure, day-to-day organiza-tion and responsibilities; previous business history; reasons for entry into the business; changes over timeon both the supply and demand sides; financing and, when appropriate, the auction process; materialssupply chains; key products and the range of products; and labor—its supply, pay, and division.

4 During a hiatus caused by a temporary ban on importing ships imposed by the High Court in May 2010, only48 of the 70 were functioning.

5 Within-industry sources suggested that some 525 RRMs are active in Bangladesh during upturns in theship-demolition market. Business reports put that figure at 250–350.

6 Vertically integrated operations are RSRM, KSRM, PHP, Rahim Steel, Sema Steel, and Giri Steel.

ECONOMIC GEOGRAPHY

48

The plan for postpartition West and East Pakistan included two state-owned and runsteel mills, one in Karachi and the other in Chittagong. Construction of the Chittagongworks did not begin until 1963, however, and it was 1967 before the plant was operational.The planned capacity was limited to about 250,000 tonnes of liquid steel per annum, butthe plant consistently operated at half capacity and imported all its raw materials fromChina and Korea. During the 1971 Liberation War, the plant was shut down. It wasrestarted in 1972. With no experience in management, no business plan, and strainedrelations between unions and management, the now Bangladeshi-run mill only everworked at one-third capacity and failed to keep pace with the increased domestic demandfor steel products. Into this gap, aided and abetted by state backing, came a secondarysteel-producing industry that saw in ship breaking a source of scrap steel for RRMs thatcould produce steel products more cheaply for the domestic market. The industry evolvedgradually and organically from the early 1980s, when a couple of pioneer ship breakersmoved from selling scrap on the spot to begin to establish their own RRMs, through aperiod of rapid expansion and emulation in the late 1980s and early 1990s, to the currentsituation.

By the mid-1990s, the steel mill at Chittagong had become increasingly unprofitable.It was closed down in 1999, eventually to be bought for scrap in 2008. At this juncture,Bangladesh became entirely dependent for its steel products on secondary steel produc-tion or on expensive foreign imports. The late 2000s, however, saw a return to primarysteel production, this time private sector led. In 2008 BSRM opened a modern melt shopin Chittagong, producing prime-quality billet to international standards. BSRM iscurrently the market leader in the production of rods in Bangladesh, with one-fifth ofthe total domestic market (“Steelmaker Up for Expansion” 2010). Its investments arebeing matched by Magnum Steel and AKS.

Notwithstanding these developments, the majority of RRMs continue to producelow-grade, uncertified rod products. Increasingly, these products, which are traded ontheir price advantage, are the target of the primary steel producers. Aggressive marketingby BSRM, for example, plays on widespread fears about the use of poor-quality rod inthe construction sector (“Bengunbari Collapse” 2010; “6 Storey Building Tilts” 2010):“there is another type of bars from the ship breaking sector based on scrap plates and opencast ingots for which there is no mill certificate or test result. These are nongradedproducts and unfit for properly designed construction” (BSRM 2008).The perception ofshoddy goods made from recycled materials is clearly being mobilized to discredit theproducts of the RRM sector. The virtues of recycling are tainted by the associations withwaste, where secondhand is second best. Countering this argument, the RRM industryclaims that the shapes and sections that are rerolled from ship steel are immune fromphysical defects and that rebar originating from high-quality ship steel lends itself well tointensive cold twisting. Nonetheless, the kernel of the symbiotic cluster is a fragileaccomplishment and one in which upgrading in processes provides both opportunities andchallenges.

Reconditioned EngineeringExamples of the products of the reconditioned goods sector range from marine

equipment for the Bangladeshi coastal shipping fleet of 8,000 registered, let aloneunregistered, vessels to kitchen equipment and utensils, and bathroom fittings andWestern toilets sold in the shops that line the Dhaka-Chittagong roadside as it passesthrough Bhatiary. The highest-value reused products from end-of-life ships, however,are engineered goods, including axles and crankshafts, gearboxes and driveshafts,compressors, generators, and motors.

Vol. 88 No. 1 2012

49

TER

RITO

RIA

LA

GG

LOM

ERAT

ION

INB

AN

GLA

DESH

The key point about the engineered products emanating from the ships on the foreshoreis that they were originally high-quality goods. All these items were manufactured byengineering firms like BMW, Daihatsu, Caterpillar, and Yanmar that are acknowledgedleading firms in their product sectors. Durability and reliability are key qualities forengineering products. So when a ship arrives on the foreshore of Sitakunda approximately25 years after its date of building, it is highly likely that its power systems will berecovered for another economic life in Bangladesh. In this case, used high-quality goodsare seen as better than new inferior-quality ones. That this is so is not only a reflection ofthe qualities of the engineered products themselves but is also indicative of Bangladesh’slack of an indigenous engineering industry. This situation has meant that Bangladesh hasbeen heavily reliant on high-cost technological imports to obtain such products. As withsecondary steel production, then, it is the gap between Bangladesh’s inability to produceits own engineered products and the high cost of such imports that has provided anopportunity for scavenging firms.

Like the kernel activity within the agglomeration, the reconditioned-goods sector is alsodifferentiated. The key distinction is between firms at the top end of the market, producingreconditioned goods for export, and the majority of firms producing purely for the domesticmarket.The sector is primarily located in the area of the CDA market, south of Bhatiary (seeFigure 1), and is widely known as a source of cheap but reliable secondhand goods.Working to international technical and labor standards, one company sells some of itsproducts back to the global market.At the other end of the spectrum, other firms reconditiongoods to sell in two domestic markets. The first is the market of the growing mass ofmiddle-class consumers, for whom standby generators provide a means to guarantee theirpersonal energy supply, a key attraction in a country where power outages are a feature ofeveryday life. The second source of demand is from business startups, particularly but notexclusively in the garment and shoe sectors. Startups and SMEs that produce cheapgarments and shoes for the burgeoning Bangladeshi consumer market, source all theirpower-related equipment from generators to cabling, except for the sewing machines andcutting equipment, from the reconditioned engineering products sector. As these SMEsgrow, so does their capacity to purchase new power equipment, but for startup businesses,the ready availability of cheap reconditioned power equipment is a key means by which theybecome established and expand.7 In a sense, then, reconditioned engineered products thatare sourced from the ship-breaking industry may be said to power entrepreneurial dreams,not only in the wider Chittagong area, but all over Bangladesh.

Remanufactured FurnitureThe furniture businesses of Bhatiary provide a paradigmatic case of industrial sym-

biosis grounded in remanufacturing. Starting in the late 1980s from a handful of pioneerbusinesses in the Fouzderhat area, the sector emerged organically, taking advantage ofmaterials that were originally seen by ship breakers as waste, to be cast aside and dumpedin the water. As one interviewee stated in 2008, “It is believed that [name] started thisbusiness in the Fouzderhat/Bhatiary area. . . . He is considered the first pioneer of thebusiness. When he started this business, most of the people thought he had gone mad andwas wasting his money on unnecessary things. Later on, many found this business veryprofitable and started following the pioneer.”

7 As one industry supplier suggested, these products sell because, while new industrial-sized generatorscommand a current price of 4 million taka, a reconditioned product with a further life of about 15 years canbe purchased for approximately 800,000 taka. The current exchange rate is U.S.$1 = 70.56 taka. Anothersupplier indicated that electrical cabling salvaged from ships retails at about half the price of new cabling.

ECONOMIC GEOGRAPHY

50

Survey work suggests that the sector is now comprised of 72 firms, in agglomerationsin the Kadum Rasal area, along the Dhaka-Chittagong highway, in Bangla Bazar, BanuBazar, Madambibr Hat, and BM Gate (see Figure 1). The growth of the sector displaysaffinities with numerous industrial districts in the South, particularly Africa, wherelow-skill, low-capital entry barriers make furniture an obvious sector for entrepreneurialactivity but one that is characterized by “survival competences” (McCormick 1999;Kaplinsky, Morris, and Readman 2002; Scott 2002, 2006; Murphy 2006, 2007). Yet, inBhatiary, there are clear lines of differentiation among firms, as well as key points ofdifference that distinguish this furniture agglomeration from others in the South.

The success of the pioneers can be seen in that the ship breakers now hold competitiveopen and box tender auctions for furniture, fixtures and fittings, and “boards” from eachvessel that is beached on the foreshore. In terms of business differentiation, the keydistinction is again about the cascade of streams of materials from those businesses thatcan raise the capital to bid in an auction to those that cannot. Raising capital is entirelydependent on credit facilities with local banks, while bidding also requires the capacity tolock up working capital for the approximately three weeks it takes to complete the auctionprocess and the employment of specialist ship surveyors. Only the largest and mostsuccessful furniture businesses can therefore bid in an auction.All such firms operate botha wholesale and a retail side of their businesses, thereby controlling the supply of boardsto the other furniture businesses.8

A single ship auction typically comprises 30 percent furniture and 70 percent “boards.”Items that are already in the form of furniture are refurbished and/or reconstituted, eitherto convert them from their previous design to furniture that is suited to the domesticsetting or to mask the patina of (over)use. However, it is the boards that are the primarymaterials input for the furniture sector. One ship typically would yield 60,000–80,000boards, which are then further classified by types of materials— plywood, partex board,particle board, and asbestos board. These materials constitute the basis for the remanu-facturing of a vast range of domestic furniture for the mass market of the lower middleclass in Bangladesh.9

With 72 firms in a sector in which there were just a handful as recently as 15 years ago,competition is fierce, and margins and profitability are tight, especially at the lower end ofthe market. This situation has created pressures about safeguarding an individual firm’slabor supply. In large part, the supply of labor has been obtained through a reliance onmigrant labor drawn from rural areas—quite the opposite of the argument that agglom-erations produce a shared pool of skilled labor and pooled knowledge founded onrelationships of trust. Anxieties about the labor supply and control of knowledge abound,as one interviewee stated in 2008: “I don’t want to share my designs with others—and Ipay higher than any other shop so that my carpenters stay with me for longer. If I mixedmy carpenters up with others my rivals may take away my workers by paying them higher

8 Data from the interviews suggest that approximately 25 percent to 33 percent of the boards from each shipare sold on to other furniture businesses, with the remainder being held as materials feedstock for the firmthat is successful at the auction.

9 Responses to the interviews indicate that at the top end of the market, firms use boards from the ships andFormica imported from India to produce cheap copies of upmarket Bangladeshi furniture brands that areaimed at the upper middle class. At the other end of the market, newer entrants to the sector rely onproducing a smaller range of relatively standard goods, including computer tables, beds, drawers, andwardrobes. Leading-edge firms sell both finished goods and boards as wholesale and retail through bothlocal and national distribution networks that span the major Bangladeshi cities. Those that are dependent onsuch firms for their raw materials supply chain sell predominantly to the local domestic market ofChittagong and the surrounding region.

Vol. 88 No. 1 2012

51

TER

RITO

RIA

LA

GG

LOM

ERAT

ION

INB

AN

GLA

DESH

wages than me.” For some businesses, therefore, too much concentration in certain partsof the Bhatiary area is seen to risk labor poaching and labor mobility, hence the clustersin a number of discrete areas that spatially segment labor markets.

A final point to note is that the majority of furniture businesses are family firms,although the leading-edge firms in the sector are partnerships. The dominance of familyfirms resonates strongly with previous research on industrial districts in South Asia(Nadvi 1999 on the surgical instrument cluster of Sialket, Pakistan; Knorringa 1999 onthe shoe agglomeration of Agra, India; and Chari 2000 on the cotton knitwear agglom-eration of Tiruppur, India). However, although furniture is an entry-level business inSitakunda-Bhatiary, many family firms moved into this business from another, havingseen the success of the pioneer furniture businesses. At the same time, it is not uncommonfor different relatives within an extended family to specialize in related ship-breakingbusinesses.10

SummarySitakunda-Bhatiary is an instance of an organically evolved secondary processing

cluster. Its kernel activity is ship breaking and secondary steel production from recoveredferrous scrap. A range of allied scavenging activities based on the wastes and by-productsof ship breaking have evolved over 20 years. These activities have constituted a networkof symbiotic exchanges that binds firms and sectors but ultimately stem from onedecomposing activity: the flow of ships to the foreshore and their breaking up.

At the macrolevel, it is no coincidence that a secondary processing complex on thescale of Sitakunda-Bhatiary should have emerged in Bangladesh. A minimal list ofcontributory factors would include the failure of state-run steel production to keep pacewith demand in a context of growth and modernization; the legacy of little or nodomestically based capital goods manufacturing capacity until recently; a weak balanceof trade with the other South Asian and East Asian economies, particularly India, China,and Korea; and a weak international currency. All this created the conditions in whicha secondary steel industry based on ship breaking could emerge and in which suchan industry could gain legitimacy. At the microlevel, the agglomeration of Sitakunda-Bhatiary rests on the conjuncture of a highly entrepreneurial culture with lax or nonex-istent regulatory controls in relation to the environment and waste management.

But symbiosis in Sitakunda-Bhatiary is more than material and economic; it is socialtoo. In Sitakunda-Bhatiary, making money from ship breaking and all its myriad wastesand by-products or using its by-products to manufacture something else underpinsinnovation and business dreams, startups and growth. It is in the Marshallian sense ofindustrial atmosphere, “in the air” (Amin and Thrift 1992). You smell it in Sitakunda-Bhatiary and the wider Chittagong area. It is what people aspire to. In the words ofone business entrepreneur who was interviewed: “The young generation [in Sitakunda]grow up keeping in mind that one day they will be big ship breakers or ship-breakingtraders. If they manage to collect a small capital, they just start a ship-breaking business.. . . Local banks are just waiting to give loans because this business is the quickest and themost profitable at the moment.” This comment signals that emulation (and its localfinancing) underpins agglomeration. But this is not a close mental distance founded in

10 Examples from our data include family-run firms that were formerly in the grocery trade, the rickshawbusiness, and the kitchen-bathroom business. An indication of the range of within-family businessportfolios is provided by one family with specializations in furniture, waste oil, and scrap, each managedby a son or uncle.

ECONOMIC GEOGRAPHY

52

cooperation. Rather, as demonstrated by the remanufactured furniture sector, it is often ahighly competitive, largely family-based form of entrepreneurship, based on price com-petition and low levels of interfirm trust (contra Hewes and Lyons 2008). A moot questiontherefore concerns the economic sustainability of this sector. At one end of the furnituresector is the capacity of emulation for self-combustion through the decline in profitabilitythat accompanies “ruinous” competition and, at the other, a question about the leadingfirms’ capacity to transcend a brand that positions its products as cheap, expendablecopies but that are remanufactured, at least in part, from the deeply problematic materialby-product of asbestos board. More profound still is the vulnerability of the entire cluster.This is a double vulnerability that impinges directly on the kernel activity. On the onehand, there is the risk that the feedstock of end-of-life ships will dry up; a consequence ofpressure by NGOs for the closure of a “dirty” industry. On the other hand, the pressureson secondary steel production consequent upon a resurgent primary steel sector pose akey challenge. The prospect of no ships, however, does not just leave the RRMs lookingfor an alternative supply of scrap steel; it also impinges on the supply side of allscavenging activities, exposing the economic precariousness of the entire agglomeration.More broadly, the case shows that when decomposition and scavenging activities prevailin a local economy, their success risks erasing the original niche and suggests thatupgrading may annihilate the very conditions of their emergence.

Learning from the South?Paying attention to places such as Sitakunda-Bhatiary in the South has implications for

academic understanding of industrial symbiosis. Recognizing the specificity of theparticular case, the intention is not to generalize from Sitakunda-Bhatiary to the entireSouth but, rather, to draw out points for future research. We make six points.

First, the recovery of secondary materials creates wastes and potential by-products asan integral part of the process of recovering materials, as any fabrication process does.Just as with primary processing, these materials can be turned into inputs and exchangesfor other industries, provided that they are either cheaper than primary raw materials orallow savings to be made on the costs of waste management. That much has beenacknowledged by work that has recognized the significance of decomposition and scav-enging activities. That symbiosis based on secondary processing has emerged in thedense, agglomerative form that it has in Sitakunda-Bhatiary, then, is precisely becausesuch activities provide cheap raw materials under conditions in which the costs of wastemanagement are negligible.

Second, rethinking symbiosis through secondary processing in the South shows that theliterature on industrial symbiosis needs to develop a more sophisticated sense of thespatialities of globalization. The field of industrial symbiosis is dominated by accountsbased on the scalar relations of geographic proximity, focusing on island economies,industrial districts, and the policy concept of the EIP.This is to neglect globalization and tooverlook that symbiosis also occurs at the global scale with movements of materials andprocessing just as complex as primary production (Kellow 1999). Symbiosis at the globalscale is particularly important with respect to the recovery and recycling of materials, forit is the countries of the South, China, and India where these activities are concentrated.Although physical symbiotic exchanges occur in relations of geographic proximity in thesecountries, they are embedded in, and dependent on, global flows in postconsumption“waste.” The point here is to recognize that symbiosis is both global and local, that is, thatit is multiscalar, not a local counterposed to a global. Furthermore, certain places are keynodes in symbiosis at the global scale, being conduits in the global exchange of secondhand

Vol. 88 No. 1 2012

53

TER

RITO

RIA

LA

GG

LOM

ERAT

ION

INB

AN

GLA

DESH

goods, wastes, and their material transformation. Sitakunda-Bhatiary is one such place, itstransformations based on the international trade in scrap metal; another similar place isGuangdong (China). Likewise, Lusaka (Zambia) and Kandla and Panipat (India) arecritical conduits in the international trade in secondhand clothing and the recycling of woolfibers (Tranberg Hansen 2000; Norris 2010).There is an entire economic geography to thisexchange, located in the South, which remains largely invisible and demands investigation.

Third, places like Sitakunda-Bhatiary show that symbiosis does not depend on anteriorterritorial agglomerations. Instead, symbiosis can emerge through pioneer entrepreneurialactivities and their emulation. As such, symbiosis can transcend the level of the singleplant exchanges emphasized in the literature and well beyond the benchmark of threeplants and two exchanges identified by Chertow (2007). Instead, entire sectors can emergeon the back of initial interfirm exchanges that are based on wastes and by-products of therecovery of ferrous metals for recycling.

Fourth, the diversity of sectors in Sitakunda-Bhatiary that are dependent for theirfeedstock on end-of-life ships and their breaking indicates the importance of the com-plexity of the foundational end-of-life product. Ships are a complex end-of-life good. Acorollary is that they have a more varied potential as future products than do other formsof postconsumer waste, such as packaging, paper, or even plastic. Fifth, the reach ofsymbiosis in Sitakunda-Bhatiary is closely related to the governing of waste materials inthis part of the world. Leaving to one side the not-inconsiderable issue that the shipsthemselves are in legal breach of international waste-shipment regulations, the wastes ofthe ferrous materials recovery process have been economized in these ways because theywere not sequestered and managed as wastes. That this agglomeration has emerged andwastes have morphed to become resources is because Bangladesh has been largelyunfettered by environmental regulation—whether that applies to ferrous metals or, andmore problematically, asbestos. This process, by which dumped and discarded wastes arerecovered and turned into new products, creating whole new industries, has its echoes inthe environmental history of 19th-century industrial capitalism in northern Europe.Comparable examples would include coal tar, formerly discarded into river courses butthen used to manufacture dyes and caustic soda (Desrochers 2002).

Sixth, symbiosis, at least as it is represented in relation to contemporary northernEuropean and North American examples, is seen as clean and green production and as anenvironmental good that works to minimize waste through closing materials loops.To turnattention to symbiosis in secondary processing in the South shows that symbiosis inconditions of minimal or nonexistent environmental regulation is far from clean and a longway from environmental understandings of green. Make no mistake: symbiosis in this partof the world now comes with huge environmental and labor costs, as it did in NorthAmericaand northern Europe in the 19th century. It is very messy. But what this mess does is to keepmaterials circulating in economies; something that classifying and sequestering materialsas wastes does not allow. As such, and notwithstanding the negatives that go alongside it,minimal conditions of environmental regulation may be an easier route to symbiosis and theminimization of wastes. The crux of this issue, however, has to be that not all end-of-lifegoods and their constituent materials are good to economize in these ways—asbestosboards turned into domestic furniture being a case in point.

ConclusionsThe global flows of wastes are a neglected field in the study of globalization and the

economies of the South, where the emphasis—almost without exception—is on primaryprocessing, global product markets, upgrading and its relation to global buyers, and

ECONOMIC GEOGRAPHY

54

immiserizing growth (Knorringa 1999; Schmitz & Nadvi 1999; Kaplinsky et al. 2002). Insuch accounts, goods dematerialize in their consumption. What actually happens is rathergrittier: stuff isexchanged throughglobalproductmarkets indiscardedgoods thatoftenmoveit toward the South (Lepawsky & McNabb 2010). There it provides paradigmatic cases ofhow, why, and where wastes can be revalued through economic activities that are grounded inrecycling, reuse, and remanufacturing but in conditions of lax environmental and laborregulation. Thus, these complexes have much to say in the debate over industrial symbiosis,which emphasizes its clean and green credentials within discourses of ecomodernization.

We close this article by making one further set of observations. At the macrolevel, itwould be easy to see the secondary processing complexes of the Global South as anotherform of immiserizing growth, in this particular instance, cementing Bangladesh’s posi-tion in “the race to the bottom.” In drawing attention to the glimmers of heterogeneitywithin this agglomeration, though, we suggest that things are not quite so simple. Whilethe race to the bottom is an appropriate epithet to affix to the shift of ship breaking fromthe West to East Asia and then South Asia, pursuing ever lower labor costs and environ-mental regulations, the economization of reuse within Sitakunda-Bhatiary looks differ-ent. Here, and at the microlevel, there is evidence that at least some businesses that drawon reconditioned engineered products, particularly those that use them in their productionprocesses and some of those involved in reconditioning activities per se, go on to upgradeboth their products and their production process. Although our research found oneinstance in which this was done to compete in global markets, the vast majority of firmsin Sitakunda-Bhatiary do so to enhance their competitiveness in domestic productmarkets. Furthermore, as we have shown (Gregson et al. 2010), consumers of the productsof the ship-breaking furniture sector purchase the products largely as part of an upwardtrajectory of social mobility. These points serve as a timely reminder that in globalization,not all consumer markets are global and not all products are produced for a global market(c.f. Parthasarathy and Aoyama 2006). Rather, alongside and as part of globalization,there are significant and growing consumer markets in certain countries in the South, ofwhich Bangladesh is an exemplar case. Whilst these consumers may lack the purchasingpower to buy global brands and thus are largely invisible as consumers in academicaccounts of globalization, they constitute a huge and important market for a parallel setof domestic-facing firms, also largely invisible in accounts of globalization. That theproducts of reuse and of remanufacturing have enabled this type of growth seems animportant counter to accounts that argue that growth in globalization in the South isinvariably immiserizing.

Amin, A., and Thrift, N. 1992. Neo-Marshallian nodes in global networks. InternationalJournal of Urban and Regional Research 16:571–87.

Basel Action Network. 2002. Exporting harm: The high tech trashing of Asia. Seattle: BaselAction Network.

——. 2005. The digital dump: Exporting re-use and abuse to Africa. Seattle: Basel ActionNetwork.

Begunbari collapse—36 dead. 2010. Daily Star (Bangladesh), June 3.Bensaude-Vincent, B., and Stengers, I. 1996. A history of chemistry. Cambridge Mass.:

Harvard University Press.BSRM. 2008. FAQs. Available: http://www.bsrm.com/faq.php

Vol. 88 No. 1 2012

55

TER

RITO

RIA

LA

GG

LOM

ERAT

ION

INB

AN

GLA

DESH

refe

renc

es

Chari, S. 2000. The agrarian origins of the knitwear industrial cluster in Tiruppur, India. WorldDevelopment 28:377–97.

Chertow, M. 2007. “Uncovering” industrial symbiosis. Journal of Industrial Ecology 11:11–30.Chertow, M.; Ashton, M.; and Espinosa, J. 2008. Industrial symbiosis in Puerto Rico: Environmen-

tally related agglomeration economies. Regional Studies 42:1299–312.Chertow, M., and Lombardi, R. 2005. Quantifying economic and environmental benefits of

co-located firms. Environmental Science and Technology 39:6535–41.Clapp, J. 2001 Toxic Exports: The transfer of hazardous wastes from rich to poor countries. Ithaca, N.Y.:

Cornell University Press.Coe, N.; Hess, M.; Yeung, HW-C.; Dicken, P.; and Henderson, J. 2004. Globalising regional

development: A global production networks perspective. Transactions of the Institute of BritishGeographers 29:468–84.

Coe, N.; Dicken, P.; and Hess, M. 2008. Global production networks: Realizing the potential.Journal of Economic Geography 8:271–95.

D’Amico, F.; Buleandra, M.; Velard, I.; and Tanase, I. 2007. Industrial ecology as “best availabletechnique”: A case study of the Italian district of Murano. Progress in Industrial Ecology 4:268–87.

Deschenes, P., and Chertow, M. 2004. An island approach to industrial ecology: Toward sustain-ability in the island context. Journal of Environmental Planning and Management 47:201–17.

Desrochers,P. 2000.Market processes and the closing of “industrial loops”: A historical approach.Journal of Industrial Ecology 4:29–43.

——. 2002. Industrial ecology and the rediscovery of inter-firm recycling linkages. Industrial andCorporate Change 11:1031–57.

——. 2004. Industrial symbiosis: The case for market coordination. Journal of Cleaner Production12:1099–110.

Desrochers, P., and Lam, K. 2007. “Business as usual” in the industrial age: (Relatively) lean, greenand eco-efficient? Electronic Journal of Sustainable Development 1:35–46.

Deutz, P., and Gibbs, D. 2008. Industrial ecology and regional development: Eco-industrialdevelopment as cluster policy. Regional Studies 42:1313–28.

Deutz, P., and Lyons, D. 2008. Editorial: Industrial symbiosis—An environmental perspective onregional development. Regional Studies 42:1295–98.

Douglas, M. 1966. Purity and danger: An analysis of pollution and taboo. London: Routledge & KeganPaul.

Ehrenfeld, J., and Gertler, N. 1997. Industrial ecology in practice: The evolution of interdepen-dence at Kalundborg. Journal of Industrial Ecology 1:67–79.

Erkman, S., and Ramaswamy,R. 2006. Industrial ecology: A new platform for developing countries.In Industrial ecology and spaces of innovation, ed. K. Green and S. Randles, 106–28. CheltenhamU.K.: Edward Elgar.

International Federation for Human Rights, YPSA (Young People Social Action), and NGOPlatform. 2008. Childbreaking yards: Child labour in the ship recycling industry in Bangladesh.Available at: http://www.fidh.org/IMG/pdf/bgukreport.pdf

Frosch, R., and Gallopoulos, N. 1989. Strategies for manufacturing. Scientific American 261:94–102.Geng, Y., and Côte, R. 2002. Scavengers and decomposers in an eco-industrial park. International

Journal of Sustainable Development and World Ecology 9:333–40.Gibbon, P. 2001. Upgrading primary production: A global commodity chain approach. World

Development 29:345–63.Gibbs, D. 2006. Prospects for an environmental economic geography: Linking ecological moder-

nization and regulationist approaches. Economic Geography 82:193–215.

ECONOMIC GEOGRAPHY

56

——. 2008. Industrial symbiosis and eco-industrial development: An introduction. GeographyCompass 2/4:1138–54.

Gibbs,D., and Deutz, P. 2005. Implementing industrial ecology? Planning for eco-industrial parks inthe USA. Geoforum 36:452–64.

Gibbs, D.; Deutz, P.; and Proctor, A. 2005. Industrial ecology and eco-industrial development.A new paradigm for local and regional development. Regional Studies 39:171–83.

Gille, Z. 2007. From the cult of waste to the trash heap of history: The politics of waste in socialist andpostsocialist Hungary. Bloomington: Indiana University Press.

Greenpeace/FIDH/YPSA. 2005. End of lifeships: The human costs of breaking ships. Availableat http://www.greenpeace.org/international/publications/en/reports/end-of-life-the-human-cost-of

Gregson, N.; Crang, M.; Ahamed, F.; Akter N.; and Ferdous, R. 2010. Following things of rubbishvalue: End-of-life ships, “chock-chocky” furniture and the Bangladeshi middle class consumer.Geoforum 41:846–54.

Hamner, B. 1997. Industrial ecology in East Asia. Journal of Industrial Ecology 1:6–8.Henderson, J.; Dicken, P.; Hess, M.; Coe, M.; and Yeung, HW-C. 2002. Global production

networks and the analysis of economic development. Review of International Political Economy9:436–64.

Hewes, H., and Lyons, D. 2008. The humanistic side of eco-industrial parks: Champions and therole of trust. Regional Studies 42:1329–42.

Johnson, J.; Pecquet,G.; and Taylor,L. 2007.Potential gains from trade in dirty industries:RevisitingLawrence Summers’ memo. Cato Journal 27:397–410.

Kaplinsky, R.; Morris, M.; and Readman, M. 2002. The globalisation of product markets andimmiserizing growth: Lessons from the South African furniture industry. World Development30:1159–77.

Kellow, A. 1999. Baptists and bootleggers? The Basel Convention and metals recycling trade.Agenda 6:29–38.

Kennedy, L. 1999. Cooperating for survival: Tannery pollution and joint action in the Palar Valley(India). World Development 27:1673–91.

Knorringa, P. 1999. Agra: An old cluster facing the new competition. World Development 27:1587–604.

Korhonen, J. 2001. Regional industrial ecology: Examples from regional economic systemsof forest industry and energy supply in Finland. Journal of Environmental Management 63:367–75.

——. 2002. Two paths to industrial ecology: Applying the product-based and geographicalapproaches. Journal of Environmental Planning and Management 45:39–57.

Lepawsky, J., and McNabb,C. 2010.Mapping international flows of electronic waste. The CanadianGeographer/ Le Géographie Canadien 54:177–95.

Lyons, D.; Rice, M.; and Wachal, R. 2009. Circuits of scrap: Closed loop industrial ecosystems andthe geography of US international recyclable material flows 1995–2005. Geographical Journal175:286–300.

McCormick, D. 1999. African enterprise clusters and industrialisation: Theory and reality. WorldDevelopment 27:1531–51.

Mirata, M., and Emtairah, T. 2005. Industrial symbiosis networks and the contribution of environ-mental innovation: The case of the Landskrona industrial symbiosis programme. Journal ofCleaner Production 13:993–1002.

Murphy, J. 2006. The socio-spatial dynamics of creativity and production in Tanzanian industry:Urban furniture manufacturers in a liberalizing economy. Environment and Planning A 38:1863–82.

Vol. 88 No. 1 2012

57

TER

RITO

RIA

LA

GG

LOM

ERAT

ION

INB

AN

GLA

DESH

——. 2007. The challenge of upgrading in African industries: Socio-spatial factors and the urbanenvironment in Mwanza, Tanzania. World Development 35:1754–78.

Murphy, J. 2008. Economic geographies of the Global South: missed opportunities and promisingintersections with development studies. Geography Compass 2/3: 851–73.

Nadvi, K. 1999. Shifting ties: Social networks in the surgical industry cluster of Sialkot, Pakistan.Development and Change 30:141–75.

Norris, L. 2010. Recycling Indian clothing: Global contexts of reuse and value. Bloomington: Universityof Indiana Press.

Park, H., and Won, J. 2007. Ulsan Industrial Park. Journal of Industrial Ecology 11:11–13.Parthasarathy, B., and Aoyama, B. 2006. From software services to R&D services: Local entre-

preneurship in the software industry in Bangalore, India.Environment and Planning A 38:1269–85.Pollard, J.; McEwan, C.; Laurie, N.; and Stenning, A. 2009. Economic geography under postcolonial

scrutiny. Transactions of the Institute of British Geographers 34:137–42.Rock,M.; Angel,D.; and Feridhanuesetyawan, T. 1999. Industrial ecology and clean development in

East Asia. Journal of Industrial Ecology 3:29–42.Schmitz, H., and Nadvi, K. 1999. Clustering and industrialisation: Introduction. World Development

27:1503–14.Schwartz, E., and Steininger K. 1997. Implementing nature’s lesson: The industrial recycling

network enhancing regional development. Journal of Cleaner Production 5:47–56.Scott, A. 2002. Regional push: Towards a geography of development and growth in low- and

middle-income countries. Third World Quarterly 23:137–61.——. 2006. The changing global geography of low-technology, labor-intensive industry: Clothing,

footwear and furniture. World Development 34:1517–36.Shi, H.; Moriguichi, Y.; and Yang, J. 2002. Industrial ecology in China, Part I: Research. Journal of

Industrial Ecology 6:7–11.Shi, H.; Chertow, M.; and Song, Y. 2010. Developing country experience with eco-industrial parks:

A case study of the Tianjin Economic-Technological Development Area in China. Journal ofCleaner Production 18:191–9.

6 storey building tilts at Begunbari. 2010. Daily Star (Bangladesh), June 4.Steelmaker up for expansion as demand rises. 2010. Daily Star (Bangladesh), August 22.Tranberg Hansen, K. 2000. Salaula: The world of second-hand clothing and Zambia. Chicago:

University of Chicago Press.Van Beers, D.; Corder, G.; Bossilkov, A.; and van Berkel, R. 2007. Industrial symbiosis in the

Australian mining industry: The cases of Kwinana and Gladstone. Journal of Industrial Ecology11:55–72.

Van Beukering, P., and Duralappah, A. 1998. The economic and environmental impact of waste-paper trade and recycling in India: a material balance approach. Journal of Industrial Ecology 2:23–42.

Yeung HW-C., and Lin, G. 2003. Theorizing economic geographies of Asia. Economic Geography79:107–28.

Young People Social Action (YPSA). 2005. Workers in ship breaking industries: A baseline survey ofChittagong (Bangladesh). Chittagong, Bangladesh: YPSA.

Zhu, Q.; Lowe, E.; Wei, Y.; and Barnes, D. 2007. Industrial symbiosis in China: A case study of theGuitang group. Journal of Industrial Ecology 11:31–42.

ECONOMIC GEOGRAPHY

58