Digital Technology And Env

Futures 36 (2004) 903–920 www.elsevier.com/locate/futures

De-materialising and re-materialising: digital technologies and the environment Frans Berkhout , Julia Hertin SPRU-Science and Technology Policy Research, University of Sussex, Falmer, Brighton, East Sussex BN1 9RF, UK

Abstract Drawing on recent literature on the environmental impact of information and communication technologies (ICTs) and the Internet, this paper identifies three main types of effects: direct impacts of the production and use of ICTs on the environment (resource use and pollution related to the production of infrastructure and devices, electricity consumption of hardware, electronic waste disposal); indirect impacts related to the effect of ICTs on production processes, products and distribution systems (de-materialisation, substitution of information goods for material goods, and substitution of communication at a distance for travel); and structural/behavioural impacts, mainly through the stimulation of structural change and growth in the economy by ICTs, and through impacts on life styles and value systems. This paper argues that the diffusion and use of ICTs are leading to both positive and negative environmental impacts. However, because the effects of ICTs on economic activity are pervasive, their impacts on the environment are difficult to trace and measure. The paper argues for a need to move beyond the dichotomy between pessimism and optimism demonstrated in much of the emerging literature. Instead the relationship must be recognised as complex, interdependent, deeply uncertain and scale-dependent. # 2004 Elsevier Ltd. All rights reserved. Keywords: Information and communication technologies; Environment; De-materialisation; Innovation

1. Introduction The relationship between the development, diffusion and use of information and communications technologies (ICTs) and the broader social goal of sustainability is 

Corresponding author. Tel.: +44-1273-877130; fax: +44-1273-685865. E-mail address: [email protected] (F. Berkhout).

0016-3287/$ - see front matter # 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.futures.2004.01.003

904

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

not well understood. Large and growing efforts have been made to understand the emergence of ICTs as a ‘general purpose technology’, and to analyse their impacts on the economy and on society. Until recently, these studies have been concerned primarily with a specific range of impacts attributed to the technology: changes in work organisation and worker productivity; structural change in the economy, and ultimately, prospects for economic growth. The rise of communication networks has also produced a flowering of research concerned with the process of ‘globalisation’ in its many facets; the reshaping of the political authority of the nation state; questions of identity, group formation and new forms of civil activism; problems about the governance of the Internet and its impacts on democracy; the ‘digital divide’ and so on [7]. Impacts on environmental sustainability have only in the last few years become the subject of systematic research.1 This is partly because compared to more traditional industries, like energy, transport and manufacturing, the overall environmental burdens linked to the production and use of ICTs have appeared to be small. But research into the links between ICTs and environmental sustainability is also limited by a lack of systematic analysis and reliable data. Although statistical sources about the diffusion and the economic contribution of ICTs exist (e.g. Ref. [39]), they naturally lag behind a fast changing reality, and medium and long-term forecasts have proven to be unreliable. Given that we cannot have a complete picture of the development of ICTs, discussion of environmental impacts arising from these future activities will also necessarily be circumscribed. This is particularly relevant because numerous studies and life cycle assessments have shown that the net environmental effects of specific technologies, such as video conferencing, online retailing and electronic directories largely depend on the specific circumstances of deployment and use (see for example, Refs. [31,43,53]). Due to these limitations, it is too early to come to a definitive assessment of the environmental impacts of information technologies. This paper aims to map out the most important linkages between ICTs and impacts on the environment, providing a qualitative assessment of the risks and opportunities stemming from the information revolution. The paper begins by describing three ICT–environment links: direct; indirect; and structural/behavioural. Each of these effects is discussed in turn in the following three sections. We conclude with an assessment of the relationship between ICTs and environmental sustainability and a reflection on effective and forward-looking policy strategies. 2. De-materialising or re-materialising? The current debate about digital technologies and the environment is characterised by a stark contrast between optimistic and pessimistic assessments. To 1

See for example recent special issues of the Journal of Greener Management International (winter 2000) and Industrial Ecology (spring 2002).

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

905

some, environmental effects of ICTs appear to be exclusively positive because ‘information’ is generally considered to be distinct from material and energy, and to act as a substitute for the use of material resources. One of the Internet’s early gurus, Kevin Kelly proclaimed ‘‘The net wins. . .displacement in our economy of materials by information. . .displacement of mass with bits. . .take away the mass of radiator, axle and drive shaft by substituting networked chips. . .’’ [29]. Coyle argued that the digital world is about as close to weightless as is possible [10]. According to Nicholas Negroponte ‘‘. . .the information superhighway is about the global movement of weightless bits at the speed of light’’ [37]. If economic value is increasingly generated by the intangible and the weightless, then perhaps postindustrial economies are further liberating themselves from the awkward imperatives of extracting, manipulating and drawing value from material goods and services. In an October 1996 speech, Alan Greenspan, Chairman of the US Federal Reserve Board said: . . .Virtually unimaginable a half century ago was the extent to which concepts and ideas would substitute for physical resources and human brawn in the production of goods and services. ... Fiber-optics has replaced huge tonnages of copper wire, and advances in architectural and engineering design have made possible the construction of buildings with much greater floor space but significantly less physical material than the buildings erected just after World War II. Accordingly, while the weight of current economic output is probably only modestly higher than it was a half century ago, value added, adjusted for price change, has risen well over threefold’ (cited in Ref. [44]). For others, however, ICTs represent a case of unsustainable production and consumption. These authors stress the wide-ranging negative environmental impacts of computers and other hardware, especially the fast-increasing waste stream of electrical and electronic equipment. Short innovation cycles in hardware and software have led to a high turnover of devices. New software applications demand more speed and larger memory, leading to rapid obsolescence of standard computer platforms. Under current conditions, it is usually less expensive and more convenient to replace PCs, printers or mobile phones to accommodate new software than to upgrade old devices. Many critics see the fast spread of ICTs as a symptom for modern materialism and short-term consumerism, rather than as a sign of de-materialisation. This paper seeks to move beyond this dichotomy, arguing that there is a complex and uncertain relationship between information technologies and environmental sustainability. To bring some clarity to the claims that are made, it is useful to draw distinctions between direct, indirect and structural/behavioural effects of ICTs and e-commerce (see Table 1). (1) Direct effects of ICTs are predominantly negative and stem from the production, use and disposal of hardware such as computers, screens, network

906

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

Table 1 ICT impacts on the environment Positive impacts Direct effects of ICT

Indirect effects of ICT

Structural and behavioural effects of ICT

Negative impacts

Environmental impacts of production, use and disposal of ICTs (e.g. electronic waste) Improved efficiency, de-materialisation Falling prices for resource inputs, and virtualisation, detection and, moni- proliferation of ‘intelligent’ devices, partial substitution (e.g. e-shopping toring of environmental change (e.g. intelligent logistics, electronic directories, as well as private shopping trips) environmental sensors) Structural and life style transitions (e.g. Stimulating growth and re-materialgrowth of ‘light’ industries, green conisation (e.g. growth of long-distance sumerism) travel)

cables, etc. They are not greatly different from the environmental effects of many other products, but pose a number of specific problems in terms of both resource use, emissions and waste management. (2) Indirect effects of ICTs are expected to be largely positive. Two main causes are put forward. First, ICTs contribute to increasing the efficiency of production processes, for example through computer-aided design (CAD), higher production speed and scale, and greater control. Second, it is expected that a wide range of products and services (insurance, access to information, music, etc.) may become fully de-materialised. On the other hand, many of the digital goods and services will come ‘in addition to’ existing goods and services, adding environmental pressures. (3) Structural and behavioural effects of ICTs relate to more fundamental processes of change and may have both positive and negative outcomes. On the positive side, the spread of ICTs contributes to a shift from an industrial economy towards a service economy, which will tend to have lower levels of resource and energy use (at the point of use). ICTs can also support behavioural changes in favour of a ‘greening’ of products and services. On the negative side, efficiency gains could be offset by a so-called ‘rebound effect’, often observed in the transport and energy sector. This occurs when efficiency gains (directly or indirectly) stimulate growing demand that balances (or even over-compensates for) positive environmental effects. The main drawback of this classification is that it does not do justice to the role of information in the shaping of knowledge and awareness about environmental issues, or in enabling responses to recognised problems. ICTs have been central to major developments in understanding environmental processes, and in providing the means for investigating and mitigating the impacts of human activities. In this sense, the ICT–environment link needs to be seen as two-way, including both impacts that may be traced back to the use of ICTs, while also providing the means for enabling a better understanding of those impacts.

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

907

3. Direct environmental effects of ICTs Although important data gaps and fast changing technology do not allow a detailed assessment, it is widely recognised that production, use and disposal of ICTs have become a serious environmental policy issue. Current ICT systems rely on a variety of different products with heterogeneous environmental characteristics [30]. These include product groups such as personal computers, network servers, mobile phones, cables and satellites, and peripheral devices (screens, printers, scanners, etc.). Although these devices have different environmental profiles, ICT hardware tends to have a short life span, uses considerable amounts of electricity, and incorporates significant quantities of materials that are harmful to the environment. 3.1. Manufacture Most ICT products consist of a large number of different components, e.g. micro-chips, semiconductors, printed circuit boards, liquid crystal displays, and batteries. The manufacture of many of these components has important environmental effects. The production of semiconductors, for example, causes significant air emissions (acid fumes, volatile organic compounds and doping gases), water emissions (solvents, cleaning solutions, acids, metals) and wastes (silicon, solvents) [16]. Overall, the manufacture of ICT equipment causes a variety of detrimental environmental effects, related to energy consumption, water use and emissions of acids, metals, volatile organic compounds, chlorinated solvents and other substances.2 Matthews shows that the ecological damage caused by computer manufacture is growing across several environmental domains (waste, energy, greenhouse gas emissions) [35]. With regard to waste, it has been found that 98% of the material used in PC production goes into the waste stream and only 2% becomes part of the product [20]. ICTs are produced through global supply chains. A typical personal computer contains 1500–2000 components sourced from around the world, and typically transported by air. The complexity and scale of the global electronics sector means that the aggregate environmental impacts of these supply chains are large, including major transport energy and greenhouse gas emissions. Electronic commerce may have the effect both of ‘shrinking’ the supply chain (by enabling greater control and reducing the number of steps) thus bringing environmental gains, but it may also have the opposite effect, allowing supply chains to become more ‘globalised’, segmented and environmentally inefficient. 3.2. Use ICT devices consume electricity. A typical medium-sized PC consumes about 1 kW h in an average working day. It has been estimated that an average office computer is in operation for more than 2000 h a year (20% in active mode and 2

A recent study estimated that the production of a PC generated 130 kg of greenhouse gases and 30 kg of total waste [2].

908

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

80% in sleep mode) [53]. There has been a major controversy over the total use of electricity by office equipment, with some estimates suggesting that 8–13% of US electricity demand would be accounted for by office, telecommunications and network equipment [24]. More recent studies have concluded that office equipment currently consumes some 2–3% of US power, although this could grow to up to 5% over the next two decades [28,45]. But residential energy use is also growing through the integration of ICTs into products and even commodities. These include advanced television equipment, mobile phones, handheld personal digital assistants and many other information appliances. For instance, a UK study projected that set-top boxes (integrated receiver decoders) could add some 6000 GW h of power demand in the UK by 2010.3 Standby losses associated with other electronic consumer goods in the home are another major source of power use. These devices account for 5–15% of residential energy use in industrialised societies [25]. 3.3. Disposal Accurate data about wastes from electronic equipment are difficult to find because national waste statistics are not standardised and because ICT-related waste is not always accounted for separately. A report by the European Environment Agency collated statistics for five electrical and electronic product groups (fridges, PCs, televisions, photo copiers and toasters) and estimated that in these categories alone the European Union generates almost 1.25 million tonnes of waste per year [14]. In 1998, electrical and electronic waste was estimated to be 6 million tonnes and thought to be increasing by at least 3–5% per year [8]. Much of this growth is accounted for by new classes of products entering the market (mobile phones, PDAs, DVDs, etc.). An average PC weighs 29.6 kg and consists of metal (43.7%), plastics (23.3), electronic components (17.3%), and glass (15%). The main area of concern for the disposal of ICT hardware is its content of metals and other toxic or harmful materials. Particularly problematic are brominated flame retardants, solders, batteries, semiconductors, plastic stabilisers and screens [14,51]. Because current product design does not usually allow the separation and recycling of these metals without further treatment, only a very small proportion of ICT hardware is currently recycled. Two recent EU directives aim to address this problem (directive 2002/95/EC on the restriction of the use of certain hazardous substances in electrical and electronic equipment and directive 2002/96/EC on waste electrical and electronic equipment). In summary, ICTs are now integrated in many ordinary consumer and commodity products. Many of these devices and components are energy-consuming and have short life cycles and are composed of toxic materials. Already the wastes generated and energy used by ICTs are significant in industrialised countries, and this seems likely to increase in future. 3

Projection by the UK Market Transformation Programme (www.mtprog.com).

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

909

4. Indirect effects: understanding of the environment, resource productivity and de-materialisation Information and communication technologies now play a fundamental part in our understanding of the environment, as well as in the development, production and distribution of products and services. ICTs are essential to the measurement, modelling and communication of environmental processes, while also having a major role in improving the productivity of labour, capital and natural resources. Although driven by the need to reduce costs, the optimisation of processes through ICTs has often benefited the environment. This is not only because of improvements in resource efficiency enabled through greater process control, but also because efficient processes tend to be relatively less polluting. Preliminary evidence to support this argument at the macro-economic scale has been produced by several US studies [32,44]. Romm argued that in 1997 and 1998, aggregate energy productivity in the US economy improved by more than 3%, reversing a trend of slow declines in the previous decade. This he saw as being an effect of the diffusion of ICTs into the economy. Similarly, Kelly [29] argued that while US gross domestic product grew by 8% between 1996 and 1998, energy use grew by less than 1%, again this was put down in part to the impact of ICTs. 4.1. Information effects Sensors, monitors and other instruments collect information about the distribution of resources (for example, seismic surveys for oil and gas), the environmental impacts of economic activities (e.g. emissions concentrations at production sites) and the state of the environment (e.g. stratospheric ozone concentrations) at all levels from the planetary to the microscopic. There has been a vast increase in environmental detection efforts by governments and international organisations over the past 50 years. Understanding of environmental problems is mediated almost universally by instruments linked to ICT networks and scientific interpretation. As environmental problems are recognised at a broader scale (biodiversity, climate change), the importance of sensing technologies in measuring the state of the environment and in enabling an assessment of environmental health is likely to increase. Perceptions of the environment are becoming increasingly separated from direct experience and more dependent on instruments and computer-based models and analysis. The processes by which sensing technologies and data analysis develop are therefore crucial in shaping public and policy debates about the environment. 4.2. Production processes ICTs enable the simulation of complex production systems to test and review costs, material use, and environmental emissions of design options. Once in operation, sensors linked through communication networks and digital controls ensure efficient and flexible operation of more integrated facilities. Modern production systems can have tens of thousands of individual microprocessors embedded in

910

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

them, controlling valves, measuring temperatures, sensing the colour of fluids, and performing other tasks [29]. These devices are a critical part of all modern production processes, improving reliability, quality, safety and reducing waste. Today up to 40% of the value of a new manufacturing plant is accounted for by digital and computer-controlled systems. Precise control is fundamental to environmental efficiency and management of production. But such improvements are not new. They have been achieved consistently in industry since computers were first introduced into manufacturing over 30 years ago. Historical trends suggest that the introduction of devices like electronic sensors and controls lead to incremental productivity improvements of between 1 and 2% per year in mature process industries like paper and plastics [5]. 4.3. Design and operation of products and services Although the specific effects of design software and simulation tools are difficult to disentangle from other developments (e.g. the use of new materials) they are recognised as reducing waste in production and operation, and to generate more efficient products. CAD has been a feature of design for products and processes for some 20 years, and is now widespread. Simple products such as packaging have been substantially de-materialised through better design and through the application of software-based environmental appraisal such as life cycle assessment. Well-known examples of this design-driven de-materialisation include reductions in the weight of drinks cans. Aluminium used in drinks cans fell by some 50% during the 1990s as a result of design improvements, enabled by CAD, testing and quality control [3]. Many complex products (cars, consumer durables) contain microprocessor controls enabling them to respond to changing conditions of performance and environmental conditions. Embedded controls improve the functionality of the product, but also influence their environmental performance. All consumer goods, and in future also less complex products, will embody microprocessors to control their behaviour under variable conditions. Amongst other functions, they control emissions to the environment, and the use of heat, water and other inputs. Other positive examples come from the service industries, many of whose environmental impacts result from the energy used in buildings. Together, residential and commercial buildings use more than two-thirds of all electricity in industrialised countries, and information systems are key to improving the efficiency of heating, cooling and lighting. Simple control systems can provide 10–15% savings with current technologies, and greater gains should be possible in future. Substitutions of information for materials and energy have also been proposed as leading to environmental benefits. The argument is made that electronic devices will increase the ease with which information can be assimilated and communicated, in the process reducing or eliminating the materiality of information-intensive goods and services. Substitutions may be partial (de-materialisation) or complete (virtualisation).

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

911

4.3.1. Print Paper-based catalogues, directories, dictionaries, encyclopaedias, newspapers and magazines are all believed to be under competitive pressure from cheaper and more easily retrieved and updated sources of information on the Internet. The Boston Consulting Group (BCG) [4] produced a comprehensive review of impacts of electronic media on paper use in the US, finding that in most cases there would be pressures to reduce demand. Online newspaper and magazines are cheaper because they do not have to support three of the major expenses of the print business: newsprint, distribution and printing costs. On the other hand, these higher costs will, to some extent, be counterbalanced by cost-savings introduced through the increasing digitalisation of production, design and print processes which contributed to a proliferation of print media in the 1980s and 1990s. These factors are expected to play a major role in the decline in paper-based catalogues, directories and information-based books. BCG predicted that substitution of on-screen information for printed matter will outweigh the contrary trend—the greater use of office papers as a result of the broader diffusion into office and home of digital printers. 4.3.2. Audio/photo/video The consumption of music offers significant potential for virtualisation. Downloading music in digital format (for example MP3) is a popular use of the Internet. Although a fast Internet connection and sufficient memory capacity are required, a large number of people exchange digital music files via specialised Internet sites. Digital music can be played with freely available software and a standard home PC. Once again, a critical unanswered question is how far these new technologies will substitute for existing media (and therefore generate environmental benefits), or stimulate new demand for devices such as MP3 players, mini disk players or CD writers. Digital cameras record and store photographic images in electronic form. The substitution of chemically processed photographic images by digitally processed images for amateur, professional and medical/industrial uses may avoid some of the environmental impacts of film manufacturing and processing [9]. A recent study suggested that digitisation of radiography would lead to savings in X-ray film, and reductions in the use of developer and fixer [26]. 4.4. Transport and distribution Markets depend on information and exchange, and ICTs have proven powerful enabling technologies in extending and deepening markets. The globalisation of supply chains has contributed to the growth in global trade and so to the growth in demand for transport and distribution services. On the other hand, information technologies have also contributed to improving the efficiency of distribution. They are opening up opportunities for replacing both goods and passenger transport through technologies such as digital telephony, e-mail, Internet, and videoconferencing. Overall, it appears that the evidence supports the view that there are both substitution (digital communication replacing physical movement) and

912

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

complementarity (digital communication stimulating new physical movement) effects.4 Four key linkages between ICTs and transport are discussed below. 4.4.1. Supply chain efficiency Increased co-ordination enabled by information technology has helped to make supply chains better organised and more efficient [23]. In the US business, the average stock turn5 was 8 a year in 1995 but reached 13.2 in 2000. The net effect of this was reduced and faster-moving inventories of goods, translating into significant reductions in storage requirements (leading to energy savings) and less obsolete stock. Market pressures may lead to further integration of supply chains. For instance, electronic business-to-business procurement exchanges reduce costs, inventory and wastage and increase utilisation of capacity. This may create environmental benefits through resource productivity gains. However, some efficiency gains could be offset by demand for high speed and just-in-time delivery by road, tending to increase congestion and reduce vehicle utilisation rates. Furthermore, the increased ordering options which e-business provides to both business and consumers will result in more geographically-extended supply patterns, and therefore higher transport intensity of goods. 4.4.2. Changing seller–buyer relationships Mass customisation—production systems that manufacture products for individual consumers—was proposed some 10 years ago as an organising principle of business, just as mass production by assembly line was the organising business principle under Fordism [41]. In the mass production model, consumer demand is predicted and a limited number of product lines are developed to meet that demand. Through branding and advertising, consumers are encouraged to buy goods that have already been manufactured and stocked. ICTs and the Internet make mass customisation possible because they enable a rich exchange of information between companies and individual customers. This allows goods to be produced that match customer demands much more specifically. Several environmental benefits have been proposed for mass customisation. First, production can be matched more tightly to customer demand, reducing the energy associated with inventories and warehousing. Second, products and services can be designed to fit the needs of the consumer more precisely, reducing waste and improving environmental efficiency. Third, the producer may be transformed into a provider of services to the final consumer ownership of the products delivering services remaining with the service provider. This may presage the development of ‘closed loop’ product and service systems in which producers have incentives to reduce resource inputs and environmental burden of goods in use, rather than simply in production [50]. 4

See Ref. [36] for a discussion of these concepts. Average stock turn is a measure of the rate at which products are sold through retail outlets. A higher stock turn denotes a more efficient operation. 5

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

913

4.4.3. Retailing Shopping accounts for a substantial part of personal travel. Complete replacement of car-based shopping trips by Internet ordering and van-based delivery could reduce the distance travelled—by 70–80% according to one simulation [6]. However, e-commerce only accounts for a very small part of retail turnover [38] and complete substitution is unlikely. Many forecasters predict that retailing will come to place greater emphasis on the ‘experience’ of shopping, with low-margin staples sourced using Internet ordering and home delivery. The limited survey evidence of consumer responses to home shopping has found that they reported less use of cars [6]. The transport implications of online retailing will depend on the rates of growth that are experienced, and on the details of logistics and transport systems. Before there is more empirical evidence about how customers respond to home shopping, and the delivery patterns used by suppliers, claims made about likely net impacts are difficult to substantiate. 4.4.4. Work organisation The major hypothesised environmental benefits of telecommuting and teleworking (T&T) are the reduced need for office space and a reduced number of journeys to work. A British study has estimated that telework has the potential to make up to 5–15% of car journeys redundant [42]. While savings may be achieved in some cases, there is also evidence that employees may be becoming more peripatetic in some service sectors. In general, studies have shown that the potential for environmental savings from T&T are likely to be small, or may be negative [18]. Evidence from an evaluation of a UK teleworking scheme suggested that 47% of people participating had reduced their commuting, with only 6% increasing their travel [21]. However, offsetting changes in terms of equipment, heating and other services in the home need to be taken into account. Equally controversial is the question of how far video conferencing technology will reduce the demand for business travel, especially by air [1]. There is recent evidence that use of teleconferencing is becoming more widespread, but this may be as a complement to business-related travel, rather than as a substitute. 4.5. Summary: realising efficiency gains While significant potential for ICT-induced environmental savings has been identified, real environmental gains may be hard to capture for a number of reasons. First, the scope for de-materialisation and virtualisation may be limited. Substitution of information for material goods and services (digital music, e-books) may occur, but the number of goods and services where complete virtualisation can be achieved is limited. For most goods, IT is integrated in the design and delivery of a product or a service, and does not substitute for it. The ‘virtual economy’ needs to be seen as intimately linked to the real, material economy, just as it is embedded in the real, political world [46]. For instance, e-commerce is likely to depend on the evolution of faster, more flexible transport infrastructures with greater capacity, while virtualisation in audio sales will depend on IPR problems being resolved.

914

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

Second, resource productivity gains tend to be slow in many technological systems. ICTs have become integrated into production processes and new products over a long period so there is good evidence on the rates and direction of environmental performance improvements that may be expected in the future. Evidence suggests a continuation along well-established evolutionary trends, rather than abrupt or radical changes in the rate of improvement. Third, while ICT and Internet-enabled services may substitute for some ‘material’ goods and services, many new digital and ‘intelligent’ goods and services will be complementary to existing ones, especially during periods of transition from one technology to another. Just as the use of computers did lead not to a ‘paperless office’, incomplete substitution of old structures by new digital systems could result in additional environmental burdens. Similarly, the behaviour of e-shoppers is likely to contain surprises, with many choosing to do a part of their shopping online while still going to the stores to buy other goods.

5. Structural and behavioural effects of ICTs 5.1. Structural change and resource use At the macro-economic level, the notion of de-materialisation (or the ‘intensity of use’ hypothesis) was first discussed in the 1970s, and produced the ‘Environmental Kuznets Curve’ (EKC) hypothesis [12,15,48]. Broadly, this argues that during early industrialisation, economies use material resources (steel, cement, energy carriers) more intensively, until a threshold is reached after which structural changes in the economy (the decline of manufacturing and the rise of services) lead to progressively less-intensive materials use. Many authors expect ICTs to facilitate a de-coupling of economic growth and environmental damage. The idea of the ‘knowledge economy’ promotes the notion that economic value is created primarily through the manipulation of ideas, rather than the exploitation of energy and materials. More developed economies are seen as growing as a result of the more intelligent use of resources to produce greater value, rather than through the addition of new resources. As evidence of this shift, more than 40% of US investment in new equipment over the past decade has been in the form of information devices and infra-structures [19]. Information technologies contribute to a long-standing structural change in the economy away from materials-intensive activity and towards more service-based and information-intensive activities. This occurs through the growth of IT-related services (e.g. software development, Internet services, new advertising and marketing services, etc.), as well as through the growth of traditional services that have been transformed by the use of ICTs (e.g. financial services). However, service sectors are supported by material infrastructures and transactions, and are not as ‘clean’ and ‘weightless’ as is often assumed [33]. There is evidence that the rapid diffusion of ICTs speeds up structural change in the economy, and therefore contributes to incremental improvements in relative resource efficiency. But empirical evidence also shows that

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

915

in absolute terms these economies are still ‘materialising’ (i.e. getting more materials-intensive), especially when the environmental burdens associated with traded goods are included [34]. 5.2. The rebound effect The rebound effect has two main stimuli: falling prices and increased capacity. As a result of induced productivity improvements and the growing scope and specialisation of markets, the prices of raw materials and energy will tend to fall. Cost pressures, one of the main drivers of eco-efficiency improvements, fall and rates of improvement in resource efficiency should slow—at least in theory. The producers of materials and energy for which demand falls will undertake efforts to innovate new products and occupy new market niches. Because many materials have a potential for substitution (windows can be made from wood, plastic or aluminium for example), new markets for environmentally damaging materials may develop. The second mechanism is the creation of new demand. Many ICT applications allow a better management of time, money, labour, transport infrastructure, and so on, thus providing scope for new demand. Whether this new demand will offset de-materialisation effects stimulated by the widespread diffusion of ICTs depends largely on the collective choices of consumers. Will the money and time potentially freed through the use of ICTs be spent in the consumption of environmentally damaging goods and services, or through demand for ‘immaterial’ services? Anecdotal evidence suggests that the rebound effect is a real threat to incremental efficiency gains. Although many digital devices (PCs, cellular phones) have been substantially miniaturised and ‘de-materialised’, they have also increased the capacity of the final consumer to consume by reducing transaction costs. By enabling the integration of markets and driving down prices, these devices (and the production–consumption systems they operate within) provide new opportunities to access and consume both material and immaterial goods and services. In particular, the Internet increases the ‘reach’ of consumers, by extending choice of goods and the range of providers. One concrete example of re-materialisation is document access through the web, allowing for a printed paper copy in just a few mouse clicks. Some US teleworking studies have found that initial reductions in car travel are partially offset over time by stimulation of new driving [22]. Confounding evidence on energy intensities from Schipper and Grubb [47] suggests that ‘rebounds’ in energy use as a result of energy price falls ‘‘. . .may have taken back some of the overall savings, but most remain. . .’’. 5.3. Information and behavioural changes Finally, new information technologies are likely to have wider impact on social values, life styles and culture. These changes and their impact on the environment remain speculative and here we highlight a few examples. A number of authors have suggested that information technologies will enhance ecological transparency [27] and give a new dimension to the phenomenon of green consumerism [49]. If

916

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

information about the environmental aspects of products, brands and companies is more easily generated and more accessible through digital media, consumers can more easily act on their green preferences, turning them into powerful market signals. At the same time, citizen-consumers [49] can also be held to account if new information systems make the environmental impacts of individual consumer choices transparent. Already today, there are websites that enable individuals to calculate their personal carbon emissions or to assess energy efficiency potentials of their household. More sophisticated IT tools for environmental management and decision-making tools will doubtless be developed for both companies and consumers (see Ref. [52]). In principle, property rights could be allocated more efficiently (a carbon budget for households, for instance) providing the basis for informational and economic incentives to induce changes in behaviour of producers and consumers [17].

6. Conclusions The difficulty of establishing a direct link between ICTs and environmental or resource productivity mirrors the ‘productivity paradox’—the longstanding debate about the apparent absence of evidence of a relationship between investment in ICTs and labour productivity [11]. At first sight, ICTs appear to have broadly positive impacts on environmental sustainability, especially through structural change in the economy and increased efficiency in production and logistics. We have sought to show that such positive impacts need to be balanced against a range of countervailing effects, including direct impacts of electronic devices and in compensating behavioural changes that may be enabled by ICTs. Economic, social, institutional and cognitive barriers are likely to prevent technical potentials for resource efficiency from being fully exploited. Over the long term, the net environmental effect of the information revolution will depend on the balance between these ‘de-materialising’ effects, and the counter ‘re-materialising’ influences of economic growth and complementarity effects. The digital economy is embedded in the material and economic world and physical infrastructures, both its own (cables, computers, networks) and those that it coordinates and motivates. Whether intelligent systems, products and services will reduce the environmental impact of the economy depends largely on how they are designed, used and supported by transport, energy and other systems. ICTs do not necessarily lead to a more environmentally-sound future, but they offer new opportunities to develop more sustainable solutions. As in many other industrial– environmental domains, the role of policy and regulation will be crucial if these opportunities are to be captured. To know where opportunities lie and how they can be exploited most effectively, we must improve our knowledge of the specific links between ICT and the environment, both empirically and in terms of understanding paths of causation and interaction. The following section sets out some future research priorities.

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

917

7. Research outlook Much of the literature on the link between ICTs and the environment is highly specific, identifying potential for environmental savings on the basis of isolated case studies. It often involves simple modelling of potential direct effects of the diffusion of a certain technology or technology-related phenomenon (if X is substituted with Y, this will lead to Z savings). Although this micro-level research is valuable, it has a number of inherent limitations and stands in the way of general findings. We agree with Mokhtarian’s [36] observation that single-application studies tend to overestimate the potential for environmental savings through ICTs. Referring specifically to the literature on telecommunications and travel she observes that: ‘. . .although direct, short-term studies of impact focusing on a single application (such as telecommuting) have often found substitution effects, such studies are incomplete and likely to miss the more subtle, indirect, and long-term complementarity effects that are typically observed in more comprehensive studies’ (pp. 53–54). In particular, micro-level studies are unable to handle feedback effects due to responses in market demand and consumer behaviour. They are inevitably based on assumptions, for example about rates of diffusion, willingness-to-pay, consumer choices and so on. In an area as technologically immature as digital communication, these kinds of assumptions remain hard to validate and test [40]. For example, medium- and long-term forecasts about the Internet and e-commerce have been shown to differ greatly from each other (sometimes by several orders of magnitude) and they have usually proven to be wrong. A more systematic analysis of the full complexity of the relationship over the longer term is now emerging, but it faces a number of methodological challenges. The empirical evidence on which to draw conclusions is still thin. This is at least in part because of the difficulty in producing explanatory models of technological change linked to models of environmental change. The development of a technology—especially a general purpose technology—is shaped by, and itself shapes, a wide spectrum of social, technical and environmental systems. As a consequence of this interdependent relationship, the development of reference scenarios (how would the sector have developed without the widespread adoption of ICTs?) and the establishment of simple causal relationships (e.g. between a technical innovation and an environmental impact) is likely to be problematic. Given these challenges, we propose three main research approaches: macro-level qualitative studies developing and assessing a range of alternative development pathways mapping the long-term relationship between environmental sustainability and ICTs (see Ref. [13]); macro-level quantitative studies of ICT-related and environmental indices and the relationship between them (e.g. investment in R&D, ICT-related GDP, energy and resource efficiency) (see Refs. [31,32]); and meso-level analysis of sectoral trends and ICT effects that look across a range of sustainability indicators to measure effects.

918

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

References [1] P. Arnfalk, Information and communications technologies and business travel, in: J. Park, N. Roome (Eds.), The ecology of the new economy, Greenleaf, Sheffield, 2000, pp. 183–192. [2] Atlantic Consulting, LCA study of the product group personal computers in the EU Ecolabel Scheme, London, 1998. [3] A to Z of Materials.com (Azom), Aluminum cans: History, development and Market, 2003. Available from (http://www.azom.com/details.asp?ArticleID=1483, accessed 21.09.03). [4] BCG (Boston Consulting Group). Paper and the electronic media, Boston Consulting Group, Boston, September 1999. [5] F. Berkhout, M. Almemark, L. Lindfors, K. Holm-Mueller, I. Hanhoff-Stemping, A. Smith, H. Stripple, Sustainability, Competitiveness and Technical Change (SCOTCH): Final Report to the European Commission, SPRU, Falmer, 2000. [6] S. Cairns, Home Delivery of Electronic Shopping: The Environmental Consequences, ESRC Transport Studies Unit, London University College, 1999. [7] M. Castells, The Information Age: Economy, Society and Culture, Blackwell, Cambridge, MA, 1996. [8] CEC (Commission of the European Communities), Proposal for a Directive of the European Parliament and of the Council on Waste Electrical and Electronic Equipment and on the restriction of the use of certain hazardous substances in electrical and electronic equipment COM (2000)347, Brussels, 2000. [9] N. Cohen, Greening the Internet: Ten ways e-commerce could affect the environment and what we can do. iMP: The Magazine on Information Impacts, October 1999. Available from (www.cisp.org/ imp/). [10] D. Coyle, The Weightless World: Strategies for Managing the Digital Economy, MIT Press, Boston, MA, 1998. [11] P. David, Digital Technology and the Productivity Paradox: After Ten Years, What Has Been Learnt? Paper Presented at Understanding the Digital Economy: Data, Tools and Research, US Department of Commerce, Washington DC, 25–26 May 1999. [12] S. de Bruyn, H. Opschoor, Developments in the throughput-income relationship: theoretical and empirical observations, Ecological Economics 20(3) (1997) 255–268. [13] M. Eames, F. Berkhout, J. Hertin, R. Hawkins, E-topia? Contextual Scenarios for Digital Futures, Final Report, SPRU, Falmer, 2000. [14] EEA (European Environment Agency), Waste from Electrical and Electronic Equipment (WEEE), Copenhagen, EEA, 2003. [15] P. Ekins, The Kuznets curve for the environment and economic growth: examining the evidence, Environment and Planning A 29 (1997) 805–830. [16] EPA, Profile of the electronics and computer industry, Office of Compliance Sector Notebook Project EPA/310-R-95-002, 1995. [17] D.C. Esty, Information Age Environmental Protection, OECD Observer, May 2001. [18] Forum for the Future, Making the Net Work: Steps Towards a Sustainable Networked World, London, 2003. [19] D. Henry, P. Buckley, G. Gill, S. Cooke, J. Dumagan, D. Pastore, The Emerging Digital Economy II. Secretariat on Electronic Commerce, US Department of Commerce, Washington, DC, 1999. [20] L.M. Hilty, T.F. Ruddy, Towards a sustainable information society, Informatik-Informatique 4 (2000) 2–9. [21] P. Hopkinson, P. James, Virtual traffic—will e-commerce mean less transport and more sustainable logistics? Report for ‘Digital Futures’ Project, University of Bradford and UKCEED, 2000. [22] P. Hopkinson, P. James, J. Selwyn, The Environmental and Social Impact of Teleworking, Paper Presented to the sixth European Conference on Telework and New Ways of Working, Aarhus, Denmark, 1999. [23] P. Hopkinson, P. James, T. Maruyama, Teleworking at BT: the economic, social and environmental impacts of its workabout scheme, SUSTEL Working Paper, 2002. Available from (www.wfh.co.uk/wfh/consultancy/implementation/Teleworkingreport_2002.pdf).

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

919

[24] P. Huber, M.P. Mills, Dig more coal—the PCs are coming, Forbes 31 May 1999. Available from (http://www.forbes.com/forbes/1999/0531/6311070a.html, accessed 21.09.03). [25] IEA (International Energy Agency), IEA Standby Power Initiative: an International Action to Improve the Energy Efficiency of Electrical Equipment while Either Off or on Standby, 2001. Available from (www.iea.org/standby/ accessed 21.09.03). [26] IST (Information Society Technologies), Case Studies of the Information Society and Sustainable Development, CEC, Brussels, 2000. [27] M. Kamppinen, P. Malaska, M. Wilenius, Citizenship and ecological modernization in the information society, Futures 33(3–4) (2001) 219–223. [28] K. Kawamoto, J. Koomey, B. Nordman, R.A. Brown, M.A. Piette, M. Ting, A. Meier, Electricity Used by Office Equipment and Network Equipment in the US.: Detailed Report and Appendices, LBNL-45917, Lawrence Berkeley National Laboratory, Berkeley, CA, 2001. [29] K. Kelly, New Rules for the New Economy, Wired, Issue 5.09, September 1997. Available from (http://www.wired.com/wired/5.09/newrules.html?pg=1&topic=, accessed 21.09.03). [30] H. Kelly, Information Technology and the Environment: Choices and Opportunities, iMP: The Magazine on Information Impacts, October 1999. Available from (www.cisp.org/imp/). [31] M. Kuhndt, J. von Geibler, V. Tuerk, S. Moll, K.O. Schallaboeck, S. Steger, Virtual de-materialisation: ebusiness and factor x, Final Report of the Digital Europe Project, Wuppertal Institute, Wuppertal, 2003. [32] J.A. Laitner, Information technology and US energy consumption: energy hog, productivity tool, or both? Journal of Industrial Ecology 6(2) (2002) 13–24. [33] C. Leadbeater, Mind over matter: greening the new economy, Green Alliance, London, 2000. [34] E. Matthews, The Weight of Nations: Material Outflows from Industrial Economies, World Resources Institute, Washington DC, 2000. [35] H.S. Matthews, The Environmental Implications of the Growth of the Information and Communications Technology Sector, paper for Environment Directorate, OECD, Paris, 2001. [36] P.L. Mokhtarian, Telecommunications and travel, Journal of Industrial Ecology 6(2) (2002) 43–57. [37] N. Negroponte, Being Digital, Vintage Books, New York, 1996. [38] OECD (Organisation for Economic Co-operation and Development), Information Technology Outlook 2000, OECD, Paris, 2000. [39] OECD (Organisation for Economic Co-operation and Development). OECD Information Technology Outlook 2002, Paris, 2002. [40] J. Park, N. Roome (Eds.), The Ecology of the New Economy: Sustainable Transformation of Global Information, Communication, and Electronics Industries, Greenleaf, Sheffield, 2002. [41] J.B. Pine, Mass Customisation: The New Frontier in Business Competitiveness, HBS Press, Boston, MA, 1993. [42] POST (Parliamentary Office of Science and Technology), Review of Telework in Britain, HMSO, London, 1995. [43] M. Reichling, T. Otto, The environmental impact of the new economy. Deutsche Telekom, telecommunication services and the sustainable future, in: J. Park, N. Roome (Eds.), The Ecology of the New Economy, Greenleaf, Sheffield, 2000, pp. 119–129. [44] J. Romm, The Internet Economy and Global Warming, Center for Energy and Climate Solutions, Washington, 1999. [45] K.W. Roth, F. Goldstein, J. Kleinman (Eds.), Energy Consumption by office and telecommunications equipment in commercial buildings, Energy consumption by office and telecommunications equipment in commercial buildings, Energy Consumptions Baseline, vol. 1. Arthur D. Little, Cambridge, MA, 2002. [46] S. Sassen, Digital networks and the state: some governance questions, Theory, Culture and Society 17(4) (2000) 19–33. [47] L. Schipper, M. Grubb, On the rebound? Feedback between energy intensities and energy use in IEA countries, Energy Policy 28(6–7) (2000) 367–388. [48] T.M. Seldon, D. Song, Environmental quality and development: is there a Kuznets curve for air pollution emissions? Journal of Environmental Economics and Management 27 (1995) 147–163.

920

F. Berkhout, J. Hertin / Futures 36 (2004) 903–920

[49] G. Spaargaren, B. van Vliet, Lifestyles, consumption and the environment: the ecological modernisation of domestic consumption, Environmental Politics 9(1) (2000) 50–76. [50] W. Stahel, From products to services: selling performance instead of goods, IPTS Journal 27 (1998). (Available from (www.jvc.es/iptsreport/vol27/english/STA1E276.htm)). [51] M. van Drunen, A.A. Olsthoorn, Environmental impacts of the computer industry, IVM-VU, Working Paper W99/06, Amsterdam, 1999. [52] I. Wernick, Environmental knowledge management, Journal of Industrial Ecology 6 (2) (2002) 7–9. [53] M. Zurkirch, I. Reichart, Environmental impacts of telecommunication services: two life-cycle analysis studies, Greener Management International 32 (2000) 70–88.