Product flow analysis using trade statistics and consumer survey data: a case study of mobile phones in Australia
Introduction
Electronic products such as mobile phones, laptops, TVs and tablets utilise the physical properties of highly specialised and geochemically scarce metals to function. These metals (e.g. Ag, Au, In and many others) must be mined and refined, sometimes at significant environmental and social cost, to be integrated into these products. Yet many electronic products are wasted at the end of their useful lives, appearing in landfill or in some cases illegally exported to developing nations, fostering further economic, environmental and social problems for these countries (Balde et al., 2015). The recovery of valuable components in electronic products has attracted significant interest in recent years as a means to reducing these risks (e.g. Li et al., 2015, Pickren, 2014), particularly amidst growing environmental impacts and regulations facing the mining industry (e.g. Mudd, 2010). However, the economic recovery of metals from e-waste requires some understanding of the location, composition and volume of products available for future extraction, so that investments into recovery operations can be properly informed. Approximations and modelling are necessary to obtain such information in the absence of direct measurement and reporting.
The materials in electronic products, and indeed all metals in society, whether active in use or dormant and not yet disposed of, are known as ‘in-use stocks’. In-use stocks have been indirectly or directly estimated through various approaches and methodologies, each with different emphases, for example in input–output accounts, national capital accounts, life-cycle assessments (LCA) and material flow analyses (MFA) (Pauliuk et al., 2015). Within the MFA studies, there are two primary approaches by which in-use stocks of products or their contained materials have historically been estimated: top-down and bottom-up. The top-down approach essentially entails the collection and analysis of data on material inputs and outputs for a specified system. The difference between inflows and outflows (e.g. imports plus domestic production minus exports) over a specified time period can indicate in-use stocks via mass balance. The bottom-up approach entails the collection of data on the number of products/commodities within a given area and summing these to estimate the total in-use stocks.
Both the top-down and bottom-up approaches have advantages and disadvantages. For example, while bottom-up studies permit the spatial distribution of in-use stocks to be estimated, they are often temporally restricted to one year. Here, top-down studies can shed more light as they permit multi-year analyses and hence trends in stock accumulation, although they often rely on highly aggregated data that do not relate to specific products, and are further problematic to disaggregate spatially. These and other methodological uncertainties are described in numerous previous studies, e.g. (Chen and Graedel, 2015a, Gerst and Graedel, 2008, UNEP, 2010).
The multi-year analyses under the top-down approach are referred to as Dynamic MFA (DMFA) and have been applied to describe historical material flows and stocks of various metal resources. Several DMFA studies have projected possible future developments and related resource flows at both national and global levels based on scenarios. Muller et al. (2006) and Wang et al. (2007) focused on the anthropogenic iron and steel cycle, Daigo et al. (2007) estimated both the in-use and the total steel stock, which includes hibernating stock in Japan, and Reck et al. (2008) analysed the nickel stock and flows at the national and global scale. While the above studies have focussed on specific metals, other MFA studies have emerged which focus on the flows of specific products. Oguchi et al. (2008) analysed the circulation of major consumer durables in Japan, Harper (2008) analysed global flows of tungsten-containing products, and Chen and Graedel (2015b) estimated in-use stocks of 91 products in the United States.
Studies in MFA have additionally used monetary Input Output (IO) tables even for relatively small flows; an example is found in the work of Nakamura et al. (2007). IO analysis is one of the most widely used tools for describing economy-wide activities and their environmental implications (Suh, 2009). IO based MFA models, for example Waste IO-MFA, analyse the compositions of the materials or substances in products and scrap. Nakamura and colleagues provided several studies on IO based MFA (Nakamura et al., 2008, Nakamura et al., 2009, Ohno et al., 2014). Nakamura et al. (2014) also provided a worthy methodological framework, MaTrace model to enable visual tracking of the fate of materials whether accumulated in in-use stocks or dissipated in waste streams. In addition, Wang et al. (2013) provided a detailed overview of different IO models used for product flow analyses and e-waste estimation.
If elements of multiple MFA methodologies are applied to the same commodity under the same system boundaries, more can be revealed about the nature of that commodity. For example, both the spatial distribution of the commodity and potential trends over time in stock accumulation could be determined, and further the uncertainty associated with each method could be compared and interpreted. Very few studies have conducted multiple assessments, e.g. both top-down and bottom-up assessments of the same commodity or product, with Hirato et al. (2009) being a notable example. This is likely due to the time taken to conduct MFA studies, and given that regardless of the specific MFA method employed, almost all MFA studies must contend with a lack of up to date and spatially relevant data. Indeed, it is relatively accepted that the contribution offered by an MFA study is that it synthesises available data to characterise the flows of a new commodity, and/or to represent previously un-studied spatial/temporal aspects, but not necessarily that it employs raw data collection.
The limited data sources which are available for MFA studies can often be re-used through multiple generations of studies, which ultimately become less spatially and temporally relevant to the source data. For several electronic products including mobile phones, we have seen increases in value and utility, and considerable hoarding behaviour developed (ACMA, 2015, Read, 2015), which affects the in-use stocks and average lifespans of the products. There are therefore considerable uncertainties for future projections of e-waste volumes associated with using fixed product lifespan and distribution parameters (e.g. Weibull function) based on previous investigations. Furthermore, the limited number of studies currently used to inform in-use stock behaviour may provide source data that is spatially explicit (i.e. reflecting usage behaviour in a certain country), making them problematic to infer for other locations.
Empirically collected data, which reflects the system boundaries of the MFA study itself, can assist in reducing these uncertainties, and hence this study focuses on how such data can be integrated into multiple methods of product in-use stocks and flows estimation. In the following sections we describe this approach in detail and demonstrate its application with the case study of mobile phones in Australia.
Section snippets
Methodology description
Estimating (waste) electrical and electronic equipment ((W)EEE) circulation is a difficult task due to often low quality and incomplete data, meaning that multiple assumptions are required for input–output modelling (Wang et al., 2013). The annual sales of EEE (in monetary value and units) are usually well recorded through national and international systems and institutions (e.g. UN Comtrade database), while the information on in-use stocks and end-of-life (EoL) products is not directly
Case study – mobile phones in Australia
In this section, we demonstrate the application of the developed methodology to mobile phones in Australia. First, the existing trade statistics and consumer survey data are compiled to estimate the number of mobile phones in stock and at the end of life. Second, the average lifespan for a mobile phone is estimated by different methods, including the Leaching model, survey based approach, and Weibull distribution fitting. Third, the Weibull distribution parameters are used to reconstruct the
Conclusion
The use of empirical data in modelling the product in-use stocks and flows can help overcome inconsistency and reduce uncertainties attributed to a lack of official information sources and statistics. The key novelty of the developed methodology in this article is the combination of top-down and bottom-up approaches, based on trade statistics and consumer survey data respectively, to assess the product lifespan, in-use stocks and flows, including reconstructing the product age structure if this
Acknowledgements
The authors would like to acknowledge the support of Wealth from Waste Research Cluster, a collaborative program between Australia's CSIRO (Commonwealth Scientific Industrial Research Organisation), University of Technology Sydney, The University of Queensland, Swinburne University of Technology, Monash University and Yale University. We also would like to thank Dr Glen Corder, Professor Damien Giurco, and Dr Ruth Lane for valuable comments and ideas for this paper during the collaborative
References (33)
The environmental sustainability of mining in Australia: key mega-trends and looming constraints
Resour. Policy
(2010)- et al.
Product flow analysis of various consumer durables in Japan
Res. Conserv. Recycl.
(2008) - et al.
Estimation of end of life mobile phones generation: the case study of the Czech Republic
Waste Manag.
(2012) - et al.
Enhancing e-waste estimates: improving data quality by multivariate input–output analysis
Waste Manag.
(2013) Communications Report 2005–06
(2006)Communications Report 2014–15
(2015)- et al.
E-Waste Statistics: Guidelines on Classifications, Reporting and Indicators
(2015) - et al.
Improved alternatives for estimating in-use material stocks
Environ. Sci. Technol.
(2015) - et al.
In-use product stocks link manufactured capital to natural capital
Proc. Natl. Acad. Sci.
(2015) - et al.
Accounting for steel stock in Japan
ISIJ Int.
(2007)
In-use stocks of metals: status and implications
Environ. Sci. Technol.
A product-level approach to historical material flow analysis
J. Ind. Ecol.
In-use stock of steel estimated by top-down approach and bottom-up approach
ISIJ Int.
Power Automation Products and Other Electrical Equipment Manufacturing in Australia
“Control-Alt-Delete”: rebooting solutions for the e-waste problem
Environ. Sci. Technol.
Exploring the engine of anthropogenic iron cycles
Proc. Natl. Acad. Sci. U. S. A.
Cited by (34)
Evaluating the circularity and multi-lifecycle green profit of product family design
2023, Resources, Conservation and RecyclingDynamic material flow analysis of tantalum in the United States from 2002 to 2020
2023, Resources, Conservation and RecyclingCitation Excerpt :Lastly, incorporating lifetime distributions into our model allows us to estimate the amount of tantalum that is coming out of use at the EoL. However, we acknowledge that the lifespan of some products may differ significantly from one country to another or from year to year (e.g., Golev et al., 2016). In this study, we used average lifetime distribution estimates from various published studies focused on highly industrialized nations (e.g., Japan, Netherlands) and assume that these are generally representative of both the U.S. consumer market and the time period covered in our model (see Table S12 in Appendix B).
Estimating the generation of waste electrical and electronic equipment in organizations: The case of a Brazilian federal agency
2021, Cleaner Engineering and TechnologyEstimation of E-waste at micro level for reverse logistics: A case of Delhi
2021, Journal of Cleaner ProductionCitation Excerpt :Based on our results, the Gompertz curve is employed to predict the number of mobile phone subscribers for the past years. Based on past literature, we consider mobile subscriptions a proxy for growth/decline in the stock of mobile phones (Golev et al., 2016; Jang and Kim, 2010). Considering that the mobile phone adoption will follow an 'S’-curve, the curve fitting parameters were obtained using the Gompertz curve.
Electronic waste generation, economic values, distribution map, and possible recycling system in Indonesia
2021, Journal of Cleaner ProductionTowards sustainable resource management: identification and quantification of human actions that compromise the accessibility of metal resources
2021, Resources, Conservation and Recycling