Resources, Conservation & Recycling 220 (2025) 108363

Contents lists available at ScienceDirect

Resources, Conservation & Recycling

Global stagnation and regional variations in steel recycling

A R T I C L E I N F O A B S T R A C T

Keywords: Steel is widely regarded as the world’s most recycled material, yet little information exists on just how circular
Circular economy the steel industry actually is, both globally and nationally. This study traces iron and steel flows across the top 30
Resource efficiency steel-producing countries, showing that the share of recycled iron inputs into global steelmaking has stagnated at
Decarbonization
~30 % over the past two decades. Although waste management practices have improved, the rapid growth of
Material flow analysis
Heavy industry
global in-use steel stock has prevented progress in making steel flows more circular. While some countries show
higher recycled content than others, ranging from <10 % to >90 %, this does not necessarily reflect leadership in
recycling practices. Rather, high circularity in some places is often supported by low circularity elsewhere
through scrap imports or ’offshoring’ the production of high-quality flat products. As long as global steel stock
continues to grow, improvements in local circularity do not necessarily lead to global progress.

1. Introduction analyses of iron and steel flows have been conducted for major econo­
mies since the 1990s (Chen and Graedel, 2012). Wang et al. (2007) were
“Steel is the most recycled material in the world and is fundamental among the first to provide a comprehensive global picture by mapping
to the circular economy”. This rather ubiquitous claim can be found steel flows across multiple countries and regions for the year 2000.
across a wide range of platforms (World Steel Association, 2023). Cullen et al. (2012) expanded this effort with a much more detailed
Indeed, thanks to its magnetic properties and structural integrity, scrap breakdown of intermediate and final steel products for the year 2008.
steel is relatively easy to separate and recycle compared to other ma­ Harvey (2022) later updated this work for the years 2011–2015 with an
terials (Reck and Graedel, 2012). Given the destructive impact of improved data handling algorithm. More recently, Gao et al. (2025)
ore-based steelmaking, which currently accounts for 5 % of global raw investigated iron and steel flows in 25 countries for the year 2017.
material extraction considering only iron ore (Plank et al., 2022) and 8 However, these studies present only isolated snapshots at different
% of CO2 emissions (Lei et al., 2023), the benefits of steel recycling are points in time and use varying system definitions. As a result, it is
indisputable. The key question, however, is: just how circular is the steel difficult to track long-term trends in steel recycling consistently at both
industry, both globally and nationally? While the question is a seem­ the global and country scales.
ingly simple one, finding the answer presents a considerable challenge. While time-series data exist for some major economies, such as China
Information on the use of recycled materials is not always available for (Song et al., 2020), the United States (Cooper et al., 2020), Japan
all countries, and even when it is, it tends to be defined differently across (Watari et al., 2023), and the European Union (Dworak and Fellner,
datasets (Graedel et al., 2011). Without consistent data and a clear 2021), meaningful cross-country comparisons require reconciling dif­
system definition, it is difficult, if not impossible, to properly monitor ferences in system definitions. Recent advances in economy-wide MFA,
the degree of progress toward a more circular steel industry and identify which tracks a range of materials across multiple countries, could help
effective intervention points. bridge this gap (Wiedenhofer et al., 2024). Yet, detailed iron and steel
Several pioneering initiatives have attempted to address this issue flows are often overshadowed by more bulky materials such as sand and
based on material flow analyses (MFA), which systematically trace the gravel. Collectively, none of the existing studies have systematically
flows of materials within a defined system (Graedel, 2019). Detailed explored the variations in recycling status between major

* Corresponding author.
E-mail address: [Link]@[Link] (T. Watari).

[Link]
Received 13 February 2025; Received in revised form 3 April 2025; Accepted 2 May 2025
Available online 11 May 2025
0921-3449/© 2025 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license ([Link]
T. Watari et al. Resources, Conservation & Recycling 220 (2025) 108363

steel-producing countries, nor have they investigated the causes of these 2.2. Recycling indicators
differences. Despite the growing interest in moving toward a more cir­
cular steel flow, the historical evolution and current state remain Progress toward a more circular steel industry can be measured using
unclear. a variety of indicators, each of which captures different aspects within a
The aim of this study is to fill this gap by systematically and system (Moraga et al., 2019). Among these indicators, two of the most
consistently mapping iron and steel flows around the world. Visualized widely used metrics are ’recycled content’ and ’end-of-life recycling
through Sankey diagrams, this representation functions as a map of the rate’, which measure the share of recycled materials used in total ma­
physical aspects of the economy (Müller et al., 2023). Our dataset and terial inputs and the share of materials in end-of-life waste that are
resulting diagrams differ from existing studies in three distinctive ways. functionally recycled, respectively (UNEP, 2011). While the ’end-of-life
First, the data cover the world’s top 30 crude steel-producing countries recycling rate’ often receives attention in policy discussions, we focus
from 2000 to 2019. These countries account for >95 % of current total more on ’recycled content’ for two key reasons. First, it is the reduction
crude steel production (World Steel Association, 2024); thus, the data of ore-based material inputs that matters, not just an increase in the use
capture nearly all steel production while maintaining national-level of recycled materials. In the context of the steel industry, the primary
resolution. Second, the data distinguish between long products (e.g., objective of improving circularity is to mitigate energy consumption
reinforcing bars, wire rods, and structural sections) and flat products (e. (Worrell and Carreon, 2017), CO2 emissions (Speizer et al., 2023), air
g., sheets, plates, and strips). This distinction is intended to reflect the pollution (Li and Hanaoka, 2022), and ecological impacts (Giljum et al.,
nature of recycling, as not all recycled steel is currently used for every 2022), all of which can only be realized when recycled materials
type of product. Currently, most recycled steel is processed into billets ‘replace’ virgin ore-based materials rather than ‘adding’ to the total
and blooms, with limited use in slab production due to contamination material inputs. Second, even perfect end-of-life recycling contributes
concerns (Daigo et al., 2021). Third, the diagrams are interactive and only marginally to reducing ore-based material inputs under high de­
open source. This feature facilitates visually intuitive and informative mand growth conditions (Haas et al., 2015). The logic is simple: as long
monitoring systems, moving beyond a series of data tables (Font Vivanco as total material inputs continue to grow, recycled materials can only
et al., 2019). Together, these features allow the resulting dataset and provide a portion of this input due to the inherent time lag between
Sankey diagrams to provide a basis for monitoring recycling practices in materials being produced and the materials becoming available as scrap.
a consistent format and to address the fundamental yet poorly under­ These factors underscore the importance of tracking ’recycled con­
stood question: How circular is the steel industry globally and tent’. The terms ’recycled content’ and ’recycling input rate’ are
nationally? sometimes used interchangeably (Chen, 2013) and sometimes have
different meanings (Espinoza and Soulier, 2018); they are also starting
2. Methods to be referred to with new terms, such as ’input socioeconomic cycling
rate’ (Wang et al., 2020) or ’circularity rate’ (Miatto et al., 2024). In this
2.1. Material flow analysis and Sankey diagrams study, we use the term ’recycled content’ in line with UNEP’s proposal
(UNEP, 2011) and define it as the share of recycled iron in total iron
Sankey diagrams for iron and steel flows are constructed based on a inputs to steelmaking. Meanwhile, we define the ’recycling input rate’ as
system definition that considers 13 product stages plus international a complementary indicator that represents the share of recycled iron in
trade (Fig. S1 in the Supplementary Information (SI)). The process be­ economy-wide material inputs, including non-iron materials in crude
gins with the extraction of iron ore and follows the journey of the iron as ore (Wiedenhofer et al., 2019). This economy-wide indicator provides a
it is transformed into various products. While iron and steel products broader view of circularity, reflecting all material flows and human
include several accompanying elements, such as carbon, phosphorus, impacts of mining, even though these flows do not always hold a direct
and alloy metals (e.g., nickel), this study tracks all the products as the economic value.
flow of iron. The primary data for this analysis are sourced from the Steel It is important to note that crude ore extraction is often not reported
Statistics Yearbook published by the World Steel Association (World by most countries as it is considered an intermediate product (Tuck
Steel Association, 2024), supplemented by statistics from the US et al., 2017). Reported figures may refer to either crude ore or usable ore
Geological Survey (U.S. Geological Survey, 2024). These sources pro­ that has been beneficiated. These reporting inconsistencies make it
vide production data on key materials, as well as data on international difficult to make fair cross-country comparisons. Therefore, this study
trade across the supply chain. The mass flows between processes are estimates the ’economy-wide recycling input rate’ only at the global
calculated using mass balance equations, grounded in the law of mass level. A similar limitation applies to the ’end-of-life recycling rate’, as
conservation. Two fundamental conditions govern these calculations: all reliable national statistics on end-of-life waste are not available. Thus,
masses must be non-negative, and process yields must not exceed 100 %. crude ore extraction and end-of-life waste are only represented in the
When discrepancies arise that violate these conditions, balancing flows global-level Sankey diagram, not in the national-level diagrams. Further
are introduced to correct the mass imbalances. In such cases, production details can be found in Section 6 of the SI.
data are always prioritized over other estimated data, such as interna­
tional trade, process yields, and iron content. A detailed description of 2.3. Underlying factors of regional variations
the mass balance equations, data reconciliation, and data sources is
provided in Sections 1–5 of the SI. To better understand the reasons behind regional variations in
Understanding the full picture of iron and steel flows can be chal­ recycled content, we analyze several indicators closely linked to recy­
lenging when flipping back and forth between numerous data tables. To cling practices. The choice of indicators is informed by existing studies
provide an intuitive grasp of the dataset, we develop interactive, open- that highlight key aspects of steel recycling. First, growing in-use stock
source Sankey diagrams using floWeaver (Lupton and Allwood, 2017). limits scrap availability and poses a barrier to closing the loop (Haas
These diagrams are accessible via a web application ([Link] et al., 2020). Second, industrialized countries are increasingly exporting
[Link]/) and a GitHub repository ([Link] scrap rather than recycling it domestically (Wang et al., 2022). Third,
[Link]/takumawatari/steel-flows-sankey), covering the last two de­ scrap steel is often downcycled into long products rather than flat
cades (2000 to 2019) for the world’s top 30 crude steel-producing products due to difficulties in contamination control (Daehn et al.,
countries. 2017).
Building on these insights, we focus on three indicators. The first and
most intuitive is domestic scrap availability, which is determined by
stock stabilization – the balance between iron entering the in-use phase

2
T. Watari et al. Resources, Conservation & Recycling 220 (2025) 108363

as goods and iron leaving as scrap. In theory, the more balanced these 3. Results
inflows and outflows are, the greater the potential for increasing the
recycled content (Pauliuk, 2018). The other two indicators capture the 3.1. Global iron and steel flows
role of international trade: a country’s import reliance for scrap steel and
finished flat products. Each is defined as the ratio of net imports to Despite steel’s inherent recyclability, just one-third of iron inputs to
domestic inputs, reflecting a country’s dependence on imports or its role steelmaking come from recycling (Fig. 1a). In 2019, the world’s top 30
as exporters. steel-producing countries required 1849 million metric tons (Mt) of iron
Beyond such simple country comparisons, we trace how the recycled inputs for steelmaking, of which 606 Mt was derived from recycled
content of different steel products evolves along global supply chains to scrap. This gives a global recycled content of 33 % – a level comparable
explore the role of international trade in regional recycled content. This to other major metals such as aluminum (International Aluminium
is done by expanding the trade dimension of our dataset using bilateral Institute, 2023), copper (Loibl and Tercero, 2021), zinc (Rostek et al.,
physical trade data. The expanded dataset functions as a multi-regional 2022), and tin (Bradley et al., 2024), when direct reuse and re-melting
physical input-output table, which enables us to explicitly track steel are included.
products passing through multiple countries during their lifecycle However, focusing solely on iron flows overlooks the broader ma­
(Wieland et al., 2022). A further detailed explanation is provided in terial needs of steelmaking associated with mining. Extracting iron ore
Sections 7–9 of the SI. involves mining crude ore, which contains both iron and non-metallic

Fig. 1. Iron and steel flows combined for the group of 30 top steel-producing countries. (a) Sankey diagram illustrating global iron and steel flows in 2019. (b) Scale-
related indicators showing trends in extraction, steelmaking, fabrication, and end-of-life waste from 2000 to 2019. (c) Recycling-related indicators showing trends in
economy-wide recycling input rate, recycled content, stock stabilization, and end-of-life recycling rate from 2000 to 2019. Boundary inputs and outputs in panel (a)
include crude ore extraction, international trade, and losses (e.g., gangue, tailings, slag, and landfill). Note that while iron and steel products include several
accompanying elements, this study tracks only the flow of iron content for all products. The exception is crude ore, which includes the mass of oxygen and gangue.
For simplicity and mass balance consistency, non-iron content is represented as a boundary-bound output immediately after mining, even though this may not always
be an accurate description of the production process. All flows are shown to scale in Mt/year.

3
T. Watari et al. Resources, Conservation & Recycling 220 (2025) 108363

minerals (e.g., tailings) that must be separated out during processing. steel stocks. Over the past two decades, the rate at which steel has
When considering these unused material inputs, including all recycled flowed into the use phase has, almost without exception, consistently
scrap and crude ore extraction, total material inputs amounted to ~3550 outpaced the availability of steel for recycling at the end of product life
Mt. In this broader context, the economy-wide recycling input rate – the cycles. As a result, even as end-of-life recycling rates have improved, the
share of recycled iron in the total material inputs – drops to just 17 %. system cannot “close the loop” fast enough to keep up with the growing
The limited circularity of the global steel flow is not due to poor demand for steel.
waste management. In fact, 85 % of end-of-life scrap was collected and This decline in circularity at the steelmaking stage has been further
recycled in 2019. Instead, the key limiting factor is the growing in-use exacerbated by increasing extraction of lower-grade crude ore, which
steel stock – the steel embedded in products, buildings, and infrastruc­ requires beneficiation to increase iron content (Tuck et al., 2017). This
ture. As demand for steel stocks continues to grow, the steel inflows into additional processing has reduced the economy-wide recycling input
society far exceed the outflows of steel available for end-of-life recy­ rate, as the need for more extensive ore processing dilutes the impact of
cling. In 2019, for instance, inflows were 2.8 times higher than outflows, recycled iron inputs. Just as global stagnation has been observed in
meaning that even with perfect recycling with no losses, recycling alone energy efficiency, in part due to rising steel demand (Wang et al., 2021),
could only have met part of the demand. our data indicates that a similar trend is emerging in steel recycling:
Steel demand is growing faster than improvements in recycling effi­
ciency, which ultimately limits the recycled content.
3.2. Stagnation in global recycled content

This imbalance between inflows and outflows is not a short-term 3.3. Variations in regional recycled content
phenomenon but rather part of a long-term trend (Fig. 1b and c). Over
the past two decades, the global steel industry has made significant ef­ While the global recycled content has stagnated at around 30 %, a
forts to recycle more end-of-life scrap. Between 2000 and 2019, the end- closer look at the world’s top 30 steel-producing countries reveals sig­
of-life recycling rate improved from 65 % to 85 %, resulting in an in­ nificant regional variations (Fig. 2). In 2019, Thailand led the world
crease in iron inputs from recycled scrap from 283 Mt to 606 Mt. Yet, with 96 % recycled content, followed by Italy (72 %), Spain (69 %),
despite this progress, the global circularity of the steel industry at the Turkey (69 %), Indonesia (68 %), the US (67 %), and Poland (54 %).
input stage has stagnated. Both the economy-wide recycling input rate Only these seven have recycled content above 50 %. On the other end of
and the recycled content in steelmaking have declined, from 21 % to 17 the spectrum, the seven countries with the lowest recycled content – Iran
% and from 35 % to 33 %, respectively. (5 %), Saudi Arabia (6 %), Ukraine (24 %), India (25 %), The
The root cause of this stagnation is, again, the expansion of in-use Netherlands (25 %), China (27 %), and Austria (27 %) – all fall short of

Fig. 2. Ranking of recycled content among the world’s top 30 steel-producing countries in 2019. The labeled percentages indicate the share of recycled steel used in
steelmaking for each country.

4
T. Watari et al. Resources, Conservation & Recycling 220 (2025) 108363

30 %. rate of stock growth alone does not effectively explain regional varia­
But does a higher recycled content necessarily indicate leadership in tions in recycled content.
circular steel practices? For instance, is Thailand’s 96 % recycled con­ A second factor helps explain why many countries deviate from the
tent truly a reflection of superiority in recycling efforts? And conversely, diagonal: scrap trade. In countries with stable in-use stock, a significant
should a country with lower recycled content be seen as lagging behind? portion of collected scrap tends to be exported rather than recycled
To unpack these nuances, we need to examine the key factors driving domestically (Fig. 3b). This trend is particularly notable in the UK,
regional differences in recycled content. where more than half of the collected scrap is exported. On the other
hand, countries that achieve relatively high recycled content despite
growing in-use stock often rely on imported scrap. Turkey, for instance,
3.4. Regional recycled content shaped by trade dynamics
is a major importer of scrap steel exported from the UK. Similarly,
countries with relatively high recycled content, such as Thailand and
Our analysis shows that while domestic scrap availability – driven by
Indonesia, act as net importers of scrap steel. These countries rely on
in-use stock growth – explains recycled content in some countries, it is
imported scrap to compensate for some of the shortfall in domestic scrap
unlikely to be the determining factor in most cases (Fig. 3a). Instead,
generation. Nevertheless, these two factors still do not fully explain the
recycled content is shaped by complex trade dynamics.
limited recycled content observed in countries with stable stock and
A simple expectation would be that countries with growing in-use
limited scrap exports. Belgium, for example, holds abundant domestic
stock rely more on ore-based steelmaking due to limited domestic
scrap and exports little, yet its recycled content remains relatively low.
scrap availability, while those with stable stock levels achieve higher
A third factor provides additional insight: industrialized countries
recycled content. This holds true in some cases: the United States, for
tend to prioritize ore-based steelmaking to produce and export flat
example, has a relatively stable stock and high recycled content, while
products (Fig. 3c). In Belgium, for instance, the domestic steel industry
India, with rapid stock growth, relies more on ore-based steelmaking.
almost exclusively produces flat products through the blast furnace-
However, this pattern does not hold across all cases. Several industri­
basic oxygen furnace (BF-BOF) process and is a net exporter of these
alized countries, such as the UK, Japan, and The Netherlands, have
products. Therefore, their main motivation for steelmaking is not do­
stable in-use stock, meaning their outflows of scrap nearly match their
mestic demand, but rather their steelmaking is driven by exports. This
inflows of new steel. In theory, these countries could meet most of their
pattern is common in other industrialized countries, such as Japan,
steel needs with just domestic end-of-life scrap, yet this is not currently
South Korea, and The Netherlands. In contrast, there is a clear trend for
the case. The opposite is also true for several countries, including
countries with higher recycled content to be more dependent on im­
Thailand, Indonesia, and Turkey, which have relatively high recycled
ports. Interestingly, the top 10 countries with respect to recycled content
content despite growing in-use stock. Many countries are, in fact, far
are all net importers of flat products. A notable example is Thailand,
from having recycled content that is consistent with domestic scrap
which has the highest recycled content, at 96 %, and is almost entirely
availability, visualized as a 1:1 diagonal in Fig. 3a. This means that the

Fig. 3. Underlying factors behind recycled content around the world. (a) Relationship between stock stabilization and recycled content. (b) Relationship between
import reliance for scrap steel and recycled content. (c) Relationship between import reliance for flat products and recycled content. All data refer to 2019. Colors are
based on the data in panel (a) and represent the distance from the diagonal. The horizontal axes of the three panels are identical. Darker blue/red indicates lower/
higher recycled content relative to domestic scrap availability, respectively.

5
T. Watari et al. Resources, Conservation & Recycling 220 (2025) 108363

dependent on imports to meet its demand for flat products. BF-BOF process is directed toward domestic manufacturing and is used
to feed the country’s growing in-use steel stock.
3.5. Four patterns of iron and steel flows The third pattern achieves a relatively high recycled content despite
limited domestic scrap availability. Thailand, for example, fits this
Collectively, iron and steel flows across the world can be broadly model, where instead of producing steel domestically, they "offshore"
categorized into four main patterns, each reflecting varying levels of the manufacturing of high-quality flat products. Thus, the central role of
recycled content for different reasons (Fig. 4). The first pattern is char­ the domestic steel industry is to recycle scrap generated in the forming
acterized by high recycled content due to abundant domestic scrap. A and fabrication processes in EAFs.
key example is the United States, where electric arc furnaces (EAFs) are Finally, the fourth pattern shows a relatively low recycled content
predominantly used to recycle domestic scrap and supply most of the despite abundant domestic scrap. Japan, for instance, prioritizes the BF-
country’s steel needs. BOF process to produce and export flat products. As a result, some of the
In contrast, the second pattern is observed in growing economies collected scrap is exported rather than recycled locally due to difficulties
such as China, where recycled content remains relatively low due to a in contamination control.
shortage of domestic scrap. In this pattern, most steel produced via the

Fig. 4. Four distinct patterns of iron and steel flows based on recycled content and domestic scrap availability. These patterns illustrate how recycled content levels
vary across countries due to different industrial structures. All data represent the year 2019. The Sankey diagrams for all target countries over a 20-year period can be
accessed via a web application ([Link] The flow shown in red is a balancing flow from the data reconciliation process. Crude ore
extraction and end-of-life waste are not included in these national-level Sankey diagrams.

6
T. Watari et al. Resources, Conservation & Recycling 220 (2025) 108363

3.6. Recycled content convergence through global supply chains 4. Discussion

The observations above suggest that trade patterns, rather than do­ 4.1. Circularity in one place, leakage in another?
mestic scrap availability alone, shape recycled content. This raises an
interesting question: How does the recycled content of steel products Overall, our results indicate that seemingly high circularity in some
evolve as they move through the global supply chain? Expanding the places is often supported by low circularity elsewhere. While global
dataset with bilateral trade data reveals a clear trend: recycled content stagnation in recycled content can be explained by growing in-use stock
converges from upstream to downstream (Fig. 5). While liquid steel and the resulting limits on scrap availability, these factors alone do not
shows wide regional variations in recycled content (ranging from 5 % to fully account for the regional differences. In fact, for most nations, in­
96 %), this range narrows to 5 %–62 % for finished steel products and 7 ternational trade patterns play a much larger role in determining the
%–50 % for end-use goods. Notably, countries with relatively high recycled content of their steel products. It is therefore an over­
recycled content at the steelmaking stage see a decline in recycled simplification to assume that higher levels of domestic scrap availability
content as steel products move downstream. For example, Thailand will automatically lead to a higher national recycled content. Similarly,
achieves the highest recycled content (96 %) at the steelmaking stage, countries with higher recycled content are not necessarily leading the
but this drops to 40 % when considering steel embedded in end-use way toward a more circular steel industry. As long as the global in-use
goods. More broadly, the top 10 countries with the highest recycled stock continues to grow, improvements in local circularity do not
content at the steelmaking stage all show a decline as steel products necessarily lead to global progress.
progress toward end-users.
These trends highlight that recycled content at the steelmaking stage
does not fully capture system-wide circularity. The steel products used 4.2. Recycled content targets must account for trade dynamics
by manufacturers, builders, and end-users have substantially different
recycled content profiles than those at the point of steelmaking. In fact, These insights hold key implications for ongoing policy discourse
once trade is accounted for, differences in recycled content between regarding recycled content targets. As countries around the world
major crude steel-producing countries become much smaller. recalibrate their strategies to transition toward a more circular economy
and achieve net-zero carbon emissions (Raabe et al., 2024), the idea of
setting national recycled content targets is beginning to gain momentum
(Systemiq, 2023). Our analysis highlights an important caveat to this

Fig. 5. Recycled content of steel products across countries and supply chain stages. The recycled content shown represents the steel products used in each country (i.
e., domestic production + imports - exports). All data refer to 2019. The dashed line shows the weighted average of the 30 countries.

7
T. Watari et al. Resources, Conservation & Recycling 220 (2025) 108363

discussion: the recycled content of steel products is shaped by global than constructing a multi-regional table annually, a simplified method,
supply chains rather than confined to national borders. As steel products such as assuming a uniform global recycled content for all imported
move through the supply chain, recycled content tends to average out products, can serve as a reasonable alternative (See Section 9 in the SI).
across countries. This means that high recycled content at the steel­ These datasets can then be effectively visualized using interactive San­
making stage does not necessarily translate into similarly high recycled key diagrams developed in this study, offering a more intuitive way of
content in finished steel products or end-use goods. communicating complex results to policymakers and industry partners.
Without careful consideration of such trade dynamics and the niche We emphasize the need for further efforts in these areas to deepen our
roles countries play in supply chains, setting recycled content targets on understanding of the physical economy and support informed
a country-by-country basis risks outsourcing ore-based steelmaking, decision-making.
increasing import dependency, and failing to enhance global circularity.
Meaningful progress requires a stabilization of global stock growth and a CRediT authorship contribution statement
clear understanding of how recycled content should be improved on a
worldwide scale. Global targets, informed by these insights, should then Takuma Watari: Writing – original draft, Visualization, Methodol­
be translated into actionable national targets to ensure that ambitious ogy, Funding acquisition, Formal analysis, Data curation, Conceptuali­
targets in one country do not lead to reduced recycled content in others. zation. Tomer Fishman: Writing – review & editing, Methodology,
Funding acquisition, Conceptualization. Hanspeter Wieland: Writing –
4.3. Stabilizing stock growth as a core strategy review & editing, Methodology, Data curation, Conceptualization.
Dominik Wiedenhofer: Writing – review & editing, Methodology,
In this context, setting effective recycled content targets must Conceptualization.
address the core barrier to circular steel flows: the continued expansion
of in-use stock. At the moment, most of the major steel-producing Declaration of competing interest
countries have national targets for materials recovery and recycling,
with China and India, for example, setting specific targets for steel The authors declare the following financial interests/personal re­
recycling (OECD, 2024). However, our analysis demonstrates that ef­ lationships which may be considered as potential competing interests:
forts to improve recycling practices have been historically counteracted Takuma Watari reports financial support was provided by the Japan
by the continued expansion of global in-use steel stock. Achieving more Society for the Promotion of Science and the Royal Society. If there are
circular steel flows thus requires addressing not only how scrap steel is other authors, they declare that they have no known competing financial
recycled but also how much new steel is continuously added to the use interests or personal relationships that could have appeared to influence
phase. the work reported in this paper.
Despite this challenge, current circular economy policies and stra­
tegies rarely include explicit targets for stock stabilization through ‘ef­ Acknowledgements
ficiency’ or ‘sufficiency’, which aim to deliver the same level of services
with less resource inputs (Rudolf and Schmidt, 2025). We emphasize the This research was supported in part by JSPS KAKENHI (24K03142),
urgency of embedding these strategies at the core of circular economy the Royal Society ISPF – International Collaboration Award (ICA/R1/
initiatives. As this study shows, ambitious recycling efforts at the local 23046), the European Union’s Horizon Programme through the CIR­
scale do not necessarily lead to global progress without directly COMOD project (101056868) and the CircEUlar project (101056810),
addressing the ongoing expansion of global in-use stock. This means that and the Austrian Science Fund (FWF) project REMASS (10.55776/
achieving a more circular steel flow is challenging at the level of indi­ EFP5). We extend our appreciation to Stefan Pauliuk and Nildem Ata­
vidual countries. Instead, this is a global challenge that demands coor­ sayar at the University of Freiburg for their insightful discussions during
dination across all countries, together with stronger governance (Ali the conceptual phase and for their assistance in developing a web
et al., 2017). application. We also thank Wataru Takayanagi and Jacob Fry at the
National Institute for Environmental Studies for helpful comments on
4.4. Limitations and steps forward graph visualization and physical input-output tables.

It is important to recognize here that recycling captures only one Supplementary materials
dimension of circularity (Worrell and Reuter, 2014). The aim of the
circular economy is much broader than merely closing the loop through Supplementary material associated with this article can be found, in
recycling (Kirchherr et al., 2017); it also includes narrowing and slowing the online version, at doi:10.1016/[Link].2025.108363.
down the material loop by reducing overall purchases, making lighter
products that last longer, and reusing products or components without Data availability
the energy-intensive step of recycling (Allwood, 2024). The dataset
developed in this study, however, cannot capture the effects of these The raw data from the World Steel Association and the US Geological
interventions. In addition, as with all modeling research, our dataset Survey can be accessed at [Link]
relies on several assumptions, which inherently involves some uncer­ bookshop/ and [Link]
tainty. While the resulting data aligns well with existing studies for information-center/iron-ore-statistics-and-information, respectively.
major economies, certain disparities exist, and validation is not possible Please note that restrictions apply to the Steel Statistical Yearbook data,
for emerging economies due to limited data availability (See Section 8 in which were acquired and used under license for this study. Processed
the SI). Addressing these issues remains a key area for future research. versions of the data are available on the ZENODO repository
Another important step forward is automating dataset construction ([Link] The Sankey diagrams are
and linking it to interactive Sankey diagrams. Continuous monitoring of accessible via a web application ([Link]
recycled content at both the product and country levels requires ongoing app/), along with the processing code on our GitHub repository
dataset updates, which demands significant time and computational ([Link] and the ZEN­
resources. It is thus useful to automate dataset construction using ODO repository ([Link]
methods such as RAS-type reconciliation algorithms (Lenzen et al.,
2009), constrained optimization (Kopec et al., 2016), or a Bayesian
approach (Lupton and Allwood, 2018). We also demonstrate that rather

8
T. Watari et al. Resources, Conservation & Recycling 220 (2025) 108363

References Loibl, A., Tercero, L.A., 2021. Current challenges in copper recycling : aligning insights
from material flow analysis with technological research developments and industry
issues in Europe and North America. Resour. Conserv. Recycl. 169, 105462. https://
Ali, S.H., Giurco, D., Arndt, N., Nickless, E., Brown, G., Demetriades, A., Durrheim, R.,
[Link]/10.1016/[Link].2021.105462.
Enriquez, M.A., Kinnaird, J., Littleboy, A., Meinert, L.D., Oberhänsli, R., Salem, J.,
Lupton, R.C., Allwood, J.M., 2018. Incremental material flow analysis with bayesian
Schodde, R., Schneider, G., Vidal, O., Yakovleva, N., 2017. Mineral supply for
inference. J. Ind. Ecol. 22, 1352–1364. [Link]
sustainable development requires resource governance. Nature 543, 367–372.
Lupton, R.C., Allwood, J.M., 2017. Hybrid Sankey diagrams: visual analysis of
[Link]
multidimensional data for understanding resource use. Resour. Conserv. Recycl. 124,
Allwood, J.M., 2024. Material efficiency—Squaring the circular economy: recycling
141–151. [Link]
within a hierarchy of material management strategies. Handbook of Recycling: State-
Miatto, A., Emami, N., Goodwin, K., West, J., Taskhiri, M.S., Wiedmann, T., Schandl, H.,
of-the-Art for Practitioners, Analysts, and Scientists. Elsevier, pp. 45–78. [Link]
2024. Australia’s circular economy metrics and indicators. J. Ind. Ecol. 28. https://
org/10.1016/B978-0-323-85514-3.00016-6.
[Link]/10.1111/jiec.13458.
Bradley, J.E., Auping, W.L., Kleijn, R., Kwakkel, J.H., Sprecher, B., 2024. Reassessing tin
Moraga, G., Huysveld, S., Mathieux, F., Blengini, G.A., Alaerts, L., Van Acker, K., de
circularity and criticality. J. Ind. Ecol. 28, 232–246. [Link]
Meester, S., Dewulf, J., 2019. Circular economy indicators: what do they measure?
jiec.13459.
Resour. Conserv. Recycl. 146, 452–461. [Link]
Chen, W.Q., 2013. Recycling rates of aluminum in the United States. J. Ind. Ecol. 17,
resconrec.2019.03.045.
926–938. [Link]
Müller, D.B., Billy, R.G., Simoni, M.U., Petavratzi, E., Liu, G., Rechberger, H., Cullen, J.,
Chen, W.Q., Graedel, T.E., 2012. Anthropogenic cycles of the elements: a critical review.
2023. Maps of the physical economy to inform sustainability strategies. Handbook of
Environ. Sci. Technol. 46, 8574–8586. [Link]
Recycling: State-of-the-Art for Practitioners, Analysts, and Scientists. Elsevier,
Cooper, D.R., Ryan, N.A., Syndergaard, K., Zhu, Y., 2020. The potential for material
pp. 27–44. [Link]
circularity and independence in the U.S. steel sector. J. Ind. Ecol. 24, 748–762.
OECD, 2024. Circular economy policies for steel decarbonisation. [Link]
[Link]
10.1787/4cfb485d-en.
Cullen, J.M., Allwood, J.M., Bambach, M.D., 2012. Mapping the global flow of steel:
Pauliuk, S., 2018. Critical appraisal of the circular economy standard BS 8001:2017 and
from steelmaking to end-use goods. Environ. Sci. Technol. 46, 13048–13055.
a dashboard of quantitative system indicators for its implementation in
[Link]
organizations. Resour. Conserv. Recycl. 129, 81–92. [Link]
Daehn, K.E., Cabrera Serrenho, A., Allwood, J.M., 2017. How will copper contamination
resconrec.2017.10.019.
constrain future global steel recycling? Environ. Sci. Technol. 51, 6599–6606.
Plank, B., Streeck, J., Virág, D., Krausmann, F., Haberl, H., Wiedenhofer, D., 2022. From
[Link]
resource extraction to manufacturing and construction: flows of stock-building
Daigo, I., Tajima, K., Hayashi, H., Panasiuk, D., Takeyama, K., Ono, H., Kobayashi, Y.,
materials in 177 countries from 1900 to 2016. Resour. Conserv. Recycl. 179,
Nakajima, K., Hoshino, T., 2021. Potential influences of impurities on properties of
106122. [Link]
recycled carbon steel. ISIJ Int. 61, 498–505. [Link]
Raabe, D., Jovičevi´c-Klug, M.J., Ponge, D., Gramlich, A., Kwiatkowski Da Silva, A.,
[Link]-2020-377.
Grundy, A.N., Springer, H., Filho, I.S., Ma, Y., 2024. Circular steel for fast
Dworak, S., Fellner, J., 2021. Steel scrap generation in the EU-28 since 1946 – Sources
decarbonization: thermodynamics, kinetics, and microstructure behind upcycling
and composition. Resour. Conserv. Recycl. 173, 105692. [Link]
scrap into high-performance sheet steel. Annu. Rev. Mater. Res. [Link]
resconrec.2021.105692.
10.1146/annurev-matsci-080222.
Espinoza, L.A.T., Soulier, M., 2018. Defining regional recycling indicators for metals.
Reck, B.K., Graedel, T., 2012. Challenges in metal recycling. Science (1979) 337,
Resour. Conserv. Recycl. 129, 120–128. [Link]
690–695. [Link]
resconrec.2017.10.022.
Rostek, L., Tercero Espinoza, L.A., Goldmann, D., Loibl, A., 2022. A dynamic material
Font Vivanco, D., Hoekman, P., Fishman, T., Pauliuk, S., Niccolson, S., Davis, C.,
flow analysis of the global anthropogenic zinc cycle: providing a quantitative basis
Makov, T., Hertwich, E., 2019. Interactive Visualization and Industrial ecology:
for circularity discussions. Resour. Conserv. Recycl. 180, 106154. [Link]
applications, challenges, and opportunities. J. Ind. Ecol. 23, 520–531. [Link]
10.1016/[Link].2022.106154.
org/10.1111/jiec.12779.
Rudolf, M., Schmidt, M., 2025. Efficiency, sufficiency and consistency in sustainable
Gao, H., Liu, J., Daigo, I., 2025. Methodology development for estimating the impact of
development: reassessing strategies for reaching overarching goals. Ecol. Econ. 227.
restriction factors to promote national steel recycling. Resour. Conserv. Recycl. 215.
[Link]
[Link]
Song, L., Wang, P., Hao, M., Dai, M., Xiang, K., Li, N., Chen, W.Q., 2020. Mapping
Giljum, S., Maus, V., Kuschnig, N., Luckeneder, S., Tost, M., Sonter, L.J., Bebbington, A.
provincial steel stocks and flows in China: 1978–2050. J. Clean. Prod. 262, 121393.
J., 2022. A pantropical assessment of deforestation caused by industrial mining 119.
[Link]
[Link]
Speizer, S., Durga, S., Blahut, N., Charles, M., Lehne, J., Edmonds, J., Yu, S., 2023. Rapid
Graedel, T.E., 2019. Material flow analysis from origin to evolution. Environ. Sci.
implementation of mitigation measures can facilitate decarbonization of the global
Technol. 53, 12188–12196. [Link]
steel sector in 1.5 ◦ C-consistent pathways. One Earth 6, 1494–1509. [Link]
Graedel, T.E., Allwood, J., Birat, J.P., Buchert, M., Hagelüken, C., Reck, B.K., Sibley, S.F.,
10.1016/[Link].2023.10.016.
Sonnemann, G., 2011. What do we know about metal recycling rates? J. Ind. Ecol.
Systemiq, 2023. Circular steel: a system perspective on recycled content targets.
15, 355–366. [Link]
Tuck, C.C., Xun, S., Singerling, S.A., 2017. USGS revision of global iron ore production
Haas, W., Krausmann, F., Wiedenhofer, D., Heinz, M., 2015. How circular is the global
data—Clarification of the reporting of iron ore production in China and application
economy?: an assessment of material flows, waste production, and recycling in the
of a uniform comparison methodology (2000-2015). Min. Eng. 69.
European union and the world in 2005. J. Ind. Ecol. 19, 765–777. [Link]
UNEP, 2011. Recycling Rates of metals – A status report, A report of the Working Group
10.1111/jiec.12244.
on the global metal flows to the International Resource Panel.
Haas, W., Krausmann, F., Wiedenhofer, D., Lauk, C., Mayer, A., 2020. Spaceship earth’s
U.S. Geological Survey, 2024. Commodity statistics and information.
odyssey to a circular economy - a century long perspective. Resour. Conserv. Recycl.
Wang, H., Schandl, H., Wang, X., Ma, F., Yue, Q., Wang, G., Wang, Y., Wei, Y., Zhang, Z.,
163, 105076. [Link]
Zheng, R., 2020. Measuring progress of China’s circular economy. Resour. Conserv.
Harvey, L.D.D., 2022. Reconciling global iron and steel mass flow datasets, with an
Recycl. 163. [Link]
update to 2011-2015 and an assessment of uncertainty in global end-of-life scrap
Wang, P., Ryberg, M., Chen, W., Kara, S., Hauschild, M., 2021. Efficiency stagnation in
flow. Resour. Conserv. Recycl. 182, 106281. [Link]
global steel production urges joint supply- and demand-side mitigation efforts. Nat.
resconrec.2022.106281.
Commun. 1–11. [Link]
International Aluminium Institute, 2023. The Global aluminium cycle [WWW
Wang, P., Zhao, S., Dai, T., Peng, K., Zhang, Q., Li, J., Chen, W.Q., 2022. Regional
Document]. URL. [Link] (accessed 9.28.24).
disparities in steel production and restrictions to progress on global decarbonization:
Kirchherr, J., Reike, D., Hekkert, M., 2017. Conceptualizing the circular economy: an
a cross-national analysis. Renew. Sustain. Energy Rev. 161, 112367. [Link]
analysis of 114 definitions. Resour. Conserv. Recycl. 127, 221–232. [Link]
10.1016/[Link].2022.112367.
10.1016/[Link].2017.09.005.
Wang, T., Müller, D.B., Graedel, T.E., 2007. Forging the anthropogenic iron cycle.
Kopec, G.M., Allwood, J.M., Cullen, J.M., Ralph, D., 2016. A general nonlinear least
Environ. Sci. Technol. 41, 5120–5129. [Link]
squares data reconciliation and estimation method for material flow analysis. J. Ind.
Watari, T., Hata, S., Nakajima, K., Nansai, K., 2023. Limited quantity and quality of steel
Ecol. 20, 1038–1049. [Link]
supply in a zero-emission future. Nat Sustain 6, 336–343. [Link]
Lei, T., Wang, D., Yu, X., Ma, S., Zhao, W., Cui, C., Meng, J., Tao, S., Guan, D., 2023.
s41893-022-01025-0.
Global iron and steel plant CO2 emissions and carbon-neutrality pathways. Nature
Wiedenhofer, D., Fishman, T., Lauk, C., Haas, W., Krausmann, F., 2019. Integrating
622, 514–520. [Link]
material stock dynamics into economy-wide material flow accounting: concepts,
Lenzen, M., Gallego, B., Wood, R., 2009. Matrix balancing under conflicting information.
modelling, and global application for 1900–2050. Ecol. Econ. 156, 121–133. https://
Econ. Syst. Res. 21, 23–44. [Link]
[Link]/10.1016/[Link].2018.09.010.
Li, Z., Hanaoka, T., 2022. Plant-level mitigation strategies could enable carbon neutrality
Wiedenhofer, D., Streeck, J., Wieland, H., Grammer, B., Baumgart, A., Plank, B.,
by 2060 and reduce non-CO2 emissions in China’s iron and steel sector. One Earth 5,
Helbig, C., Pauliuk, S., Haberl, H., Krausmann, F., 2024. From extraction to end-uses
932–943. [Link]

9
T. Watari et al. Resources, Conservation & Recycling 220 (2025) 108363

and waste management: modeling economy-wide material cycles and stock World Steel Association, 2023. Steel-the permanent material in the circular economy.
dynamics around the world. J. Ind. Ecol. [Link] Belgium.
Wieland, H., Lenzen, M., Geschke, A., Fry, J., Wiedenhofer, D., Eisenmenger, N., Worrell, E., Carreon, J.R., 2017. Energy demand for materials in an international context.
Schenk, J., Giljum, S., 2022. The PIOLab: building global physical input–output Philos. Trans. R. Soc., A 375. [Link]
tables in a virtual laboratory. J. Ind. Ecol. 26, 683–703. [Link] Worrell, E., Reuter, M.A., 2014. Handbook of Recycling : State-of-the-art for
jiec.13215. Practitioners, Analysts, and Scientists. Elsevier Inc. [Link]
Association, WorldSteel, 2024. Steel Statistical Yearbooks. 0-07046-1.

10