Professional Documents
Culture Documents
Edited by B. Turner, Arizona State University, Tempe, AZ; received July 18, 2022; accepted May 12, 2023
From 2000 through 2020, demand for cobalt to manufacture batteries grew 26-fold.
Eighty-two percent of this growth occurred in China and China’s cobalt refinery Significance
production increased 78-fold. Diminished industrial cobalt mine production in the
early-to-mid 2000s led many Chinese companies to purchase ores from artisanal cobalt The linkage between abuse to
miners in the Democratic Republic of the Congo (DRC), many of whom have been artisanal cobalt miners—
found to be children. Despite extensive research on artisanal cobalt mining, fundamen- including children—in the
tal questions about its production remain unanswered. This gap is addressed here by Democratic Republic of the
estimating artisanal cobalt production, processing, and trade. The results show that, Congo (DRC) and use of cobalt in
while total DRC cobalt mine production grew from 11,000 metric tons (t) in 2000 advanced batteries has
to 98,000 t in 2020, artisanal production only grew from 1,000 to 2,000 t in 2000 prompted global supply chain
to 9,000 to 11,000 t in 2020 (with a peak of 17,000 to 21,000 t in 2018). Artisanal
reviews, responsible sourcing
production’s share of world and DRC cobalt mine production peaked around 2008
initiatives, and research to
at 18 to 23% and 40 to 53%, respectively, before trending down to 6 to 8% and
9 to 11% in 2020, respectively. Artisanal production was chiefly exported to China remove cobalt from battery
or processed within the DRC by Chinese firms. An average of 72 to 79% of artisanal cathodes. However, no study has
production was processed at facilities within the DRC from 2016 through 2020. As systematically evaluated how
such, these facilities may be potential monitoring points for artisanal production and large of a role artisanal mining
its downstream consumers. This finding may help to support responsible sourcing plays in world cobalt mine
initiatives and better address abuses related to artisanal cobalt mining by focusing production. Doing so reveals
local efforts at the artisanal processing facilities through which most artisanal cobalt artisanal supply trends and
production flows. potential monitoring points for
Downloaded from https://www.pnas.org by 45.226.238.191 on July 6, 2023 from IP address 45.226.238.191.
of an estimation methodology in the left column. Rather than relying Rather, it is to gather all available estimates from the literature into
on an estimation method, many literature references provide rough a single analysis; compare the reported estimates to the results of the
percentages of total DRC cobalt mine production. The lack of trans- two estimation methods developed here (i.e., accounting estimates);
parent estimation methods for literature estimates motivates the compare multiple results within a single year; and identify artisanal
development of the two estimation methods presented below. Any production trends, destinations, and drivers.
Methods and data used to estimate annual DRC artisanal cobalt mine production
minus plus
Industrial mine production Artisanal imports
Sum of production reported for Total Chinese ore/concentrate imports
industrial mines in the DRC from DRC minus industrial imports from
industrial mines
Data sources: Mining company Data sources: China Customs and DRC
financial reports Ministry of Mines
Artisanal production estimates
Artisanal-production-A Artisanal-production-B
Results of this method are also Results of this method are
Literature estimates referred to as also referred to as
‘nonindustrial production’ ‘artisanal processing & imports
(i.e., artisanal production)’
Fig. 1. Methods and data used to estimate artisanal cobalt mine production in the Democratic Republic of the Congo (DRC): unknown methods for literature
estimates in the left column, artisanal-production-A, also referred to as “nonindustrial production” in the middle column; artisanal-production-B, also referred
to as “artisanal processing and imports (i.e., artisanal production)” in the right column.
2 of 11 https://doi.org/10.1073/pnas.2212037120 pnas.org
For context and general interest, estimates of the total number of Results
artisanal miners, the number of children working or present at arti-
sanal cobalt sites, and the share of children estimated to be at such Comparison of Results and Literature Estimates. Fig. 3A shows
sites are gathered from the literature. For the same reasons that it is the artisanal cobalt production estimates reported in the literature,
difficult to estimate artisanal cobalt production, it is difficult to esti- the results generated by each method depicted in Fig. 1, and the
mate how many people are working at artisanal mine sites. It is annual average price of cobalt metal deflated to 2020 dollars (8).
especially difficult to estimate the number of children who are dig- Fifty-three artisanal cobalt-mining estimates are available in the
ging, tunneling, washing ore, sorting ore, transporting ore, running literature (23–25, 27, 29–41). Except for 2007 and 2008, multiple
errands, watching younger siblings, or being watched by parent estimates are reported for each year. The reported estimates from
miners (10, 20–25). The estimates from the literature are presented the literature are presented visually as dots as a scatter plot in
as found in the original studies in Fig. 2 with abbreviated references Fig. 3A by reporter type (government, industry, or NGO). Fig. 3B
in the figure caption. Notably, these estimates have not been con- compares artisanal-production-A and artisanal-production-B by
firmed in this analysis. Complications in estimating the number of quantity and by their share of total DRC cobalt mine production.
total and child artisanal cobalt miners include limited site visits, Fig. 3C then shows artisanal-production-A as a percentage of
colocated living/mining spaces, and misreporting of combined esti- world and DRC cobalt mine production.
mates that include both artisanal cobalt miners in the southern cop- The results for artisanal-production-A (also referred to as nonin-
per–cobalt belt with artisanal tin, tungsten, tantalum (i.e., coltan), dustrial production) are presented in Fig. 3A as the solid purple line.
and gold miners in northern Katanga (10, 20–25). Unlike these They indicate that artisanal-production-A increased from 2,000
so-called 3TG commodities, cobalt has not been designated as a metric tons (t) in 2000 to 12,000 t in 2005. This increase was largely
conflict mineral (26) and recent studies have found no signs of forced due to increasing Chinese demand for cobalt refinery feed; at the
labor (20, 23, 24, 27). See SI Appendix, Supporting Discussion 2 for same time, increases in industrial DRC cobalt mine production were
information about the studies that estimate the number of artisanal hampered by a lack of foreign investment (8). Except a dip to 6,000
miners in the DRC and Dataset S13 for numerical estimates. t in 2013, artisanal-production-A fluctuated between 9,000 and
Artisanal cobalt mine site studies and interviews provide an indi- 17,000 t per annum (tpa) from 2005 through 2020. During this
cation of the types of work children may be involved in at such sites. period, artisanal-production-A peaked at 16,000 t in 2010, then
These estimates are not confirmed here and are presented as found again at 17,000 t in 2015 and 2018. Only the 2018 peak corresponds
in the original studies. Of children working at cobalt mine sites, site with a spike in the cobalt price. But each peak roughly corresponds
studies have estimated that roughly 50% are fifteen to seventeen with the discovery of a surface (heterogenite) deposit that artisanal
years old, roughly 40% are ten to fourteen, and the remainder are miners can exploit (31, 32, 42, 43) and an annual increase in Chinese
younger than ten (20, 22). Similarly, of children working at cobalt refinery production of 8,000 t or more (8). From 2005 through
mine sites, roughly 20% clean ores that have been extracted, 25% 2020, artisanal production hit its lowest points at 6,000 t in 2013
sort cleaned ores, another 25% are working in other activities at the and 9,000 t in 2016 and 2020. Although the causes of such declines
Downloaded from https://www.pnas.org by 45.226.238.191 on July 6, 2023 from IP address 45.226.238.191.
surface, and the remaining 30% are transporting ore or performing are currently unknown, these years do correspond with troughs in
other tasks (20, 22). Of children who are excavating ore under- the price of cobalt.
ground or at the surface, between 60% and 70% are 15 or older The results for artisanal-production-B (also referred to as
(22). Artisanal cobalt mine site studies have estimated that children artisanal processing and imports) are presented in Fig. 3 A and
numbered 24,000 or 40% of total artisanal cobalt miners in 2002 B as the dotted lavender line. Like artisanal-production-A,
(9), 60,000 or 40% in 2007 (25), and 35,000 or 14% in 2017 (28). artisanal-production-B peaks in 2005, 2010, 2015, and 2018,
360 60%
For the years
child cobalt miners, '000
Number of artisanal and
production,'000 metric
24 $28
18 $21
12 $14
6 $7
- $-
2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020
80%
18
60%
12
40%
6
20%
- 0%
2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020
Downloaded from https://www.pnas.org by 45.226.238.191 on July 6, 2023 from IP address 45.226.238.191.
Artisanal-production-A (results)
Artisanal-production-B (results)
Artisanal-production-A's share of total DRC cobalt mine production (results)
Artisanal-production-B's share of total DRC cobalt mine production (results)
C
160 60%
48%
120
Cobalt mine
36%
80
24%
40
12%
0 0%
2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020
as well as bottoming out in 2013, 2016, and 2020. The annual for 17 of the 20 y (Dataset S11d). From 2000 through 2020,
trends for artisanal- production- B were similar to those of the average difference between artisanal-production-A and -B
artisanal-production-A, with the sign of annual change (positive was 800 tpa (1%) with a minimum difference of zero (0%) and
or negative) for artisanal-production-A and -B being the same a maximum of 5,000 tpa (67%) (Dataset S11d).
4 of 11 https://doi.org/10.1073/pnas.2212037120 pnas.org
Similar to artisanal-production-A, artisanal-production-B with a range from 58% lower to 100% higher. Industry references
increased from 2,000 t in 2000 to 11,000 t in 2005 due almost were 19% lower on average, with a range from 55% lower to 25%
entirely to the increase in artisanal Chinese imports from 1,000 t higher. NGOs were 209% higher on average, with a range from
to 10,000 t (Dataset S10). Artisanal-production-B then increased 41 to 800% higher. These results indicate that, on average, NGOs
to 17,000 t in 2008, again due to an increase in artisanal Chinese report cobalt production 2.2 to 3.1 times larger than that of other
imports to 14,000 t. From 2009 through 2020, this relation references. On the contrary, government, industry,
switched with artisanal processing exceeding artisanal Chinese artisanal-production-A, and artisanal-production-B estimates have
imports for all but 1 y. For example, while artisanal Chinese no more than an average difference of 27%. The references above
imports fell 11,000 t from 2008 to 2009, artisanal processing have the same sign of annual changes from 80 to 100% of available
increased from 3,000 t to 10,000 t as the result of investments by years, while NGO references range from 36 to 64%. These results
predominantly Chinese and Indian companies in processing (but may indicate that government and industry references are using sim-
not) mining facilities (Dataset S4b). The peaks in 2015 and 2018 ilar methods of estimation to artisanal- production- A and - B.
were due to increases in both artisanal processing and artisanal Alternatively, it may mean that government and industry have access
Chinese imports, while nadirs in 2016 and 2020 were both due to authoritative references whose production information has been
to decreases in artisanal processing and artisanal Chinese imports. approximated by artisanal-production-A and -B.
Like the decline in artisanal-production-A from 2018 to 2020,
artisanal-production-B fell from 21,000 t to 11,000 t. From 2000 Results and Literature Estimates as a Percentage of Total World
through 2006, the results of artisanal-production-B were on aver- and DRC Cobalt Mine Production. Fig. 3C presents artisanal-
age 1,400 tpa (31%) lower than those of artisanal-production-A. production-A as estimated artisanal production in the context of
Given the low quantities for these early years, this difference may total DRC and world cobalt mine production. The quantity of
be immaterial. From 2007 to 2020, artisanal-production-B was world, DRC, and artisanal-production-A is presented inFig. 3C as
equal to or greater than artisanal-production-A for all years except unstacked cobalt blue, sky blue, and dark purple areas, respectively,
one, with an average of 1,900 tpa (17%) higher. This could be on the Left axis. Artisanal-production-A, as a percentage of total
due to unreported Chinese imports from industrial operations, world and DRC production, is presented as dark and light rose
unreported artisanal feed for industrial mining operations, or some lines, respectively, on the Right axis. Fig. 3C shows that artisanal-
combination thereof. production- A’s share of total world and DRC cobalt mine
As for the artisanal production estimates from the literature, production increased from 5% and 18% in 2000 to 18% and 49%
the government estimates from 2000 through 2005 come from in 2005, respectively (Dataset S11b and S11c). During this period,
the U.S. Geological Survey (USGS) (36, 37). The estimate for China’s cobalt refinery production increased more than ten-fold.
2006 is from the World Bank (25). No government estimates were Industrial mines were unable to keep pace with this demand, so
located for 2007 or 2008. Estimates from 2009 through 2019 most of the rise in total DRC mine production was comprised
from the Bundesanstalt für Geowissenschaften und Rohstoffe or of artisanal mining. Then, increasing production from industrial
Downloaded from https://www.pnas.org by 45.226.238.191 on July 6, 2023 from IP address 45.226.238.191.
BGR (23, 24). The industry estimates from 2009 through 2020 mines in the DRC (and elsewhere) reduced artisanal-production-
are reported by CRU, a commodities research company (44, 45), A’s percentage of world and DRC production, which fell from
as well as from Darton Commodities (a cobalt research and trading 18% and 49% in 2005 to 6% and 11% in 2013, respectively. With
company) for the years 2013 through 2020. The NGO references the discovery of new surface heterogenite deposits at the end of
providing artisanal production estimates include TRACeability of 2014, artisanal-production-A’s percentage increased to 14% and
hEterogenite or TRACE (34), Global Witness (9), the OECD1 24% in 2015 before trending down to 6% and 9% in 2020. The
(27), PACT (35), and Save the Children (46). See Dataset S1 for causes of this reduction were that industrial production nearly
details. Despite the absence of an estimation methodology, gov- doubled from 2015 to 2020 and artisanal production fell from
ernment, industry, and NGO estimates have (until now) been the 17,000 t to 9,000 t.
only artisanal mine production estimates available to the public. In comparison, government estimates, as a percentage of total
Given that these estimates seek to measure the same quantity of world and DRC production, peaked in 2004 at 12% and 35%,
artisanal cobalt production as the results of the two methods, these respectively. Government estimates then fell over time to 8%
estimates are compared with the results below. and 12% in 2019, which are the same percentages as artisanal-
In relation to artisanal-production-A, the annual difference production-A and close to artisanal-production-B results of 10%
between production estimates reported by government references and 15% in 2019. Industry estimates peaked at 17% and 30% in
is 2,800 t (i.e., 27%) lower on average, with a range from 7,000 t 2010 (as opposed to 10% and 18% for both artisanal-production-A
(−67%) lower to 1,000 t (17%) higher. The largest differences were and -B). In 2020, the average of industry estimates is 6% and 9%,
for the peak years of 2005, 2010/2011, and 2015. Annual changes equal to artisanal-production-A and like artisanal-production-B’s
for government estimates and artisanal-production-A are the same 7% and 10%. In other words, the trends and quantities of gov-
sign for 12 of the 15 available years, indicating similar overall ernment and industry estimates were largely in line with the results
trends. Industry references are 700 t (7%) lower on average than for artisanal-production-A and -B.
those of artisanal-production-A, with a range from 6,000 t (−42%) In contrast, the average of NGO estimates, as a percentage of
lower to 7,000 t (41%) higher. Like government estimates, the total world and DRC production, peaked in 2004 at 31% and
largest differences were for the peak years of 2015 and 2018. 93%, respectively. The average of 93% is from two NGO esti-
Annual changes for industry estimates and artisanal-production-A mates, one of 74% of DRC production (34) and the other of
are the same sign for all 11 of the 11 available years. In contrast to 111% (9). It is likely that these are overestimates, given that 65%
government and industry references, no NGO references were (as opposed to 26%) of total DRC cobalt mine production in
lower than those of artisanal-production-A. Estimates from NGO 2004 can be accounted for by just four industrial mining opera-
references were 8,900 t (120%) higher on average, with a range tions (i.e., 700 t from Gécamines, 2,900 t from Big Hill, 5,000 t
from 4,000 t (27%) higher to 22,000 t (300%) higher. from Luiswishi, and 4,000 t from Kababankola) (47–49). Neither
In relation to artisanal-production-B, production estimates of these references report a methodology associated with these
reported by government references were 18% lower on average, estimates, so it is not possible to evaluate the causes of such wide
of artisanal feed sourcing, and references. declining to an average of 8,500 tpa from 2012 through 2016,
Artisanal processing, as a percentage of artisanal-production-A these imports fell to 0 t in 2020. For references, numerical esti-
from 2000 through 2008, remained at or below 23% as most mates, and the companies reported to provide industrial imports,
artisanal production was exported due to the lack of domestic see Dataset S8a for 2000 through 2008 and Dataset S9 for 2009
processing facilities. With the increase in artisanal processing and through 2020.
the decrease in artisanal-production-A from 2008 to 2009, artisanal Artisanal (i.e., nonindustrial) imports, on the contrary, are com-
processing’s share jumped from 23 to 91%. Artisanal processing’s prised of total Chinese imports from the DRC minus industrial
share then fluctuated between 50% and 64% from 2010 through Chinese imports. The results for artisanal imports in Fig. 5 show
2015, except for a spike to 100% in 2013 when the result for an increase from 1,000 t in 2000 to 14,000 t in 2008, which
artisanal-production-A dipped to 6,000 t. As domestic processing reflects increasing Chinese demand for cobalt refinery inputs while
capacity expanded, Chinese imports shifted to crude cobalt the DRC’s industrial mine production struggled to recover.
20 100%
Thousand metric tons,
16 80%
cobalt content
12 60%
8 40%
4 20%
0 0%
2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020
Artisanal-production-A
Artisanal processing
Artisanal processing's share of artisanal-production-A
Fig. 4. Artisanal-production-A in the Democratic Republic of the Congo (DRC), unstacked dark purple area on the Left axis; processing of artisanal production
into intermediate cobalt materials (i.e., artisanal processing) in the DRC, blue gray unstacked area, Left axis; and artisanal processing as a percentage of artisanal-
production-A, light blue line on the Right axis.
6 of 11 https://doi.org/10.1073/pnas.2212037120 pnas.org
30 100%
cobalt content
tons, 18 60%
12 40%
6 20%
0 0%
2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020
Artisanal imports then decreased to an average of 6,600 tpa from assumptions, and numerical estimates, see Dataset S8 for 2000
2009 through 2015 before falling farther to an average of 3,800 through 2008 and Dataset S9 for 2009 through 2020.
tpa from 2016 through 2020. The decline in artisanal imports The artisanal-production-B results from the artisanal processing
reflects both increased artisanal processing and increasing crude plus artisanal imports (i.e., artisanal production) are presented in
cobalt hydroxide imports from the newly recapitalized industrial Fig. 6, along with artisanal processing’s share of artisanal-production-B.
mining industry. For references, calculations, assumptions, and Given that the results for artisanal-production-B and China’s arti-
numerical estimates, see Dataset S8 for 2000 through 2008 and sanal imports have already been discussed, this section focuses on
Dataset S9 for 2009 through 2020. the percentage comprised of artisanal processing. From 2000
As a percentage of total Chinese imports from the DRC, arti- through 2008, artisanal processing’s share of artisanal-production-B
sanal imports of the years 2000, 2007, and 2020 stand out. As averaged 16%, in comparison to 13% for artisanal-production-B.
mentioned, in 2000, the DRC’s mining industry was just restarting This share increased to 77% in 2009 as artisanal processing increased
Downloaded from https://www.pnas.org by 45.226.238.191 on July 6, 2023 from IP address 45.226.238.191.
and little industrial concentrate supply was available, which is why from 3,000 t to 10,000 t and artisanal imports fell from 14,000 to
Chinese imports for 2000 were artisanal. This trend continued 3,000 t. From 2010 through 2015, artisanal processing and imports
from 2000 through 2007 as artisanal imports averaged 77% of were balanced, even despite significant fluctuations, resulting in an
total Chinese imports from the DRC. Artisanal imports’ share average share for artisanal processing of 53%. These results reflect
spiked again to 100% in 2007 because almost all exports from a transition period from artisanal imports to artisanal processing
industrial operations were prevented during that year by the DRC that is reflected in artisanal processing’s average share of 72% from
raw material export ban. This share then fell to an average of 39% 2016 through 2020.
from 2008 through 2017 as industrial operations expanded ore/
concentrate production. It then increased to 100% in 2020 as Boss Key Data and Methodological Uncertainties. Assumptions,
Mining, Metal Mines, and COMIKA ceased industrial ore/con- uncertainty, and data issues that may impact the results of the two
centrate exports, which were replaced by crude cobalt hydroxide estimation methods are briefly presented here. For more detail, see
from industrial mines in 2020. Again, for references, calculations, Materials and Methods below, as well as SI Appendix, Supporting
25 100%
Thousand metric tons,
20 80%
cobalt content
15 60%
10 40%
5 20%
0 0%
2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020
8 of 11 https://doi.org/10.1073/pnas.2212037120 pnas.org
and 5). From 2016 through 2020, between 72% and 79% of artisanal-production-A). Total DRC cobalt mine production for year t is reported at
artisanal production (on average) passed through these facilities. the country level. Industrial production for year t is the sum of industrial mine i’s
From these facilities, one may track artisanal cobalt trade upstream reported production—from its own deposits—for 1 to n industrial mines. Industrial
to artisanal mine sites to evaluate questions such as where are the production is calculated in Eq. 1. The remaining portion of total DRC mine produc-
miners, what are the conditions, how many are there, and how tion is comprised of nonindustrial production for year t. Nonindustrial production
many are children. In the other direction, one may track the sale is calculated in Eq. 2, and presented as the middle column of Fig. 1.
of intermediate materials downstream to answer other questions
such as what refiners are purchasing these materials, where are Industrial productiont
they, and who are their customers. Such an approach would be n
[1]
∑
similar to that employed in US regulations regarding the DRC’s = Reported production for DRC industrial cobalt mineit
artisanal production of the so-called conflict minerals tin, tanta- i=1
tion in the DRC and elsewhere in the world. Artisanal cobalt of the USGS MYB, which collect most of this information from industrial mining
mining, as a percentage of DRC cobalt mine production in company financials and data from the DRC Ministry of Mines. This mine-level
2020, was estimated to be between 9% and 11%. In other information can often be corroborated by industry publications written by cobalt
words, if a cobalt-refining firm were to randomly source cobalt traders and analysts with a deep understanding of the market. Within this context,
mining material from the DRC, there was roughly a 90% chance overestimation of artisanal-production-A could result if there were industrial mine
that the material would be nonartisanal. If, as was estimated production that was not captured in Dataset S2. Alternatively, underestimation
(24), one in four artisanal cobalt mine sites had children work- could result if industrial mining operations supplemented their mined material
with artisanal feed. Both possibilities are more likely for the earlier years (i.e.,
ing—and all artisanal sites produced the same amount of
2000 through 2008) than the later years (i.e., 2009 through 2020).
cobalt—there would be a 98% chance that randomly selected
cobalt mine material from the DRC was not linked to child Method B: Overview. In method B, artisanal production is assumed to be
labor in 2020. comprised of one portion that is processed into intermediate cobalt mate-
Although these results will change with future artisanal produc- rials at facilities in the DRC and another portion that is sent to China (right
tion estimates and artisanal mine site studies, the current results column of Fig. 1). Due to China’s rapidly increasing cobalt refinery production
suggest that, rather than being a “cobalt problem,” child labor in from 2000 through 2020, China imported most of the world’s cobalt ore and
artisanal cobalt mining may be more of a “development problem.” concentrate trade (56). At the beginning of this period in 2000, in contrast
Many downstream OEMs are finding it difficult or impossible to to the newly developing refiners in China, existing cobalt refiners in other
substitute cobalt without increasing the risk of batteries catching countries had already established long-term supply agreements and did not
on fire or without reducing performance (such as long-range elec- need to seek out new feed supply sources. Given the lack of sufficient cobalt
tric vehicles) to an extent that could impact full deployment mine production in China and the dearth of available refinery feed supplies
(19, 41). Cobalt is often portrayed as being mined by child miners from operational industrial mines, Chinese trading firms came to the DRC in
the early 2000s to secure cobalt materials and export them to China (25, 38).
to produce the lithium-ion batteries in our smartphones and elec-
Due to the collapse of the DRC economy and mining industry in the 1990s,
tric vehicles (42, 50–52)—resulting in calls to remove cobalt from
artisanal ores and concentrates were the most readily available cobalt feed
batteries entirely (18). In terms of the river metaphor, this would available to be exported to China (25, 38). Government, industry, and NGO
be equivalent to trying to divert 63% (in 2020) of the water at the references all report artisanal production to predominantly be shipped to
river’s estuary, rather than addressing the problem at its source. China (9, 10, 12, 25, 41).
Materials and Methods Method B: Artisanal Processing. As mentioned above, many facilities in the
DRC are reported to process artisanal cobalt ores and concentrates into interme-
Method A: Total DRC, Industrial, and Nonindustrial DRC Cobalt Mine diate cobalt materials including concentrate, cobalt copper alloy, crude cobalt car-
Production. Method A estimates artisanal production by subtracting the pro- bonate, and crude cobalt hydroxide. To estimate artisanal processing, this analysis
duction of industrial mines from total reported DRC cobalt mine production (mid- multiplies the quantity produced at the facility by the percentage of feed sourced
dle column of Fig. 1). The result is an estimate of nonindustrial production (i.e., from artisanal production (often 100%), as calculated in Eq. 3,
cial reported statistics. Such skepticism has largely pertained to official statistics
regarding GDP growth (58–64), but to a lesser extent, some have also questioned
i=1 [4]
trade statistics (61, 64). As such, references are provided here for the reader that
may be of interest in this literature, which has shown that China’s trade statistics
Downloaded from https://www.pnas.org by 45.226.238.191 on July 6, 2023 from IP address 45.226.238.191.
broadly align with trade statistics from other countries (61, 64). To illuminate Artisanal importst = Total importst − Industrial importst [5]
this question within the context of this analysis, several Supporting Datasets are
provided for specialists to compare China Customs’ data with that from other
reporters. Dataset S6 presents reported cobalt ore/concentrate exports to China
Artisanal processingt = Artisanal production Bt
from the customs ministries of the DRC and likely transit countries. The DRC
[6]
Customs Ministry has historically been underfunded (25) and was not able to = Artisanal processingt + Artisanal importst
publish such export data until 2015. Dataset S7 presents cobalt ore/concentrate
imports from the DRC as reported by the customs ministries of the DRC’s trading
partners. A comparison between the totals of Datasets S6a and S7a may indicate
that the customs ministries of the DRC and transit countries have not accounted Data, Materials, and Software Availability. All study data are included in the
for all cobalt ore/concentrate exports to all their trading partners, let alone China. article and/or Supporting Appendices and Datasets.
A final comparison of China Customs’ reported cobalt ore/concentrate imports
from the DRC is provided in Dataset S9, which contains the DRC’s exports of ACKNOWLEDGMENTS. I would like to thank Kim B. Shedd—the cobalt special-
cobalt ores/concentrates by company, as reported by the DRC Ministry of Mines ist—and Thomas R. Yager—the Congo (Kinshasa) mineral industry specialist—at the
provincial divisions for Katanga, Haut Katanga, and Lualaba. These data, however, USGS National Minerals Information Center for their invaluable experience and
are only available from 2009 through 2020. As a result, this analysis relies on time consulting with me on this article.
1. Roskill Information Services Ltd., Lithium-Ion Batteries–Fourth Edition (Roskill Information Services 11. UN Security Council, “November 2001 Addendum to the report of the panel of experts on the
Ltd., 2021). illegal exploitation of natural resources and other forms of wealth of the Democratic Republic of the
2. A. L. Gulley, E. A. McCullough, K. B. Shedd, China’s domestic and foreign influence in the global Congo” (2001).
cobalt supply chain. Resour. Policy 62, 317–323 (2019). 12. Pact, “PROMINES study–Artisanal mining in the Democratic Republic of Congo” (Pact, Inc.,
3. A. L. Gulley, N. T. Nassar, S. Xun, China, the United States, and competition for resources that enable Washington, DC, 2010).
emerging technologies. Proc. Natl. Acad. Sci. U.S.A. 115, 4111–4115 (2018). 13. C. Banza Lubaba Nkulu et al., Sustainability of artisanal mining of cobalt in DR Congo. Nat. Sustain.
4. N. T. Nassar et al., Evaluating the mineral commodity supply risk of the U.S. manufacturing sector. 1, 495–504 (2018).
Sci. Adv. 6, eaay8647 (2020). 14. Cutting cobalt. Nat. Energy 5, 825–825 (2020).
5. The Executive Office of the President, Executive order 13953 of September 30, 2020: Addressing 15. OECD, “Due diligence guidance for responsible supply chains of minerals from conflict-affected and
the threat to the domestic supply chain from foreign adversaries and supporting the domestic high-risk areas” (Organisation for Economic Co-operation and Development, Paris, France, 2013)
mining and processing industries. 85, 62539–62544 (Federal Register, 2020). 10.1787/9789264185050-en.
6. The Economist, “Electric leapfrog” (Econ, London, UK, 2020). 16. Z. Yangfan, “Responsible Cobalt Initiative: Facing challenges, sharing responsibility, and achieving
7. Darton, “Cobalt market review: 2008-2009” (Darton Commodities Ltd., Guildford, United Kingdom, win-win” (Chinese Chamber of Commerce for Metals, Minerals & Chemicals (CCCMC) Importers
2009). Exporters, Beijing, China, 2016).
8. A. L. Gulley, One hundred years of cobalt production in the Democratic Republic of the Congo. 17. Entreprise Générale du Cobalt, “EGC responsible sourcing standard” (Entreprise Générale du Cobalt,
Resour. Policy 79, 103007 (2022). Kolwezi, Democratic Republic of the Congo, 2021).
9. Global Witness, “Rush and ruin: The devastating mineral trade in Southern Katanga, DRC” (AFR 18. K. Turcheniuk, D. Bondarev, V. Singhal, G. Yushin, 10 years left for redesign of Li battery. Nature 559,
62/3183/2016, Amnesty International Ltd., London, UK, 2004). 467–471 (2018).
10. Amnesty International, “This is what we die for: Human rights abuses in the Democratic Republic of 19. U.S. Department of Energy, “Battery critical materials supply chain opportunities” (U.S. Department
the Congo power the global trade in cobalt” (2016). of Energy, Washington, DC, 2020).
10 of 11 https://doi.org/10.1073/pnas.2212037120 pnas.org
20. N. Tsurukawa, S. Prakash, A. Manhart, Social impacts of artisanal cobalt mining in Katanga, 48. Roskill Information Services Ltd., “The economics of cobalt” (Roskill Information Services Ltd.,
Democratic Republic of Congo. Öko-Institut eV–Inst. Appl. Ecol. Freibg. 49, 65 (2011). London, UK, 2007).
21. S. Goethals, J.-P. Okenda, R. Mbaya, “Chinese mining operations in Katanga” (Rights & 49. S&P Global Market Intelligence, “S&P Global Metals & Mining mineral asset database” (S&P Global
Accountability in Development (RAID), London, UK, 2009). Inc., New York City, NY, 2020).
22. B. Faber, B. Krause, R. S. De La Sierra, “Artisanal mining, livelihoods, and child labor in the cobalt 50. T. C. Frankel, The cobalt pipeline: Tracing the path from deadly hand-dug mines in Congo to
supply chain of the Democratic Republic of Congo” (2017). consumers’ phones and laptops. Washington Post, 2016.
23. BGR, “Mapping of the artisanal copper-cobalt mining sector in the provinces of Haut-Katanga and 51. A. Y. Mulimbi, Hunt for the ‘blood diamond of batteries’ impedes green energy push. New York
Lualaba in the Democratic Republic of the Congo” (2019). Times, 2022.
24. BGR, “Mining conditions and trading networks in artisanal copper-cobalt supply chains in the 52. Associated Press, Lawsuit claims Apple, Microsoft have profited from child miners. New York Post,
Democratic Republic of the Congo” (2021). 2019.
25. World Bank, “Democratic Republic of Congo: Growth with governance in the mining sector” (2008). 53. U.S. Department of Commerce, Minerals Yearbook 1932–1933 (U.S. Department of Commerce,
26. SEC, “Disclosing the use of conflict minerals” (2017). 1933).
27. L. Maiotti, B. Katz, T. Gillard, H. Koep-Andrieu, “Interconnected supply chains: A comprehensive look 54. National Research Council, Minerals, Critical Minerals, and the U.S. Economy (National Academies
at due diligence challenges and opportunities sourcing cobalt and copper from the Democratic Press, 2008), 10.17226/12034.
Republic of the Congo” (Organization for Economic Co-operation and Development (OECD), Paris, 55. S. M. Fortier et al., “Draft critical mineral list—Summary of methodology and background
France, 2019). information” (2018). published 30 September
28. S. Kara, “Is your phone tainted by the misery of the 35,000 children in Congo’s mines?” (Guard, 56. K. B. Shedd, E. A. McCullough, D. I. Bleiwas, Global trends affecting the supply security of cobalt.
London, UK, 2018). Min. Eng. 69, 37–42 (2017).
29. Darton, “Cobalt market review: 2018-2019” (Darton Commodities Ltd., Guildford, UK, 2019). 57. IHS Markit, "Global Trade Atlas" (IHS Markit Ltd, London, UK, 2022). published 18 December
30. Darton, “Cobalt market review: 2017-2018” (Darton Commodities Ltd., Guildford, UK, 2018). 58. A. Young, Gold into base metals: Productivity growth in the People’s Republic of China during the
31. Darton, “Cobalt market review: 2016-2017” (Darton Commodities Ltd., Guildford, UK, 2017). reform period. NBER Working Paper Series (National Bureau of Economic Research, Cambridge, MA,
32. Darton, “Cobalt market review: 2015-2016” (Darton Commodities Ltd., Guildford, UK, 2016). 2000).
33. Darton, “Cobalt market review: 2014-2015” (Darton Commodities Ltd., Guildford, UK, 2015). 59. G. Chow, Are Chinese official statistics reliable? CESifo Econ. Stud. 52, 396–414 (2006).
34. Y. Vanbrabant et al., “TRACE (TRACeability of hEterogenite) project report” (2009). 60. L. R. Klein, S. Özmucur, The estimation of China’s economic growth rate. J. Econ. Soc. Meas. 28,
35. PACT, “Confronting modern slavery in DRC-UK mineral supply chains” (Pact, Inc., Washington, DC, 2020). 187–202 (2011).
36. K. B. Shedd, “Cobalt” in 2005 Minerals Yearbook (US Geological Survey, 2008). 61. X. Xu, Understanding China’s official statistics. China Econ. J. 6, 63–79 (2013).
37. K. B. Shedd, “Cobalt” in 2003 Minerals Yearbook (U.S, Geological Survey, 2006). 62. J. N. Koch-weser, “The reliability of China’s economic data: An analysis of national output” (U.S.-
38. C. Berke, J. Pulkowski, N. Martin, “Natural resources in the Democratic Republic of the Congo–A China Economic and Security Review Commission, Washington, DC, 2013).
development potential?” (2007). 63. J. Fernald, I. Malkin, M. Spiegel, “Economic Letter: On the reliability of Chinese output figures”
39. S. Al Barazi et al., Cobalt from the DRC–Potential, risks and significance for the global cobalt market. (Federal Reserve Bank San Francisco, San Francisco, CA, 2013).
DERA Commod. TopNews 53, 1–15 (2017). 64. D. Plekhanov, Quality of China’s official statistics: A brief review of academic perspectives.
40. Darton, “Cobalt market review: 2020-2021” (2021). Copenhagen J. Asian Stud. 35, 76–101 (2017).
41. Darton, “Cobalt market review: 2019-2020” (2020). 65. K. B. Shedd, “Cobalt” in 2002 Minerals Yearbook (U.S. Geological Survey, 2005).
42. N. Niarchos, The dark side of Congo’s cobalt rush. New Yorker, 5 2021. published 24 May 66. K. B. Shedd, “Cobalt” in 2008 Minerals Yearbook (U.S. Geological Survey, 2011).
43. J. K. Musao, “The problem of artisanal mining in the province of Katanga (case of the district of 67. A. L. Gulley, Valuing environmental impacts of mercury emissions from gold mining: Dollar
Kolwezi) in in French” (University of Lubumbashi, Lubumbashi, Democratic Republic of the Congo, per troy ounce estimates for twelve open-pit, small-scale, and artisanal mining sites. Resour.
2009). Policy 52, 266–272 (2017).
44. G. Heppel, “Artisanal mining will balance the cobalt market” (CRU Group, London, UK, 2018). 68. D. Van Brusselen et al., Metal mining and birth defects: A case-control study in Lubumbashi,
45. CRU, “Cobalt market report 2021” (2022). Democratic Republic of the Congo. Lancet Planet. Heal. 4, e158–e167 (2020).
46. C. Gaffar, I. Kaempfer, “Opportunities for businesses to promote child rights in cobalt artisanal and 69. P. Landrigan et al., Reducing disease and death from Artisanal and Small–Scale Mining (ASM)–The
small-scale mining” (Save the Children, Berlin, Germany, 2021). urgent need for responsible mining in the context of growing global demand for minerals and
47. K. B. Shedd, “Cobalt” in 2004 Minerals Yearbook (U.S. Geological Survey, 2007). metals for climate change mitigation. Environ. Heal. 21, 78 (2022).
Downloaded from https://www.pnas.org by 45.226.238.191 on July 6, 2023 from IP address 45.226.238.191.