1 Introduction

Common wisdom suggests that it is possible to manage things that can be measured (Lowenstein 1996). The same phenomenon has been identified in relation to the management of sustainability in companies (Elgie et al. 2021; Lammerant et al. 2021). Quantified measures help to set measurable targets, which helps to manage change. Therefore, in addition to qualitative methods, quantitative methods are important in addressing social sustainability from a management perspective. Qualitative social science methods, such as interviews and participatory observation, work well when an analyst wants to study social implications locally or in an otherwise bounded setting (Tracy 2019). However, the situation becomes more complex when there is a need to understand similar implications at the level of whole product supply chains, which typically combine numerous socio-cultural environments (Härri et al. 2023). In such situations, quantitative methods like S-LCA may be more useful.

Life cycle thinking (LCT) and life cycle management (LCM) frameworks aim at covering the whole life cycle of a product from raw material extraction to end-of-life to avoid shifting the impact from one phase to another (UNEP/SETAC 2009; Massari et al. 2023). Successful sustainability management often requires the integration of diverse forms of social, environmental and economic information (Seuring and Müller 2008). Compared to environmental and economic sustainability, social sustainability is still the least understood aspect of sustainability in general (Ajmal et al. 2018; Alexander et al. 2020). However, there have been continuous efforts to develop S-LCA methodology for 20 years (Ramos Huarachi et al. 2020). According to UNEP (2020), S-LCA can be defined to assess positive and negative social and socio-economic impacts related to products and services. Social impacts in LCM framework can be also studies in an organizational level (SO-LCA) to, e.g., improve the social performance of the company (D’Eusanio et al. 2022). One of the goals of S-LCA development is to connect it to more comprehensive sustainability LCA (LCSA) together with environmental LCA and life cycle costing (Valdivia et al. 2021). The current implementation of S-LCA is based on the UNEP (2020) Guidelines for Social Life Cycle Assessment of Products and Organizations and ISO 14040 (ISO 2006) framework, but a new ISO 14075 standard for S-LCA is under development (ISO 2023).

There are several ways to conduct the impact assessment in S-LCA (S-LCIA). The S-LCIA is typically divided into two main types: impact pathway and reference scale, both of which often aim to quantify social impacts (Chhipi-Shrestha et al. 2015). In addition, there are fully qualitative approaches that draw on social sciences methodologies (Baumann and Arvidsson 2020). The inventory and impact assessment indicators in S-LCA can be quantitative, semi-quantitative, or qualitative or combine several different types of data (UNEP 2020). The impact pathway S-LCIA is used to assess the potential short- and long-term social consequences of an activity using cause–effect model chains (Iofrida et al. 2018; Sureau et al. 2020). Human wellbeing is considered as the area of protection (Schaubroeck and Rugani 2017). Disability-adjusted life years, quality-adjusted life years, Social Metric for Life Cycle, the Wilkinson pathway and the Preston pathway are some examples of impact pathway S-LCIA (Weidema 2006, 2018, 2023; Feschet et al. 2013; Bocoum et al. 2015; Arvidsson et al. 2018, 2022; de Araujo et al. 2021). Frameworks for assessing human wellbeing in impact pathway S-LCIA has also been proposed because the current S-LCIA has not proceeded beyond health impacts to realize the full potential impacts on human wellbeing (Lindkvist and Ekener 2023). Reference scale S-LCIA uses qualitative scales to quantify potential social risks or social performance in the product chain (Russo Garrido et al. 2018). It focuses more on the immediate activities and their social impacts (Iofrida et al. 2018; UNEP 2020). There are different existing reference scales such as the subcategory assessment method (SAM) and Product Social Impact Assessment, or it can be defined based on the research needs and socio-political context (Ramirez et al. 2014, 2016; Fontes et al. 2018; UNEP 2020; Tsalidis et al. 2023; Luna Ostos et al. 2024). Secondary data for reference scale S-LCA can be obtained using different databases (UNEP 2020). Databases that have been specifically developed for S-LCA are the Product Social Impact Life Cycle Assessment Database (PSILCA) and the Social Hotspot Database (SHDB) (Benoit-Norris et al. 2012; Maister et al. 2020). Both of these databases use Multi-Regional Input–Output modelling (MRIO) (NewEarth B 2019; Maister et al. 2020). The largest benefit of using S-LCA databases is the relatively easy data collection compared to other approaches (Schlör et al. 2018). S-LCA databases use ordinal scales to quantify qualitative impacts, which can lead to challenges since the distance between different values is unknown (Arvidsson 2019). SHDB can be used to assess different social risks and opportunities at the sector and country levels (Benoit-Norris et al. 2012), but it does not offer accurate quantified S-LCA results. However, it can be used to identify social hotspots and the most important stakeholder groups to include in the data collection, estimation and analysis of the final results (UNEP 2022). The S-LCIA used in SHDB has also been further developed in non-database context regarding labour risks (Blackstone et al. 2023).

This study focuses on EV battery material mining and CAM production processes which are upstream processes for automotive and mobility sectors through EV battery manufacturing. It has been studied that S-LCA can support responsible sourcing of raw materials, e.g., metals, in Europe (Di Noi et al. 2020). Some previous S-LCA research focusing on the EV battery material mining have been conducted (Mancini et al. 2021; Orola et al. 2022; Roche et al. 2023; Agusdinata et al. 2023). In addition, the transportation sector has been studied from multiple perspectives to identify different improvement opportunities related to S-LCA from the sectoral perspective (Ekener-Petersen et al. 2014; Zanchi et al. 2018; Karlewski et al. 2019; Gompf et al. 2020, 2022). There is an ongoing work to develop S-LCA frameworks for different sectors like circular economy, bioeconomy, construction and tourism sector (Backes and Traverso 2023; Luthin et al. 2023; Rebolledo-Leiva et al. 2023; Miralles et al. 2024). A framework for mobility scenarios has already been developed (Bouillass et al. 2021). Developing sectoral frameworks is important since different stakeholder groups, subcategories and impact categories should be prioritized differently for distinct sectors (Bouillass et al. 2021; Lehmann et al. 2024). Framework for the battery sector has not been researched yet.

If going beyond indicators, one potentially important aspect of battery and battery material sector is price volatility. Price volatility refers to variation in the trading prices of products. Some minerals are more volatile to price changes than others (García and Guzmán 2020; Sohag et al. 2023). Battery materials and related intermediate product prices have varied significantly in recent years. During the last 5 years, the prices of cobalt, lithium and nickel have fluctuated more than 50% from the average price (Fig. 1) (USGS 2023c, 2023d, 2023b). Occasionally, the prices of the intermediate EV battery material products have even been lower than the prices of the raw materials (Greenfield and Scott 2023).

Fig. 1
figure 1

The mineral spot prices have varied significantly in recent years (USGS 2023a, d, b, c)

Full size image

Since the inventory analysis in SHDB is based on material prices (Benoit-Norris et al. 2012), the volatility of prices might affect the results significantly. SHDB uses 1 USD as a reference flow. The social risks related to different processes are scaled based on the worker hour activity variable and different income levels in different countries (Benoit-Norris et al. 2012). This means that a more expensive product consumes more working hours and therefore causes more social risks. In addition, if the same price is used for two different countries, the country with a lower income level will have higher social risks. The benefit of this approach is that it allows approximating the worker hours without having to collect data about the actual hours as worker hours can also be used as single activity variable in S-LCA without connecting it to monetary values (Ugaya et al. 2011; Petti et al. 2018). The S-LCA modelling with monetary units can be broken into two parts: the activity variable and the price of the product. The activity variable was developed to connect economic and occupational variables (Norris 2006). In general, the challenge in using worker hours as an activity variable is that not all indicators used in S-LCA are related to worker hours, as S-LCA includes stakeholders other than workers (e.g. local communities and society) (Ciroth et al. 2019). However, worker hours are also used with other stakeholder groups, such as local communities, in the SHDB and PSILCA databases (Benoit-Norris et al. 2012; Maister et al. 2020). Therefore, results based on the activity variable might be difficult to interpret (Tragnone et al. 2023). The challenge with using monetary value as a reference flow is that the price of a mineral product is based on more than just the value of work, including the supply, demand and energy prices (Sverdrup et al. 2015). In addition, natural disasters, conflicts, global economic crises, market manipulation and bans on mineral exports before processing can impact mineral prices (Lim et al. 2021; USGS 2013).

Koese et al. (2023) recognized the possible impacts of price volatility in S-LCA in the context of vanadium redox batteries. Popien (2023) assessed the impacts on the total cost of batteries if the added value of different materials was changed. Added value is an activity variable used in S-LCA to scale social impacts (UNEP 2020). Otherwise, mineral price volatility has not been previously addressed in battery material studies using S-LCA databases. Previous battery material studies have focused on global production chains in which the manufacturing phase occurs in Asia or Germany (Thies et al. 2019, 2021; Arvidsson et al. 2022; Shi et al. 2023; Popien et al. 2023) The social impacts of mineral raw material extraction phase of EV battery materials in Europe have not been considered previously. Accordingly, the aim of this study is to determine how price volatility impacts the S-LCA results in monetary unit-based modelling in order to take those impacts into account in S-LCA method development. The aim is also to offer some insight for the potential battery sector product category development. EU has recently released a regulation concerning batteries which emphasizes the importance of studying their social impacts and mentions S-LCA as one of the instruments to assess the social impacts (European Commission 2023). Therefore, it would be important to develop specific product category rules for the S-LCA of batteries and pay attention to the possible special features of battery industry sector, e.g., the high price volatility of the battery materials. SHDB is typically used for social hotspot identification in product supply chain and in preliminary research to identify the most important processes for a more site-specific data collection in S-LCA (Benoit-Norris et al. 2012; UNEP 2020, 2022) or in upstream data collection if site-specific data is not available (EPD INTERNATIONAL AB 2023). There is a potential risk that the impacts of price volatility could lead to incorrect decisions. Cathode active material (CAM) production supply chains in Finland and globally were compared to determine whether price volatility would have different impacts in different regions. Finland and the largest battery material production countries have different income levels (World Population Review 2023), and therefore, price volatility may have different impacts due to using worker hours as an activity variable.

2 Materials and methods

2.1 Goal and scope definition

The goal of this S-LCA was to compare theoretical battery CAM supply chains in Finland and globally (in largest producer countries) to identify the impacts of price volatility on the use of reference scale approach based on monetary units. The target audience of this S-LCA study is academia aiming for further S-LCA method development since the implications of price volatility can have significant effect on the results of S-LCA studies. The S-LCA supply chains were modelled using SHDB.

NMC 811 CAM was selected as a case example since it contains multiple minerals that have recognized social impacts as well as high price volatility. The largest production countries for cobalt, nickel, lithium and manganese according to USGS (2023d, a, b, c) are the Democratic Republic of the Congo (DRC), Indonesia, Australia and South Africa. China is the largest EV battery manufacturing country (International Energy Agency 2023). Finland has ambitious plans for increasing its battery material industry (Työ-ja elinkeinoministeriö 2021). Both cobalt and nickel are mined in Finland, and Finland also has economically viable lithium reserves, which it plans to utilize in the future (Työ-ja elinkeinoministeriö 2022). There are plans to build NMC precursor CAM (pCAM) and CAM factories in multiple regions in Finland (Aluehallintovirasto 2023; Finnish Battery Chemicals 2021). In the case of the Finnish supply chain, the S-LCA model is based on published plans since the battery material industry in Finland is still in the developmental phase. As many of countries that mine battery minerals are attempting to move away from exporting mineral ores or mineral concentrates on further processed products in the global supply chain (International Energy Agency 2022b; Reid 2021), it is assumed that the mineral products will be refined in their country of origin.

The function of the studied system is to produce CAM for an EV battery. The functional unit of the study is 1 USD worth of CAM because it is an intermediate product, and the function of intermediate products is often difficult to define precisely (Marmiroli et al. 2022). The inventory data for NMC 811 single-crystal CAM production was collected according to Kallitsis et al. (2020), Ellingsen et al. (2014) and Majeau-Bettez et al. (2011). NMC battery technology that uses single-crystal technology instead of polycrystal has been studied as a way to increase the durability of EV batteries (Saunders et al. 2022). In addition, Ecoinvent 3.8 database was utilized for data collection to identify material and energy flows in physical terms. The monetary values of the physical flows needed for SHDB were calculated using different reports and the World Bank’s World Integrated Trade Solution (WITS) database. The inventory results were scaled according to the functional unit of 1 USD, and the SHDB S-LCIA was used to calculate social risks assigned to that value (Fig. 2).

Fig. 2
figure 2

The inventory was first collected in physical units and then modelled in monetary units

Full size image

The studied product system includes the mining of the raw materials, intermediate production, CAM production and related electricity production and fuel extraction for thermal energy production. The system boundary is described in Fig. 3. Life cycle phases from cathode production to end of life are excluded. The construction of the CAM production plant was excluded since the construction materials are complex and often include multiple input and outputs (Hosseinijou et al. 2014; Backes and Traverso 2023) which also makes the estimation of price of the building challenging. Similarly, the distribution of fuel for energy production was excluded since the price estimation of the pipeline needed for natural gas would have added too much uncertainty to the results. Specific processes related to extraction of raw materials other than mining and fuel for thermal energy are excluded. However, since SHDB uses MRIO, it contains data about related supply chains through purchases between different sectors and geographical locations (NewEarth B 2019; Tarne et al. 2018). The countries of origin for material inputs for studied processes are obtained from WITS. The global supply chain is based on the largest separate production countries of these minerals and the largest countries of origin that import to the countries where minerals are processed to make other materials. It has been assumed that the intermediate production happens in mining countries since these countries have increasingly banned the export of unprocessed raw materials (Nguyen and Liu 2023; Reid 2021).

Fig. 3
figure 3

The system boundary of CAM production

Full size image

It was assumed that wood chip fuel is used in steam production in Finland. The other option would have been natural gas which is mainly imported from Russia (Finnish Battery Chemicals 2021; International Energy Agency 2022a). However, due to the current conflict between Russia and Ukraine, it was assumed that the need for self-sufficiency affects decision-making in projects that are still in the developmental phase (International Energy Agency 2022a). It was further assumed that the nickel grade for nickel matte is 70% and that for nickel ore/concentrate is 15% (British Geological Survey 2008). Nickel matte was used as the nickel intermediate in the model since the prices for nickel ore and concentrate could not be separated. It has been assumed that allocation is not needed since the study includes social impacts that are not measured on product level therefore making the co-production irrelevant (UNEP 2020).

SHDB includes the following stakeholder groups which were also included in this study: workers, local community and society (Benoit-Norris et al. 2012). Stakeholders were not involved in the process even though it is recommended in the S-LCA guidelines (Benoit-Norris et al. 2012; UNEP 2020). This is because the focus was more on the method itself than on the exact results. SHDB aggregates social impacts by considering different stakeholders and social impacts in different social impact categories, which differ slightly from UNEP S-LCA guidelines (Benoit-Norris et al. 2012; UNEP 2020), which include labour rights and decent work, health and safety, society, governance, community and socio-economic contributions (Benoit-Norris et al. 2012; NewEarth B 2019).

A sensitivity analysis was performed to assess the impacts of mineral and energy price volatility on the model. Data quality and uncertainty was assessed qualitatively using a data quality matrix, which is a tool developed for qualitative uncertainty analysis in LCA (UNEP 2021). Lower data quality was occasionally accepted because the main aim of the study was to focus on impacts of price volatility.

2.2 Inventory analysis

The inventory (Supplementary material 1 Tables 1 and 2) was collected based on Kallitsis et al. (2020). The inventory analysis was first performed in physical units (kg, MJ, kWh). Then, the value of different materials was calculated in monetary units (USD) and connected to the generic country and sector level data in SHDB (Supplementary material 1 Tables 3 and 4) using activity variable to get the data from social hotspots of the CAM supply chains. Since the SHDB was used, the data about activity variable, inventory indicators and weighting were provided by the database. SHDB does not include data from the DRC (Table 1). Instead, data from Zambia was used. Zambia is located in the same cobalt-mining area (Copperbelt) as the DRC (Crundwell et al. 2020). Water used in CAM production processes was excluded due to lack of data about industrial water prices.

Table 1 Location of processes in the global supply chain
Full size table
Table 2 Reference scale S-LCIA used in SHDB (Benoit-Norris et al. 2012)
Full size table
Table 3 Included social impact categories and social impact subcategories (NewEarth B 2022)
Full size table
Table 4 The values used in the mineral price volatility sensitivity analysis
Full size table

No site-specific or primary data was collected or used since the aim of the study is merely assess the impacts of the price volatility to S-LCA results. The value of different materials was based on global market prices and global product import and export value statistics (Supplementary material 2). Five-year averages were used, if available, to balance the model since there have been multiple global events in recent years that have affected the global markets and prices. The five-year averages were also used in exchange rates when changing to USD from other currencies (Supplementary material 1 Table 5). The challenge with import and export prices is that they also include other costs in addition to the product values. These are insurance and freight prices in case of import prices. Export prices are based on freight-on-board price (The World Bank 2013). Using the producer price data which is recommended by the SHDB (Benoit-Norris et al. 2012) was not possible due to poor availability. Therefore, import and export data has been applied. Additional data from literature and news reports were collected for result interpretation. The processes with the highest social risks were prioritized.

Table 5 The impacts of price volatility on social risks impact contributions in the global supply chain
Full size table

Mining phase was modelled by estimating the mineral content of the mined rock and calculating the needed amount of rock for mineral content of the produced material (Supplementary material 1 Table 6). Transportation was assumed to be done by truck and cargo ship to and from nearest port. The assumed road transportation distances are presented in the supplementary material (Supplementary material 1 Table 7). The price of water transportation was assumed to be 2034 USD/container (UNCTAD 2022) and max load 24,000 kg/container (Menon 2022; iContainers 2023). The calculations for thermal energy production and fuels used in different processes are presented in the supplementary material (Supplementary material 1 Table 8). Different energy production fuels were used in different countries depending on the main energy production technology (Supplementary material Tables 3 and 4) (Ratshomo and Nembahe 2021; Secretariate General of the National Energy Council 2021; Australian energy regulator 2023).

2.3 Impact assessment

SHDB’s Social Hotspot 2022 Subcategory Method/Equal was used as S-LCIA method. The results were calculated as medium risk hours equivalent (mrheq) and characterized according to the method (Table 2). The study included workers, local communities and society as stakeholder groups, but the results were not aggregated between the groups. All indicators, subcategories and stakeholders available in SHDB were included except for socio-economic contributions because they were excluded from the weighting method (Table 3). Weighting method from SHDB was used. The focus was on the impacts of price volatility on the total result without addressing different social impact subcategories.

3 Results

The results indicate that there are less social risks in CAM production in Finland compared to global production (Fig. 4). The weighted results for the global supply chain were 0.66 mrheq, while the results for the Finnish supply chain 0.098 mrheq. Health and safety was the largest social impact category in both supply chains. Labour rights and decent work and governance were the second largest social impact categories. In Finnish supply chain, the smallest impact category was society, whereas in global supply chain, it was community.

Fig. 4
figure 4

The social risks in global and Finnish CAM supply chains

Full size image

The results for 15 processes with the highest social risks in the global and Finnish supply chains are presented in Fig. 5. The processes are named according to the material flows, but the results are based on the country and sector level data. Due to high number of processes, only those with the highest risks are included in the figure. In both supply chains, CAM production is the highest contributor to the total results even though there are not any specific social risks related to the CAM production. However, the CAM production constitutes over 50% (26 USD) of the total inventory in monetary units. In the global supply chain, cobalt hydroxide and cobalt sulphate production have high social risks. Cobalt ore mining is only sixth highest on the list which was unexpected because the social issues related to artisanal cobalt mining are often brought up by media, human rights organizations, and research (Amnesty International 2016; Kara 2018; Banza Lubaba Nkulu et al. 2018; Sovacool 2019; Katz-Lavigne 2020) whereas the cobalt processing is less discussed. However, the price of the cobalt ore is significantly lower compared to cobalt hydroxide and cobalt sulphate which probably causes these results. In addition, nickel matte and nickel sulphate production have high social risks. The social impacts of nickel industry in Indonesia has also gained attention in research community (Hudayana et al. 2020; Camba 2021; Glynn and Maimunah 2023); therefore, the results are in line with that. In the Finnish supply chain, nickel matte production, lithium ore mining, nickel sulphate production, lithium carbonate production and nickel ore mining have the highest social risks. There is no research indicating that there are significant social issues related to, e.g., working conditions in the mining or processing these materials even though there is some general resistance against mining in industry in Finland from local communities mainly due to environmental and recreational reasons (Lassila 2021; Mononen and Sairinen 2021).

Fig. 5
figure 5

Top 15 processes with highest social risks in the a global and b Finnish supply chains based on medium risk hours

Full size image

4 Discussion

4.1 Impacts of mineral price volatility on the results

Sensitivity analysis was conducted to assess the impacts of mineral price volatility. The monetary values for cobalt, nickel, lithium and manganese ores and intermediate products were changed (Table 4). The monetary values used in sensitivity analysis are based on import and export value changes from the 5-year average in the global supply chain. However, the data for nickel ore from Indonesia were changed to Philippines since there has been an export ban on unprocessed ores from Indonesia for the last few years, which impacts the quality of statistical data. Additionally, the value for manganese oxide was used instead manganese sulphate due to the lack of 5-year data.

Different sensitivity analyses were conducted (Fig. 6). The prices for the materials are presented as follows:

  • China: Minerals are completely processed in China using natural gas as thermal energy source except for nickel due to the ore exportation ban in Indonesia.

  • Min: All values changed to 5-year minimum values.

  • Max: All values changed to 5-year maximum values.

  • Lithium min: Only lithium products changed to minimum values.

  • Lithium max: Only lithium products changed to maximum values.

Fig. 6
figure 6

Sensitivity analysis of mineral price volatility in the global and Finnish supply chains

Full size image

The value of CAM was adjusted by calculating how much the value of the materials presented in Table 4 increased and changing the CAM value accordingly.

The sensitivity analysis, in which the mineral processing is located in China, had lower social impacts compared to global supply chain (Fig. 6). The result of the sensitivity analysis implies that price volatility might have a larger impact on the social risks on the global supply chain than on the Finnish supply chain. This is probably because there is smaller difference between social risk levels in Finnish supply chain compared to global supply chain. Thus, when, e.g., the lithium price is changed, the relative share of social impacts of different processes scaled according to the functional unit changes less in Finnish supply chain. The biggest challenge with the S-LCA model and the price volatility is that when the mineral price is decreased in most of the sensitivity analyses, the social impacts increase. This also goes other way around. A possible explanation for this is that cobalt has a significantly higher social risks but lower price volatility compared to some other materials. If the relative share of cobalt decreases, the overall social risks decrease. Lithium has the highest price volatility of the studied mineral, so it alone can cause this impact as seen in the sensitivity analyses where only the price of the lithium materials are changed.

4.2 Impacts of energy price volatility on the results

In addition to mineral prices, a sensitivity analysis of energy prices in the global and Finnish supply chains was also conducted by changing all the energy prices by 50% (Fig. 7). The value of CAM and CAM materials was adjusted by calculating how much the value of electricity increased, changing the value of CAM and different electricity-demanding materials accordingly. The increase in energy price decreased the social risks in the global supply chain and other way around. This is likely because the relative share of cobalt products from the DRC decreased in the global supply chain even though the social risks did not increase in reality. The social risk in the Finnish supply chain increased slightly when increasing or decreasing the energy prices.

Fig. 7
figure 7

The impacts of energy price volatility on the global and Finnish supply chains

Full size image

4.3 Impacts of price volatility on the significance of the different processes

It was also studied whether the mineral and energy price volatility might impact different processes and how they contribute to the results in the global supply chain in the sensitivity analyses described above (Table 5). This is important because one of the main uses of SHDB is to identify the hotspots in the supply chain for company supply chain management or to find the most important data collection points for research. It has been perceived that the fluctuation in price levels between different countries or price volatility can impact the results of S-LCA using monetary values as reference flow even though the risk level itself does not change (Thies et al. 2019; Springer et al. 2024). The ranking of social risks of some processes changed, but the top 20 most significant processes were mainly the same despite the price volatility changes. However, the ranking of social risks of the processes changed in different sensitivity analyses. The changes in mineral prices affected the ranking of lithium ore and lithium carbonate significantly. The social impacts of lithium extraction in Chile are often reported (Agusdinata and Liu 2023). However, the social impacts of lithium mining in Australia have been less studied (Agusdinata et al. 2018). Lithium mining possesses some potential social risks for indigenous communities since some of the mines are located on the aboriginal land (The Government of Western Australia 2019; Butler et al. 2021). Many of the other materials were slightly affected but usually just by a rank or two. The energy price changes affected the ranking of energy-related processes, but also, the ranking of ocean transportation of nickel was affected.

4.4 Comparison with other battery-related S-LCA studies

Comparing the results of the study with other S-LCA battery studies is difficult because of different system boundaries and methodological choices because common framework does not exist yet. In monetary unit-based S-LCA, price volatility and overall changes in prices can also impact comparability. Most of the other S-LCA battery studies that have utilized databases have used added value as the activity variable or included their own additional weighting method (Popien et al. 2023; Shi et al. 2023; Thies et al. 2021) According to Shi et al. (2023), the social risks for materials in developed countries are usually 1–5 mrheq/USD, and they are greater than 5 mrheq/USD for developing countries in SHDB.

4.5 Implications and limitations

The aim of the study was to study the impacts of price volatility on monetary-based modelling, which is used in both the SHDB and PSILCA S-LCA databases. The PSILCA database utilizes worker hours as the activity variable, similarly to SHDB (Maister et al. 2020), which means that the results of this study can be applicable to the PSILCA database to some extent. The study implies that using S-LCA in sectors with high price volatility may cause some challenges for S-LCA result interpretation when using S-LCA databases. Monetary values are not commonly used as reference flow in other types of S-LCA.

The second aim of this study was to provide some insights for product category rules development for battery sector especially in EV context. Mineral price volatility is one of features related to EV battery material markets. Recognizing the special features regarding different sectors in S-LCA is important. This matters if for example, there is an aim at creating specific product category rules as used in E-LCA in environmental product declarations. Since the EU battery act is encouraging research of social impacts of batteries using S-LCA, creating product category rules for batteries is important. The first product category rules for social product declarations have been released, and they recommend using hotspot analysis in a data collection if site-specific data cannot be obtained from suppliers (EPD International AB 2023). According to the product category rules of the rolling stock when using generic data, the reference data should be as new as possible (EPD International AB 2023); however, with price volatile materials, this may cause challenges when comparing SPD’s from different years with each other. The main implication of the study is that price volatility should be considered if developing the product category rules for battery and mining sectors. In addition, this might affect other price volatile industry sectors like energy production. The limitation of the study is that the CAM production is just part of the battery production. Therefore, it is difficult to predict how extending the research to include the whole EV battery, automotive or mobility system would affect the results.

Based on the results the high price volatility can affect the relative ranking of processes by their contribution of the total results. A study by Springer et al. (2024) indicated that price change of component with high price volatility also impacted the share of different social risk in relation to total risk level and affected whether the component manufacturing was regarded as social hotspot or not. If hotspot assessment is used in prioritizing the data collection among large amounts of processes, this may lead to prioritizing wrong processes in, e.g., site-specific data collection if the impacts of price volatility are not considered. Some studies also use S-LCA databases as basis for their own methods and the price volatility may affect their results as well especially if they aim at measuring social impacts or comparing two product systems with each other if one of them includes materials with high price volatility. The study implies that all S-LCA studies especially focusing on the sectors with high price volatility like minerals should pay special attention to sensitivity analyses when using monetary flow-based modelling. Using sensitivity analyses to recognize the impact of price volatility in result interpretation is also suggested by Springer et al. (2024).

If the monetary values are preferred as a reference flow in the S-LCA databases, the impacts of price volatility in different sectors should be studied. Based on the results considering ways to reduce the impacts of price volatility in monetary value-based S-LCA models or avoiding using monetary values as reference flow in very volatile sectors would be recommended. Some S-LCA studies have utilized added value as an activity variable instead of product price (Benoit-Norris et al. 2012; Thies et al. 2019). However, using added value does not decrease the price fluctuation between countries and the impacts on S-LCA results (Thies et al. 2019). Therefore, it is difficult to estimate whether using added value as activity variable would affect the challenges related to global price volatility. In addition, using added value as an activity variable might lead to identifying only countries and sectors with higher added value (Costa et al. 2022). The challenges caused by price volatility has been recognized in hotspot assessment and one option is to use mass-based reference flow as in E-LCA (Blackstone et al. 2023). Based on the results, it is also recommended to consider this option in price volatile sectors when collecting data about worker hours from primary sources is possible.

5 Conclusions

The study aimed at assessing the implications of price volatility on monetary value-based S-LCA modelling which is used in SHDB and PSILCA databases. Global and Finnish supply chains were compared to see if the results were distinct in countries with different income levels and how that affected the results of the comparative S-LCA. Price volatility caused smaller variation to the results of Finnish supply chain compared to global supply chain in the S-LCA model. However, the outcome of the comparison, which indicated that there were less potential social risks in Finnish supply chain compared to the global one, stayed the same.

The sensitivity analyses were the main focus of this study. The sensitivity analyses indicate that the changes in mineral and electricity prices affected the S-LCA results and the relative contribution of different processes. The price volatility caused significant variation to results especially in case of lithium due to high price fluctuation. This may cause some challenges for the interpretation of the results, e.g., in hotspot identification for more detailed data collection. To avoid this, the importance of sensitivity analyses should be emphasized in price volatile sectors.

The limitation of this study is that it only focuses on one part of the EV battery material life cycle, and therefore, it is difficult to predict what the implications could be on wider scope. The insights of this study should be considered if product category rules for mining and battery-related industrial sectors as well as automotive and mobility-related sectors were developed. In that case, using other that monetary units, e.g., physical units, as reference flow should be examined. More research is needed to recognize other special features related to battery material S-LCA for SPD’s.