Monitoring the consumption footprint of countries to support policy‐making: An assessment of data availability in Germany

The European Green Deal and the German Resource Efficiency Programme both aim at decoupling resource consumption and associated environmental burdens from economic growth. Monitoring the progress of such policies requires robust estimates of environmental pressures and impacts, both from a domestic and a footprint perspective. Building on the life cycle assessment‐based consumption footprint (CoF) indicator, developed by the European Commission Joint Research Centre, we assess the environmental impacts of Germany's consumption in the areas of food, mobility, housing, household goods, and appliances during the period 2010–2018. A comparison between European and national consumption statistics revealed some differences in terms of data composition, granularity, consumption intensities, and calculated environmental impacts. Using national data sources results in slightly lower environmental impacts (e.g., due to differences in the assessment scope of national statistics) and requires some data preparation to match the CoF indicator. Emerging consumption trends can be highlighted using national data. Both data sources converge on main trends: Germany transgresses the safe operating space for several impact categories, with consumption of food, household goods, and mobility being the main drivers. Domestic impacts have decreased over time at the expense of outsourcing environmental pressures and impacts to other countries. The CoF indicator could complement resource monitoring frameworks and might be further aligned to the national context using country‐level consumption statistics and life cycle inventory data.

the 17 Sustainable Development Goals (SDGs) represent the results of international consensus building on the environmental, social, and economic dimensions of sustainable development (United Nations, 2015).SDG 12 focuses particularly on ensuring sustainable consumption and production patterns as part of the broader sustainability agenda.
In the European Union (EU), the European Green Deal aims at ensuring climate neutrality by 2050, decoupling economic growth from resource use, and leaving no person and place behind (EC, 2019a).Sustainable consumption and production is a transversal issue among multiple ambitions of the EU Green Deal, such as sustainable food production (Farm to Fork Strategy) (EC, 2020a), biodiversity conservation (EU Biodiversity Strategy for 2030) (EC, 2020b), circular economy (EC, 2020c), zero pollution (EC, 2021a), or sustainable consumption choices (EC, 2020d).The European Green Deal highlights the need to address societal challenges holistically by considering the entire value chain of products and by taking a consumptionbased approach.In this context, quantifying and monitoring the environmental impacts associated with the consumption of EU citizens is of utmost importance (European Parliament, 2021 ).
This paper aims at assessing the environmental impacts of Germany's household consumption by building upon the EU consumption footprint (CoF) indicator (Sala & Sanyé-Mengual, 2023) and focusing on the potential use of national data sources.The CoF is increasingly relevant for ensuring evidence-and science-based support to policies, for example, it is one of the headline indicators in the 8th EU Environment Action Program (EC, 2022a).Our study is novel as it is the first attempt to examine in depth the availability of national consumption data for the CoF indicator.We also test the use of the existing CoF indicator in Germany including possible adaptations to the national-level context.
The specific objectives include: 1. Identify and compare German national data sources with EU statistics used in the CoF in terms of their composition and differences in household consumption intensities, including the identification of missing products based on national statistics, 2. Evaluate the differences in consumption intensities and resulting environmental impacts between a national data-based CoF and the original EU-wide CoF, 3. Provide an overview of the CoF of Germany from decoupling and planetary boundaries perspectives, 4. Compare results with other footprint estimates and discuss the potential use of the CoF indicator in the German policy context.
The structure of the paper is as follows: First, we provide a brief background of studies assessing the environmental impacts of consumption (Section 2).Second, we describe the methods used in this study to adapt the CoF indicator for Germany (Section 3).Finally, we discuss results (Section 4) and present conclusions for future work (Section 5).

ASSESSING THE ENVIRONMENTAL IMPACTS OF CONSUMPTION
Several studies have estimated the environmental impacts of consumption at macro-scale.For example, multi-regional input-output (MRIO) models have been used to assess the environmental pressures associated with trade flows between countries and at the level of economic sectors.
The EXIOBASE database (Wood et al., 2015) has been used to evaluate environmental footprints of countries (Tukker et al., 2014), household consumption (Ivanova et al., 2016), and with a focus on individual environmental issues (e.g., land use; Bjelle et al., 2020).The Eora database (Lenzen et al., 2013) (and more recently GLORIA; IEL, 2021) are used by the UN to examine environmental footprints for various countries (Piñero et al., 2019).However, using "top-down" approaches based on the economic flows between sectors is also associated with certain limitations such as, for example, low product granularity or a small number of environmental impacts that can be examined to date (due to the more limited coverage of elementary flows compared to process-based life cycle assessment; LCA) (Beylot et al., 2019;Castellani et al., 2019a).
An LCA-based approach (ISO, 2006a(ISO, , 2006b) ) for capturing the environmental impacts of consumption has been developed by the European Commission (EC) Joint Research Centre (JRC) in the form of the CoF indicator (Sala et al., 2019;Sala & Castellani 2019;Sala & Sanyé-Mengual, 2023), which is now listed as one of the headline indicator for the 8th European Environment Action Program (EC, 2022a).In contrast to MRIObased approaches, the CoF follows a full "bottom-up" approach by quantifying the environmental impacts of consumption based on a basket of representative products (BoP), for which the consumption intensities and environmental impacts are estimated using official consumption statistics and LCA calculations, respectively.This indicator includes around 165 representative products in five key EU consumption areas, namely food (Castellani et al., 2017b;Notarnicola et al., 2017;), mobility (Castellani et al., 2017a), housing (Baldassarri et al., 2017;Lavagna et al., 2018), household appliances (Hischier et al., 2020;Reale et al., 2019), and household goods (Castellani et al., 2019b(Castellani et al., , 2021)).A process-based life cycle inventory (LCI) is compiled for each representative product, which allows for a high granularity at product, life cycle stage, and environmental pressure level.
Compared to MRIO approaches, the LCI has a resolution of thousands of environmental pressures (i.e., elementary flows representing resource use or emissions to the environment at the environmental sub-compartment level) supporting a comprehensive environmental impact assessment with a broad coverage of environmental impacts (Beylot et al., 2019;Castellani et al., 2019a).However, the CoF is currently limited to mass-based consumption (i.e., physical products), excluding the full coverage of services (e.g., food services are accounted for but not additional aspects such as specific packaging or restaurant locations), which might account for around 10% of the impacts of final consumption (Castellani et al., 2019a).
The CoF indicator has multiple potential policy uses, from monitoring purposes including the assessment of environmental impact decoupling along time (Sanyé-Mengual et al., 2019), to supporting the identification of relevant environmental hotspots and assessing policy scenarios (Sala & Sanyé-Mengual, 2023).Furthermore, Sala and colleagues suggested to link the CoF to the planetary boundaries framework thereby supporting the policy ambition of living within the limits of our planet (Sala et al., 2020).At the EU level, this indicator is being employed in monitoring frameworks (EC, 2022b(EC, , 2022a) ) and often in combination with the domestic footprint with the aim of comparing the CoF with impacts of domestic activities (Sanye-Mengual et al., 2022).The CoF was originally designed to monitor SDG 12 and aimed at complementing the SDG indicator on material footprint (Sanyé-Mengual & Sala, 2022).
In Germany, progress with regard to SDG 12 is monitored by tracking the footprints for raw materials, energy, and CO 2 caused by private household consumption (see "indicator 12.1.b:Global environmental impact") (German Federal Government, 2021).The conservation of natural resources is targeted specifically in the German Resource Efficiency Programme (ProgRess) which was adopted in 2012 to determine targets, guiding principles, and policy measures for the more sustainable management of raw materials along their full life cycle (BMU, 2012(BMU, , 2020;;BMUB, 2016).For this, material footprints are among the reported indicators.

METHODS
This paper aims at applying the CoF indicator for Germany using national-level data on household consumption.The scope of the analysis includes the consumption of (average) German citizens considering the areas of consumption food, mobility, housing, household goods, and appliances.The system boundaries of the LCA of each individual representative product are cradle-to-grave, thereby including foreground and background activities from the extraction of raw materials to the end-of-life (EoL) management.Data were collected for the time period 2010-2018.While the EU Consumption Footprint Platform now includes data up to year 2021 (EC-JRC, 2023), our analysis presented in this paper ends in 2018 due to national-level data availability at the time of writing.
The CoF indicator was chosen because of its: (a) comprehensive coverage of environmental pressures (including both resource uses and environmental emissions) to assess a broad range of impact categories, (b) life-cycle-wide perspective, (c) assessment at impact level (with the possibility to also assess at pressure level), (d) granularity at product, life cycle stage, and environmental pressure level, (e) data availability at national and EU level (comparability), (f) possibility to benchmark with the planetary boundaries (reference value), and (f) use in official monitoring indicators systems proposed at EU level.These are also relevant aspects for resource monitoring in Germany (Nuss et al., 2021).
The environmental impacts of German consumption were quantified with the CoF indicator based on representative products (i) (Equation 1).
This LCA-based indicator includes three main steps (Figure 1): a. selection of the representative products, currently including around 164 products of five areas of consumption (Excel SM 2.1), b. calculation of the consumption intensity (i.e., consumption per year) of each representative product via modeling of the apparent consumption (i.e., based on production and trade data) or by modeling of the entire sector (i.e., by disaggregation into representative products), c. assessment of potential environmental impacts of each representative product (i.e., environmental impact per product and year), based on a LCA model to express the 16 impact categories of the environmental footprint (EF 3.0 reference package) method (EC-JRC, 2021).

Selection of representative products
The composition of the CoF indicator by BoP was aligned to the original indicator (see a detailed list of the 164 representative products in Excel SM 2.1).Note that for housing, the 30 building archetypes from the original CoF are limited to the 10 archetypes corresponding to the moderate climate zone of Germany.Representative products target those products with the highest share in the market (i.e., most consumed based on their mass and economic value), associated with specific environmental impacts (e.g., palm-oil products and deforestation), showing emerging trends (e.g., electric vehicles), or associated with sustainable lifestyles (e.g., products composing healthy diets like nuts).

Calculation of consumption intensities
This study focused on the calculation of consumption intensities using two different data types: (a) EU-wide country-level data sources from the original CoF indicator (from now on: "EU-wide CoF"), and (b) national data sources (from now on: "national CoF").
EU-wide country-level data sources correspond to those employed in the original CoF indicator.Data were retrieved from the Consumption Footprint Platform (EC-JRC, 2023).
The selection of national data sources was performed by screening and identifying national data sources for each consumption area and representative product (Figure 1).A comparison of national data with the data from the E-Uwide CoF was carried out and the final data for the calculation selected (including identification of data gaps).In the case of missing national data for a representative product, data from the E-Uwide CoF were used.
When choosing data for the national CoF, data from the Federal Statistical Office (Destatis) was given highest priority, followed by data from other official sources (e.g., government ministries and agencies), industry and other literature data, and EU-wide CoF data collected by JRC (Figure 2).The lowest level of priority was only considered in cases of missing data, incompatibility with the CoF composition (e.g., due to different granularity), or in case of high uncertainty of national estimates.
Table 1 provides an overview of data sources employed for estimating the consumption intensity of German consumption for both EU-wide and national CoFs (Figure 1).Detailed data tables and an overview of overall consumption data and units are provided in the accompanying Excel file.
Housing: Data on building stocks in Germany (i.e., number of flats and building year) were taken from Destatis (2018aDestatis ( , 2019)).Data classification with regard to building age did not always match the CoF structure, which employs data from the EU buildings database (EC, 2021b) (see SM Figure S2 for the assumed correspondences).The majority of residential buildings in 2018 are single family houses (SFHs) (Destatis, 2019).The total final energy use of private households consists of electricity for cooling (e.g., refrigerators and freezers), heating (e.g., cooking), mechanical energy (e.g., electric motors), information and communication (e.g., telephones and internet), lighting, and energy for room heating and warm water.In line with the EU-wide CoF, BoP housing includes the energy used in cooking, household appliances, and goods.Data on final energy use of German households

Household goods
PRODCOM database (Eurostat, 2021a) COMEXT database (Eurostat, 2021g) Household goods data from the EU-wide CoF could not always be mapped to Destatis consumption statistics of private households (Destatis, 2020a(Destatis, , 2021a)).However, Destatis data include detailed consumption intensities in monetary units which might be transformed into physical units and used in future studies (while this was tested, the additional effort involved was beyond the scope of our study).

Household goods:
The consumption intensities of household goods belonging to nine overarching product groups (detergents, sanitary products, personal care products, furniture, bed mattresses, footwear, clothing, paper products, and plastic products) can be partly estimated based on a detailed survey of consumer spending for different household goods performed every 5 years among a large number (Destatis, 2015(Destatis, , 2020a) ) and annually among a smaller number (Destatis, 2021a) of representative households.For each representative product, the monetary spending (€) was converted into the number of units (i.e., mass or pieces) by employing calculated producer prices from PRODCOM (from the French PRODuction COMmunautaire (Community Production)) (Eurostat, 2021a), similarly to the approach from Genta et al. ( 2022).Since PRODCOM reports prices at production, these were converted into estimated consumer prices by considering the value-added taxes at retail, i.e., 19% apart from books, newspapers, and hygiene products (7%) (Federal Ministry of Finance, 2021) (Equation 2).
Consumption intensity RP (mass or pieces) = Expenses RP ( Food: Product groups covered include meat, fish and seafood, dairy, eggs, cereal-based products, sugar, oils, tubers, vegetables, legumes, fruit, nuts and seeds, coffee and tea, beverages, confectionery products, and pre-prepared meals.Data are taken from national statistics and are crosschecked using survey and business data (BMEL, 2021;Destatis, 2016;Deutscher Brauer Bund, 2019;Essential Foods, 2020;FIZ, 2020;STATISTA, 2020aSTATISTA, , 2020b)).Missing data for specific products were complemented with EU-wide CoF estimates (EC-JRC, 2023).The composition considered the most relevant food product groups based on their importance by mass and economic value (Notarnicola et al., 2017), emerging consumption trends, and relevance on biodiversity impacts.
Appliances: Representative products include, for example, dishwashers, washing machines, electronic condenser tumble dryers, combined refrigerators-freezers, air conditioners, electric ovens (built-in), compact fluorescent lamps, halogen lamps, incandescent lamps, light emitting diodes (LEDs), notebooks, and LCD TV screens.The number of representative products (in pieces) was calculated from survey statistics on the existing stock (Excel SM2.6) (Destatis, 2020b).Additional data for electric ovens were taken from Destatis (2018c).Data for lighting products were taken from the EU-wide CoF (EC-JRC, 2023).For each product the whole stock present in German households was divided by the number of service life years as defined in Sala et al. (2019) (i.e., no changes to the lifetimes used at EU level were made for Germany).

Life cycle inventory
The life cycle phases considered for each representative product include: upstream activities, production, packaging, logistics, use phase, maintenance, and EoL.The LCI modeling is based on the EU-wide CoF indicator: food (Castellani et al., 2017b;Notarnicola et al., 2017), mobility (Castellani et al., 2017a), housing (Baldassarri et al., 2017;Lavagna et al., 2018), household appliances (Hischier et al., 2020;Reale et al., 2019), and household goods (Castellani et al., 2019b(Castellani et al., , 2021)).The LCI data for housing were adjusted with regard to electricity inputs to represent the German electricity mix regarding the use phase of products (note that all electricity consumption in the use phase of products is accounted for in the housing area of consumption to prevent double counting).For this, the 2018 electricity mix was modeled using data for the annual electricity mix by country (Eurostat, 2021h;IEA, 2021) including electricity domestic production and trade (Eurostat, 2021h, 2021i, 2021j).

Life cycle impact assessment
The EF 3.0 method (EC-JRC, 2021) was used for the life cycle impact assessment (LCIA) stage to calculate the environmental impacts associated The EF method can be assessed against the planetary boundaries through an adaptation of the planetary boundary framework (Rockström et al., 2009;Steffen et al., 2015) to the metrics of the EF impact categories (Sala et al., 2020).The adaptation to the metrics included also an interpretation of the planetary boundary control variables to the impact level (as some control variables are rather state or pressure level) according to the causeeffect chain of the LCIA model underpinning each impact category.As well, a factor 2 concept was considered to derive a threshold for the resource use categories, since this is not covered by the nine ecological processes under the original framework.

RESULTS AND DISCUSSION
Robust methods and approaches to capture Germany's natural resource use and environmental impacts both inland and considering trade are important to inform national resource and sustainability policies (BMU, 2020;German Federal Government, 2021).However, to date environmental policies often focus on single environmental pressures (e.g., raw materials or GHGs or air pollution) while considering a broader range of environmental impacts (which integrate the potential impact of a number of pressures) within one assessment approach such as the CoF indicator is important to highlight possible burden shifting between impact categories and geographical areas (i.e., outsourcing of environmental burdens).
The results are presented focusing on: (i) the comparison of national consumption intensities with the EU wide statistical data used as input to the EU-wide CoF (Section 4.1); (ii) the environmental impacts due to Germany's consumption, per area of consumption (Section 4.2.1),examining the level of decoupling from gross domestic product (GDP), and comparing impacts with the planetary boundaries (Section 4.2.2); and (iii) the comparison between the results calculated with national and EU estimates (Section 4.3).Detailed data tables are given in the accompanying Excel file.

Alignment of national data with the CoF's composition and granularity
Housing: Data were available for calculating the three main elements of the housing area of consumption: number of dwellings per archetype (according to climatic area, year, and type), energy consumption and water consumption.Main differences were found in the data classification per construction year (Figure S2).
Mobility: Available national mobility data were not detailed enough to match the representative products of the CoF structure, which requires breakdowns according to detailed characteristics of the 34 representative vehicle types (e.g., fuel type, age, and motor size).Bicycles, e-bikes, and trams were identified as relevant modes of transportation currently missing in the CoF structure.Particularly, the growth of e-bikes is not yet reflected in the CoF despite the relevant environmental burden of batteries.On the contrary, tram mobility might be approximated using the LCI entry for electric trains in the future.

Household goods:
The national household consumption statistics in monetary terms provide detailed consumption data with several hundreds of product codes covering, for example, food and drinks, clothing and shoes, living and energy, interior household goods, health, mobility, post and telecommunication, leisure/entertainment/culture, education, and restaurants (Destatis, 2020a(Destatis, , 2021a)).CoF product categories can be mapped to the national data, except for plastic products (e.g., furniture from plastic, plastic household articles, and so on) and jeans for which a clear distinction is not always possible.The national data highlight various products which are not yet covered in the CoF such as, for example, pharmaceutical and medical products, carpets, jewelry and watches, or household spending on leisure activities and financial services (Figure 3).
Food: German data sources include mainly the Ministry of Food and Agriculture (BMEL, 2021) and surveys carried out by Destatis (Destatis, 2016).Several items with a high per-capita consumption were found which are not yet covered in the CoF such as, for example, juices (apple juice, orange juice, multivitamin juice), pre-prepared meals (frozen items) (i.e., pizza and other non-meat-based pre-prepared meals), and others (e.g., honey).Note that some products are covered at the product group level through an upscaling process but are not represented with a specific product (e.g., the product group "nuts and seeds" includes only "almonds" and "cashew" as products, while representing the full group and thereby including products such as peanuts or walnuts).Other relevant food items in terms of their consumption intensities (e.g., fruits such as pears, peaches, and grapes; or vegetables such as onions, cucumbers, lettuce, cabbage, and mushrooms) could be included in future iterations of the German CoF, especially for those items which are found to be significant also from an environmental impact perspective (Clark et al., 2022;Poore & Nemecek, 2018).
Appliances: A comparison of both datasets highlights that a number of additional products are used by average households which are not yet covered in the CoF structure (Figure 3).Among them are emerging items such as e-bikes (see Mobility) or a detailed breakdown of notebooks into tablet PCs and other laptops (Destatis, 2018c(Destatis, , 2020b)).Such information could increasingly complement data collected in the EU-wide CoF and help to better model emerging national consumption trends.

Consumption intensities using national and EU-wide data sources
Housing: Data available for building stocks in Germany (number of residential dwellings) (Destatis, 2018a(Destatis, , 2019) ) match the data used in the EUwide CoF (EC, 2021b) in 2018.National building stock data by construction year did not fully match the EU-wide CoF structure (SM Figure S2).
Energy (RWI, 2020; UBA, 2020) and water consumption (BDEW, 2020;Destatis, 2018b) are lower by 13% and 3%, respectively, which might be the result of data gaps in the EU buildings database for 2018.We consider the national statistics to be more accurate and use them in the calculations.
Mobility: German data for the use of different modes of transportation (i.e., private vehicles, public transport, trains, and airplanes) are aligned with data in the EU-wide CoF (Excel SM2.3_BoP_Mobility).Slight data differences can be due to different vehicle classifications or occupancy factors used.Note that the data source employed in the EU-wide CoF (EU Statistical Pocketbook) relies on data provided from Eurostat which is coming from countries, but these might go through additional data quality checks and refinements contributing to slight differences (around 5%).
Furthermore due to limited granularity of national data hampers the use of national data, we have used data from the EU-wide CoF in the calculations.Regarding the missing modes of transportation identified, the number of bicycles, and more recently e-bikes/pedelecs, have been constantly increasing over time with 78.8% of German households owning a bicycle and 11.4% a pedelec (e-bike) in 2020 (Destatis, 2020c).
Household goods: A comparison of consumption intensities requires conversion of household spending data (monetary) (Destatis, 2018d(Destatis, , 2020a) ) into the number of items (physical) using unit prices (here PRODCOM price data; Eurostat, 2021a) (Equation 2).However, this approach is prone to errors due to (1) problems in unambiguously linking national statistics to the representative products of the CoF, (2) challenges in selecting appropriate price data (consumer vs. producer), and (3) changes in product classifications in national statistics over time.Carrying out a comparison for a single year (2017) highlighted that data are approximately within the same order of magnitude for most products, but large discrepancies exist, for example, for breast pads, liquid soap, tampons, wooden tables, and work and waterproof footwear (Excel SM2.4_BoP_HHGoods).Therefore, the consumption intensities from the EU-wide CoF are used in this study.However, in the future the national statistics could become an important data source for (a) identifying missing products and household consumption trends and (b) cross-checking consumption intensities and improving the overall representativeness of the product basket.
Food: National statistics on food consumption are generally well aligned with data of the EU-wide CoF (Excel SM 2.5_BoP_Food_BMEL).Note that the type of data sources employed in the EU-wide CoF consider the full consumption of food (i.e., not only at the household but also at food services, and including government and NGOs consumption) using apparent consumption calculations (i.e., production + import − export).Instead, national data sources are based on consumption surveys representing actual consumption limited to the household.Differences in consumption intensities can be due to data gap filling using EU production averages which might over-or underestimate the national production (e.g., for oranges, olive oil).National data sources were employed for most food products, apart from avocados, cashew, chickpeas, lentils, meat-based pre-prepared meals, quinoa, soy drink, and tofu.
Appliances: Detailed household surveys for appliances (Destatis, 2018c(Destatis, , 2020b) cover a wide range of representative products of the CoF.
Data highlight a gradual increase in the overall number of appliances used by average private households between 2008 and 2020.Since 2019 the number of smartphones used has significantly increased and is found at around 70 million items in 2020.The use of air conditioners by private households in Germany is so far negligible due to the moderate climate.National and EU-level data are generally well aligned (Excel SM 2.6_BoP_Appl_total).The physical stock of appliances present in German households (Destatis, 2018c) was divided by the number of service life years and total population to obtain the per person stock in a year.On the contrary, consumption intensity data employed in the EU-wide CoF are based on a dynamic approach, where the stock considers the apparent consumption of the entire lifespan of the product (e.g., for a product with a 5-year lifespan, the average apparent consumption of years 2014−2018 would be considered as the stock per year of the given product).

Differences in environmental impacts between the national and EU-wide CoFs
The environmental impacts differ between the national and the EU-wide CoF (except for household goods and mobility which are using the same consumption intensity values and impact factors) (Figure 4).For food and appliances, the slightly lower impacts using national data are due to lower consumption intensities.For instance, for food the lower per-capita use of environmentally damaging products such as pig meat or beef using national data explains part of these differences.For housing, environmental impacts differ by impact category and are partly explained by the use of a different electricity mix (and associated impact).The national CoF uses the environmental impact factor for the German electricity mix, while the EU-wide CoF employs an EU-level electricity mix factor.This affects particularly the categories FEU, AC, IR, and WU due to a different share of energy sources, for example, higher emissions associated with coal combustion or lower emissions of radionuclides due to nuclear power.Differences in other categories, such as HTOX_c are associated with lower energy consumption per dwelling using national statistics.Uncertainties of the CoF calculation are due to uncertainties of its components (i.e., consumption and LCA data, the LCIA model, and the normalization and weighting steps).

4.2
The life-cycle-wide environmental impacts of Germany's consumption

Contribution by area of consumption
The life-cycle-wide environmental impacts of German consumption are dominated by food, followed by housing and mobility (Figure 5).The impact contribution varies by environmental impact category (see Section 3.3.2for the full names of each impact category), with food dominating impacts associated with primary production (e.g., LU, MEU, TEU, or AC) or housing and mobility being linked to impacts due to energy consumption (e.g., IR,

Decoupling assessment and benchmarking against planetary boundaries
The impacts (single score) of German consumption considering upstream impacts (i.e., those occurring beyond Germany's borders) are larger than those of domestic activities (Figure 6).When considering impacts over time from 2010 to 2018, the domestic footprint decreased by −14% while the CoF shows a stable value along the assessed period.In other words, Germany's environmental impacts were largely outsourced to other countries.
These results are similar to those of the EU-wide CoF (+4% increase for the same period).Within the same period, GDP has increased by 31%, meaning that absolute decoupling took place considering the domestic environmental impacts (which decreased), while for the CoF (considering impacts of trade) only weak decoupling can be found.Outsourcing of environmental impacts over time is also found for the EU-27 (Sanyé-Mengual et al., 2019).
One central question of German policy is whether national consumption-related impacts are still within the safe "operating space for humanity" (German Federal Government, 2021).For this, the national CoF is compared against the set of planetary boundaries defined for the EF method, which are allocated equally among the global population (Sala et al., 2020) (Figure 7).(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018).The assessment includes the domestic footprint (production approach).The underlying data for this figure can be found in Supporting Information S2, "RESULTS" tab.
F I G U R E 7 Assessment of the German consumption footprint (CoF, national data) (2018) against the planetary boundaries defined for the environmental footprint (EF) method, per impact category and area of consumption. 1 See Section 3.3.2for the full names of each impact category.
The underlying data for this figure can be found in Supporting Information S2, "RESULTS" tab.
In 2018, the environmental impacts associated with German consumption are found outside the safe operating space for humanity for several impact categories.This transgression of planetary boundaries is in line with current trends at the EU level (EC-JRC, 2023; Sala & Sanyé-Mengual, 2023).Most alarming categories are associated with ECOTOX (e.g., due to use of pesticides in food production), PM (e.g., due to emissions to the environment from energy production for housing, mobility, or household goods), and CC (e.g., due to energy consumption or direct emissions like methane from livestock).These three impact categories are impacting around 10 times the defined boundaries and strong efforts in the short-term to reduce such impacts toward a safe operating space for humanity would be required.Other categories such as FEU (e.g., due to organic emissions during wastewater treatment) or FRD (e.g., due to energy consumption) show a closer gap with the planetary boundaries being around a factor 3.
Moreover, it has to be noted that some planetary boundaries are set not only at global level but also at regional level with a level of detail up to the watershed.In this study, the global boundaries where allocated to Germany following a per-capita allocation (i.e., egalitarian principle), and the analysis of further spatial resolution considering local thresholds.
When assessing the impacts due to German consumption against the planetary boundaries framework, one should consider the challenges and uncertainties associated with the needed adaptation of the original framework to the EF method.This adaptation is detailed in Sala et al. (2020) and required two main steps: (a) a mapping between the nine ecological processes addressed in the framework to the 16 impact categories of the EF method, and (b) an adaptation of the control variables to the metrics of the impact indicators (e.g., from ppm to kg CO 2 equivalent for climate change).Such adaptations added a layer of uncertainty to the analysis, beyond the uncertainty of the control variables of the original planetary boundaries framework (Rockström et al., 2009;Steffen et al., 2015) as well as of the CoF results (as noted in Section 4.1.3).

Comparison between national and EU estimates
Total GHG emissions triggered by German consumption equaled 807 million metric tons CO 2 equivalent in 2018 (CoF using national data) (Excel

SM Results
).This compares to 885 million tons CO 2 in 2017 (latest available date; at the time of writing no estimates for other GHGs were available) (Destatis, 2021b).For energy use, the national CoF estimates the fossil energy demand at 9.74 million TJ (terajoule) in 2018 while Destatis provides a slightly higher estimate of 12.76 million TJ for 2017 (+30%).Differences are due to a number of reasons.First, Destatis bases the estimates upon environmentally extended-input-output (EEIO) analysis following the systems of environmental economic account (SEEA) and as such they are covering a larger range of economic activities (e.g., including also services).Second, the environmental extensions generally follow the "residence principle," implying that emissions or energy use due to activities of German residents are accounted for, regardless of where they actually take place geographically.On the other hand, the CoF calculations cover all GHGs while the Destatis number only includes CO 2 emissions to date.For energy use, Destatis figures refer to the overall energy use while the CoF focuses specifically on fossil energy carriers.
Overall, the CoF approach provides reasonable estimates for both the carbon and energy footprint considering its focus in terms of sector coverage.The CoF indicator was already compared to EEIO approaches outlining differences, pros and cons, including convergence in messages of trends and hotspots (Castellani et al., 2019a).While EEIO allows for a more comprehensive coverage of economic activities, there is a lack of granularity at the product level which is essential to monitor evolution of production and consumption systems and as a basis for testing improved scenarios.
Furthermore, an assessment of impacts (rather than pressures) faces certain limitations associated with the environmental extensions of IO tables (e.g., in EXIOBASE) including, for example, the limited number of environmental pressures (e.g., pesticides) that can lead to the exclusion of impact categories not currently covered (e.g., ODP) or the underestimation of impacts (e.g., toxicity) (Beylot et al., 2019;Castellani et al., 2019a).This can affect the robustness of the environmental impact assessment step.In addition, some reported environmental pressures have a limited level of detail (e.g., PM 2.5 , mineral resource use), leading to a mismatch with characterization factors in the impact assessment methods.On the other hand, carrying out a bottom-up assessment of environmental impacts via LCAs can be more resource intensive than using EEIO-based approaches.
A comparison with the CoF for the EU-27 shows that Germany has similar environmental hotspots due to consumption.Food (especially animalbased products), housing (mainly energy consumption for heating purposes), and mobility (particularly due to transport with private cars) are the main drivers of environmental impacts of the EU-27.Evolution along time shows only a relative decoupling, with several planetary boundaries being transgressed (Sala & Sanyé-Mengual, 2023).

CONCLUSIONS
This study showed that national consumption statistics in Germany provide a rich source of data and knowledge about the consumption intensities of representative products included in the CoF.Missing products could be identified and including these in future CoF updates could better capture the national situation (e.g., better mirroring socio-cultural trends) and highlight emerging consumption trends (e.g., increased use of e-bikes or tablet PCs).However, consumption intensities can also vary due to different scopes of the data sources used (e.g., apparent vs. actual consumption), different modeling approaches (e.g., stock modeling for appliances vs. household's ownership), or the difficulty of translating monetary units-common in expenditure surveys-into physical quantities.Using national data sources led to slightly lower environmental impacts when compared to the EU-wide CoF, mainly due to a broader scope of the EU-wide data sources, which include not only households but also overall consumption.
Some of the limitations of our study include that, with the exception of adaptations to the energy mix, no further changes were made to the LCI datasets.Furthermore, the CoF is currently limited to mass-based consumption and does not provide for full coverage of services.Finally, when assessing impacts against the planetary boundaries, it is important to consider the challenges and uncertainties associated with the adaptation of the EF method.
However, the CoF indicator provides a possible LCA-based approach for highlighting different environmental pressures and subsequent impacts at a high level of granularity (at the level of products, and further to life cycle stages).This could complement existing EEIO-based assessments for Germany (Destatis, 2021b;Lutter et al., 2022;UN, 2020), provide linkages to consumption-based accounting approaches (UBA, 2022), and find application in resource monitoring frameworks (Nuss et al., 2021).
Future research should focus on further translating and connecting national statistics (especially for household goods) to the CoF model and ensuring that missing products and services are increasingly included in the assessment.National-level LCIs could be tested (or existing inventories adapted to the national circumstances) to better regionalize environmental impact estimates.Given the high granularity of the CoF model,

F
I G U R E 1 Overview of the consumption footprint (CoF) indicator, including the focus of this study on comparing data sources to calculate the consumption intensity at representative product level and the followed process (adapted fromSala & Sanyé-Mengual, 2023).BMEL, Federal Ministry of Food and Agriculture; Destatis, Federal Statistical Office; FAO, Food and Agricultural Organization; UBA, German Environment Agency (Umweltbundesamt).

F
Prioritization pyramid for the selection of national-level data sources for the national consumption footprint (national CoF).BDEW, Bundesverband der Energie-und Wasserwirtschaft e.V.; BMEL, Federal Ministry of Food and Agriculture; BMVI, Federal Ministry of Transport and Digital Infrastructure; RWI, Leibniz-Institut für Wirtschaftsforschung. TA B L E 1 Overview of consumption intensity data sources for the EU-wide consumption footprint (CoF) and national CoF, by area of consumption.(EC, 2021b) Hotmaps Project (IWU, 2016) TABULA web tool (Pezzutto et al., 2019) -Building stocks for flats and buildings (long time series from 1969-2019; Destatis, 2018a, 2019) -Energy use: (RWI, 2020; UBA, 2020) -Water use: Destatis Public water supply and waste water treatment, BDEW Use of drinking water in households (BDEW, 2020; Destatis, 2018b), and other sources.

FRD
, and POF).Mineral resource depletion (MRD) is associated largely with appliances and mobility.These results are in line with the analysis of the CoF for the EU-27 and for other EU countries (EC-JRC, 2023; Sala & Sanyé-Mengual, 2023).

F
Ratio between national CoF and EU-wide CoF per area of consumption and impact category 1 , considering also the inter-temporal variability(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018).Mobility and household goods are 100% for all cases. 1 See Section 3.3.2for the full names of each impact category.The underlying data for this figure can be found in Supporting Information S2, "Figure4" tab.F I G U R E 5 Contribution analysis of the areas of consumption to the overall Germany's consumption footprint (2018), by impact indicator 1 . 1 See Section 3.3.2for the full names of each impact category.SWS, single weighted score (Pt).The underlying data for this figure can be found in Supporting Information S2, "RESULTS" tab.

F
Decoupling assessment for Germany: Comparison of the evolution along time between the consumption footprint (CoF) (both German national data and EU-wide data) and the economic growth (in terms of gross domestic product, GDP) and population evolution The latest edition of ProgRess calls specifically for the further development EU data for individual countries are available in the EU statistical pocketbook (EC, 2019b) and were selected also for Germany due to two main reasons.First, German statistical offices provide data to Eurostat and the Statistical Pocketbook, in which data gap filling procedures and Use of food products per capita and other sources from(BMEL, 2021).Missing product groups are covered with EU-wide CoF data.Mobility:data quality checks are implemented thereby increasing data robustness.Second, publicly available data regarding private and public vehicles were not granular enough to align with the modes of transportation implemented in the BoP mobility.Data were validated using national estimates of motorized traffic available from (BMVI, 2021) in person-kilometers.