The most common way to calculate consumption-based emissions for a country is to use Environmentally Extended Multi-Regional Input-Output Analysis (EE MRIOA), either on its own, or in an approach in which it is a part of the method. The basis for EE MRIOA is the Input-Output Tables for a country. These are typically collected by statistical offices as a part of the National Accounts of a country. A central part of the Input Output (IO) Tables are matrices with industrial sectors on the axes. The numbers in the tables (usually in monetary units) describe how different sectors trade with each other. For example, the food industry is a sector that produces food products. In order to do so, the food industry needs to buy products from the agricultural sector, but the industry also needs transport services, energy, machinery etc. The tables describe how much each sector buys from the other sectors. In principle, each sector buys products from all other sectors in order to produce the goods and services it produces. In addition, the Input-Output tables also include information about the proportion of production that goes to private consumption, public consumption, investments (both private and public) and exports. There are also tables that show imports.
By connecting Input-Output tables for several countries and regions, a Multi-Regional Input Output Analysis can be made. Careful work is then needed to make sure that imports and exports in the different tables match each other.
Environmental information can be added to the calculations using tables for Environmental Extensions. These are tables that describe emissions (or resource uses) that are associated with a specific industry sector in a specific country. These are typically described as intensities, expressed as, for example, the mass of an emission per monetary unit, for the specific product group in the specific country.
Using these tables and matrix algebra, it is possible to calculate the emissions from consumption in a specific country (e.g. Wood et al., 2015). The calculations cover the whole supply chain regardless of where the production takes place.
Several models are available for EE MRIOA (e.g. Dawkins et al., 2019). The differences between them include, for example: how many, and which, countries and regions are included; how many, and which, industries and product groups are included; which environmental impacts are included; and which data sources are used. There are also different types of modelling choices that have to be made, for example, when data from different data sources needs to be combined and matched. That means that it may be difficult to compare results, for example between two different countries, if different methods for EE MRIOA have been used in analysing each country.
An advantage of EE MRIOA is that it covers all sectors in the world. The method can be described as a top-down method, since a starting point is the total emissions of a country, totals which are allocated to the different sectors in the country in the Input-Output tables. But this also means that uncertainty typically increases with the level of disaggregation. Results for specific sectors, product groups, societal groups or sub-national regions are typically more uncertain than data for a whole country. Another advantage is that it is based on data from the System of Environmental and Economic Accounting (SEEA) which is internationally standardised and harmonised (United Nations et al., 2014).
Greenhouse gases are probably the most common group of emissions that are studied with EE MRIOA. Results are sometimes presented for CO2 (which is the most important greenhouse gas) or for CO2 equivalents, other greenhouse gases typically aggregated based on their Global Warming Potentials for a time frame of 100 years. Depending on the data availability, results can also be calculated for a broad range of emissions and resources (e.g. Palm et al., 2019; Persson et al., 2019; Brown et al., 2022). One example of interest for this study is different types of air pollutants such as NOx, SO2 and particulates. Data for these emissions are typically collected in the SEEA and so data availability is, in general, reasonably good.
Consumption-based emissions are sometimes calculated for only private consumption, and sometimes for all consumption including public consumption and investments (including both private and public investments in buildings, infrastructure, factories etc.).
Calculation of consumption-based emissions has developed significantly during the last decade(s) and can be expected to develop further in future. A likely trend is that calculation of consumption-based emissions will be incorporated into the normal work of statistical offices. Statistics Sweden currently produces statistics on consumption-based emissions of both greenhouse gases and other air pollutants as a part of Sweden’s official statistics. Eurostat also regularly produces statistics on consumption-based CO2 emissions. It seems likely that more statistical offices will follow suit and produce statistics for consumption-based emissions of greenhouse gases, but also other air pollutants.
Exiobase is probably the most used EE MRIOA model, both in general, and specifically by statistical offices in the Nordic countries (Wood et al., 2015). This model has been developed and updated by a research consortium. However, several actors have suggested it would be advantageous if an international organisation would develop and update a model that could be used by many countries. Eurostat has developed, and keeps developing, Figaro (Remond-Tiedrez and Rueda-Cantuche, 2019) which is an EE MRIOA model. Figaro may become a model preferred by institutional actors. Exiobase and other EE MRIOA models can then continue to be developed for research purposes.
Statistics Sweden recently made a comparison between Exiobase and Figaro (Cederholm et al, 2024). The study showed that the overall results for the consumption-based emissions of Sweden were 8.5% higher when Figaro was used instead of Exiobase, and that larger differences occurred for more disaggregated data.
Cederholm et al. (2024) also compared how Exiobase and Figaro are updated. Both models include data from a number of different sources. When time series are produced, data are updated for the relevant years. In some cases actual data are used, and in some cases data are modelled using updated macro-economic data. Data on emissions are typically more updated than the IO tables and data for OECD countries are typically more updated than non-OECD countries. Figaro has the advantage of using real IO tables, while the last year data for intermediate use matrices in Exiobase is 2011 and data for more recent years are modelled. Efforts are currently ongoing between several international actors to coordinate the compilation of IOA (Input-Output Analysis) tables and faster updates with consistent quality can be expected in the future.
An aspect of EE MRIOA is that all products within a product group are assumed to have the same emission intensity on a monetary basis. So, to give an example, a product that costs twice as much as another one in the same product group is assumed to cause double the emissions. This can lead to overestimations of the environmental impacts of both expensive products and the footprints of higher income groups (André et al., 2024; Leferink et al., 2023). This is also one reason why EE MRIOA is less suitable for analysis of products. Instead, other methods such as Life Cycle Assessment (LCA) (e.g. Hellweg et al., 2023) may be more suitable in those cases. Hybrid models have also been developed in which some data comes from EE MRIOA and some from LCA or process data (e.g. Andersson, 2020; Carlsson Kanyama et al., 2021; Heinonen et al., 2022) which can avoid some of the limitations of EE MRIOA. The Joint Research Center has also developed methods for calculating consumption and consumer footprints based on LCA bottom-up methodology (Castellani et al., 2019). In these approaches, the environmental impacts of a limited number of products calculated with LCA are assumed to represent the whole product group. Although the development of EE MRIOA models has been strong, there is still a need for future research. Examples of questions that can be addressed include: