Environmental Input-Output Analysis / Environmentally Extended Input-Output

Wikis > Formulation Tools > Formulation Tools in Support of Life Cycle Assessment (LCA) > Environmental Input-Output Analysis / Environmentally Extended Input-Output
Primary Source: Suomalainen, K. (2006) EIOA/EEIOA Swat Evaluation in Report on the SWOT analysis of concepts, methods, and models potentially supporting LCA. Eds. Schepelmann, Ritthoff & Santman (Wuppertal Institute for Climate and Energy) & Jeswani and Azapagic (University of Manchester), pp 62-71

Level of analysis: Mainly macro, can be applied to meso-level in some cases depending on the level of detail of available data and purpose of assessment. Bottom-up approaches of IO-based LCA allow micro level assessment.

Assessed aspects of sustainability: Environmental, generally national environmental accounting.

Main purpose of the assessment: To account for environmental burdens and understand interdependencies of environmental impacts at a national level, related to transactions between national industries/sectors.

Description of the methodology: EIOA/EEIOA are expansions of conventional IOA introducing the environmental dimension to the conventionally monetary analysis. EIOA uses physical units to describe the activities between sectors, added as extra lines to the conventional IO-table. EEIOA defines a separate so called intervention matrix, which shows the amount of pollutants emitted and natural resources consumed to produce one unit monetary output of each industry. In addition to flows (input materials) corresponding to monetary trade, quantifiable externalities such as pollution are typically also included.

Detailed description

Input-output analysis accounts for the level of output of each sector corresponding to that level of activity in all the other sectors in a national economy. By-products, mostly negative from the environmental view point (e.g. pollution), are often linked directly to the network of physical flows that build up an IO-table. This technical interdependence can be described in terms of structural coefficients similar to those already used to trace the interdependence of production between the sectors of the economy (Leontief, 1970).


In EIOA pollutants can be integrated to the conventional IO-tables as their own sectors; each sector in its production may cause a certain amount of pollution that is accounted for as an input to that sector. Thus the vectors and matrices are grouped by whether the figure refers to goods or pollutants;

1,2,3, .. . i. .. j .. . m, m+i,m+2,. .. g.. . k .. . n

         Goods                     pollutants

The technical coefficients become;

aij – input of good i per unit of output of good j

aig – input of good i per unit of eliminated pollutant g (eliminated by sector g)

agi – output of pollutant g per unit of output of good i (produced by sector i)

agk – output of pollutant g per unit of eliminated pollutant k (eliminated by sector k)

rgi, rgk – proportion of pollutant g generated by industry i or k eliminated at the expense of

that industry.

Variables are denoted by;

xi – total output of good i,

xg – total amount of pollutant g eliminated,

yi – final demand of good i,

yg – final delivery of pollutant g,

pi – price of good i,

pg – price (cost) of eliminating one unit of pollutant g,

vi – value-added of industry i per unit of good i produced by it,

vg – value-added in anti-pollution sector g per unit of pollutant g eliminated by it

With this notation the matrices and vectors become;

A11 = [aij] i, j = 1, 2, 3,…, m

A21 = [agi] g, k = m+1, m+2,…, n

A12 = [aig]

A22 = [agk]

Q21 = [qgi] qgi = rgi agi

Q22 = [qgk] qgk = rgk agk


Once the appropriate technical input and output coefficients have been identified, pollution and its mitigation can be included in the IOA as integral parts of the economic processes. The physical input-output balance becomes;


The input-output balance between prices and values-added (in monetary units) is described as;



In EEIOA, on the other hand, a separate interventions matrix is compiled to represent environmental interventions for each industry involved. As for technologies in IO-tables, linearity is assumed also for environmental interventions, i.e. that the amount of environmental intervention associated with an industry is proportional to the amount of output of that industry. If B (dimension qxm) is the intervention matrix, showing the amount of pollutants emitted and natural resources consumed to produce one monetary unit of output of each industry, then the total direct and indirect pollutant emissions and natural resources corresponding to satisfying a certain amount of final demand y is given by

m = B(I-A)-1y

where m (dimension q) is the total domestic direct and indirect vector of environmental burdens and A is the conventional Leontief technology matrix (Tukker at al, 2006).

This is closely related to National Accounting Matrices with Environmental Accounts (NAMEAs), which were introduced by the Dutch Central Bureau of Statistics in the 1990’s and were soon after adopted by Eurostat and spread to all EU countries. NAMEAs integrate economic accounts with environmental data providing a picture of where exactly in the economy emissions, waste and resources are used or generated.

In Europe EEIOA has been used with NAMEAs to identify hotspots within the European production and consumption patterns and reductions in per capita resource use and emission outputs. This provides a tool for developing policies for sustainable consumption and production (SCP). Such a tool can answer questions on which features of European consumption and production are mainly responsible for the high energy and resources use, and where the greatest gains in sustainability can be made. EEIOA with NAMEAs is a method for identifying the environmental pressures resulting both from national production and consumption; it may be used to identify key intervention points and to help EU and ational policy makers to take SCP a step further. EEIOA and NAMEAs offer complement to existing

inventories on air emissions and energy, and allows for a consumption-oriented view of the economy; EEIO methods present real possibilities for estimating indirect pressures associated with national consumption.

However, the NAMEAs are still limited in scope and are yet to be developed to include several more categories such as land-use, waste flows, water flows and toxic emissions. Also timeseries of NAMEAs are not yet widely available.

Other similar methods

Several other IO-approaches have been developed to support traditional LCA studies. Since all processes in an economy are directly or indirectly linked with each other, a process analysis based LCI is always truncated to a certain degree, limited by the practically and viably set scope of the study. This problem has led to the use of IOA in LCA. IO-based LCI uses basically the structure of EEIOA for tracking material flows and pollutants. It may be argued that the system boundary of an IO-based LCI is generally more complete than that of a process analysis, since it includes all transaction activities within the country. However, the IOA method can provide LCIs only for the pre-consumer stages, thus not including use and end-oflife phases. Also the IO approach is suitable only if the amount of imported commodities is negligible in the studied system. Finally, the available data for IO-based LCIs is normally older than that available for process-based analysis, since it typically takes several years to collect and publish IO tables (Suh & Huppes, 2005).

The best choice for future seems to be the so called hybrid analysis, which links process-based and IO-based analysis to combine their strengths. Three main ways of doing it are tiered hybrid analysis, IO-based hybrid analysis and integrated hybrid analysis. Tiered hybrid analysis uses process-based analysis for use, end-of-life and the most significant upstream phases and IO-based LCI for the remaining input requirements. This offers an easy extension on simple partial LCA systems in filling in the gaps. IO-based hybrid analysis increases the IO resolution by disaggregating industry sectors and uses this augmented IO-table for the upstream processes. Process based analysis is used for use and end-of-life phases. This IO approach avoids the problem of overlapping calculations for upstream processes. Both tiered and IObased LCI use external linking between the IO approach and the process based LCA approach. In integrated hybrid analysis the IO table is interconnected with the matrix representation of the physical product system both at upstream and downstream cut-offs, where detailed, case specific data are not available. This appears to be the most seamless approach of the hybrid alternatives. It has an advantage in its quality of results especially when is comes to system completeness (Suh & Huppes, 2005).


The input-output approach has been applied on many different types of environmental analysis, but at a national level the general question to be answered is of the form “how much production is needed in various parts of the economy to satisfy a final demand of a certain product worth x monetary units”. From an environmental perspective, EIOA can be used e.g. to calculate energy consumption or emissions related to the production of a certain good, in total or per a monetary unit of production, or to analyse emissions due to imports (Miller & Blair 1985) or more generally the environmental impacts of trade (Hubacek & Giljum 2003). EIOA has also been used to evaluate the emissions of CO2, SO2 and NOx in relation to a governmental long-term scenario for the national economy (Östblom, 1996), or to build indicators for emissions per product group, e.g. identifying the CO2-intensities for different goods (Alfredsson, 2002). At the process level IO-approaches have generally been used as a complementing tool for LCA and provide systematic tools that can be used for integrated environmental analysis and planning (Pan & Kraines, 2001).


Relevance of assessment (what is being assessed, usefulness)

EIOA/EEIOA are useful for identifying the flows that account for the most significant environmental impacts at the level of study, and serve as a sound basis for mitigation policy.

These approaches bring together economic and environmental data in a consistent, related sectoral framework.

EEIOA may be used:

–          to identify key intervention points,

–          to help EU and national policy makers to take SCP a step further,

–          as a complement to existing inventories on air emissions and energy, allowing for a consumption-oriented view of the economy,

–          to present real possibilities for estimating indirect pressures associated with national consumption,

–          to evaluate environmental performance at the meso and macro dimensions, especially where higher resolution at sectoral levels has been achieved.


EIOA/EEIOA is computationally compatible with LCA and flexible at integrating other data sources. It has the computational advantage of being a transparent tool, all flows can easily be tracked to their sources. EIOA also has the additional flexibility to include all relevant flows (not based on existing standardized formats) and can be applied on all levels, if relevant data is available. Thus as long as the used data is valid for the studied flows, the method is highly robust and reliable.

EEIOA has the methodological advantage of a coherent framework where environmental, economic and social data can be inventoried. Such standardized data sets can be used for multiple purposes from environmental monitoring to policy analysis.


Relevance of assessment (what is being assessed, usefulness)

EIOA/EEIOA cannot be used to draw very detailed conclusions from a macro-level study, but more specific analysis is usually required.

EIOA/EEIOA is typically based on static accounting, which makes it less suitable for long term sustainability analysis. It is difficult to include consequences of technological change in the system unless a more complex model is built e.g. in time-series. I addition, EIOA/EEIOA gives the average effect of a change in the system, which may be an understatement. It is the marginal effect that would give the real effect of a given change, and thus allow identifying where in the economy responding to that change would be least expensive.


Currently the data is often indirect and/or based on partial measurements of resource extraction and emissions corresponding to economic inputs and output, and thus statistical procedures have to be used for producing the data sets required for deriving EEIO tables. Also the underlying IO tables used still differ considerably among countries both in structure and level of differentiation, making comparison of EEIO analyses difficult. At a European level EEIO tables are still limited to few emissions to air (GHG emissions) and comprehensive EU-wide EEIO tables are still non-existent (Tukker et. al., 2006). It will take several years until EEIO tables with a few dozen emissions to air and water, and the most important extractions of natural resources for EU-25 will be available.


At macro-level much detail is omitted and the assumption of homogeneity between products of a sector may be a disadvantage when trying to identify specific sources of pollutants. Also all interdependencies are assumed linear, which brings the problem of scaling – linear interdependencies are typically an acceptable approximation only for small changes. Data collection and model construction may be cost-intensive and time-consuming tasks.

There is also a gap between theoretical and practical application of the methodology. Data gathering on products are currently in different and only partially linked classification systems; one for domestic product flows and another for imports and exports. Data gathering and processing becomes unnecessarily complex.

Opportunities for broadening and deepening LCA

EIOA/EEIOA can be used at EU or national level to identify spots where a more detailed study is necessary with a well-defined LC approach. It can be used to help identify areas for policy-targeting. Whereas IOA informs us about the economic interconnections of an economy, a general framework for EIOA/EEIOA could inform policy makers about the general environmental performance of that economy.

EU research can take advantage of standardized EU statistics to identify coherent EU level patterns and construct an entire snapshot of physical flows (including pollutants) within the EU or beyond. Such results can be used as input for Sustainable Consumption and Production (SCP) policies as well as for knowledge gathering for the Thematic Strategy on Sustainable Use of Natural Resources.

In order to work through the inconsistency between IO- and EIO-tables among countries and regions, a combination of ESA95 and NAMEAs to a consolidated EU table has been proposed (Tukker et. al., 2006). Three pathways have been identified differing in structure, data and costs, of which the medium resolution EEIO tables , so called IO/NAMEA++, was suggested as most favourable for short-term developments. This approach is a European EEIO model based on existing procedures and using only European data with easier updatability than what current procedures allow for.

IO-LCI and hybrid analysis clearly offer systematic ways of broadening the scope of LCA, both in completeness and number of aspects taken in to account. Whereas traditional LCA inevitably has to cut-off the scope at some level, hybrid analysis allows all the national interconnectedness to be included. Also the matrix form allows easy inclusion of additional aspects; an economic evaluation of the life cycle becomes quite straight-forward.

Risks for broadening and deepening LCA

e.g. risks due to insufficient data, cost-intensiveness, lack of salience or credibility

Gathering data from all EU member countries has the risk of lack of validation of data. Using data only from established data sources may be limited. For comprehensive long-term conclusions the collection and analysis of time-series is necessary, which is a cost-intensive and time-consuming task.

In addition EIOA lacks clear definition of the methodology, especially in scope, which may hinder broader use of the tool. EU research can avoid risks that are harmful to EIOA by defining a standard model and the statistical data that should be provided by member countries to build a robust and reliable tool for sustainability assessment at the EU-level.

Whereas EIOA is quite flexible in defining its scope, it is essential to standardise these procedures for EEIOA used at EU-level. Not doing so risks the credibility of this methodology. A standardised “all-inclusive” EU-wide table also avoids the risk of omitting some relevant environmental aspects.

Although EEIOA methods and models have been developed in a unified framework much due to the UN-directed standardisation and classifications of sectors and products, it is not compatible with the currently used European ESA95 data sets. Due to the lack of European standardised classification systems and procedures, data gathering is currently much more expensive than necessary – by better coordination and alignment the current effort in manpower could produce greatly improved data for supporting sustainability policies.

Literature/Internet links

Alfredsson, E., 2002, Green consumption, energy use and carbon dioxide emissions. Thesis, Department of Social and Economic Geography, Spatial Modelling Centre, Umeå University, Umeå.

Finnveden, G., Moberg, Å., 2005, Environmental systems analysis tools – an overview, Journal of Cleaner Production 13, 1165-1173.

Grêt-Regamey, A., Kytzia, S., 2007, Integrating the valuation of ecosystem services into the Input-Output economics of an Alpine region, Ecological Economics 63, pp. 786- 798.

Hoekstra, R., Bergh, J.C.J.M van den, 2006, Constructing physical input-output tables for environmental modeling and accounting: Framework and illustrations, Ecological Economics 59, pp. 375-393.

Hubacek, K., Giljum, S., 2003, Applying physical input-output analysis to estimate land appropriation (ecological footprint) of international trade activities, Ecological Economics 44, pp 137-151.

Leontief, W., 1970, Environmental Repercussions and the Economic Structure: An Input-Output Approach, The Review of Economics and Statistics, Vol. 52, No. 3., pp. 262-271.

Leontief, W., 1986, Input-Output Economics, Oxford University Press.

Miller, R., Blair, P., 1985, Input-Output analysis – foundations and extensions, Prentice-Hall, Englewood Cliffs NJ.

Pan, X., Kraines, S., 2001, Environmental Input-Output Models for Life cycle Analysis, Environmental and Resource Economics 20, pp. 61-72.

Suh, S., Huppes, G., 2005, Methods for Life Cycle Inventory of a product, Journal of Cleaner Production 13, pp. 687-697.

Tukker, A., Huppes, G., Oers, L. van, Heijungs, R., 2006, Technical Report Series: Environmentally extended input-output tables and models for Europe, Institute for Prospective Technological Studies, European Commission, Directorate-General Joint Research Centre.

Östblom, G., 1996, Emissions to air and the allocation of GDP: Medium term projections for Sweden in conflict with the goals of CO2, SO2 and NOx emissions for year 2000. Working Paper no.54, Konjunkturinstitutet, Stockholm.



Register| Forgot Password?

Policy Tools Wiki