| Front page | | Contents | | Previous | | Next |

Model for selection of future target areas in the Danish Program for Cleaner Products

1 Methodology description

The chapter describes the methodology developed in the project "Model for selection of future target areas in the Program for Cleaner Products". The description focuses on a number of new elements in environmental assessments and how they are combined with other types of statistical information. Not all details are given in the paper, and the interested reader is therefore referred to the original report "Model til udpegning af fremtidige indsatsområder inden for Program for renere produkter" ("Model for selection of future target areas in the Danish Program for Cleaner Products"), which contains some more details.

1.1 Purpose of the study

The purpose of the project was to develop a preliminary model for screening/ identification of possible and relevant areas for future environmental efforts towards products and product groups as a part of the Danish Integrated Product Policy. The purpose signals a shift from a sector-orientated focus to a product focus and as a consequence it requires that new information sources are explored.

The model should be used to identify 3-5 relevant areas (sectors or product groups) for the product orientated environmental efforts in Denmark in 2002. Subsequently, the model should be developed further in order to support the Danish EPA in its future efforts for cleaner products.

The focus for the development of the model was on products and product groups rather than on sectors. In order to be able to implement the future efforts it is however necessary to know which sectors produce the product groups that are identified in the prioritization. In order to create and maintain this overview, all basic information and intermediate calculations have been stored in a database in Access, allowing for fast data retrieval as well as new calculations.

1.2 Methodology

The screening and prioritization is done in a six-step procedure that is outlined in Table 1.

Table 1.:
The six step procedure in the selection of product groups

1.3 Step 1 - Coupling of economic information between sectors and product groups

On the economic product level, the core information source is the Danish Statistics of Goods ("Varestatistikken"). The statistics contains economic information about the value of product groups (95 groups in all on 2-digit KN-nomenclature level which is the chosen level in this project) being produced and/or used in Denmark. The respective values for production, import and export are combined in order to find the Danish supply of a given product group (Supply = Production + Import – Export). For import and export data, foreign trade information related to the Statistics of Goods are used to provide the requested information.

On the sector level, the core information regards 106 sub-sectors with a production. The sub-sectors are identified by a 3-digit DB-93 code. The DB-93 code system is a Danish parallel to the NACE code system, the first four digits in the two systems being identical while the two last digits in the DB-93 system are Danish subdivisions. Additionally, 40 sub-sectors from four general sectors (supply of electricity, gas, water and heat, building and construction, trade (retail and wholesale), and transportation) are identified.

Statistics Denmark provided information on the turnover of goods related to the specific production sectors. This coupling is made by using the Statistics of Goods that is based on information from companies. The companies are in turn characterized by belonging to one sector only, i.e. their main business area. With this information a coupling between the goods and the sectors is made, revealing which goods are being produced in which sectors and the value of the production in each sector and of each product group.

1.3.1 Assessment of economic importance

The economic importance is assessed by the Danish supply of a given product group rather than the Danish production of the same product group.

By taking import and export into consideration the shift from sector to product orientation is stressed. An obvious implication of this in the economic overview is that Danish sectors with a large export will be less important than sectors with a large import of certain product groups. In other words, the focus is shifted from "what can Danish industry do to produce cleaner products?" to "what can the Danish society do to secure that the products used in our economy are as clean as possible?"

It can be argued that the two approaches are supplementary to each other, i.e. that the combined knowledge is more suitable for prioritization of future efforts. This is probably true, and the current pilot project shall therefore be seen as a first step towards creating such an overview. Until it is created, the prioritization must be based on the new knowledge produced by the current methodology in combination with existing information from many years of experience with environmental efforts in Danish industrial sectors.

In practice, nine product groups were identified as having a "low" priority in relation to the supply figures, but at the same time having a high economic importance because most of the produced product groups are exported. Such product groups will always have a low ranking in the combined assessment, and it is therefore essential to examine these product groups manually in more details.

1.4 Step 2 - Environmental assessment of product groups

The assessment and ranking of the environmental impacts of product groups is based on input/output analysis. Obviously, the most precise result would be achieved if a (very) large number of life cycle assessments were available. This is not the case and instead environmental input-output analysis was used as the carrying element in the assessment. A number of options were available at the time of the study:

1. The EIOLCA (Environmental Input Output Life Cycle Assessment) software developed by the Carnegie Mellon Green Design Initiative in USA².
2. A Swedish IO-study with a relatively limited number of product groups and sectors, and with a limited number of environmental interventions. The Swedish approach³ calculates the impacts (CO₂, SO₂, NO_x, industrial waste, consumption of chemicals) from 46 product groups. The report summarizes the results and makes priorities that are similar to those established in the current project.
3. An older Danish study⁴, addressing a large number of product groups, but only using resource and energy consumption as environmental indicators.
4. The Danish NAMEA (National accounting matrices including environmental accounts) and PIOT (Physical Input Output Tables). Currently, the Danish NAMEA includes 40 types of energy, the reserves of natural gas and oil in the North Sea, emissions to air of eight types of substances, and trans-boundary flows of these substances to and from Denmark. PIOT tables exist for all products taken together and for various individual groups of products (animal and vegetable products, stone gravel and building materials, wood and paper, metals and machinery, and chemical products and fertilizers).

Option 4 is probably the best choice of model on the long term, because it relates to Danish conditions. It was however disregarded in the present study because it was not possible to determine whether the available information could be made operational at a sufficient level of detail during the very short period of time for the study. Options 2 and 3 were excluded due to the limited details of information in the reports.

1.4.1 The EIOLCA software

The assessment of the environmental impacts from product groups was therefore done using the EIOLCA (Environmental Input Output Life Cycle Assessment) software developed by the Carnegie Mellon Green Design Initiative in the USA. The basic function of the software is that it calculates the environmental impacts when purchasing for a given amount of money from a sector, and it is thus possible to compare different product groups by the same "functional unit", e.g. environmental impacts per 1 million dollars worth of products within the product group.

Basically, an IO-model gives an overview of the trade in a national economy. It shows how products are being sold from producers either to final consumers or to other sectors for further processing. It can be visualized as a set of large tables (or matrices) with one column and one row for each sector. The tables can represent total sales from one sector to others, purchases from one sector, or the amount of purchases from one sector to produce a dollar of output for the sector. The tables are a result of an iterative calculation, i.e. that production in one sector is based on inputs from all sectors, and the production of these inputs is in its turn based on production from all sectors, etc. An economic IO-model is linear, so that the effects of a €1000 purchase from a sector will be ten times greater than the effects of a €100 purchase from the same sector.

The IO-model in the EIOLCA-software is based on the 1992 goods/goods input-output matrix of the US economy as developed by the US Department of Commerce. The matrix includes 485 groups of goods/economic activities and is among the most detailed in the world. The buyers and suppliers on the market are grouped in production sectors and sectors for final use.

The economic IO-data are supplemented with information on environmental interventions (energy consumption, waste generation, water consumption, emissions of pollutants, etc.) from a number of sources that all are based on measurements and reporting of US conditions. The data for environmental interventions are divided by the annual economic output from each sector to derive average pollution coefficients for each sector. These coefficients are used with the supply chain computations to estimate supply chain pollution upstream of each product group.

The number of environmental interventions in the EIOLCA-model is large, 72 in total. This allows for a very detailed examination of the environmental profile of sectors, but is not operational in a screening procedure as in this project.

Therefore, a number of important interventions were selected and used in the further procedure. The following parameters were selected:

Table 2.
Environmental parameters selected for the prioritization of products.

The parameters that are omitted from the further procedure represent primarily an increased level of detail. Examples are energy consumption, where the software makes a distinction between 11 types of energy sources, use of non-renewable resources where the software calculates the consumption of 8 types of metals and alloys, and weighted toxics where the software makes a distinction point and non-point sources as well as between releases to air, water, land and underground. The only general omission from the procedure is fertilizers, where the software calculates the consumption of four different types. This omission is justified by the assumption that most fertilizer is consumed in agriculture and thus not will provide much additional information when a broad range of sectors is compared.

1.4.2 The pros and cons of EIONET-software

The use of the EIOLCA-software data to assess Danish environmental impacts is of course associated with inherent as well as practical problems, some of which are discussed in the subsequent sections.

The validity of the data to perform an analysis of Danish conditions depends mainly on two assumptions:

1. The US sectors are comparable to Danish sectors with respect to the products produced within the sector.
2. The relative environmental impacts per produced value unit are the same in Denmark and the United States

Ad 2. It is outside the scope of this study to verify that Danish and US production is comparable with respect to environmental impacts per produced unit of value. It seems, however, reasonable to assume that the technology used in the two countries in general is comparable, and hence also the environmental interventions from the processes.

There may however be exceptions that are overlooked in the automated calculation procedure. As an example, efforts to reduce the environmental impacts from a given sector through implementation of cleaner technology may have been initiated in one country and not in the other. This is presumably the case for steel production, where the only Danish producer after implementation of several cleaner technology projects claims to be more energy-efficient and less polluting than its European competitors.

It is an open question how subsidies and taxation of specific sectors and products affect the results. Obviously it has an influence of the economy of the sectors, but how this is reflected in the economic IO-tables has not been investigated in the present study. If such subsidies are included in the economic overview in the IO-model, the affected sectors will be comparably underrated with respect to environmental impacts, because the value of the products has been "artificially" increased.

The second assumption must thus be regarded as questionable, because there are potentially large differences between the Danish and US economy. It should, however, be remembered that the prioritization and selection in the project is not based on absolute values, but on impacts per produced value unit. The basic assumption is thus that the difference between the impacts from producing products with a certain value in different sectors is comparable in Denmark and the United States.

A third problem that is inherent to IO-analysis is that the use and disposal stages are not included in the calculations. The only way to include the two stages is to conduct a life cycle assessment of relevant or selected product groups, but this is outside the scope of this study. Obvious examples of the importance of the use stage are energy-consuming products like cars and electronic equipment, where life cycle assessments have shown that the use stage may cause environmental impacts that are more than ten times greater than in the production stage. An example of the importance of the disposal stage is lead-containing products that have a large potential for impacts if not disposed in the best possible way.

1.4.3 Results of step 2

1.4.3.1 Environmental ranking

The 95 product groups are rated in three categories, "high" "medium" or "low" importance in each of the impact parameters examined (see Table 2). For each of the parameters the 95 product groups have been sorted by their contribution to the parameter and subsequently been divided into three groups of equal size.

1.4.3.2 Combined environmental/economic ranking

Subsequently, the economic importance (measured as the Danish supply of a product group) is combined with the environmental importance (measured in impacts per value unit) for each of the impact parameters by a simple multiplication. Finally, the resulting figures are divided into three groups of approximately the same size by using the same criteria as applied in the environmental rating.

The following table gives an overview of the distribution of the ranking of the 95 product groups, both based on the environmental impact parameters alone, and in combination with their economic importance.

Table 3.
Distribution of the ranking with respect to environment alone, and a combined environmental and economic ranking.

As it can be seen from the table, the three groups in the combined environmental/economic ranking are not of exactly the same size. The reason for this is that several product groups performed equally in the relatively simple ranking system and therefore were placed in the same group.

Table 4 shows which of the 96 product groups that have been ranked as high with respect to environmental importance, combined environmental and economic importance or in both. Please note that the description of the product groups is very summarily and therefore only gives an indication of the actual products included under the heading.

Table 4.
Overview of the products that have been ranked as high in the environmental assessment, the combined economic and environmental assessment, or in both.

1.4.4 Other product characteristics

In order to provide a more detailed overview of the relation between sub-sectors and product groups, additional statistical information was requested from Statistics Denmark.

Firstly, the Danish Statistics of raw materials was used to create an overview of which raw materials (specified on 2-digit KN-nomenclature level) are used in which sectors (specified on 3-digit NACE-code level). This exercise gives a good indication of which sub-sectors that use a specific product group as raw material. Secondly, the Statistics of foreign trade was used to create an overview at the same level of detail regarding which product groups that are imported and exported to and from specific (sub-)sectors.

This statistical information can be used to give an indication of how the product groups (as defined in the statistical information) are positioned in larger product chains and accordingly also to indicate the environmental "properties" of such product chains. This was investigated in more detail for the 14 selected product groups (see below) and included in the total description of these 14 product groups.

1.4.4.2 Environmental labeling and green purchasing guidelines

1.5 Step 3 – Selection of relevant product groups

The third step – Selection of relevant product groups – was conducted by a combined search in the established database for product groups with a "high" environmental ranking and a "high" environmental/economic ranking at the same time. The products identified in this way are in the procedure regarded as those that potentially are relevant for future efforts.

The selected product groups were subsequently described with respect to their relation to the sectors that are most important for their presence in Denmark, characteristics of relevant sectors and their companies, important environmental parameters, and the position of the products in relevant product chains.

1.6 Step 4 – Previous efforts related to sub-sectors

The information compiled in the previous steps and stored in the database is assessed in step 5. Firstly, an overview of all previous efforts is created and secondly, suggestions for the targets in 2002 were developed. Both of these assessments are reported in separate chapters, but are not described in detail here.

1.8 Step 6 – Assessment of the action potential in relevant sub-sectors

1.9 Discussion of the methodology

The methodology presents a new approach to environmental assessment of products at a macro-economic level. As indicated previously, this is associated with a number of (large) uncertainties that must be kept in mind when the results are used for decision-making.

In the following sections, some of the most important uncertainties are addressed, and suggestions for future improvements are given. Firstly, however, it is stressed that precise information on environmental impacts from products and services can only be established by using very detailed life cycle assessments (LCA), and this can take months or even years for just a single product. As there are literally thousands of different product groups on the market - and several suppliers of each product group – the LCA approach is not possible on this level of decision-making.

1.9.1 Use of information from the United States in the assessment and selection of product groups

The major uncertainty in the methodology, at least on the psychological level, is probably that the assessment is based on environmental interventions in the United States. The economies in Denmark and the United States are very different in many respects, and this may also apply to the environmental impacts associated with the economic activities.

1.9.2 No inclusion of final use and disposal

A major limitation of input-output analysis to examine the environmental impacts of products is that they do not include the use of the products or their final disposal.

There are no short cuts to handle this problem. The environmental impacts from energy-consuming products are in IO-analysis alone related to their production, taking all upstream interventions into account, but omitting downstream interventions that may be significantly higher. The only "numeric" solution seems to be to use knowledge obtained from LCAs as supplementary information, but as already mentioned this information is seldom readily available. Therefore, the only viable way at the present time seems to be to combine the information from the IO-analysis with "common sense knowledge" from experienced persons.

1.9.3 The level of detail in the environmental assessment is not satisfactory

It can be argued that the environmental interventions used for prioritization and selection in the current procedure are not sufficiently broad. Although 72 different types of interventions can be calculated, a full picture is not obtained.

The argument is true in the sense that the level of detail is less than in LCA, where a well-established methodology allows for inclusion of an infinite number of interventions and still creates a relatively operational overview.

It is possible to include more interventions in the current selection procedure, but it will of course require more resources to do so. In fact, some of the impact parameters are similar to those used in LCA, with global warming potential as a prominent example. It is also possible to aggregate other interventions and produce results that are similar to those in a LCA. An example is acidification, where SO₂ and NO_x are the dominant contributors in almost all LCAs. This information is readily available from the IO-analysis, and it is very easy to aggregate these into sulfur dioxide equivalents as it is done in LCA.

It was a deliberate choice in the current project to only use a limited amount of impact parameters. Within the short project period, about two months, a methodology should be established that at the same time could address a total range of product groups in the Danish economy and provide an overview of the impacts that could be used for decision-making. If the full possible range of impact parameters were used, a matrix of 96 (product groups) times 72 (interventions) would have been the result. It was the opinion of the project team that this would not be operational, and it was therefore suggested to reduce the number of interventions to eight parameters, that each identified important environmental properties of a product.

It must be recognized that in doing so, important information may be missing. Water-borne emissions are only included as an element under the heading "total toxic emissions" and it is accordingly not possible to give an assessment of the eutrification potential. Likewise, only consumption of copper is included in the present study as an indicator of consumption of non-renewable materials. The EIONET-software gives the possibility of including a broader range of metals and alloys, but in the current context, copper was used because of its short supply adequacy.

In a further development it may also be possible to use Danish statistical information on some impact categories that are currently integrated at a low level of detail.

Statistics Denmark is currently developing information on Direct Material Input (DMI) and Total Material Requirements (TMR). DMI and TMR are inventories for the draw on non-renewable resources, distributed on product groups and sectors in Denmark and globally. The sector definitions are identical to those used in the national accounting system, and are therefore different from those used in the present project. It is, however, possible to convert these with access to the basic statistical information and knowledge about how to convert.

Based on the report "Status og perspektiver på kemikalieområdet" (Miljøstyrelsen 1996) (Status and perspectives in the chemical area, the Danish Environmental Protection Agency 1996), it is possible to relate the compounds on the "List of unwanted substances" to a number of specific products. These can again be related to specific sub-sectors by their KN-codes. The work must be done manually and is assumed to be relative demanding on human resources.

Danish waste statistics is rather detailed and is based on the information that companies in different sectors are obliged to register and report to the authorities. It may be possible to extract information on sector-related waste production and transfer these to the product level. Another possibility is to use information from the project "Affaldstunge brancher" (industries with large amounts of dangerous wastes) that includes a mapping of amounts and types of waste in selected sectors that are known to produce relatively large amounts of waste. This latter approach can, however, not be used consistently for all sectors.

1.9.4 Lack of detail on the product level

The model operates with only 95 product groups (on 2-digit KN-code level) that are related to 106 production sectors and 40 trade and service sectors. This is an intentional choice, based on the request for a consistent assessment of all product groups.

The consequence of the choice is that the calculated impacts from some product groups cover a broad range of products. The group "Products of iron and steel" thus includes products ranging from nails to stoves and bridges.

To increase the level of detail for products, it is necessary to use information on the 4-digit level for KN-codes. This will increase the number of product groups to about 1200 and thereby also increase the practical work with the assessment significantly. Furthermore, the EIONET-software only includes 485 product groups, and some additional work with relating the two lists of product groups to each other must be anticipated.

An increased level of detail for the product groups may cause problems when relating the information to the trade in some sectors. Even on the 3-digit DB-93 code level used in the current model it is necessary for reasons of confidentiality to aggregate information for several sub-sectors. Experience shows that it is possible to establish and use information on a 4- or 5-digit DB-93 code level for some sectors, but it is not possible to increase the level of detail in a consistent way. The effort will also require special extracts from Statistics Denmark to replace the current statistical background material.