resource efficiency metrics – initial...
TRANSCRIPT
1 Resource Efficiency Metrics – initial findings
Resource Efficiency Metrics – initial findings Anne Owen, Jannik Giesekam and John Barrett for the University of Leeds 02.08.18
Executive summary
This study aims to develop a set of indicators to understand the changing resource productivity
in the UK in a way that has more policy relevance. This document sets out initial findings and
suggests how the research can be further developed. In particular, it:
develops and calculates a carbon-based metric of resource efficiency;
looks into decomposition of the metric by sector to enable comparisons across sectors;
shows how well the metric indicates the extent of decoupling of raw material consumption from economic activity.
The document also sets out a work plan for a comparison of carbon intensity of materials and
products over their lifetime.
Analyses undertaken
Calculated the raw material footprint (MF) associated with UK consumption using a multi-regional input output database for the period 1997-2014. This MF includes all material extraction associated with the supply chains servicing UK final demand.
Compared the calculated MF with results from the ONS, Eora and Global Resource Accounting Model (GRAM) databases.
Conducted a decomposition analysis to understand the effect on UK carbon emissions of the changing carbon intensity of materials; material intensity; and final demand.
Conducted a secondary decomposition analysis on selected product groups.
Proposed an approach to identify key sector products and a means of compiling a supplementary Index of Product Resource Efficiency.
Findings
In 2014, the UK’s MF was 1,092 million tonnes of raw material.
81% of the UK’s MF is from materials extracted abroad.
The UK’s MF is made up of construction minerals (45% by weight), biomass (27%), fossil fuels (21%) and metal ores (7%).
Measured as extracted tonnes of material used by the economy per £ of GVA, the material intensity of UK production and imports reduced throughout the analysis period. The same is true of carbon intensity (carbon emissions per £ of GVA). This change in intensity reflects changes in efficiency, the structure of the economy and trade.
Between 1997 and 2012 the carbon intensity of materials (carbon emissions per tonne of extracted materials) produced in the UK was increasing. This is because the carbon intensity of production (tonnes CO2e per £’000 of GVA from UK production) fell but at a slower rate than the material intensity. Post 2012 the carbon intensity of materials produced in the UK reduced.
The carbon intensity of materials imported has reduced slightly over the time period 1997-2014. This is because, for imported materials, the carbon intensity of production fell (improved) at a faster rate than the material intensity of production.
Results from the decomposition analysis reveal final demand to be a positive driver of the emissions increase in non-recession years. The carbon intensity of materials acts as a negative driver of emissions as the proportion of imports increases (because the carbon intensity of imports has fallen).
Results from the product group decomposition analysis show that decreasing material intensity has contributed to emissions reductions for agriculture, forestry & fishing products and finance & business services but is a positive driver for energy & water products (where material intensity has increased). Effects of final demand and carbon intensity of materials on carbon emissions are similar to the total UK results.
2 Resource Efficiency Metrics – initial findings
There is a strong link between resource use and carbon emissions, with a small number of sectors accounting for the majority of the carbon and material footprints. For instance, the 30 sectors accounting for 80% of the total carbon footprint in 2014 also accounted for 62-85% of each material footprint. This suggests that an index which tracks key sectors and products, could indicate progress in improving resource efficiency and reducing carbon emissions.
Recommendations
We recommend the annual production of a UK material footprint account with refinements to that which is currently available. Key metrics, such as the carbon intensity of materials, could be tracked using this account.
Further project work packages should develop a supplementary indicator of progress across key products, such as the proposed Index of Product Resource Efficiency.
An additional project should adopt alternative econometric modelling approaches to deliver the desired secondary and tertiary outputs, including a deeper understanding of how prices might affect consumption of materials.
3 Resource Efficiency Metrics – initial findings
Terminology
When describing the results of the analyses in this report a common terminology is adopted.
The following pages are intended to provide a handy reference summarising the model
regions, materials, product groups and commonly used terms.
The model used in this report is a UK multi-region input-output database with a materials
extension. The model contains 106 defined sectors. The output of each of the model’s defined
sectors is referred to as a ‘sector product’. The output of a sector located in a particular region
is referred to as a ‘regionalised product’; and the particular goods or services produced in any
given sector are simply referred to as a ‘product’. The model regions and resources are:
Model resources
Code Resource Code Resource
GHG Greenhouse gases COAL Coal
BANI Biomass animals CNST Construction minerals
BFEE Biomass feed GAS Gas
BFOO Biomass food OIL Oil
BFOR Biomass forestry ORES Metal ores
Model regions
Code Region Code Region
UK United Kingdom KOR South Korea
AUS Australia NAM North America
CHN China CSA Central and South America
IND India EU European Union
IDN Indonesia REU Rest of Europe
JPN Japan MID Middle East
RUS Russia ROW Rest of World
Where results are classified by product groups these sectors are grouped into:
Product group
Manufacturing Agriculture, forestry & fishing
Transport & communication Construction
Energy & water Financial & business services
Public admin, education & health Other services
Wholesale and retail trade
Definitions
The following terms are used throughout the report and the definitions that we use are
explained below with accompanying notes:
Carbon footprint: the full amount of greenhouse gas emissions required to meet a nation’s
final demand for all goods and services.
Note that the carbon footprint for the UK is an official statistic (Defra, 2016). The carbon
footprint in this report was calculated using a slightly different model (one with a greater
disaggregation of trade regions) and therefore the values reported are slightly different. The
carbon footprint includes emissions arising from a product’s production, use and disposal and
so includes emissions arising from the process of recycling, incineration and landfill. If a
product is manufactured using a greater proportion of recycled material as opposed to virgin
material compared to previous years, the model will identify that waste material is part of the
product’s supply chain. If the emissions associated with waste (recycling output) are smaller
than emissions associated with processing the virgin material, we will observe a reduction in
the carbon footprint of the final good compared to the previous year. However, the present
4 Resource Efficiency Metrics – initial findings
industrial classifications in the model are quite aggregated and there are not separate sectors
for recycling of different types of material. This means that it is not possible to accurately
analyse the effects of recycling.
Carbon intensity: tonnes of greenhouse gas emissions per £’000 of gross value added. The
carbon footprint divided by gross value added.
Carbon intensity of materials: tonnes of greenhouse gas emissions per tonne of material
extraction. The carbon footprint divided by the material footprint.
Consumption-based account: a method for allocating resource use or emissions to the point
of consumption rather than where the materials were extracted or emissions released.
Footprints are calculated using consumption-based approaches.
Domestic extraction: materials extracted within a country’s border.
Domestic material consumption: domestic extraction plus imported materials minus
materials that are exported.
Embodied emissions: emissions released as part of the supply chain of producing a final
demand product
Embodied materials: materials extracted as part of the supply chain of producing a final
demand product.
Note that this includes: materials that make up the final product; materials that were waste
products during the production stages; and the materials that were used to make machinery
and transport the product during its production.
Gross value added: a measure of the value of goods and services produced in the UK. This
value is calculated in the UK national accounts and is used in the calculation of Gross
Domestic Product.
Material footprint: the full amount of raw materials required to meet a nation’s final demand
for all goods and services.
Material intensity: tonnes of materials per £’000 of gross value added. The material footprint
divided by gross value added.
Note that where separate figures are presented for the material intensity of imports and
domestic production, these are based upon a calculation of the embodied value added by
source. Therefore the GVA of domestic production is the portion of UK GVA that remains in
the country associated with UK goods bought by UK consumers. Meanwhile, the imported
GVA is foreign GVA embodied in imports to satisfy UK final demand.
Material intensity of carbon: tonnes of material extraction per tonne of greenhouse gasses
emitted. The material footprint divided by the carbon footprint.
5 Resource Efficiency Metrics – initial findings
1. Introduction
Desired project outputs
This initial paper is part of a larger project with the following desired outputs:
Primary outputs:
1. A carbon-based metric of resource efficiency based on a full life cycle and full supply chain approach and which includes embodied emissions.
2. A comparison of carbon intensity of materials and products used over their lifetime. 3. An appraisal of feasibility of decomposition of the metric by sector to enable
comparisons across sectors. 4. An analysis of how well the metric robustly indicates the extent of decoupling of raw
material consumption from economic activity.
Secondary outputs:
1. A critical appraisal of supplementary metrics based on monetised carbon. 2. A supplementary value-based metric that would take into account the price of
materials.
Tertiary outputs:
1. Identification of material flow interrelationships which facilitate economic modelling. 2. An appraisal of the implications of any new metrics for modelling, e.g. for forward
looking models for energy demand, resource efficiency and growth, including policy impacts.
This initial report provides calculations for primary outputs 1, 3 and 4, sets out a work plan for
output 2, and next steps more generally.
Approaches to accounting for resource use
The study aims to develop metrics that can be used to measure resource efficiency in the UK.
It is important that these metrics can be tracked over time and that they can be used as policy
relevant evidence. Before developing a metric of resource efficiency, we first examine many
ways in which a country’s resource use can be accounted for. This is widely known as the
study of material flow accounting (MFA).
Domestic extraction (DE) takes a fully territorial perspective on materials use and accounts
only for those materials extracted within the country’s border (Hirshnitz-Garbers et al., 2014).
This indicator does not account for materials that are imported to or exported from a nation.
Domestic material consumption (DMC) is calculated by taking the DE and adding imported
materials and subtracting those that are exported (Fischer-Kowalski et al., 2011). The DMC is
described as “the most prominent indicator in MFA and accepted as a headline indicator for
resource use and resource efficiency” (Eisenmenger et al., 2016, p178). The DMC Trade flows
are measured according to the mass that crosses the country’s border. However, as
Eisenmenger et al. (2016) point out, because traded commodities can be at differing stages
of processing, their mass on crossing the border differs from the initial mass of extracted
material required to produce them meaning that the full resource use is not captured.
Raw material equivalents (RME) account for the full upstream material requirements needed
to produce traded goods (Wiedmann et al., 2015). If the RME of imports is added to the DE
and the RME of exports removed, we arrive at a measure of full material consumption. This
measure will account for the full mass of material required in the production of traded goods.
The terms raw material consumption (RMC) and material footprint (MF) are seemingly
used interchangeably (Wiedmann et al., 2015) with both meaning the full resource use
associated with a country’s consumption. However, in order to account for the full supply chain
resource use in traded goods we would need an understanding of the production processes
6 Resource Efficiency Metrics – initial findings
involved in making the product. Eisenmenger et al. (2016) describes two approaches to
quantifying RME.
1) Using material coefficients from life cycle inventories (LCI). 2) Using environmentally extended input-output analysis (EE-IOA).
The main issues with the LCI approach is that there is a possibility of double-counting if
products pass multiple borders; truncation errors are introduced when the full material supply
chain is not accounted for; and conversion factors are sourced from multiple years making it
difficult to assess improvements over time. An IO approach to calculating raw materials in
imports and exports avoids these problems but brings its own unique issues. IO tables tend to
use monetary data to describe the production processes required in making a finished good.
Using money as a proxy for material use can lead to some misallocation of impact. In addition,
the sectors involved in IO tables tend to group many products together which can cause issues
when price to weight ratios differ across the aggregated products. For further discussion of
the advantages and disadvantages of LCI and EE-IOA approaches see Hirshnitz-Garbers et
al. (2014).
Since this work on Resource Efficiency Metrics aims to be policy relevant and used alongside
the existing greenhouse gas (GHG) consumption-based account (CBA) for the UK1, we will
be using a similar IOA approach to calculating the resource use of the UK. There is some
discussion in the literature as to whether the full material use calculations should follow the
description of:
𝑅𝑀𝐶 = 𝐷𝐸 + 𝑅𝑀𝐸𝑖𝑚𝑝𝑜𝑟𝑡𝑠 − 𝑅𝑀𝐸𝑒𝑥𝑝𝑜𝑟𝑡𝑠
(the ONS approach) or whether MRIO (multiregional input-output) analysis should be used to
calculate the full materials used to satisfy a nation’s final demand. While these approaches
seem similar they are subtly different. The first, a Trade Adjusted Inventory (TAI) (Kanemoto
et al., 2012; Peters & Solli, 2010) removes exported materials by using IO tables to consider
the full material requirements needed to produce that product in the exporting country. Import
impacts are calculated by determining the full material requirements needed to produce that
product in the import country. This method does not allow for re-imports to be accurately
accounted for; where a semi-finished product is exported from the source country, processed
abroad and then re-imported. The MRIO approach takes the DE figures and reallocates them
to measures of final demand. In essence calculating the materials extracted to satisfy a
country’s final demand. The MRIO approach can quantify flows that are re-imported, which is
a key advantage.
The UK GHG CBA, uses the MRIO approach. So an additional advantage with using the MRIO
approach in this study is that it will ensure that the two metrics (UK GHG CBA and the metrics
recommended for adoption later in this paper) are consistent and can be used in combination
to further understand resource efficiency. In addition, the most recent and extensive work in
this field, the ‘material footprint of nations study’ (Wiedmann et al., 2015) uses an MRIO
approach. In this paper, the term MF is used rather than RMC. Since we are following the
same method, we will refer to the RMC as MF. In the supporting information accompanying
the study, Wiedmann et al. (2015) also provide a method for calculating 𝑅𝑀𝐸𝑖𝑚𝑝𝑜𝑟𝑡𝑠 and
𝑅𝑀𝐸𝑒𝑥𝑝𝑜𝑟𝑡𝑠 using an MRIO approach.
Resource use calculations for the UK
Domestic Material Consumption basis
The UK reports a Material Flows Account (ONS, 2016a) which includes the domestic
extraction (DE) of biomass, metal ores, non-metallic minerals and fossil energy materials. This
1 https://www.gov.uk/government/statistics/uks-carbon-footprint
7 Resource Efficiency Metrics – initial findings
account is compiled using a methodology developed by Eurostat and the estimates are
classed as experimental. In addition to DE, this account also includes a measure of the total
DMC and DMC by material type. DE and DMC accounts for the UK are also constructed by
Vienna University (WU, 2016). Table 1, below, shows that the figures for both DE and DMC
are slightly higher in the WU account compared to the ONS figures. DE is also described in
the Eora database (which was used in Wiedmann et al's. 2015 ‘material footprint of nations
study’) as a territorial material account. Eora’s DE figures are lower than both the ONS and
WU and the DE has not been updated past 2008.
Material Footprint basis
Turning now to calculations on a raw material basis, in 2016, the ONS also reported a measure
of RMC (ONS, 2016b). ONS uses the 𝑅𝑀𝐶 = 𝐷𝐸 + 𝑅𝑀𝐸𝑖𝑚𝑝𝑜𝑟𝑡𝑠 − 𝑅𝑀𝐸𝑒𝑥𝑝𝑜𝑟𝑡𝑠 approach and
so will not take account of re-exports and re-imports. In addition, the Eurostat model used to
calculate the materials intensity of traded goods assumes that UK imports have the same
profile as the European average, when in reality the UK’s trade partners will be different. This
is important because production practices vary worldwide and knowing exactly where the UK
imports come from will give a more accurate estimate of the material footprint. The final three
columns in Table 1, below, compare the ONS results (on a raw material basis) with both the
Eora MF figures and the 2005 estimate from Bruckner et al.'s 2012 study, which uses the WU
DE figures with the Global Resource Accounting Model (GRAM) which is an MRIO database.
As the final three columns show, the ONS MF calculations are significantly lower than those
calculated using Eora and GRAM. However, Eora calculations post 2008 are likely to be
inaccurate since they use 2008 DE figures in the calculations.
Table 1: A comparison of DE, DMC and MF from the ONS, WU and Eora and GRAM MRIO databases. All figures in million tonnes.
Domestic extraction (DE) Domestic material consumption (DMC) Material footprint (MF)
ONS Eora WU ONS WU ONS Eora GRAM
1995 735 518 783 799 967
1996 726 525 761 777 1,103
1997 727 522 763 782 1,169
1998 723 516 752 776 1,276
1999 733 527 766 777 1,318
2000 718 507 767 739 780 896 1,332
2001 698 481 748 743 788 955 1,271
2002 683 477 729 722 763 895 1,252
2003 664 459 706 723 759 867 1,332
2004 655 437 705 752 787 876 1,406
2005 622 410 673 733 774 871 1,442 1,166
2006 601 367 655 724 764 842 1,512
2007 590 388 649 719 760 859 1,562
2008 558 371 600 675 703 796 1,470
2009 490 371 523 593 625 683 1,294
2010 474 371 504 577 604 698 1,311
2011 460 371 500 582 693 1,276
2012 421 371 459 563 696 1,272
2013 419 371 454 570 712 1,257
2014 435
This study aims to calculate the UK’s MF using the best available data. For this, we require a
database built using the most accurate model of the UK economy; an understanding of how
the UK trades with other countries and how these countries trade with each other; and a
dataset of DE data for the UK and all other countries. Section 2 explains the source data and
methods used to calculate the UK’s MF.
8 Resource Efficiency Metrics – initial findings
2. Data and methods
Choice of domestic extraction data
As Figures 1 and 2 show, the pattern of DE is similar across all three datasets, with WU
reporting slightly higher than the ONS and significantly higher than Eora.
Figure 1: Comparison of databases of domestic extraction
9 Resource Efficiency Metrics – initial findings
Figure 2: Comparison of databases of domestic extraction - construction minerals and fossil fuels
For subsequent analysis in this paper we have decided to construct the UKMRIO using the
ONS DE for the UK and the WU data for every other country’s DE. We made this decision
because the Eora data has not been updated since 2008 and the UKMRIO model philosophy
is to use UK Government data wherever possible.
Construction of the input-output database
The University of Leeds (UoL) calculates the UK’s officially reported CBA (consumption based
account) for CO2 and all other GHG emissions (Defra, 2016). To calculate the CBA, UoL has
constructed the UKMRIO database. Since the CBA is a National Statistic2, the MRIO database
is built using IO data produced by the UK’s Office of National Statistics (ONS). This data is
supplemented with additional data on UK trade with other nations and how these other nations
trade between themselves from the EXIOBASE MRIO database (Tukker et al., 2013; Wood et
al., 2015). The ONS produces Supply and Use tables (SUT) on an annual basis at a 106
sector disaggregation (ONS, 2016). These SUTs are the data that underpins the UKMRIO
database. The use tables are combined use tables, meaning that the inter-industry transaction
table is the sum of both domestic transactions and intermediate imports, and the final demand
table shows the sum of both domestic and imported final products. On a 5-yearly basis, the
2 https://www.gov.uk/government/statistics/uks-carbon-footprint
10 Resource Efficiency Metrics – initial findings
ONS produces a set of analytical tables where the use table is of domestic use only. Final
demand is also split to show domestic purchases separately. Taking proportions of domestic
versus imports from the analytical tables, we are able to extract domestic and import data from
the annual SUT tables. Imports to intermediate industry is now a single row of data and exports
to intermediate and final demand forms a single column of data.
Data from the EXIOBASE MRIO database (Tukker et al., 2013; Wood et al., 2015) is used to
further disaggregate the import and export data to sectors from other world regions. Data from
EXIOBASE is also used to show how foreign sectors trade with each other, but first the data
must be converted to Great Britain Pounds (GBP). The EXIOBASE MRIO database is mapped
onto the UK’s 106 sector aggregation. Once this step has been performed, the data can be
further aggregated by region. Since EXIOBASE contains data from 49 regions, we are able to
select the most appropriate regional grouping for the trade data.
2.2.1. Choice of regions
For this MRIO study, we construct 14 regions (see Table 2). These regions were chosen
because they represent a combination of the UK’s most important trade partners and show
the regions which extract large volumes of particular materials. For example, fossil fuels from
the Middle East and metal ores from Australia.
Table 2: Regional groupings for the materials extended UK MRIO database
Region code Region Region code Region
UK United Kingdom KOR South Korea
AUS Australia NAM North America
CHN China CSA Central and South America
IND India EU European Union
IDN Indonesia REU Rest of Europe
JPN Japan MID Middle East
RUS Russia ROW Rest of World
2.2.2. Allocation of domestic extraction data to UKMRIO sectors
The ONS and WU materials databases report 10 types of material extraction. Each of these
must be mapped to one or more of the 106 sectors in the UKMRIO database. Table 3 shows
how the sectors correspond to the material extension.
Table 3: Mapping extracted materials to extraction sectors
Products of agriculture, hunting and related services
Products of forestry, logging and related services
Fish and other fishing products; aquaculture products; support services to fishing
Coal and lignite
Extraction of Crude Petroleum And Natural Gas & Mining of Metal Ores
Other mining and quarrying products
Biomass animals 1 0 1 0 0 0
Biomass feed 1 0 0 0 0 0
Biomass food 1 0 0 0 0 0
Biomass forestry 0 1 0 0 0 0
Coal 0 0 0 1 0 0
Construction 0 0 0 0 0 1
Gas 0 0 0 0 1 0
Industrial materials 0 0 0 0 1 1
Oil 0 0 0 0 1 0
Ores 0 0 0 0 1 0
11 Resource Efficiency Metrics – initial findings
Input-output analysis
Consider a transactions matrix Z , showing sales by each sector (rows) and the purchases by
each sector (columns). Reading across a row reveals which other sectors a single industry
sells to and reading down a column reveals who a single sector buys from in order to make its
product output. A single element, 𝐳𝐢𝐣, within 𝐙 represents the contributions from the ith supplying
sector to the jth producing sector in an economy. The 𝐙 matrix is in monetary units.
Reading across the table, the total output (𝑥𝑖) of a particular sector can be expressed as:
𝑥𝑖 = 𝑧𝑖1 + 𝑧𝑖2 + ⋯ + 𝑧𝑖𝑛 + 𝑦𝑖 (1)
where 𝑦𝑖 is the final demand for that product produced by the particular sector. The IO
framework shows that the total output of a sector is the sum of its intermediate and final
demand. Similarly if a column of the IO table is considered, the total input of a sector is the
sum of its intermediate demand and value added in profits and wages (𝐡).
If each element, 𝑧𝑖𝑗 , along row 𝑖 is divided by the output 𝑥𝑗, associated with the corresponding
column 𝑗 it is found in, then each element in 𝐙 can be replaced with:
𝑎𝑖𝑗 =𝑧𝑖𝑗
𝑥𝑗 (2)
forming a new matrix 𝐀, known as the direct requirements matrix. Element 𝑎𝑖𝑗 is therefore the
input as a proportion of all the inputs in the production recipe of that product.
(2) can be re-written as:
𝑧𝑖𝑗 = 𝑎𝑖𝑗𝑥𝑗 (3)
Substituting for (3) in (1) forms:
𝑥𝑖 = 𝑎𝑖1𝑥1 + 𝑎𝑖2𝑥2 + ⋯ + 𝑎𝑖𝑛𝑥𝑛 + 𝑦𝑖 (4)
Which, in matrix notation is + 𝐲 . Solving for 𝐱 gives:
𝐱 = (𝐈 − 𝐀)−𝟏𝐲 (5)
(5) is known as the Leontief equation and describes output 𝐱 as a function of final demand 𝐲.
𝐈 is the identity matrix, and 𝐀 shows the inter-industry requirements. (𝐈 − 𝐀)−𝟏 is known as
the Leontief inverse (denoted hereafter as 𝐋). (5) can be re-written as:
𝐱 = 𝐋𝐲 (6)
Consider a row vector 𝐟 of annual material extraction from each production sector. It is possible
to calculate material intensity (𝐞) by dividing the total emissions of each sector by total sector
output (𝐱). 𝐞 can be any of the raw materials that we have data for, or the sum of all extracted
materials.
𝐞 = 𝐟�̂�−𝟏 (7)
In other words, 𝐞 is the coefficient vector representing extraction per unit of output.
Multiplying both sides of (6) by 𝐞 gives:
𝐞𝐱 = 𝐞𝐋𝐲 (8)
and from (7) we simplify (8) to:
12 Resource Efficiency Metrics – initial findings
𝐟 = 𝐞𝐋𝐲 (9)
However, we need the result (𝐟) as a flow matrix (𝐐), showing the source sector and region of
extraction embodied in all products, and so we use the diagonalised �̂� and �̂�:
𝐐 = �̂�𝐋�̂� (10)
𝐐 is the consumption based material extraction account, or Material Footprint calculated by
pre-multiplying 𝐋 by extraction per unit of output and post-multiplying by final demand.
Materials are reallocated from production sectors to final products. We can use 𝐐 in a number
of ways. If we sum the columns, it tells us the material footprint of products, with products
classified as the belonging to the ‘country of final purchase’ i.e. where we imported the finished
product from. If the rows are summed we calculate the material footprint by source sector and
nation.
2.3.1. Adjusting the economic data to constant prices
In order to make calculations of the material and carbon intensity of the UK’s resource use
and greenhouse gas consumption-based account we divide the environmental impact (either
in mass of material use or emissions generated) by the global value added (GVA) associated
with UK consumption. Before we can do this, we must convert all economic data in the UK
MRIO to constant prices to allow for year-on-year comparisons to be made. We chose a base
year of 2010 and adjust the tables for the years 1997-2009 and 2011-2014 to 2010 prices.
The adjustments follow the double deflation method explained in (Lan et al., 2016) using
Producer Price Inflation data from the ONS.
3. Results
In this section, we first compare the MF for the UK, as calculated by UoL’s MRIO database,
with the results from the ONS study, Eora and the GRAM databases. We then use the UoL
MRIO database to present the UK’s MF in greater detail, showing the impact by material type
and by the region of extraction. We move on to show the material intensity (kg material per £
GDP) and carbon intensity (kg CO2e per £ GDP) of UK consumption, again using UoL MRIO
data for calculating MF, before introducing metrics of material intensity of carbon and carbon
intensity of materials. The section concludes with an investigation into the drivers of change
in carbon emissions using a decomposition analysis to understand the role of changes in
carbon intensity of materials, material intensity and changes in final demand.
Throughout this section a common terminology is adopted, whereby the output of each of the
model’s 106 defined sectors is referred to as a ‘sector product’. The output of a sector located
in a particular region is referred to as a ‘regionalised product’; and the particular goods or
services produced in any given sector are simply referred to as a ‘product’. Results are
presented for the resources shown in Table 4.
Table 4: Model resources
Resource code Resource Resource code Resource
GHG Greenhouse gases COAL Coal
BANI Biomass animals CNST Construction minerals
BFEE Biomass feed GAS Gas
BFOO Biomass food OIL Oil
BFOR Biomass forestry ORES Metal ores
13 Resource Efficiency Metrics – initial findings
UK raw material footprint – comparison of results
Figures 3 and 4 compare results obtained from the UoL’s MRIO database with those from the
ONS study, Eora and GRAM databases.
Figure 3: Comparison of material footprint between databases
14 Resource Efficiency Metrics – initial findings
Figure 4: Comparison of material footprint between databases - construction minerals and fossil fuels
Calculating the MF of a nation is cutting edge science and as such, there is considerable
variation in estimates for the UK. We believe that the results calculated using the UKMRIO
database provide the most accurate estimate of the UK’s MF. We take DE for the UK from the
ONS (unlike Eora) but the MRIO approach to calculating the MF allows for a more
sophisticated treatment of the materials embodied in imports. We use the MF as calculated
by the UKMRIO database for all further analyses in this report.
15 Resource Efficiency Metrics – initial findings
UK’s raw material footprint by region of extraction
Figure 5 shows the contribution of extracted material to the UK material footprint between
1997 and 2014, broken down by the region of extraction of all materials (see ‘Terminology’ on
page 2 for regions key).
Figure 5: Raw material footprint for the UK between 1997 and 2014, broken down by region of extracted material
UK’s raw material footprint by material
Figure 6 shows the contribution of extracted material to the UK material footprint between
1997 and 2014, broken down by material.
Figure 6: Raw material footprint for the UK between 1997 and 2014, broken down by material
16 Resource Efficiency Metrics – initial findings
UK’s raw material footprint by material and region of extraction
Figures 7 and 8 show the contribution of extracted material to the UK material footprint
between 1997-2014, broken down by region of extraction for each of the materials.
Figure 7: Raw material footprint for the UK between 1997 and 2014, broken down by region of extracted material
17 Resource Efficiency Metrics – initial findings
Figure 8: Raw material footprint for the UK between 1997 and 2014, broken down by region of extracted fossil fuels
18 Resource Efficiency Metrics – initial findings
UK’s raw material footprint by source region and material
Figures 9-13 show the contribution of extracted material to the UK material footprint broken
down by source region.
Figure 9: Raw material footprint for the UK between 1997 and 2014 by source region
19 Resource Efficiency Metrics – initial findings
Figure 10: Raw material footprint for the UK between 1997 and 2014 by source region
20 Resource Efficiency Metrics – initial findings
Figure 11: Raw material footprint for the UK between 1997 and 2014 by source region
21 Resource Efficiency Metrics – initial findings
Figure 12: Raw material footprint for the UK between 1997 and 2014 by source region
22 Resource Efficiency Metrics – initial findings
Figure 13: Raw material footprint for the UK between 1997 and 2014 by source region
Trends in material and emissions intensity
In this section we consider the changing material and carbon intensities of UK consumption
over the time period 1997-2014 with a view to developing an indicator linking carbon and
materials.
3.6.1. Materials intensity
Dividing the material footprint by the GVA of the UK gives a measure of material intensity (in
tonnes of material per £’000 GVA). This can be split to show the intensity of domestic and
imported material (see Figure 14a overleaf). The material intensity of imports (in tonnes of
imported material per £’000 GVA) increased until 2007, after which it decreased. The material
intensity of domestic consumption has gradually decreased over time.
Similarly, dividing the GVA by the material footprint of the UK gives a measure in GVA per
tonne of material (see Figure 14b). The GVA per tonne of material of imports is quite stable,
and the GVA per tonne of material for domestic increases until 2012 before decreasing slightly.
23 Resource Efficiency Metrics – initial findings
Figure 14a: Material intensity of UK consumption (MF/GVA) broken down by domestic and imported sources
Figure 14b: Inverse material intensity of UK consumption (GVA/MF) broken down by domestic and imported sources
Table 5: Reduction in material intensity between 1997 and 2014
Region Reduction in intensity (MF/GVA)
All 22%
UK 61%
Imports 9%
Figure 15: Material intensity (MF/GVA) of UK consumption broken down by import sources
24 Resource Efficiency Metrics – initial findings
3.6.2. Carbon intensity
Dividing the carbon footprint by the GVA of the UK gives a measure of carbon intensity (in
tonnes of CO2e per £’000 GVA). This can be split to show the intensity of domestic and
imported emissions (in tonnes of imported emissions per £’000 GVA). Both the carbon
intensity of imports and domestic consumption has decreased.
Figure 16: Carbon intensity of UK consumption broken down by domestic and imported sources
Table 6: Reduction in carbon intensity between 1997 and 2014
Region Reduction in intensity (CF/GVA)
All 50%
UK 45%
Imports 38%
Figure 17: Carbon intensity of UK consumption broken down by import sources
25 Resource Efficiency Metrics – initial findings
3.6.3. Carbon intensity of materials
Dividing the carbon footprint by the material footprint gives a measure of the carbon intensity
of material measured in tonnes of CO2e per tonne material used. This can be done for each
of domestic production, exports and imports, by using the materials (tonnes) and carbon
emissions (tonnes CO2e) attributed to each of the three elements.
In UK production, the level of carbon emissions associated with each tonne of material used
increased until 2012 and has decreased in the last two years. Imports to the UK are less
carbon intensive in material terms. Over time, for every tonne of imported materials, there are
less emissions involved.
Figure 18: Carbon intensity of materials associated with UK consumption broken down by domestic and imported sources
For UK production over the period until 2012, material intensity (tonnes per £’000 GVA) fell
(improved) faster than carbon intensity, so the carbon intensity of materials (carbon emissions
per tonne of materials) increased. In the last two years, material intensity has increased,
reversing the trend in the carbon intensity of materials.
In our imports, carbon intensity is improving at a faster rate compared to material intensity so
the carbon intensity of materials decreased.
Figure 19: Carbon intensity of materials associated with UK consumption broken down by import sources
26 Resource Efficiency Metrics – initial findings
3.6.4. Material intensity of carbon
Dividing the material footprint by the carbon footprint gives a measure of the material intensity
of carbon measured in tonnes of material per tonne CO2e. This can be done for each of
domestic production, exports and imports, by using the materials (tonnes) and carbon
emissions (tonnes CO2e) attributed to each of the three elements.
In the UK, in terms of what we domestically produce, less and less material use is associated
with each tonne of CO2e emitted but our imports are more material intensive in carbon terms.
Over time, for every tonne of imported CO2e, there are more materials involved.
Figure 20: Material intensity of carbon associated with UK consumption broken down by domestic and imported sources
With UK production, material intensity is improving faster than carbon intensity, so the material
intensity of carbon (materials used per tonne of carbon emissions) decreases (dark blue line).
In our imports, material intensity is improving at a slower rate compared to carbon intensity so
the material intensity of carbon increases.
Figure 21: Material intensity of carbon associated with UK consumption broken down by import sources
27 Resource Efficiency Metrics – initial findings
3.6.5. Material intensity of carbon versus carbon intensity of material
The material intensity of carbon is the reciprocal (inverse) of the carbon intensity of material.
Therefore, the rate of increase in one is equal and opposite to the rate of decrease in the
other. It does not make sense to prioritise improvement in one metric over the other since they
essentially measure the same thing. We recommend the latter indicator (carbon intensity of
materials) be annually reported as talking about emissions associated with making materials
used in products makes more sense from a supply chain perspective and it allows for
calculations to consider the role of materials in reducing emissions.
Results by product group
In this section we argue that the carbon emissions associated with UK consumption will be
high if one or more of the following is true:
1. The carbon intensity of the material used to make the product is high; 2. The product is materially intensive; 3. The UK consumes large quantities of the product.
We consider the three factors above across a set of broad product groups (grouping results
for the 106 products from the UK’s supply and use table into 9 overarching groups shown in
Table 7).
Table 7: Product groups
Product group
Manufacturing Agriculture, forestry & fishing
Transport & communication Construction
Energy & water Financial & business services
Public admin, education & health Other services
Wholesale and retail trade
We compare results across the product groups for 2014 (the latest year of data) and also
present trends in material intensity, carbon intensity and carbon intensity of materials over
1997-2014.
3.7.1. Material intensity
Figure 22 overleaf reveals that the group of agriculture, forestry & fishing products is the most
materially intensive, containing over 6 tonnes of material per £’000 GVA. Figure 23 shows how
the material intensity of each product group has changed over time. There have been
substantial reductions in material intensity across all product groups over the analysis period,
but this trend has reversed in recent years for some product groups, such as construction.
28 Resource Efficiency Metrics – initial findings
Figure 22: Material intensity of product groups consumed by the UK (2014)
Figure 23: Material intensity of product groups consumed by the UK (1997-2014)
29 Resource Efficiency Metrics – initial findings
3.7.2. Carbon intensity
Figure 24 shows the relative carbon intensity of each product group. The most carbon
intensive product groups are agriculture, forestry & fishing and energy & water. Figure 25
shows how the carbon intensity of each product group has changed over time. There have
been substantial reductions in carbon intensity across all product groups over the analysis
period.
Figure 24: Carbon intensity of product groups consumed by the UK (2014)
Figure 25: Carbon intensity of product groups consumed by the UK (1997-2014)
30 Resource Efficiency Metrics – initial findings
3.7.3. Carbon intensity of materials
Figure 26 shows that transport & communication, followed by energy & water, are product
groups with the most carbon intensive materials involved in their production. Agriculture,
forestry & fishing and construction products have the least carbon intensive materials involved
in their production.
Figure 26: Carbon intensity of materials for product groups consumed by the UK (2014)
Figure 27 shows a mixed picture, in terms of the changing carbon intensity of materials over
the analysis period. Some product groups, such as public admin, education & health, have
significantly declined, whereas others, such as transport & communication have increased in
recent years.
Figure 27: Carbon intensity of materials for product groups consumed by the UK (1997-2014)
31 Resource Efficiency Metrics – initial findings
3.7.4. Final demand
When comparing product groups, it is important to consider not just the carbon intensity and
material intensity, but also total spending on each group. Figure 28 indicates that the most
money is spent on financial & business services, followed by public administration, education
& health services and manufactured products.
Figure 28: Final demand spend on product groups consumed by the UK (2014)
3.7.5. Material footprint
The total material footprint of the product groups is compared in Figure 29 for 2014.
Manufacturing has the largest total material footprint, whilst wholesale and retail trade has the
smallest.
Figure 29: Total material footprint of product groups consumed by the UK (2014)
32 Resource Efficiency Metrics – initial findings
The total UK material footprint and the relative contribution of each product group has changed
substantially over the analysis period (Figure 30). We see an increase in the footprint of all
product groups in the decade until 2007. This is followed by a general reduction through to
2009. Since 2009 the picture has been mixed with some product groups, such as construction,
substantially increasing, whilst others, such as agriculture, forestry & fishing have been
declining.
Figure 30: Changing UK material footprint by product group (1997-2014)
3.7.6. Carbon footprint
The carbon intensity of materials, material intensity and volume of spend (final demand) can
be multiplied together to form the carbon footprint of the product groups (Figure 31). The
product groups with the highest carbon footprint are manufactured goods, transport &
communication and agriculture, forestry & fishing.
Figure 31: Total carbon footprint of product groups consumed by the UK (2014)
33 Resource Efficiency Metrics – initial findings
The total UK carbon footprint and the share associated with each of these product groups has
changed over time (Figure 32). We see a general increase in emissions between 1997 and
2007, followed by a dip to 2009, followed by a period of minimal change from 2009 to 2014.
But how much has the carbon intensity of materials, material intensity and volume of spend
respectively contributed to these changes? Section 3.8 introduces a technique called
decomposition analysis to begin to explore this.
Figure 32: Changing UK carbon footprint by product group (1997-2014)
Decomposition analysis
In this section we use decomposition analysis to determine the role that each of the three
factors identified in Section 3.7 has on the UK’s changing carbon footprint.
𝐶𝑝 is the carbon footprint of product 𝑝
𝑀𝑝 is the material footprint of product 𝑝
𝑌𝑝 is the final UK expenditure (in 2010 prices) of product 𝑝
Then
𝐶 = 𝐶
𝑀.𝑀
𝑌. 𝑌
For simplicity, let
𝐶𝐼𝑀 =𝐶
𝑀 carbon intensity of material
𝑀𝐼 =𝑀
𝑌 material intensity of spend
Now the difference in the carbon emissions in time 𝑡 and time 0 is
𝐶𝑡 − 𝐶0 = 𝐶𝐼𝑀𝑡𝑀𝐼𝑡𝑌𝑡 − 𝐶𝐼𝑀0𝑀𝐼0𝑌0
Using a mathematical technique called decomposition analysis (Dietzenbacher & Los, 1998),
we can calculate the difference in carbon emissions as the product of three effects:
𝐶𝑡 − 𝐶0 = 𝐶𝐼𝑀𝑒𝑓𝑓𝑒𝑐𝑡𝑀𝐼𝑒𝑓𝑓𝑒𝑐𝑡𝑌𝑒𝑓𝑓𝑒𝑐𝑡
34 Resource Efficiency Metrics – initial findings
These are: the effect that the carbon intensity of material has on the changing UK carbon
footprint; the effect that the materials per £ of final demand spend has on the changing UK
carbon footprint and finally the effect that changes in demand has on the changing UK carbon
footprint. This can also be written as:
𝐶𝑡 − 𝐶0 = ∆𝐶𝐼𝑀 (𝑀𝐼0. 𝑌0 +1
2(𝑀𝐼0. ∆𝑌 + ∆𝑀𝐼. 𝑌0) +
1
3(∆𝑀𝐼. ∆𝑌))
+ ∆𝑀𝐼 (𝐶𝐼𝑀0. 𝑌0 +1
2(𝐶𝐼𝑀0. ∆𝑌 + ∆𝐶𝐼𝑀. 𝑌0) +
1
3(∆𝐶𝐼𝑀. ∆𝑌))
+ ∆𝑌 (𝐶𝐼𝑀0. 𝑀𝐼0 +1
2(𝐶𝐼𝑀0. ∆𝑀𝐼 + ∆𝐶𝐼𝑀. 𝑀𝐼0) +
1
3(∆𝐶𝐼𝑀. ∆𝑀𝐼))
We use the equation above to calculate the effect of each factor. Figure 33 overleaf shows
the drivers of change in the UK’s Carbon Footprint. Changes in final demand spend is
generally a positive driver of change. The recession (2007-2009) was a period of reduced
spend and this had the effect of reducing GHG emissions. The material intensity of products
has a smaller positive effect between 1997 and 2007. During the recession there was a
change in the material intensity of products bought by the UK which contributed to reduced
emissions. During the recession, fewer goods were purchased from abroad and, as shown in
Figure 14, the UK’s material intensity is lower than our import partners’ and is reducing. The
falling carbon intensity of the materials used in the products purchased by the UK acted to
reduce the UK’s GHG emissions from consumption between 1997 and 2007. As Figure 18
shows, the carbon intensity of domestic production increased during this period because
material intensity in UK production was reducing at a faster rate than carbon intensity. The
trend of reducing carbon intensity of materials is due to the portion of imports increasing in
this time period, over which imports showed a declining carbon intensity of materials (Figure
18).
Figure 33: Change in the UK’s carbon footprint decomposed by 3 factors: carbon intensity of materials, material intensity, and total final demand (1997-2014)
Next we decompose each factor by the broad product groups introduced in Section 3.7. Rather
than look at all nine broad groups, we take the groups that are most important in terms of their
material intensity (agriculture, forestry & fishing), carbon intensity of materials (energy & water
and transport & communication), total spend (financial and business services) and carbon
footprint (manufacturing).
35 Resource Efficiency Metrics – initial findings
3.8.1. A focus on the group of agriculture, forestry and fishing products
Figure 34: Change in emissions associated with the group of agriculture, forestry & fishing products decomposed into three factors
Figure 34 shows that the emissions associated with the group of agriculture, forestry and
fishing products peaked in 2002, driven by an increase in final demand spend on these
products. Between 2002 and 2008, final demand acted as a negative driver and emissions
reduced. The effect of changes in the carbon intensity of materials and the material intensity
were minimal until after 2008. Post-recession we see a reduction in emissions driven by
reduced material intensity of agriculture, forestry and fishing products. Between 2008 and
2014, final demand acted as a positive driver once more but its effect was not large enough
to cause emissions increases.
3.8.2. A focus on the group of energy and water products
Figure 35: Change in emissions associated with the group of energy & water products decomposed into three factors
Figure 35 shows that the emissions associated with the group energy and water products
increased between 1997 and 2008 driven partly by an increase in final demand for these
goods (until 2004) but mainly due to the material intensity increasing. At the same time, the
carbon intensity of energy and water products reduced. Post 2008, emissions reduced, initially
36 Resource Efficiency Metrics – initial findings
driven by changes in the material intensity and most recently (2012-2014) by a change in the
carbon intensity of materials.
3.8.3. A focus on the group of manufactured products
Figure 36: Change in emissions associated with the group of manufactured products decomposed into three factors
Figure 36 shows that the emissions associated with the group of manufactured products
increased between 1997 and 2004 driven entirely by an increase in final demand for these
goods. Emissions then stabilised until the start of the recession in 2007. The large reduction
in the emissions associated with these products during the recession was driven by both a
reduction in final demand and the change in material intensity. Post-recession, the emissions
stabilised with the positive final demand driver cancelled out by the effect of the change in
carbon intensity of materials and material intensity.
3.8.4. A focus on the group of transport and communication products
Figure 37: Change in emissions associated with the group of transport & communication products decomposed into three factors
Figure 37 shows a steep increase in the emissions associated with the group of transport and
communication products between 1997 and 2007 driven by an increase in final demand for
these goods and the effect of a change in material intensity (until 2004). The large reduction
in the emissions associated with these products during the recession was driven by both a
37 Resource Efficiency Metrics – initial findings
reduction in final demand and the change in material intensity. Post-recession, the emissions
stabilised with the positive final demand driver cancelled out by the effect of the carbon
intensity of materials and material intensity.
3.8.5. A focus on the group of financial and business services
Figure 38: Change in emissions associated with the group of finance & business services decomposed into three factors
Figure 38 shows that the emissions associated with the group of finance and business
services increased between 1997 and 2001 driven by an increase in final demand for these
goods and the change in material intensity. Between 2001 and 2007 emissions stabilised with
the positive effect of the change in final demand cancelled out by the change in material
intensity. The reduction in the emissions associated with these products during the recession
was driven by both a reduction in final demand and further change in material intensity. Post-
recession, the emissions stabilised with the positive final demand driver cancelled out by the
effect of the carbon intensity of materials and material intensity.
Proposed suite of metrics
We recommend that the carbon intensity of materials be tracked for UK consumption. This
requires the production of an annual UK material footprint calculation.
The results presented in Section 3.8 demonstrate the potential for using decomposition
measures to understand the drivers of change in emissions that are related to the carbon
intensity of materials, the material intensity and the volume of spend. The level of analysis
shown is quite coarse and we recommend further work which considers an analysis as to the
UK’s changing trade partners. Much of the change in material intensity and carbon intensity
of materials may be explained by where we are sourcing goods and services from. We also
recommend analysis at the individual product level rather than considering product groups.
Section 4 develops this idea further.
4. Understanding the materials and emissions footprints of key products
Preceding sections of this report have proposed a suite of high level metrics based upon an
economy wide analysis using a time series formed from a number of static snapshots. Such
metrics indicate aggregate trends in materials and emissions intensity and highlight how each
sector and region is contributing to the overall trend. However, these metrics do not distinguish
the tangible products within each sector that are driving these trends, nor do they indicate the
whole life impacts of any product entering the stock in a given year.
38 Resource Efficiency Metrics – initial findings
All products incur material and carbon impacts in their production, use and end-of-life
treatment. Therefore a key objective of product resource efficiency strategies is to reduce
carbon emissions or material use on a whole life basis. In some instances this is best achieved
by expending additional carbon in the production phase to either reduce emissions in the
product’s use phase or to extend the expected product lifetime - thus minimising the need for
further production and delaying end-of-life treatment. In any given year an IO based metric
gathers the production impacts of new products entering the stock, the use phase impacts of
products in the current stock and the end-of-life impacts from products exiting the stock that
year. A metric based on such an approach does not indicate if new products entering the stock
are expected to yield reductions on a whole life basis.
For these reasons we would advise accompanying the proposed economy wide metrics with
a metric that tracks improvements in the whole life carbon and resource efficiency of key
sectors and products. The following sections propose a means of distinguishing ‘key sector
products’ and suggest one approach to developing such a metric. The scope of this report is
limited to a methodological proposal based upon preliminary analysis. The full development
of such a metric would be undertaken as part of a future work package.
Identification of key sector products
There are a number of factors to consider when identifying key products. Principal amongst
these are:
The emissions and materials intensity of the product;
The aggregate production impacts of the sector which produces the product;
The opportunity to mitigate product impacts through the adoption of resource efficiency
strategies.
Some products incur a high impact in the production of each unit, are produced in large
volume, and have impacts that may be effectively reduced using one or more resource
efficiency strategies. By contrast, other products are produced with very low impacts, in low
volume, and these impacts may be better addressed by other mitigation strategies. In reality,
the products of very few sectors fall in either extreme, and capturing the bulk of an economy’s
impacts in a single metric requires the selection of a range of representative sectors and
products.
Figure 39 shows the proportion of the total GHG emissions footprint that is captured by the
inclusion of a given number of regionalised products. The inclusion of a few hundred
regionalised products captures the overwhelming majority of the total emissions footprint.
Figure 40 presents the same results aggregated by sector product, i.e. ignoring the region of
origin. Figure 41 indicates the location of the top 100 regionalised products, which collectively
account for 94% of the total GHG footprint.
39 Resource Efficiency Metrics – initial findings
Figure 39: Percentage of 2014 GHG footprint captured by inclusion of regionalised products ranked from highest to lowest GHG footprint
Figure 40: Percentage of 2014 GHG footprint captured by inclusion of sector products ranked from highest to lowest GHG footprint
Figure 41: GHG footprint attributable to top 100 regionalised products ranked from highest to lowest GHG footprint in 2014
40 Resource Efficiency Metrics – initial findings
Figure 42 presents the proportion of each total material footprint captured by the inclusion of
a given number of sector products ranked by GHG footprint. This suggests that sector
products that account for the majority of the GHG footprint also account for the majority of
each material footprint. This relationship holds well for most materials, with the exception of
biomass forestry (BFOR). This is largely due to a small number of sectors with a low GHG
footprint but very high forestry product use e.g. furniture production.
Figure 42: Proportion of total footprint accounted for by material – based on sector products ranked by GHG footprint in 2014
These results suggest that products from a minority of sectors account for the majority of both
material and emissions impacts. This suggests that an agreed cut off value could be used to
determine key sectors. For instance, products of the 30 sectors with GHG footprints greater
than 7 MtCO2e account for 80% of the total GHG footprint and 62-85% of each material
footprint in 2014 (see Figure 43).
Figure 43: Proportion of total footprint accounted for by 30 sector products with GHG footprint exceeding 7 MtCO2e in 2014
41 Resource Efficiency Metrics – initial findings
The 30 sector products used for Figure 43 can be categorised into the aforementioned broad
product groups, as in Table 8 overleaf. Figure 44 shows the distribution of the GHG footprint
amongst these groups. Many of these groups, such as construction and manufactured
products, can be effectively targeted with resource efficiency strategies. However, some, such
as energy & water, will predominantly be addressed through alternative mitigation strategies
such as fuel switching. Therefore, when developing a metric that can capture ongoing
improvements in product resource efficiency, it may be desirable to limit the metric to certain
sectors.
Figure 44: Distribution of GHG footprint amongst product groups containing the 30 sector products with GHG footprint exceeding 7 MtCO2e in 2014
42 Resource Efficiency Metrics – initial findings
Table 8: Categorisation of 30 sector products with GHG footprint exceeding 7MtCO2e in 2014
Product Group Sector products with footprint exceeding 7 MtCO2e in 2014
Manufacturing Computer, electronic and optical products
Motor vehicles, trailers and semi-trailers
Machinery and equipment n.e.c.
Coke and refined petroleum products
Furniture
Textiles
Preserved meat and meat products
Dairy products
Other food products
Wearing apparel
Basic pharmaceutical products and pharmaceutical preparations
Fabricated metal products, excl. machinery and equipment and weapons & ammunition - 25.1-3/25.5-9
Other manufactured goods
Transport & communication
Air transport services
Land transport services and transport services via pipelines, excluding rail transport
Water transport services
Food and beverage serving services
Accommodation services
Public admin, education & health
Human health services
Public administration and defence services; compulsory social security services
Residential Care & Social Work Activities
Education services
Energy & water Electricity, transmission and distribution
Gas; distribution of gaseous fuels through mains; steam and air conditioning supply
Waste collection, treatment and disposal services; materials recovery services
Financial & business services
Owner-Occupiers' Housing Services
Real estate services, excluding on a fee or contract basis and imputed rent
Agriculture, forestry and fishing
Products of agriculture, hunting and related services
Construction Construction
Other services Services of households as employers of domestic personnel
43 Resource Efficiency Metrics – initial findings
Proposal for additional metric
One means of capturing change over time across sectors with many products is to monitor a
‘basket of representative products’ such as in the calculation of consumer price inflation (CPI)3.
In this approach a number of representative items are selected for each sector, price data is
gathered periodically and outputs of each sector are weighted to produce an aggregate metric.
Table 9, below, shows the current allocation of items for the calculation of consumer price
inflation. An index based upon a similar approach could track improvements in product
resource efficiency.
Table 9: Allocation of items to CPI divisions in 2016 (ONS, 2016c)
CPI weight (per cent)
Observed variation in price changes
Representative items (per cent of total)
1 Food & non-alcoholic beverages
10.3 Medium 24
2 Alcohol & tobacco 4.2 Medium 4
3 Clothing & footwear 7.1 Medium 11
4 Housing & household services 12.0 Medium 4
5 Furniture & household goods 5.9 Medium 10
6 Health 2.8 Low 3
7 Transport 15.3 Medium 6
8 Communication 3.2 High 2
9 Recreation & culture 14.8 High 17
10 Education 2.5 High 1
11 Restaurants & hotels 12.3 Low 7
12 Miscellaneous goods & services
9.6 Medium 11
The development of such an index would proceed in four stages as follows.
Stage 1: Identification of key sectors and index weightings.
The key sectors could be identified by a means similar to that described in the preceding
pages, with the exclusion of sectors that are best addressed through alternative mitigation
strategies e.g. electricity transmission and distribution. The index weightings for each sector
would be apportioned based on each sector’s relative contribution to the total GHG footprint
that year.
Stage 2: Selection of representative products.
A set of representative products would be determined for each key sector in collaboration with
stakeholders from each sector. These products would represent the majority of each sector’s
output, with a greater number of products required for sectors with a highly diverse output.
These representative products would be periodically updated with new models and new
products as consumer preferences change, through a process similar to the annual review of
the CPI basket of goods. This would require the formation of an expert group representing
each key sector, which would convene annually.
Stage 3: Life cycle analysis of representative products.
Life cycle analysis (LCA) data for each representative product would be gathered in a manner
that ensures transparency and future replicability, for example through the use of published
3https://www.ons.gov.uk/economy/inflationandpriceindices/articles/ukconsumerpriceinflationbasketofgoodsandservices/2017 describes the current basket of goods.
44 Resource Efficiency Metrics – initial findings
Environmental Product Declarations (EPD) which are prepared in accordance with defined
Product Category Rules. For products where no transparent data or common format for impact
reporting currently exists, initial life cycle analyses would be undertaken to provide a baseline.
These initial analyses would be periodically updated as part of the annual product basket
review and would use a frequently updated source of life cycle inventory data, such as the
ecoinvent or GaBi databases. These initial analyses would be replaced over time with product
data that conforms to sectoral standards as this data becomes available. At first these LCA
could be restricted to one impact category, such as GHG emissions, but could be extended in
future to cover material use or multiple weighted impacts.
Stage 4: Calculation of an Index of Product Resource Efficiency.
The relative changes in the life cycle impacts over time would be multiplied by the sector
weightings and captured in an Index of Product Resource Efficiency. This index would track
the relative change in whole life impacts of new products entering the stock each year.
Simple worked example
To illustrate how this Index could be calculated, consider the following simplified example.
Stage 1: Identification of key sectors and index weightings.
The key sectors included in this example are those previously identified as part of the
manufacturing, construction, agriculture, forestry and fishing product groups in Table 8. The
impact of all of these product groups could be significantly reduced through the implementation
of resource efficiency strategies. The 15 sectors included each had product footprints
exceeding 7 MtCO2e in 2014, and a combined footprint of some 256 MtCO2e. The weightings
attributed to each sector are based on their contribution to the total GHG footprint in 2014 and
are as shown in Table 10.
Table 10: Key sectors and index weightings
Product Group Sector GHG footprint in 2014
(ktCO2e)
Index weight
(%)
Agriculture, forestry and fishing
Products of agriculture, hunting and related services 51932 20.3
Construction Construction 48548 19.0
Manufacturing
Coke and refined petroleum products 22029 8.6
Computer, electronic and optical products 20633 8.1
Motor vehicles, trailers and semi-trailers 17486 6.8
Machinery and equipment n.e.c. 13585 5.3
Preserved meat and meat products 12745 5.0
Furniture 9866 3.9
Other manufactured goods 9552 3.7
Wearing apparel 9087 3.6
Basic pharmaceutical products and pharmaceutical preparations
8533 3.3
Other food products 8397 3.3
Fabricated metal products, excl. machinery and equipment and weapons & ammunition - 25.1-3/25.5-9
8123 3.2
Textiles 7908 3.1
Dairy products 7430 2.9
45 Resource Efficiency Metrics – initial findings
Stage 2: Selection of representative products.
For this simplified example let us consider the sector ‘Computer, electronic and optical
products’. Within this sector there are a wide variety of products that could be considered as
representative. For instance the CPI basket includes the following:
Table 11: Related products included in the CPI basket
08.2/3 Telephone and Telefax Equipment and Services Telephone Smartphone handset Fixed line telephone charges Mobile phone charges Cost of directory enquiries Mobile phone applications Subscription to the internet Mobile phone accessory Bundled communication services
09.1 Audio-Visual Equipment and Related Products 09.1.1 Reception and Reproduction of Sound and Pictures
Flat panel televisions DVD player Blu-ray disc player Digital television recorder/receiver Digital (DAB) radio Audio systems Personal MP4 player Headphones
09.1.2 Photographic, Cinematographic and Optical Equipment Digital compact camera Interchangeable lens digital camera Digital camcorder
09.1.3 Data Processing Equipment PCs – desktop and laptop PC peripherals Tablet computer Computer software
09.1.4 Recording Media CDs, including CDs purchased over the internet Pre-recorded DVDs, including DVDs purchased over the internet Pre-recorded Blu-ray discs, including discs purchased over the internet Recordable CD Music downloads Portable digital storage device
09.1.5 Repair of Audio-Visual Equipment and Related Products
Rather than assemble a full product list, for the sake of example, let us consider a few common
products, a smartphone handset, tablet computer and laptop PC. The market for each of these
example products is dominated by a small number of producers. For instance, Apple’s mobile
operating system represented nearly 50% of the UK market in May 2017. The models of each
product change frequently and the market share of different brands will change over time. For
each product it may be necessary to take more than one model or brand to represent the
market. Each generation of products may be more or less resource efficient in their production,
operation and end of life treatment. Therefore at each annual review the basket of products
would need to be updated to include products and models that represent a high market share
at that time. To consider a simple historical example let us take one model as representative
of each product in the years 2010-2016 as shown in Table 12 below.
Table 12: Example representative products and models
Smartphone Tablet Laptop
2010 iPhone4 iPad (1st gen) MacBook MC207
2011 iPhone4 iPad 2 MacBook MC207
2012 iPhone4S iPad (3rd gen) MacBook MD102
2013 iPhone5S iPad (4th gen) MacBook MD102
2014 iPhone6 iPad (4th gen) MacBook MD102
2015 iPhone6S iPad (4th gen) MacBook MF865
2016 iPhone7 iPad (5th gen) MacBook MLH82
These representative products must be assigned a weighting representing their share of
sector output. For the sake of example let us assign weightings as per Table 13. These
weightings would be reviewed annually by the expert group as purchasing patterns evolve.
46 Resource Efficiency Metrics – initial findings
Table 13: Product weightings
Product Group
Sector Representative product
Weight (%)
2010 2011 2012 2013 2014 2015 2016
Manufacturing Computer, electronic and optical products
Smartphone 40 45 50 50 50 50 50
Tablet 10 10 15 20 20 25 25
Laptop 50 45 35 30 30 25 25
Stage 3: Life cycle analysis of representative products.
To illustrate how the life cycle impacts of a product may vary over time as new models are
adopted, consider the three example products. Apple have been publicly reporting the
environmental impacts of each of their products to common boundaries since 20084. This
allows for easy comparison over time, with life cycle GHG emissions as shown in Table 14.
Indexing these impacts to a 2010 baseline gives Figure 45. Albeit this example has focussed
on Apple models, other producers of these products - such as Lenovo, Dell, and LG - also
undertake product footprinting.
Table 14: Product life cycle GHG emissions
Product Life cycle GHG emissions (kgCO2e)
Indexed by product 2010 = 100
iPhone4 45 100
iPhone4S 55 122
iPhone5S 65 144
iPhone6 95 211
iPhone6S 61 136
iPhone7 63 140
iPad (1st gen) 130 100
iPad 2 105 81
iPad (3rd gen) 180 138
iPad (4th gen) 170 131
iPad (5th gen) 135 104
MacBook MC207 350 100
MacBook MD102 580 166
MacBook MF865 470 134
MacBook MLH82 430 123
Figure 45: Life cycle GHG emissions of selected products indexed against 2010
4 See https://www.apple.com/uk/environment/reports/ for full range of product environmental reports.
47 Resource Efficiency Metrics – initial findings
The underlying calculations reveal an increasing share of life cycle emissions attributable to
the production phase (now 78-86% of total), with changes in body and display materials and
battery size driving increases. Albeit these products are becoming more energy efficient in
operation, the overall life cycle emissions have increased between models.
Stage 4: Calculation of an Index of Product Resource Efficiency.
Following selection of key sectors, representative products and models, and collection of the
associated LCA data, an index of product resource efficiency can be calculated based upon
the designated weightings.
For this simplified example, let us calculate the change in ‘Computer, electronic and optical
products’ and assume all other sectors have remained the same throughout 2010-2016. For
‘Computer, electronic and optical products’ the result would be as in Table 15.
Table 15: Calculation for computer, electronic and optical products
Computer, electronic and optical products
Representative products Total
Smartphone Tablet Laptop
Model Weight (%)
Index Model Weight (%)
Index Model Weight (%)
Index
2010 iPhone4 40 100.0 iPad
(1st gen) 10 100.0 MacBook
MC207 50 100.0 100.0
2011 iPhone4 45 100.0 iPad 2 10 80.8 MacBook MC207
45 100.0 98.1
2012 iPhone4S 50 122.2 iPad
(3rd gen) 15 138.5 MacBook
MD102 35 165.7 139.9
2013 iPhone5S 50 144.4 iPad
(4th gen) 20 130.8 MacBook
MD102 30 165.7 148.1
2014 iPhone6 50 211.1 iPad
(4th gen) 20 130.8 MacBook
MD102 30 165.7 181.4
2015 iPhone6S 50 135.6 iPad
(4th gen) 25 130.8 MacBook
MF865 25 134.3 134.0
2016 iPhone7 50 140.0 iPad
(5th gen) 25 103.8 MacBook
MLH82 25 122.9 126.7
Let us assume that the sector weightings (shown in Table 10) remain the same throughout
2010-2016. In reality, these could also be adjusted annually to reflect the relative importance
of each sector. Under this assumption, the overall index of product resource efficiency would
be as in Table 16.
Table 16: Calculation for index of product resource efficiency
Products of agriculture, hunting and related services
Computer, electronic and optical products
Preserved meat and meat products
...[other sectors] Index of product resource efficiency
Weight
(%) Index Weight (%) Index Weight
(%) Index
2010 20.3 100 8.1 100.0 5.0 100 100.0
2011 20.3 100 8.1 98.1 5.0 100 99.8
2012 20.3 100 8.1 139.9 5.0 100 103.2
2013 20.3 100 8.1 148.1 5.0 100 103.9
2014 20.3 100 8.1 181.4 5.0 100 106.6
2015 20.3 100 8.1 134.0 5.0 100 102.8
2016 20.3 100 8.1 126.7 5.0 100 102.2
48 Resource Efficiency Metrics – initial findings
Translation into practice
In practice, the process would be significantly more complex than indicated by this simple
example and would not be based upon historic data. Central to the process would be the
annual review wherein the expert group representing each sector would determine changes
to the representative products and weightings, and review updates to the product LCA data.
When initially convened, each group would also need to establish the current status of product
footprinting within their sector; identify relevant sector specific standards, product category
rules and so forth; collate compliant product declarations; identify gaps where no declarations
are available, and commission additional baseline assessments where necessary. For some
sectors this will be significantly easier than others. For instance, EPD are well established
within the construction sector, with common product category rules for many products and a
growing body of published declarations. By contrast few LCA have been undertaken for
pharmaceutical products. Ideally these expert groups would also convene as part of a broader
initiative encouraging companies in these sectors to share best practice in the footprinting of
their products and to identify opportunities to implement resource efficiency strategies.
The next steps in development of such an index include: establishment of formal criteria for
selection of key sectors; assessment of the current status quo of product footprinting within
each sector; and a trial attempt to establish a baseline within one sector. These steps would
require a more formal feasibility study that incorporated a broader programme of stakeholder
engagement.
5. Next steps
We have outlined the next steps in delivering a suite of policy relevant resource efficiency
indicators, followed by a brief discussion of future research needs in relation to understanding
resource productivity.
Producing resource efficiency metrics
This report provides a top-down assessment of the UK’s historical resource requirements.
Going forward, we recommend that the material footprint account should be replicated
annually with progress of key metrics, such as the carbon intensity of materials, monitored.
Our analysis also demonstrates the benefit of understanding the resource use and carbon
intensity of key products to the UK economy. This report provided one example of how a
supplementary ‘product level’ resource efficiency indicator could be developed. Such an
indicator is essential to guide the UK’s industrial strategy and highlight the possible resource
productivity gains to the UK economy. A partnership with relevant industries and consultancies
with specific sector experience is required to further develop such a metric, building on the
approaches identified in this initial report. We recommend that this metric is the subject of a
future work package.
Understanding resource productivity
Future research is required to gain a more complex understanding of the role of resource
productivity in the UK. This requires working closely with BEIS and Defra to align the goals of
the industrial strategy with resource productivity as well as understanding future projections
and policy interventions. This initial report has delivered on a number of the desired primary
project outcomes. However, delivering the secondary and tertiary outputs requires a deeper
understanding of the role of prices in shaping the consumption of materials. This will
necessitate other forms of modelling - requiring econometric expertise - that could be
undertaken as part of an additional project.
49 Resource Efficiency Metrics – initial findings
References
Bruckner, M., Giljum, S., Lutz, C., & Wiebe, K. S. (2012). Materials embodied in international trade - Global material extraction and consumption between 1995 and 2005. Global Environmental Change, 22(3), 568–576. doi:10.1016/j.gloenvcha.2012.03.011
Defra. (2016). UK’s Carbon Footprint. Retrieved from https://www.gov.uk/government/statistics/uks-carbon-footprint
Dietzenbacher, E., & Los, B. (1998). Structural Decomposition Techniques : Sense and Sensitivity. Economic Systems Research, 10(4), 307–323.
Eisenmenger, N., Wiedenhofer, D., Schaffartzik, A., Giljum, S., Bruckner, M., Schandl, H., … Koning, A. (2016). Consumption-based material flow indicators — Comparing six ways of calculating the Austrian raw material consumption providing six results. Ecological Economics, 128, 177–186. doi:10.1016/j.ecolecon.2016.03.010
Fischer-Kowalski, M., Krausmann, F., Giljum, S., Lutter, S., Mayer, A., Bringezu, S., … Weisz, H. (2011). Methodology and indicators of economy-wide material flow accounting: State of the art and reliability across sources. Journal of Industrial Ecology, 15(6), 855–876. doi:10.1111/j.1530-9290.2011.00366.x
Hirshnitz-Garbers, M., Srebotnjak, T., Gradman, A., Lutter, S., & Giljum, S. (2014). Further Development of Material and Raw Material Input Indicators – Methodological Discussion and Approaches for Consistent Data Sets Input paper for expert workshop. Retrieved from http://www.ecologic.eu/sites/files/publication/2014/input_indicator_project_inputpaper_workshopjune2014.pdf
Kanemoto, K., Lenzen, M., Peters, G. P., Moran, D., & Geschke, A. (2012). Frameworks for comparing emissions associated with production, consumption, and international trade. Environmental Science & Technology, 46(1), 172–179. doi:10.1021/es202239t
Lan, J., Malik, A., Lenzen, M., McBain, D., & Kanemoto, K. (2016). A structural decomposition analysis of global energy footprints. Applied Energy, 163, 436–451. doi:10.1016/j.apenergy.2015.10.178
ONS. (2014). Supply and Use Tables. Retrieved from http://www.ons.gov.uk/ons/taxonomy/index.html?nscl=Supply+and+Use+Tables
ONS. (2016a). Material Flows Account, United Kingdom. Retrieved December 14, 2016, from https://www.ons.gov.uk/economy/environmentalaccounts/datasets/ukenvironmentalaccountsmaterialflowsaccountunitedkingdom
ONS. (2016b). UK Environmental Accounts: How much material is the UK consuming? Retrieved December 14, 2016, from https://www.ons.gov.uk/economy/environmentalaccounts/articles/ukenvironmentalaccountshowmuchmaterialistheukconsuming/ukenvironmentalaccountshowmuchmaterialistheukconsuming
Peters, G. P., & Solli, C. (2010). Global Carbon Footprints: Methods and import/export corrected results from the Nordic countries in global carbon footprint studies.
Tukker, A., de Koning, A., Wood, R., Hawkins, T. R., Lutter, S., Acosta-Fernández, J., … Kuenen, J. (2013). Exiopol – Development and Illustrative Analyses of a Detailed Global MR EE SUT/IOT. Economic Systems Research, 25(1), 50–70. doi:10.1080/09535314.2012.761952
Wiedmann, T., Schandl, H., Lenzen, M., Moran, D., Suh, S., West, J., & Kanemoto, K.
50 Resource Efficiency Metrics – initial findings
(2015). The material footprint of nations. Proceedings of the National Academy of Sciences of the United States of America, 112(20), 6271–6276. doi:10.1073/pnas.1220362110
Wood, R., Stadler, K., Bulavskaya, T., Lutter, S., Giljum, S., Koning, A. De, … Tukker, A. (2015). Global Sustainability Accounting—Developing EXIOBASE for Multi-Regional Footprint Analysis. Sustainability, 7, 138–163. doi:10.3390/su7010138
WU. (2016). The online portal for material flow data. Retrieved March 3, 2016, from http://www.materialflows.net/home/