See discussions, stats, and author profiles for this publication at: [Link]

net/publication/370075506

Introduction to Input-Output Analysis

Method · April 2023

DOI: 10.13140/RG.2.2.11634.99526

CITATIONS READS

2 6,528

2 authors:

Simon Mair Angela Druckman

University of York University of Surrey
43 PUBLICATIONS 793 CITATIONS 97 PUBLICATIONS 5,442 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Simon Mair on 18 April 2023.

The user has requested enhancement of the downloaded file.

 Introduction to Input-Output Analysis
Originally written by Simon Mair and Angela Druckman, October 2018

Updated by Simon Mair, March 2021

[Link]@[Link]

1. Introduction

Input-output analysis is an economic model built around a description of how

different economic sectors buy and sell from each other. Originally developed by

Wassily Leontief (1936, 1966) as a way to investigate economic problems, it was first

extended to include environmental problems in 1970 (Leontief 1970). Since then it has

been widely applied to analysis of environmental issues (Brizga, Feng, and Hubacek

2016; Druckman and Jackson 2009; Tukker et al. 2014) social issues (Simas et al.,

2014, Zamani et al., 2016, Mair et al., 2017) as well. Several studies have combined

Economic, environmental and social analysis to produce input-output based

sustainability assessments (Onat et al., 2014, Mair et al., 2016). Input-output analysis

is also used for forecasting and planning, taking into account economic

interdependencies (Füllemann et al. 2020; Ulrich and Lehr 2020; Whealan George

2019).

These notes outline the form and principals of economic input-output analysis

before describing how it can be extended to analysis of environmental or social

impacts. The purpose of these notes is to provide an introduction to the essential

 2

elements of input-output analysis. For more detail, and an introduction to more

advanced topics in input-output analysis, there are several good textbooks available

online (particularly, Hara, 2008, and Miller and Blair, 2009).

3. Input-Output Tables

Input-output analysis was developed as a tool to analyse the connections

between the various consuming and producing sectors of national economies. At the

core of input-output analysis is the use of a set of linear equations to map the

interdependence of economic sectors. This map of sector interdependencies is a useful

way of describing supply chains. The coefficients in input-output equations are

determined empirically from input-output tables. Here we introduce the organisation

of an input-output table, anyone wanting to take a closer look at its structure, or

wondering how it is compiled in real like should consult the handbook of the System

of National Accounts (European Comission et al., 2008).

An input-output table consists of data on the transactions between industries

for a given region and period of time (usually a year). Table 1 is a simplified,

hypothetical input-output table for a simple 3 sector economy over the period of a year.
 3

Table 1 An input out table for a hypothetical, 3 sector economy. All values are in dollars.

Intermediate Consumption
Agriculture Manufacturing Services Final Total Output
Demand

Agriculture 60 40 100 15 220

Manufacturing 80 30 150 30 300

Services 50 180 40 15 295

Value Added 30 50 5

Total Inputs 220 300 295

Columns are inputs to production in a sector, and rows are outputs from that sector.

So, Table 1 shows that the inputs to agricultural production were:

$60 of agricultural goods;

$80 of manufacturing goods;

$50 of services;

$30 of value added.

These inputs were used to produce $220 of agricultural goods (the output), of which:

$60 was sold to agriculture;

$40 was sold to manufacturing;

$100 was sold to services;

$15 was sold to final consumers.

 4

The last row and column of the table show the annual input and output of each sector

respectively. Note that for each sector inputs equals outputs.

In an input-output table, the inputs to a sector are split into two categories:

intermediate inputs and factors of production. Intermediate inputs are inputs purchased

by one sector from other economic sectors and used as part of the production process.

In our example table, the Manufacturing sector sells $60 of goods to Agriculture,

which combines these with other inputs to produce $220 of Agricultural output.

Likewise, Manufacturing buys $40 worth of goods from Agriculture and uses these to

produce $300 worth of goods. More specifically, imagine a clothing manufacturer. In

order to produce a woolen jumper, the clothing manufacturer has to purchase wool

(from the Agricultural sector), machinery to knit the wool into a jumper (from the

Manufacturing sector), and banking and insurance services (from the Service sector).

These are all intermediate inputs, therefore we call this intermediate consumption, and

call the producing sectors of the economy the intermediate sectors. We call the part of

the input-output table describing the transactions between the intermediate sectors the

transactions matrix.

Below the transactions matrix is a row labelled value added. This represents

the ‘factors of production’: inputs to the production process that are not supplied by

intermediate sectors. In the system of national accounts, factors of production are

represented by gross value added. Gross value added is defined as a residual: it is

calculated by subtracting the intermediate consumption of a sector from the total

output of the sector. It should be equal to the sum of compensation earned by labour,
 5

capital and government in the production process. That is, value added is all the money

paid to workers and government, plus all the money collected by owners of capital.

Value added also includes the value of imports1. Most input-output databases provide

detailed breakdowns of the components of value added, and multi-regional databases

have very detailed import data. In extended input-output analysis, we also think of

impacts (carbon emissions, jobs created, biodiversity loss, etc.) as factors of

production. We will discuss this in more detail later.

The column on the right of the transactions matrix is the final demand vector

(𝐲). ‘Final’ demand is so named because it comes from consumers who are largely

outside the production cycle of the economy: they purchase goods for their own

purposes, rather than as inputs to a production process. When a household purchases

food, this is to sustain themselves, or for the experience of eating or cooking, not to

produce something new that will be sold in the formal economy. For simplicity, in our

example final demand is shown as a single sector. However, in the system of national

accounts final demand is split into household consumption, non-profit organisations

serving households, government consumption, gross fixed capital formation2 and

exports.

1
Note that different databases will provide different levels of detail in how they breakdown
and represent the components of value added.
2
Gross fixed capital formation is a measure of expenditure on fixed assets, such as machinery
and buildings. It includes both the purchasing of new fixed assets and major repairs on existing assets.
 6

4. Input-Output Equations

We can describe the input-output table algebraically. In the input-output

literature, we usually to refer to economic output as 𝑥, the sale of intermediate goods

from one sector to another as 𝑧 and final demand as 𝑦. Here, we’ll call value added

𝑣. This notation is shown for our example economy in Table 2, where subscript

numbers refer to sectors. So 𝑧12 represents sales of intermediate goods from sector 1

(agriculture) to sector 2 (manufacturing), 𝑦3 represents final demand for goods from

sector 3 (services) and 𝑥1 represents economic output from sector 1 (agriculture).

Table 2 An input-output table for a hypothetical, 3 sector economy, expressed in algebraic

form.

Intermediate Consumption
Agriculture Manufacturing Services Final Total Output
Demand

Agriculture 𝑧11 𝑧12 𝑧13 𝑦1 𝑥1

Manufacturing 𝑧21 𝑧22 𝑧23 𝑦2 𝑥2

Services 𝑧31 𝑧32 𝑧33 𝑦3 𝑥3

Value Added 𝑣1 𝑣2 𝑣3

Each row represents the sales from one sector, so to find the economic output

of each of our sectors we sum across the rows of the table. In our example the economic

output of Agriculture (𝑥1 ) is,

𝑥1 = 60 + 40 + 100 + 20

or,
 7

𝑥1 = 𝑧11 + 𝑧12 + 𝑧13 + 𝑦1 (1)

For our full input-output table we have a system of three linear equations:

𝑥1 = 𝑧11 + 𝑧12 + 𝑧13 + 𝑦1

𝑥2 = 𝑧21 + 𝑧22 + 𝑧23 + 𝑦2 (2)

𝑥3 = 𝑧31 + 𝑧32 + 𝑧33 + 𝑦3

Each of these equations tells us that the total economic output of a sector is equal to

the sum of all the sales from that sector.

4.1 Introducing Matrices

The system of equations in equation (2) can be written in matrix form, where
we use a letter to represent each part of Table 2:

𝐱 = 𝐙𝐢 + 𝐲 (3)

where,

𝑥1 𝑧11 𝑧12 𝑧13 1 𝑦1

𝐱 = [𝑥2 ] , 𝐙 = [𝑧21 𝑧22 𝑧23 ] , 𝐢 = [1] and 𝐲 = [𝑦2 ] (4)
𝑥3 𝑧31 𝑧32 𝑧33 1 𝑦3

Bold lowercase letters represent a ‘column vector’, and bold uppercase letters

represent a ‘matrix’. Note that when we multiply matrices and vectors we do it slightly

differently than in linear algebra. We don’t have space here to go into the mechanics

of matrix algebra, but it is worth pointing out that post multiplication of 𝐙 by 𝐢 (a

 8

column vector of ones) has the effect of summing the row entries of 𝐙 (producing a

column vector). Horst (1963) is a useful primer on matrix algebra with a social-science

focus (available in the library), and there are lots of resources on the internet that

provide useful introductions (or reminders!) on the rules of matrix algebra. I

particularly like: [Link] which

provides a very simple guide to matrix algebra.

Representation as in Equation 3 is useful as matrix techniques simplify the

process of solving linear sets of equations (particularly useful when working with very

large datasets, such as [Link] which has 15,000 rows and 15,000

columns!).

Equation (3) offers very little analytical opportunity, rather it provides a

snapshot of the interdependencies of economic sectors in a given timeframe. Equation

(3) is a representation of the input-output table, it provides an estimate of all the

consumption and production in an economy at a given point in time. In order to

uncover the analytical power of input-output analysis we have to introduce a series of

assumptions known as the Leontief production function.

4.2 Assumptions and the Leontief Production Function

A production function relates the inputs to a sector to the maximum output that

could be produced by that sector. The Leontief production function makes three key

assumptions.
 9

1) Sectors use inputs in fixed proportions (so one input cannot be

swapped for another, and we don’t have any changes in efficiency). This is

often taken to mean that we assume ‘constant production technology’, as

swapping one good for another implies a change in how goods are produced.

For example, if a firm starts producing bicycle frames made from carbon fibre

where it used to use steel, we would describe this as a change in production

technology. This is in contrast to other types of production function. In some

cases, these go to the other extreme and assume perfect substitution between

all resources. Both positions are problematic. The advantage of a Leontief

production function is that it allows us to use simple mathematics, which mean

that we can capture the inputs used in a relatively high level of detail.

Production functions with more substitution between resources typically have

more complex mathematics which means we have to look at fewer inputs.

Whether this trade-off is acceptable will depend on your application of the

model.

2) There is a linear relationship between input and outputs. This means

that there are no ‘economies of scale’ – we don’t get more or less efficient as

we produce more of a good. To illustrate this, consider a car manufacturer. It

takes four tyres to produce one car, and forty tyres to produce ten cars. So there

is a linear relationship between the number of tyres used as inputs and the

number of cars produced. This seems entirely logical, but when dealing with

monetary units this is a cause for inaccuracy. Often there is a discount for

purchasing larger quantities of goods - for example, at a personal consumer

 10

level, supermarkets frequently offer “3 for the price of 2” deals. This is not

taken into account in input-output tables, and is a limitation of the approach.

3) Sectors produce a single homogenous good. So, Agriculture

produces just one good and this is sold to all customers (whether intermediate

or final). This is why many input-output tables have hundreds of sectors, and

why the level of detail is important. This assumption is taken to mean that the

input-output table represents the average good produced by a sector. An

example of the problems caused by this limitation can be illustrated by

considering the power generation sector. In some countries, the national

accounts include only one power generation sector, and this one sector thus

includes electricity produced by burning fossil fuels and electricity produced

by renewable energy sources such as wind power. For some analyses this may

not be problematic: however, if we are interested in the carbon emissions of

electricity production this is a drawback.

Often, these assumptions do not hold, but they provide an approximate estimate of

average relationships between sectors at a macro-economic level.

We operationalise the Leontief production function by dividing each input to a

sector by the output of that sector. For inputs from the intermediate sectors, this ratio

is known as the technical coefficient. In our example economy we estimate the

Agricultural sector’s technical coefficients by dividing each of the intermediate inputs

to the agricultural sector (𝑧11 , 𝑧12 , 𝑧13 ) by the total output from the Agricultural sector

(𝑥1 ) (for now we’re ignoring factors of production):

 11

𝑍11 60
𝑎11 = = 220 = 0.27
𝑋1

𝑍21 80
𝑎21 = = 220 = 0.36 (5)
𝑋1

𝑍31 50
𝑎31 = = 220 = 0.23
𝑋1

Repeating this process for remaining intermediate sectors gives us the values in Table

3.

Table 3 Technical coefficients for our hypothetical 3 sector economy

Intermediate Consumption
Agriculture Manufacturing Services

Agriculture 0.27 0.13 0.34

Manufacturing 0.36 0.10 0.51

Services 0.23 0.60 0.14

Each of these technical coefficients tells us the amount of the required input

needed to produce 1 unit of output from the respective sector. Reading down each

column in Table 3 we can see the direct requirements from the intermediate sectors

that are needed to produce one unit of output from that sector. To produce one unit of

Agricultural goods we need 0.27 units of Agricultural goods, 0.36 units of

manufacturing goods and 0.23 units of services.

Because we’re assuming that inputs rise linearly with outputs, we can use these

coefficients to estimate what inputs are needed to produce whatever level of output we

like. To do this we multiply the technical coefficient by the desired output. If we

 12

multiply the technical coefficient by the original output of the sector, we will arrive

back at the original transactions matrix (if interested, you can work this through and

check it for yourself). For our example economy,

𝑥1 = 𝑎11 𝑥1 + 𝑎12 𝑥2 + 𝑎13 𝑥3 + 𝑦1

𝑥2 = 𝑎21 𝑥1 + 𝑎22 𝑥2 + 𝑎23 𝑥3 + 𝑦2 (7)

𝑥3 = 𝑎31 𝑥1 + 𝑎32 𝑥2 + 𝑎33 𝑥3 + 𝑦3

where 1, 2, and 3 represent agriculture, manufacturing and services respectively, 𝑥 is

economic output, , 𝑎 is the technical coefficient, and 𝑦 is final demand.

The set of equations in (7) demonstrates that the total output of each sector is

dependent on the level of output from the other sectors. To produce its own outputs

each sector has to buy inputs from other sectors. We can use this to estimate how total

impact of production is driven by final demand. This allows us to examine how the

impacts of final consumption, 𝐲, are embedded in their supply chains. We might also

be interested in examining the effects of forecast or hypothetical changes in final

demand, 𝐲, on economic output, 𝐱.

To be able to do this we have to define a relationship between economic output

and final demand. For our example economy this is done by rearranging (7). First we

take the technical coefficients and economic output terms from the right-hand side

away from the economic output term on the left hand side:

𝑦1 = 𝑥1 − 𝑎11 𝑥1 − 𝑎12 𝑥2 − 𝑎13 𝑥3

 13

𝑦2 = 𝑥2 − 𝑎21 𝑥1 − 𝑎22 𝑥2 − 𝑎23 𝑥3 (8)

𝑦3 = 𝑥3 − 𝑎31 𝑥1 − 𝑎32 𝑥2 − 𝑎33 𝑥3

We can then simplify this by grouping like terms:

𝑦1 = (1 − 𝑎11 )𝑥1 − 𝑎12 𝑥2 − 𝑎13 𝑥3

𝑦2 = −𝑎21 𝑥1 − (1 − 𝑎22 )𝑥2 − 𝑎23 𝑥3 (9)

𝑦3 = −𝑎31 𝑥1 − 𝑎32 𝑥2 − (1 − 𝑎33 )𝑥3.

4.3 Using Matrices

We solve this system of equations for economic output using matrix algebra. To do

this we define two new matrices, the technical coefficients matrix, 𝐀, and the identity

matrix, 𝐈:

𝑎11 𝑎12 𝑎13 1 0 0

𝑎
𝐀 = [ 21 𝑎22 𝑎23 ] and 𝐈 = [0 1 0] (10)
𝑎31 𝑎32 𝑎33 0 0 1

𝐀 contains all the technical coeffcients. The columns of 𝐀 represent the intermediate

inputs required for a sector to produce one unit of output. 𝐈 is used to represent the 1s

in the system of linear equations (9). We can now write final demand (𝐲) in terms of

technical coefficients and economic output (𝐱), as follows:

𝐲 = (𝐈 − 𝐀)𝐱 (10)
 14

Lastly, because matrix division is represented by multiplication by the inverse, we can

solve (10), to find gross output (𝐱) as a function of final demand (𝐲),

𝐱 = (𝐈 − 𝐀)−1 𝐲 (11)

We can use equation (11) to estimate how much economic output is associated

with the production of a given set of final demand, assuming the input requirements

of each sector are fixed. In other words, (11) tells us how much gross output is

generated in each economic sector for a give pattern of final demand assuming fixed

supply chain structures (as represented by (𝐈 − 𝐀)−1 the ‘Leontief Matrix’).

Conceptually, (11) can be thought of as:

Output (£) =
supply chain production requirements per unit final demand (unitless) * final demand (£)

Therefore, if we change the final demand level we can estimate how the scale and

distribution of economic output will change. Likewise, if we only include a subset of

final demand in 𝐲 (for example, final consumption in the agricultural sector) we can

estimate how much economic output is generated in each sector of the economy to

produce that subset of final demand.

5. Extended Input-Output Analysis

To extend input-output analysis to environmental or social impact analysis, we

start by adding a satellite account to our input-output table. This satellite account is an

additional row (or rows) underneath the input-output table. They are usually measured

in physical units (e.g. hours worked, or tonnes of carbon emitted) and not compiled by
 15

national governments. Table 4 shows how adding a carbon satellite account might look

in our example economy.

Table 4 An environmentally extended input-output table for a hypothetical, 3 sector

economy. All values are in $, apart from Carbon which is in tonnes.

Into Intermediate Consumption

Agriculture Manufacturing Services Final Total Output
Demand

Agriculture 60 40 100 15 220

Manufacturing 80 30 150 30 300

Services 50 180 40 15 295

Value Added 30 50 5

Total Inputs 220 300 295

Carbon 10 20 5

We can treat satellite accounts as factors of production: inputs to production

processes from outside the intermediate sectors of the economy. This might seem

counter-intuitive for carbon. Consider an alternative satellite account, such as hours

worked, or land used: these are inputs that do not come from the producing sectors of

the economy. We think of carbon and other impacts in the same way. For the year

represented by our table, the agricultural sector could not have produced $220 of

output without also producing 10 tonnes of carbon. The only way for agriculture to

reduce this number would be to produce less, or to change its production methods,

therefore it is useful to think of carbon as an input.

 16

To incorporate our satellite account in to our input-output model, we estimate

a vector, 𝐮′, whose elements have units of impact per unit economic output,

𝐮′ = [𝑒1 ⁄𝑥1 𝑒2 ⁄𝑥1 𝑒2 ⁄𝑥1 ]. (12)

where, 𝑒1 is the total direct impact produced by sector 1. For ease, we write the

elements of 𝐮′ as 𝑢1 , 𝑢2 and 𝑢3 . In the case of our example economy 𝑢1 = 𝑒1 ⁄𝑥1 , and

this tells us the amount of carbon emitted per unit of economic output from the

agricultural sector.

Denoting total greenhouse gas production from our economy as 𝑐 we can see

that total carbon emitted by the economy is equal to the carbon coefficient (factor

intensity) multiplied by the output of that sector,

𝑐 = 𝑢1 𝑥1 + 𝑢2 𝑥2 + 𝑢3 𝑥3 = 𝐮′𝐱 (13)

So, if we change the level of output in one or all of the sectors we change the carbon

emitted. Likewise, if we can estimate the share of a sectors output that was generated

in the supply chain of particular good then we can estimate how much carbon was

associated with production of that good.

In equation (11), we estimated the amount of output in each sector associated

with production of a particular set of final demand. Therefore we can now substitute

(11) into (13) to get,

c = 𝐮(𝐈 − 𝐀)−1 𝐲 (14)

 17

Equation (14) tells us the amount of carbon emitted in the supply chains required to

produce a given set of final demand. This is the standard model of environmentally

and socially extended input-output analysis. Conceptually, we are updating the

Leontief Matrix (our supply chain map) with ‘impact’ data. Thus instead of every unit

of final demand stimulating a given amount of economic activity, a unit of final

demand stimulates a given quantity of impact.

6. Multiplier Analysis

Multipliers3 are a key analytical statistic derived from input-output models.

Multipliers are a ratio describing the total impact of production per unit of final

demand. Final demand is almost always measured in monetary terms, i.e. dollars or

pounds sterling. In an economic analysis, impact might be all the employment required

to produce a dollars worth of goods. So the multiplier would have units of hours

worked per dollar spent. In an environmental analysis, we might look at carbon

emissions produced, or biodiversity loss. In these cases units might be tonnes of carbon

per dollar, or number of species lost per dollar.

Multipliers can be derived from equation (14). To see this, recall how we

interpret the parts of equation (14). 𝐮 is a vector of environmental or social impacts

per unit of economic output. (𝐈 − 𝐀)−1 is a matrix representing all the supply chains

in an economy. 𝐲 is a vector of final demand.

3
As with all the topics here, a more detailed treatment of multipliers is found in Miller and
Blair (2009).
 18

In (14) we multiply our map of supply chains by the vector of impact per unit

of economic output to get a three terms together we get the total impact in an economy.

What we want is to find the environmental impact per unit of final demand for different

sectors in the economy.

To get this we divide the right hand side of (14) by final demand which gives:

𝐦 = 𝐮(𝐈 − 𝐀)−1 (15)

Equation (15), multiplies the map of industry supply chains by the impact per unit of

economic output, to give us a vector of the impacts stimulated by production of a single

unit of final demand. The elements of 𝐦 are our multipliers.

7. Uncertainty and error in input-output.

Input-output models are inherently uncertain, but this is often hidden in input-

output analyses where a systematic examination of uncertainty is still rare (see Bullard

and Sebald, 1988, Weber, 2008, Wiedmann et al., 2008 for rare exceptions to the rule).

Uncertainty in input-output analysis comes from the assumptions we make in the

Leontief production function (fixed input ratios, linear relationships between inputs

and outputs, and homogeneous outputs). While these assumptions may hold for an

average production process, they do not hold for individual processes.

Another source of uncertainty is in the data itself. This is normally compiled

from surveys (and a few other techniques). Therefore it is a mistake to think of input-
 19

output tables as observations of what is happening in an economy. Rather, they are our

best guess of what is happening in an economy.

Moreover, input-output is perhaps most powerful when used at a global scale

(as it captures the whole supply chain). But, global input-output introduces new

uncertainties (Weber, 2008). For example, when we convert from local currencies to

a single currency should we use market exchange rates or purchasing power parity?

Our choice will affect our results, and neither is perfect (Callen, 2007, Mair et al.,

2016). Likewise, we often lack the very detailed trade data that we would like to have

when building global input-output databases. Different databases take different

approaches to dealing with the issues that arise (Dietzenbacher et al., 2013). There is

no right or wrong way to do it. But, just as with any model, choices have to be made,

and these will shape our results.

8. Price concepts and actually using input-output analysis.

The application of input-output analysis (including multiplier analysis), to real

world questions always introduces additional complications. Here we will address one

of these: the issue of different price concepts.

National accountants (the people most often tasked with compiling input-

output tables), use many price concepts for the different kind of statistics they put

together. For our purposes there are two that are most relevant: basic prices and

purchaser’s prices.
 20

1. Basic prices are the price received by the producer, minus any taxes and

plus any subsidies.

2. Purchaser’s prices are the price payed by the consumer. In addition to basic

prices they include any retail and wholesale margins, non-deductible taxes

and transport costs.

The difference between these two is important because most input-output tables are

available in basic prices, but we often want to use them to explore the impact of a

change in spending, and we most often have access to this in in purchaser’s prices.

To see the problems this might cause, consider a simple example. Suppose we

want to estimate the carbon impact of the average UK household food consumption.

We go out and collect information on how much money a household spends on food

in a year. Say it comes to £1000. We can use a carbon multiplier for the food sector to

estimate the impact of this. Say the multiplier is 1.2 tonnes of carbon per £ spent on

food (these numbers are all made up!). 1.2 multiplied by $1000 gives us 1200 tonnes

of carbon per year. But, that multiplier was estimated at basic prices, while the food

bill is valued at purchaser’s prices. To convert the food bill in basic prices we have to

take away the wholesale and retail margins, and any taxes/subsidies and transport

costs. For ease of the maths, let’s say that the total of the margins, taxes and transport

is $200 (it could be much more!). What we now have is a food bill of £800 multiplied

by a carbon multiplier of 1.2, which gives us a carbon impact of only 960 tonnes per

year. If we hadn’t converted our estimate from purchaser’s prices to basic prices we

would have overestimated the carbon footprint by 20%.

 21

Unfortunately, in practice it can be hard to come by all the information we need

to move between purchasers and basic prices. However, some databases do provide

data in both price concepts. For example the OECD provide data for a number of

countries in both purchaser’s prices and basic prices:

[Link] There are still

choices to be made in how we move between the concepts. A relatively straightforward

approach to convert a price in purchasers’ prices to basic prices is to simply calculate

the ratio (b/p) of basic prices (b) to purchasers’ prices (p) from the appropriate dataset,

and multiply the value in question by this ratio.

9. Sources of Data

The International Input-Output Association maintains a comprehensive list of

national and international input-output tables. This can be found at:

[Link] This includes links to national statistical

offices, as well as some of the big global databases with environmental and social

extensions such as Eora (also available at [Link]) and the World Input-

Output Database ([Link]).

Finding multiplier data can be tricky as most databases assume that users are

happy to calculate their own datasets. However, if you do not wish to do the matrix

manipulation yourself, there are a few multiplier datasets available that are ‘ready for

use’. Their suitability for your project must be considered carefully before use. Also,

if you use multipliers in an assignment, you must show that you understand their

derivation.
 22

Some countries provide multiplier datasets specifically for their country. For

example, the UK provides them here: [Link]

carbon-footprint - see “Table 13 - Indirect emissions from the supply chain”. An

advantage of this dataset is that it is provided in purchasers prices, and so no

manipulation of prices is required. Please read the information provided at the top of

the spreadsheet carefully before making use of the data.

A comprehensive set of multipliers for a wide selection of countries is provided

at [Link] This website enables you to

choose the data-source, and you will notice that the results vary depending on the data-

source used. More information each source can be found at

[Link] The multipliers on this website are

provided in impact per million USD using basic prices4. Therefore before use you may

need to convert your project specific expenditures to purchaser prices.

References

Brizga, J., Feng, K. and Hubacek, K. (2016) 'Household carbon footprints in the Baltic
States: A global multi-regional input–output analysis from 1995 to 2011',
Applied Energy.

Bullard, C. and Sebald, A. (1988) 'Monte Carlo Sensitivity Analysis of Input-Output

Models', The Review of Economics and Statistics, 70(4), pp. 708-712.

Callen, T. (2007) 'PPP Versus the Market: Which Weight Matters?', Finance &
Development, 44(1), pp. 1-4.

4
To be more precise, the multipliers on this website are provided in impact per million USD
using market exchange rates in current (that is, not accounting for inflation) prices, using basic prices.
 23

Dietzenbacher, E., Los, B., Stehrer, R., Timmer, M. and G, d. V. (2013) 'The
Construction of World Input–Output Tables in the WIOD Project', Economic
Systems Research, 25(1).

Druckman, A. and Jackson, T. (2009) 'The carbon footprint of UK households 1990-

2004: A socio- economically disaggregated, quasi- multi- regional input-
output model', Ecological Economics, 68(7), pp. 2066-2077.

Druckman, A. and Jackson, T. 2010. An Exploration into the Carbon Footprint of UK

Households, RESOLVE Working Paper Series 02-10, University of Surrey,
November 2010.

European Comission, IMF, OECD, UN and World Bank (2008) System of National
Accounts. Available at:
[Link] (Accessed:
17/11/2015.

Feenstra, R. and Hanson, G. (1996) Globalization, Outsourcing, and Wage Inequality.

NBER Working Paper 5424. Available at: [Link]
(Accessed: 07/08/2014.

Füllemann, Yann, Vincent Moreau, Marc Vielle, and François Vuille. 2020. ‘Hire
Fast, Fire Slow: The Employment Benefits of Energy Transitions’. Economic
Systems Research 32(2): 202–20.

Hara, T. (2008) Quantitative tourism industry analysis: introduction to input-output,

social accounting matrix modeling and tourism satellite accounts. Routledge.

Horst, P. (1963) Matrix algebra for social scientists. New York: New York : Holt,
Rinehart and Winston.

Jackson, T. (2016) Working Paper No. 2: Beyond Consumer Capitalism. Centre for
the Understanding of Sustainable Prosperity Working Paper Series. Guildford,
UK: University of Surrey. Available at: [Link]/publications
(Accessed: 10/10/2016.

Jorgensen, A., Finkbeiner, M., Jorrgensen, M. and Hauschild, M. (2010) 'Defining the
baseline in social life cycle assessment', The International Journal of Life
Cycle Assessment, 15(4), pp. 376-384.

Lenzen, M. (2000) 'Errors in Conventional and Input-Output-based Life-Cycle

Inventories', Journal of Industrial Ecology, 4(4), pp. 127-148.

Lenzen, M. and Dey, C. (2000) 'Truncation error in embodied energy analyses of basic
iron and steel products', Energy, 25(6), pp. 577-585.
 24

Leontief, W. (1970) 'Environmental Repercussions and the Economic Structure: An

Input- Output Approach', The Review of Economics and Statistics, 52(3), pp.
262-271.

Los, B., Timmer, M. P. and de Vries, G. J. (2014) 'HOW GLOBAL ARE GLOBAL
VALUE CHAINS? A NEW APPROACH TO MEASURE
INTERNATIONAL FRAGMENTATION', Journal of Regional Science, pp.
n/a-n/a.

Mair, S., Druckman, A. and Jackson, T. (2016) 'Global Inequities and Emissions in
Western European Textile and Clothing Consumption', Journal of Cleaner
Production, 132, pp. 57-69.

Mair, S., Druckman, A. and Jackson, T. (2017) 'Investigating fairness in global supply
chains: applying an extension of the living wage to the Western European
clothing supply chain', The International Journal of Life Cycle Assessment.

Miller, R. and Blair, D. (2009) Input output analysis, foundations and extensions. New
Jersey: Prentice-Hall.

Onat, N., Kucukvar, M. and Tatari, O. (2014) 'Integrating triple bottom line input–
output analysis into life cycle sustainability assessment framework: the case
for US buildings', The International Journal of Life Cycle Assessment, 19, pp.
1488–1505.

Simas, M., Golsteijn, L., Huijbregts, M., Wood, R. and Hertwich, E. (2014) 'The “Bad
Labor” Footprint: Quantifying the Social Impacts of Globalization',
Sustainability, 6(11), pp. 7514-7540.

Small World Consulting (2016) BT’s supply chain carbon emissions – a report on the
approach and methodology. Available at:
[Link]
hodology/SWC_BT_supplychainemissions2016.pdf (Accessed: 09/11 2016).

Steinberger, J., Friot, D., Jolliet, O. and Erkman, S. (2009) 'A spatially explicit life
cycle inventory of the global textile chain', International Journal of Life Cycle
Assesment, 14, pp. 443-455.

Tukker, A., Bulavskaya, T., Giljum, S., de Koning, A., Lutter, S., Simas, M., Stadler,
K. and Wood, R. (2014) The Global Resource Footprint of Nations. Carbon,
water, land and materials embodied in trade and final consumption calculated
with EXIOBASE 2.1. Lieden: The Netherlands Organisation for Applied
Scientific Research.

Weber, C. 'Uncertainties in Constructing Environmental Multiregional Input-Output

Models', International Input Output Meeting on Manageing the Environment,
Seville, Spain.
 25

Wiedmann, T., Lenzen, M. and Wood, R. 2008. Uncertainty Analysis of the UK-MRIO
Model – Results from a Monte-Carlo Analysis of the UK Multi- Region Input-
Output Model (Embedded Emissions Indicator); Report to the UK Department
for Environment, Food and Rural Affairs by Stockholm Environment Institute
at the University of York and Centre for Integrated Sustainability Analysis at
the University of Sydney. Defra, London, UK.

Zamani, B., Sandin, G., Svanström, M. and Peters, G. (2016) 'Hotspot identification
in the clothing industry using social life cycle assessment—opportunities and
challenges of input-output modelling', The International Journal of Life Cycle
Assessment, pp. 1-11.

View publication stats