Project: Bram Edens, Rutger Hoekstra, Daan Zult, Harry Wilting

(PBL), Ronghao Wu (intern)

Carbon footprints – reconciling

academic and statistical work

Overview

– Overview work within statistical community

– Overview work within academic community

– State the problem

– Possible solution: SNAC footprint

– Conclusions and discussion

2

Overview of footprint calculations at NSI’s and other government agencies

NS

I/O

the

r

In

stit

ute

Ty

pe

Co

un

try

spe

cifi

c IO

Ye

ars

En

viro

nm

en

ta

l

Re

gio

ns

Ind

ust

rie

s

National

Statistical

Institutes

Australian Bureau of

Statistics

SRIO Y 2007/ 2008 GHG 1 40

Statistics Canada MRIO Y 2002&2006 GHG 4 ?

Statistics Denmark Partial Y 2005 CO2 13 60

INSEE Partial Y 2005 CO2 ±15 60

DESTATIS Partial/ hybrid Y 2007 Energy, CO2 14 73

Statistics Netherlands Partial N 2009 GHG (4) 17 60

Statistics Sweden SRIO N 1993-2008 Energy; materials; air

emissions

2 134

GSO Vietnam SRIO N 2005, 2007 CO2 1 5

Other

government

agencies

PBL Netherlands

Environmental Assessment

Agency

Partial and MRIO Y 2001 GHG (3) and land 13 57

DEFRA MRIO Y 1990-2009 CO2 and GHG 4 123

International

institutes

OECD MRIO Y 1995, 2000, 2005, 2008,

2009

CO2 (only emissions from

fuel combustion)

57 18

Eurostat SRIO N 2000-2007 8 pressures 2 64 3

Statistical community

– There is wide range of methods being used

‐ NSIs often use simpler models

– Focus broader than carbon

– Clear interest in additional breakdowns

‐ Household characteristics such as income

– Dissemination practices of the institutes show that the

results are not always presented as “official statistics”

4

Overview of MRIO databases that are currently publically available

GTAP EXIOBASE WIOD Eora

Acronym Global Trade Analysis

Project

EXIOPOL: Externality data and

input-output tools for policy analysis

World Input-Output Database -

Institute Purdue University EXIOBASE: FP6 project (EXIOPOL)

led by FEEM Database created by

NTNU, TNO, SERI, CML

FP7 project lead by the University of

Groningen

University of Sydney

Years 1990, 1992, 1995, 1997,

2001, 2004, 2007 (years are

not comparable)

2000 1995-2011 1990-2009

Prices of previous

year

- - 1995-2009 -

Countries/

Regions

15-129 (depends on year) 43

(27 EU, 16 non-EU)

(95% of the global GDP)

40

(27 EU and 12 non-EU)

(80% of world GDP in 2006)

187

Number of

industries

57 industries 130 industries 35 industries 100-500 industries

Environmental

data

Greenhouse gases (CO2,

NO2, CH4)

Energy use

Land use (split agro-

ecological zone)

Emissions (56)

Materials (96)

Land use (15)

Water use (14)

Energy use / several energy carriers

Water consumption

Land use

Emissions of greenhouse gases

Air pollutants

Resource use/extraction

Generation and treatment of various

types of waste

Greenhouse gases

Air pollution

Water use

Ecological Footprint

5

Academic work

– Difference between MRIOs:

‐ Aggregation (industries and/or countries)

‐ Construction method: IO based, SUT based, or trade

based

‐ Assumptions RoW or ITMs

‐ Emission data (modeled or not)

– Aim of MRIOs

‐ Information about global developments

‐ No claim to be 100% correct at national level

‐ Focus on consistent method (rather than best country

data)

6

Carbon footprints for the Netherlands from various MRIO databases

7

0

50

100

150

200

250

300

350

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

Em

issi

on

s (M

tCO

2) PNAS

NCC

ESSD

EORA

GRAM

WIOD

OECD

Data provided by Glen Peters and Nori Yamano

Carbon footprints for the Netherlands: WIOD and Eora

8

Stating the problem

– Growing policy interest in footprints, but no clear answers

– MRIOs have set the standard, but outside NSIs capabilities

• Labour intensive

• Assumptions

– MRIOs vs. official statistics

• Always inconsistent due to integration/balancing

required:

• Trade asymmetries

– Can we reconcile statistical and academic work in area of

footprint analysis?

9

A SNAC footprint

– Produce a footprint, based on MRIO, that is consistent

to official statistics of the Netherlands

‐ Single-country National Accounts consistent (SNAC)

– Main approach: “Adjust WIOD to be consistent to Dutch

data”

– Why WIOD?

‐ Transparancy

‐ Time series availibility

– Gain insight why results could be so different

10

Method

– Follow WIOD procedure, but overrule with improved Dutch int

SUT:

‐ Improved allocation of imports/exports to countries:

• Trade in goods: Bilateral trade data (re-exports and

domestic trade) from micro data

• Trade in services: Trade in services (confidential)

‐ National Accounts

• Used IO database to isolate re-exports

• Used IO database for valuation layers -> basic prices

‐ Expand from 35 to 72 industries (CO2 only)

– Balancing using the WIOD procedure but keeping the Dutch data

fixed

11

Results: SNAC vs. WIOD

12

– Overall difference in footprint: 4-6%

– Mainly due to lower foreign emissions

– NL becomes net exporter of emissions

2003 2009

Name SNAC-footprint WIOD SNAC-footprint WIOD

Absolute/Percentage MtCO2 MtCO2 % MtCO2 MtCO2 %

Total Footprint 206 220 -6% 202 210 -4%

Domestic indirect emissions 81 78 4% 80 71 11%

Domestic direct emissions 41 40 3% 40 39 5%

Total Domestic 122 118 3% 120 109 9%

Total Foreign 85 102 -17% 83 101 -22%

Resident emissions 209 205

Results 2: per capita

13

Source: CBS 2013

Results for top 10 countries/regions

14

Source: CBS 2013

– China at 19%: largest foreign emissions followed by

Germany

– WIOD: 21%

Why do SNAC and WIOD results differ?

WIOD

intermediate consumption

Final consumption expenditure

Gross fixed capital

formation

Changes in inventories

and valuables

Gross capital formation

Exports Total use at basic prices

use 362 344 78 -3 75 310 1,091 imports (incl ITM) 187 59 22 0 22 0 268 Total intermediate inputs 549 403 100 -3 97 Taxes less subsidies on products 33 21 9 0 9 62 Intermediate Inputs adjusted 582 425 109 -4 105 Value added at basic prices 510 0 Output at basic prices 1,091 0

CBS use 411 368 75 -1 75 244 1,096 imports 159 37 20 -2 20 0 214 Total intermediate inputs 570 405 95 -3 95 Taxes less subsidies on products 15 32 14 0 14 62 Total intermediate inputs adjusted 585 437 109 -3 109 Value added at basic prices 511 Output at basic prices 1,096

15

Why do SNAC and WIOD results differ? 2

CBS/WIOD intermediate consumption

Final consumption expenditure

Gross fixed capital formation

Changes in inventories and valuables

Gross capital formation Exports

Total use at basic prices

use 114% 107% 96% 33% 100% 79% 100%

imports 85% 63% 91% 91% 80%

Total intermediate inputs 104% 100% 95% 100% 98%

Taxes less subsidies on products 45% 152% 156% 156%

Total intermediate inputs adjusted 101% 103% 100% 75% 104%

Value added at basic prices 100%

Output at basic prices 100%

16

Differences due to: • Re-exports • Valuation layers More detailed information

Conclusions and discussion

1. Shift: uncertainty within MRIO to between MRIOs 2. MRIOs are produced for global questions, a SNAC-footprint is

more relevant for national policy makers due to consistency to country statistics

3. SNAC makes a difference! (at least for the Netherlands) 4. MRIO producers could quite easily make a footprint for

individual countries using “SNAC-philosophy”; key issue is sharing data

5. Need for enhanced cooperation especially in light of 2008 SNA ‐ Between statistical offices ‐ Between MRIO producers ‐ Statistical and academic community

17