Articles

Modeling household solid fuel use towards

reporting of the Millennium Development
Goal indicator
Sumi Mehta
Health Effects Institute, 120 Second Avenue, Boston, MA 02129, USA
E-mail: [email protected]
Fiona Gore, Annette Prss-stn, and Eva Rehfuess
Protection of the Human Environment, World Health Organization (WHO), 20, Avenue Appia
CH-1211 Geneva 27, Switzerland
Kirk Smith
Environmental Health Sciences, University of California, Berkeley CA 94720, USA

Household use of solid fuels, such as dung, wood, agricultural residues, charcoal, and coal, is likely
to be the largest indoor source of air pollution in developing countries. Combustion of solid fuels
in inefficient stoves under poor ventilation conditions can result in large exposure burdens, particularly for women and young children, who spend the major part of their time at home. The
importance of this public health and environmental issue is reflected in the inclusion of the percentage of households using solid fuels as an indicator towards the achievement of the Millennium
Development Goal (MDG) for environmental sustainability. This article outlines the model used for
completing missing country data for household use of solid fuels. Out of a total of 181 country
data points reported, data were available for 94 countries. These included estimates from 42 countries where household solid fuel use was previously unknown, largely from the recently conducted
World Health Survey. Based on the data available for these countries, using step-wise regression,
a model to predict household solid fuel use based on rural population, gross national income (GNI)
and geographic regional variables was developed. Thirty-five data points were estimated using this
model. In general, household solid fuel use seems to be lower in 2003 than in 2000. Yet, even with
increases in economic development and urbanization, drastic reductions in household solid fuel use
are unlikely to occur in the absence of targeted programs to promote cleaner fuels.

for exposure to indoor air pollution for burden of disease

estimation, as well as to indicate progress towards the attainment of the Millennium Development Goal (MDG)
for environmental sustainability.
In cases where data coverage is low or non-existent, it
is necessary to predict solid fuel use. Previous estimates
modeled missing country data for 2000 when survey data
were available for only 52 countries [Smith et al., 2004].
This paper presents modeled estimates of household solid
fuel use for countries where solid fuel use was unknown
in 2003, now based on known data of 94 countries, including data from the recently conducted World Health
Survey.

1. Background
Cooking and heating with solid fuels, such as dung, wood,
agricultural residues, charcoal, and coal, is likely to be
the largest indoor source of air pollution in developing
countries. Combustion of solid fuels in inefficient stoves
under poor ventilation conditions can result in large exposure burdens, particularly for women and young children, who spend the major part of their time at home.
Because there is relatively little information on the indoor concentrations or exposures of these pollutants from
solid fuel use, or on the baseline concentrations within
similarly constructed households not using solid fuels,
epidemiological evidence on the health effects of solid
fuel use is based mainly on binary classifications of exposure, for example whether solid fuels were used for
cooking or heating. Similarly, in the absence of detailed
information on exposure to indoor air pollution from solid
fuel use at the national level, binary classifications of
household solid fuel use are used as a practical surrogate
36

Energy for Sustainable Development

2. Data and methods

2.1. Available data on household solid fuel use at the
national level
In many countries where large proportions of the population cook with solid fuels, household energy data are
!

Volume X No. 3

September 2006

Articles

2.2. Statistical modeling to predict household solid fuel

use at the national level
Using available national estimates of household solid fuel
use, a stepwise linear regression model was used to predict national solid fuel use. A wide range of development
indicators were tried as parameters in the modeling exercise, including average annual growth rates, per capita
electricity and petroleum consumption, income (represented by gross national income[1] (GNI) per capita), Gini
coefficient, adult female illiteracy, per capita fuelwood
production, urbanization, and national traditional fuel consumption, and WHO Region (Table 2). All analysis was
done using SPSS Version 8.0 (SPSS Inc., USA) and
STATA 7.0 (Stata Corporation, USA).
For those countries where no census or survey data
were available, estimates of household solid fuel use at
the national level are based on the following assumptions.
1. As countries develop, people gradually shift up an energy ladder from solid fuels to cleaner fuels and
cleaner cooking technologies. Although the picture
may be more complex at local and household levels,
it is assumed that this generally holds true over the
long term on a regional scale.
2. In the 2000 round, Brazil, with a 1999 GNP per capita
of US$ 4420, was the richest country in the original
database to have a substantial proportion of households using solid fuels. Thus, to avoid extrapolating
the model to areas where it may be inappropriate, estimates were only made for countries with 1999 GNP
per capita less than $5,000. In other words, all countries with a GNP per capita greater than $5,000 are
assumed to have made a complete transition to clean
household cooking systems, either with cleaner liquid
or gaseous fuels or electricity or, where solid fuel is
still used for cooking/heating, to fully ventilated appliances. As recent surveys have indicated that much
higher-income countries may still use significant proportions of solid fuels, estimates here are made for all
countries with a 1999 GNP per capita less than US$
10,500.
3. After a certain level of economic growth, it is assumed
that countries will entirely shift away from using solid
fuels in conditions with poor ventilation. Solid fuel
use for heating may continue, especially in coal-rich
and wood-rich regions. At higher levels of economic
development, however, as this is essentially done using flues and chimneys, associated indoor air pollution
exposures are limited.
The final model for the year 2000 (Table 3) included percentage rural population, location within the Eastern
Mediterranean Region (EMR) and GNI per capita (logtransformed), and per capita petroleum use. As this was
a linear regression model, beta are the regression coefficients for the variables included in the model. Location
within EMR is likely important due to the fact that this
encompasses a region with substantial oil and gas resources, and limited availability of solid fuel resources.
To test for violations of regression assumptions, residuals
were analyzed and plotted. Plots did not indicate any

widely, although not universally, available. In some cases,

the data come directly from national census information
or energy statistics, which state explicitly the number or
fraction of households that rely predominantly on solid
fuels for their energy. For example, information on the
main fuel used for cooking is collected during the house
listing of the decadal Census of India [GOI, 2001]. These
data, disaggregated into urban and rural sectors, are available at the district level. (In India, a district, the administrative level below that of state, contains on average
about two million people.)
In some countries, where censuses are infrequent and/or
do not contain residential energy data, household surveys
are an important information source. Some, such as
Demographic and Health Surveys, are routinely conducted
and report the main household cooking fuel. Less frequently conducted or one-time surveys are also a valuable
source of information. For example, primary household
energy estimates for 22 African countries, based on household welfare surveys with sample sizes ranging from
1,000 to over 14,000, are included in a database of African
development indicators compiled by the World Bank
[World Bank, 2000].
In other countries, including Indonesia and Thailand,
data are available in the form of aggregate annual residential fuel consumption. In these cases, the percentage
of households using solid fuels was estimated on the basis
of the quantity of fuel consumed. We recognize that exposures from cooking and heating can be quite different
because of different cooking technologies. In most cases,
however, it was not possible to distinguish between the
two end-uses.
The original global database of household fuel use used
in the burden of disease project [Smith et al., 2004] was
updated through a review of national census data and
other household surveys. The updated database compiled
by the World Health Organization, which includes information from 1990 to 2004, provides data for 94 countries.
A bulk of this additional information comes from preliminary data of the World Health Survey conducted in 2003.
All household fuel use is expressed as percent of households using each fuel type. (See Table 1.)
Admittedly, the compilation of data from various
sources, with different methodologies and sample sizes,
makes it difficult to characterize the uncertainty associated with the estimates for each country and region.
Where more than one survey was available for a given
country, the latest survey was selected. This was the case
even for preliminary data (such as available from the
World Health Survey), provided that the second latest survey was three years old or older. The differences between
preliminary data and final estimates are likely to be minor
compared to possible changes in solid fuel use over a
three-year period. Where other survey results were available within two years of the latest preliminary survey results, however, the second latest survey was selected. No
obvious inconsistencies were found where surveys were
available for various years for the same country. Only nationally representative surveys were considered.
Energy for Sustainable Development

Volume X No. 3

September 2006

37

Articles

Table 1. Available data on national estimates of household solid fuel use, 2003
Region[1]

Country

AfrD

Algeria

AfrD

Data source

Year

Reference

4%

National energy
statistics

1999

WRI, 2003

Angola

100 %

National energy
statistics

1999

IEA, 1999

AfrD

Benin

95 %

Household survey

2001

WBAHSD Demographic and Health Surveys (DHS):

Benin 2001

AfrD

Burkina Faso

96 %

Household survey

2003

WHS: Burkina Faso, 2003 (Preliminary data from WHS)

AfrD

Cameroon

83 %

Household survey

2001

WBAHSD Enqute Camerounaise Auprs des Mnage

(ECAM)

AfrD

Chad

97 %

Household survey

2003

WHS: Chad, 2003 (Preliminary data from WHS)

AfrD

Comoros

76 %

Household survey

2003

WHS: Comoros, 2003 (Preliminary data from WHS)

AfrD

Gabon

28 %

Household survey

2000

Macro, 2003

AfrD

Gambia

96 %

Household survey

1998

WBAHSD National Household Poverty Survey, Gambia

1998

AfrD

Ghana

88 %

Household survey

2003

WHS: Ghana, 2003 (Preliminary data from WHS)

AfrD

Guinea

99 %

Household survey

1994/5

WBAD, 2000 citing Household Integrated Survey, Enqute

intgrale budget et consummation, 1994/95

AfrD

Guinea-Bissau

95 %

Household survey

1993/4

WBAD, 2000 citing Inquerito ligeiro junto s familias, 1992

AfrD

Madagascar

98 %

Household survey

1999

WBAHSD Enqute Prioritaire auprs des Mnages (EP)

AfrD

Mali

98 %

Household survey

2001

WBAHSD Demographic and Health Survey (DHS)

AfrD

Mauritania

65 %

Household survey

2003

WHS: Mauritania, 2003 (Preliminary data from WHS)

AfrD

Mauritius

1%

Household survey

2003

WHS: Mauritius, 2003

AfrD

Niger

98 %

Household survey

1995

WBAD, 2000 citing Household Priority Survey, Enqute

permanente de conjuncture conomique et sociale, 1995

AfrD

Nigeria

67 %

Household survey
and census data

1992

WBAD, 2000 citing Consumer Expenditure Survey, 1992;

Government of Nigeria, Social Statistics in Nigeria: 1990,
Federal Office of Statistics, Lagos, Nigeria, 1990, p. 61

AfrD

Senegal

41 %

Household survey

2003

WHS: Senegal, 2003

AfrD

Sierra Leone

92 %

Household survey

1989/90

WBAD, 2000 citing Survey of Household Expenditure and

Household Economic Activities (SHEHEA), 1989/90

AfrE

Botswana

65 %

National census

1991

GOB, 1991

AfrE

Burundi

100 %

Household survey

1998

WBAHSD Enqute Prioritaire (EP)

AfrE

Central
African Republic

99 %

Household survey

1993

WBAD, 2000 citing Household Priority Survey, 1993

AfrE

Congo

84 %

Household survey

2003

WHS: Congo, 2003 (Preliminary data from WHS)

AfrE

Cte dIvoire

74 %

Household survey

2003

WHS: Cte dIvoire, 2003 (Preliminary data from WHS)

AfrE

Dem. Rep. of
Congo

100 %

National energy
statistics

1999

WRI, 2003

AfrE

Eritrea

80 %

Household survey

1995

WBAHSD Demographic and Health Survey (DHS)

AfrE

Ethiopia

96 %

Household survey

2003

WHS: Ethiopia, 2003

AfrE

Kenya

81 %

Household survey

2004

WHS: Kenya, 2004

AfrE

Malawi

98 %

Household survey

2003

WHS: Malawi, 2003

AfrE

Namibia

63 %

Household survey

2003

WHS: Namibia, 2003

AfrE

Rwanda

100 %

Household survey

2000

WBAHSD Demographic and Health Survey (DHS)

AfrE

South Africa

18 %

Household survey

2003

WHS: South Africa, 2003

AfrE

Swaziland

68 %

Household survey

2003

WHS: Swaziland, 2003 (Preliminary data from WHS)

38

Households
using solid
fuel

Energy for Sustainable Development

Volume X No. 3

September 2006

Articles

Table 1 (contd.)
Region[1]

Country

Households
using solid
fuel

AfrE

Tanzania

AfrE

Data source

Year

Reference

96 %

Household survey

1992/3

WBAD, 2000 citing Human Resources Development

Survey, 1993 (World Bank 2000).

Uganda

97 %

Household survey

2000/2001

WBAHSD Demographic and Health Survey (DHS)

AfrE

Zambia

86 %

Household survey

2001/2002

Macro, 2003

AfrE

Zimbabwe

73 %

Household survey

2003

WHS: Zimbabwe, 2003 (Preliminary data from WHS)

AmrB

Brazil

12 %

Household survey

2003

WHS: Brazil, 2003

AmrB

Columbia

15 %

Household survey

2000

Macro, 2003

AmrB

Dominican Republic

14 %

Household survey

2003

WHS: Dominican Republic, 2003

AmrB

Mexico

12 %

Household survey

2003

WHS: Mexico, 2003

AmrB

Uruguay

2%

Household survey

2003

WHS: Uruguay, 2003

AmrD

Bolivia

25 %

Household survey

1998

Macro, 2003

AmrD

Ecuador

3%

Household survey

2003

WHS: Ecuador, 2003 (Preliminary data from WHS)

AmrD

Guatemala

62 %

Household survey

2003

WHS: Guatemala, 2003 (Preliminary data from WHS)

AmrD

Haiti

96 %

Household survey

2000

Macro, 2003

AmrD

Nicaragua

58 %

Household survey

2001

Macro, 2003

AmrD

Peru

33 %

Household survey

2000

Macro, 2003

EmrB

Iran

2%

National Energy
Statistics

1999

WRI, 2003

EmrB

Lebanon

0%

National energy
statistics

1996/7

WRI, 2003

EmrB

Libya

3%

National energy
statistics

1996/7

IEA reports 1996/7 that the residential sector in Libya

consumed 5236 TJ of biomass, but no coal [IEA,1999]

EmrB

Tunisia

5%

Household survey

2003

WHS: Tunisia, 2003 (Preliminary data from WHS)

EmrB

United Arab
Emirates

0%

Household survey

2003

WHS: United Arab Emirates, 2003 (Preliminary data

from WHS)

EmrD

Afghanistan

98 %

National energy
statistics

1999

WRI, 2003

EmrD

Djibouti

6%

Household survey

1996

WBAD, 2000 citing Household Priority Survey,

Enqute Djiboutienne auprs des mnages. Indicateurs
sociaux, 1996

EmrD

Egypt

2%

Household survey

2000

Macro, 2003

EmrD

Iraq

2%

National energy
statistics

1999

WRI, 2003

EmrD

Morocco

4.8 %

Household survey

2003

WHS: Morocco, 2003

EmrD

Pakistan

72 %

Household survey

2004

WHS: Pakistan, 2004 (Preliminary data from WHS)

EmrD

Sudan

100 %

National energy
statistics

1999

IEA, 1999

EmrD

Yemen

42 %

Household survey

1998

WBHBS, 2004

EurA

Croatia

12 %

Household survey

2003

WHS: Croatia, 2003

EurA

Czech Republic

2%

Household survey

2003

WHS: Czech Republic, 2003

EurA

Slovenia

8%

Household survey

2003

WHS: Slovenia, 2003 (Preliminary data from WHS)

EurA

Spain

2%

Household survey

2003

WHS: Spain, 2003 (Preliminary data from WHS)

EurB

Armenia

26 %

Household survey

2000

WBDDHS

EurB

Bosnia and
Herzegovina

51 %

Household survey

2003

WHS: Bosnia and Herzegovina, 2003 (Preliminary

data from WHS)

Energy for Sustainable Development

Volume X No. 3

September 2006

39

Articles

Table 1 (contd.)
Region[1]

Data source

Year

Reference

42 %

Household survey

2003

WHS: Georgia, 2003

23 %

Household survey

2002

WBFBS

Slovakia

2%

Household survey

2003

WHS: Slovakia, 2003 (Preliminary data from WHS)

EurB

Tajikistan

75 %

Household survey

1999

WBLSMS

EurB

Turkey

11 %

National energy
statistics

1999

WRI, 2003

EurB

Turkmenistan

0.2 %

Household survey

2000

Macro, 2003

EurC

Estonia

15 %

Household survey

2003

WHS: Estonia, 2003

EurC

Kazakhstan

5%

Household survey

2003

WHS: Kazakhstan, 2003 (Preliminary data from WHS)

EurC

Latvia

10 %

Household survey

2003

WHS: Latvia, 2003 (Preliminary data from WHS)

EurC

Russian Federation

7%

Household survey

2003

WHS: Russian Federation, 2003

EurC

Ukraine

6%

Household survey

2003

WHS: Ukraine, 2003 (Preliminary data from WHS)

SearB

Indonesia

72 %

Agricultural census

2003

WBAC

SearB

Sri Lanka

67 %

Household survey

2003

WHS: Sri Lanka, 2003

SearB

Thailand

72 %

National energy
statistics

1997

FAO, 1997

SearD

Bangladesh

88 %

Household survey

2003

WHS: Bangladesh, 2003

SearD

India

74 %

Household survey

2003

WHS: India, 2003

SearD

Myanmar

95 %

Household survey

2004

WHS: Myanmar, 2004

SearD

Nepal

80 %

Household survey

2003

WHS, 2003

WprB

Papua New Guinea

90 %

Household survey

1996

World Bank Household Survey (HHS) found under

the LSMS website
http://www.worldbank.org/html/ prdph/lsms/l

WprB

Cambodia

96 %

Household survey

2000

WBDDHS

WprB

China

80 %

National energy
statistics

1996

GPRC, 1996

WprB

Lao Peoples
Democratic Republic

97 %

Household survey

2003

WHS: Lao Peoples Democratic Republic, 2003

(Preliminary data from WHS)

WprB

Malaysia

1%

Household survey

2003

WHS: Malaysia, 2003 World Health Organization

WprB

Philippines

47 %

Household survey

2003

WHS: Philippines, *2003 World Health Organization

WprB

Vietnam

70 %

Household survey

2003

WHS: Vietnam, 2003

Country

Households
using solid
fuel

EurB

Georgia

EurB

Romania

EurB

Note
1. World Health Organization epidemiologic regions are defined by geographic location and patterns of adult and child mortality. A more detailed description is provided in Appendix A.

seems to be limited. See Table 5.

violation of assumptions of normalcy and constant variance.

Collinearity and multicollinearity were assessed, and tolerance statistics were not suggestive of any multicollinearity.
Newly available household fuel use data and more recent values of development parameters were used in the
model for the year 2003 to test whether the original
model, based on limited data, still holds with a larger
number of data points. The initial model turned out to be
relatively robust and resulting estimates of the 2003 modeling were not very different from the 2000 predictions
(Tables 3 and 4). Per capita petroleum use, however, was
no longer a significant parameter in the 2003 predictions
and was therefore dropped from the model. The importance of information on per capita petroleum use thus
40

Energy for Sustainable Development

3. Results
This model was used to predict percentages of household
solid fuel use in all countries with unknown solid fuel
use patterns. For the recent MDG reporting on solid fuel
use, for a total of 181 countries, 94 country data were
compiled from surveys and 52 country data were assumed
to have made the transition to clean fuels with less than
5 % of solid fuel use. For 35 countries missing data were
estimated by the model described in this article. Missing
values of explanatory variables (e.g., GNI per capita) were
replaced with mean values for each variable. In order to
force the percentage of solid fuel use to lie between 5 %
!

Volume X No. 3

September 2006

Articles

Table 2. Development parameters used to model household solid fuel use

Indicator

Source

Proportion of households using solid fuels (dependent variable)

Household Fuels Database [Smith and Desai, 2001]

Adult female illiteracy, 1998

[World Bank, 2001]

Average annual growth rate

[World Bank, 2003]

Dummy variables for six WHO sub-regions

World Health Organization

Electricity consumption, per capita, 1997 (kWh/year)

[World Bank, 2001]

Annual fuelwood production per capita (kg)

Transformation of fuelwood production, 1996

Annual fuelwood production (kg)

[United Nations, 1996]

Population 2000

[United Nations, 2003]

Gini coefficient

[World Bank, 2003]

GNI per capita ($, PPP)

[World Bank, 2003]

ln (GNI per capita, 1999)[1]

Transformation of GNI per capita

Petroleum use per capita (kg)

[United Nations, 1996]

ln (petroleum use per capita)

Transformation of petroleum use per capita

Rural population (%)

[United Nations, 2003]

Note
1. ln = natural logarithm

Table 3. Fuel prediction model, 2000[1]

Unstandardized coefficients
Beta

(Constant)

Std. error

95 % CI for beta
t

Sig.

Lower

Upper

1.12

0.350

3.19

0.0025

0.414

1.82

0.661

0.214

3.09

0.0033

0.231

1.09

WHO Eastern Mediterranean Region

-0.248

0.0709

-3.50

0.0010

-0.390

-0.105

ln (GNI)

-0.104

0.0405

-2.56

0.0136

-0.185

-.0224

-0.0003

0.0001

-2.55

0.0143

-.0006

-.0001

Rural

Per capita petroleum use

Note

1. Dependent variable is the percentage of households using solid fuels.

Table 4. Fuel prediction model, 2003[1]

Unstandardized coefficients
Beta

Std. error

95 % CI for beta
t

Sig.

Lower

Upper

(Constant)

1.543

0.2595

5.95

0.000

1.03

2.06

Rural

0.857

0.1243

6.90

0.000

0.61

1.104

WHO Eastern Mediterranean Region

-0.156

0.0568

-2.75

0.007

-0.27

-0.044

ln (GNI)

-0.181

0.0267

-6.76

0.000

-0.23

-0.128

0.0000339

-1.13

0.260

Per capita petroleum use

-0.000453

-0.000124

0.0000341

Note
1. Dependent variable is the percentage of households using solid fuels.

Energy for Sustainable Development

Volume X No. 3

September 2006

41

Articles

Table 5. Fuel prediction model without petroleum use, 2003[1]

Unstandardized coefficients
Beta

Std. error

95 % CI for beta
t

Sig.

Lower

Upper

(Constant)

1.534

0.260

5.90

0.000

1.017

2.050

Rural

0.891

0.1214

7.38

0.000

0.651

1.131

WHO Eastern Mediterranean Region

-0.169

0.0559

-3.02

0.003

-0.280

-0.0580

ln (GNI)

-0.184

0.0267

-6.89

0.000

-0.237

-0.131

Note
1. Dependent variable is the percentage of households using solid fuels

Figure 1. Proportion of households using solid fuels: known vs. predicted values of national household solid fuel use

and 95 %, all predicted values less than 0 or greater than

100 were converted to 5 % and 95 % respectively.
3.1. Comparison of known versus predicted values
The following graph of known versus predicted values of
solid fuel use suggests that the model performs reasonably
consistently across different proportions of solid fuel use
(Figure 1). In other words, the model estimates fuel use
42

Energy for Sustainable Development

as well for countries with low proportions of solid fuel

use as it does for countries with high proportions of solid
fuel use. There does, however, seem to be a tendency to
slightly under-predict at higher proportions of solid fuel
use, and to slightly over-predict at lower proportions. In
addition, there does not seem to be a bias towards overor underestimation of fuel use for countries where around
!

Volume X No. 3

September 2006

Articles

households using solid fuels is an imperfect indicator of

indoor air pollution, selected for convenience and current
availability of data. This indicator does not address the
issue of ventilation/household characteristics and most
importantly, does not necessarily provide a good reflection of potential exposures resulting from solid fuel use.
A useful indicator of indoor air pollution should not only
report the main cooking fuel but also include aspects of
ventilation, such as stove type and cooking location.
Moreover, in many countries, particularly in sub-Saharan
Africa, cooking is mostly done outdoors.
Indoor air pollution from solid fuel use remains a widespread and important risk factor in developing countries.
Over 1.6 million deaths have been attributed to indoor air
pollution from solid fuel use in 2000 alone [Smith et al.,
2004]. A majority of these deaths are from acute respiratory infections in young children under five years of age.
Shifting up the energy ladder towards modern or
cleaner fuels is a way not only to achieve better health,
but also to improve economic development prospects. In
addition, the increased combustion efficiency that makes
these fuels cleaner and more energy-efficient can also actually result in lower greenhouse gas emissions compared
to poorly burned biomass fuels, even when renewably harvested [Smith and Desai, 2001]. In the absence of interventions or policies to increase access to cleaner fuels,
however, drastic reductions in household solid fuel use
are unlikely to occur. Governments and international development agencies should identify cost-effective, locally
appropriate interventions, including access to cleaner fuels and improved ventilation [Mehta, 2004], to achieve
better health and other important objectives related to sustainable development [Smith, 2002].

half the population uses solid fuels.

3.2. Comparison of 2000 and 2003 estimates
As the model used for 2003 incorporates year-specific information on annual income per capita and urban/rural
population, it enables the statistical estimation of changes
in household solid fuel use predicted to occur with annual
changes in income and population growth. In general,
known estimates of solid fuel use seem to be lower in
2003 than in 2000. It should be noted, however, that these
differences are probably largely due to improvements in
the underlying household solid fuel use data, and therefore do not necessarily reflect a decline over the threeyear period from 2000 to 2003. In general, countries
which have experienced economic development have
slightly lower estimates of solid fuel use in 2003.
4. Discussion
Within just a few years, the knowledge base about solid
fuel use has significantly increased, mainly due to the recent information gathered through the World Health Survey providing data for 42 countries with previously very
scarce information. In 2000, fuel use was only available
for 52 countries. In contrast, in this new round of reporting on the situation in 2003, solid fuel use estimates had
to be modeled for only 35 countries.
Statistical estimations are not absolute reflections of reality. Small changes in estimates generated by the model
may not necessarily indicate small changes in household
solid fuel use. As previously mentioned, the compilation
of data from various sources, with different methodologies and sample sizes, makes it difficult to characterize
the uncertainty associated with the estimates for each
country and region. Thus, while the model is a useful
tool, there is no substitution for routine assessments of
household solid fuel use.
Many of the nations with poor predictions (including
Algeria, Afghanistan, the Congo, and Sudan) are experiencing serious civil unrest or war. This suggests that the
economic and development indicators used here, and/or
the known solid fuel use estimates, may not be robust
under circumstances of political instability or insufficient
governance.
If the prevalence of household solid fuel use is relatively high in a country, environmental health specialists
should lobby for the addition of household fuel-related
questions to national censuses or household surveys, preferably distinguishing among the principal categories, i.e.,
electricity, gas, coal, wood, charcoal, dung, and crop residues. For some locations, response categories may also
need to include modern biofuels, such as biogas or ethanol, and solar energy. Although quantities of fuel and enduse information are often difficult to assess by minimally
trained surveyors, experience from several Demographic
and Health Surveys suggests that accurate information on
the main fuel used for cooking and heating is easy to
ascertain. The use of these questions will dramatically improve the reliability and consistency of information on
household fuel use on a global scale.
It should be noted, however, that the percentage of
Energy for Sustainable Development

Addendum: a request for help

The authors recognize that they have likely missed national surveys or other reliable sources
of information on solid fuel use, as many of these are not available in the indexed literature
archives, but in government and NGO reports and other gray literature. Thus, we ask
readers who know of such information missing in our database to share relevant references.
Please send such information to Kirk Smith at [email protected] and Eva Rehfuess
at [email protected]. Thank you.
Acknowledgements
The authors are thankful to the Department for the Protection of the Human Environment,
World Health Organization, for supporting this work. We much appreciate the comments,
advice, and leads to sources provided by Robert Bailis, Jonathan Sinton, and Maria Josefina
Figueroa.
The views expressed in this paper are those of the authors and do not necessarily reflect
the position of the World Health Organization, the Health Effects Institute or its sponsors.
Note
1. Unlike gross domestic product (GDP), gross national income (GNI) is a quantitative
measure of the size and performance of an economy calculated using a three-year
average of exchange rates, which makes it less sensitive to short-term fluctuations in
exchange rates and policies. For more detailed information on how the GNIs used here
were calculated, visit www.worldbank.org.
Prefatory note on references
The following abbreviations, some of which have been used in repeated references of which
they form a part, are listed here for convenience to enable the reader to easily find the
references of which they form part.
WBAD, 2000 = World Bank, 2000. Africa Database: 2000
WBAHSD = World Bank. World Bank Africa Household Survey Databank
WBAC = World Bank. World Bank Agricultural Census

Volume X No. 3

September 2006

43

Articles

WBDDHS = World Bank. World Bank Databank Demographic and Health Survey

WBAD, 2000 citing Household Priority Survey, Enqute Djiboutienne auprs des mnages,
Indicateurs sociaux, 1996, 2000.

WBFBS, 2004 = World Bank. World Bank Family Budget Survey (FBS), 2004

WBAD, 2000 citing Inquerito ligeiro junto s familias, 1992, 2000.

WBHBS, 2004 = World Bank. World Bank Household Budget Survey (HBS), 2004

WBAD, 2000 citing Survey of Household Expenditure and Household Economic Activities
(SHEHEA), 1989/90, 2000.

WBLSMS = World Bank. World Bank Living Standards Measurement Study (LSMS)
WHS = World Health Organization. World Health Survey

World Bank, 2001. World Development Indicators, World Bank, Washington, DC.
World Bank, 2003. World Development Indicators. World Bank, Washington, DC.

References

World Health Organization. World Health Survey [hereafter WHS], 2003.

Food and Agricultural Organization (FAO), 1997. Review of Wood Energy Data in RWEDP
Member Countries, FAO Field Doc. No. 47, 1997.

WHS: Bangladesh, 2003.

Government of Botswana (GOB), 1991. Population; Housing Census, CSO, Gabarone, Botswana, Table H10, p. 483.

WHS: Bosnia and Herzegovina, 2003 (Preliminary data from WHS).

Government of India (GOI), 2001. Census of India.

WHS: Burkina Faso, 2003 (Preliminary data from WHS).

Government of Peoples Republic of China (GPRC), 1996. China Energy Statistical Yearbook: 1991-1996, Zhongguo Tongji Chubanshe (Chinese Statistical Publishing House), beijing.

WHS: Chad, 2003 (Preliminary data from WHS).

International Energy Agency (IEA), 1999. Energy Statistics of Non-OECD Countries 19961997, Paris, The Organization for Economic Cooperation and Development (OECD) (International Energy Agency).

WHS: Congo, 2003 (Preliminary data from WHS).

WHS: Brazil, 2003.

WHS: Comoros, 2003 (Preliminary data from WHS).

WHS: Cte dIvoire, 2003. (Preliminary data from WHS)
WHS: Croatia, 2003.

Macro International Inc. (Macro), 2003. Household Energy Use in Developing Countries: A
Multi-country Study (Source: Demographic and Health Survey (DHS) specially tabulated by
Macro International Inc. for the World Bank, Joint UNDP/World Bank Energy Sector Management Assistance Programme, 2003.

WHS: Czech Republic, 2003.

WHS: Dominican Republic, 2003.
WHS: Ecuador, 2003 (Preliminary data from WHS).

Mehta, S., and Shahpar, C., 2004. The health benefits of interventions to reduce indoor
air pollution from solid fuel use: a cost-effectiveness analysis, Energy for Sustainable Development, VIII(3), September, pp. 53-59.

WHS: Estonia, 2003.

WHS: Ethiopia, 2003.

Smith, K.R., 2002. Indoor air pollution in developing countries: recommendations for research, Indoor Air, 12(3), pp. 198-207.

WHS: Georgia, 2003.

Smith, K., and Desai, M., 2001. The contribution of global environmental factors to illhealth, in Martens, P., and McMichael, A., (eds.), Environmental Change, Climate, and
Health: Issues and Research Methods, Cambridge University Press, Cambridge, UK.

WHS: Guatemala, 2003 (Preliminary data from WHS).

WHS: Ghana, 2003 (Preliminary data from WHS).

WHS: India, 2003.
WHS: Kazakhstan, 2003 (Preliminary data from WHS).

Smith, K., Mehta, S., and Maeusezahl-Feuz, M., 2004. Indoor air pollution from household
use of solid fuels, in Ezzati, M., Lopez, A.D., Rodgers, A., and Murray, C.J.L., Comparative
Quantification of Health Risks: Global and Regional Burden of Disease Due to Selected
Major Risk Factors, Vol. 2, World Health Organization, Geneva, pp. 1435-1493.

WHS: Kenya, 2004.

WHS: Lao Peoples Democratic Republic, 2003 (Preliminary data from WHS).
WHS: Latvia, 2003 (Preliminary data from WHS).

United Nations, 1993. Energy Statistics Yearbook, United Nations, New York.

WHS: Malawi, 2003.

United Nations, 2003. United Nations World Population Prospects, United Nations, New
York.

WHS: Malaysia, 2003.

World Bank, 2000. Africa Database: 2000 (CD-ROM) [hereafter WBAD, 2000]

WHS: Mauritania, 2003 (Preliminary data from WHS).

WBAD, 2000 citing Consumer Expenditure Survey, 1992; Government of Nigeria, Social
Statistics in Nigeria: 1990, Federal Office of Statistics, Lagos, Nigeria, 1990, p. 61.

WHS: Mauritius, 2003.

WHS: Mexico, 2003.

WBAD, 2000 citing Household Integrated Survey, Enqute intgrale budget et consummation, 1994/95, 2000

WHS: Morocco, 2003.

WBAD, 2000 citing Human Resources Development Survey, 1993, 2000.

WHS: Namibia, 2003.

WHS: Myanmar, 2004.

World Bank. World Bank Africa Household Survey Databank Demographic and Health
Survey (DHS), http://www4.worldbank.org/afr/poverty/databank/default.cfm, accessed 2004
[hereafter WBAHSD].

WHS: Pakistan, 2004 (Preliminary data from WHS).

WHS: Paraguay, 2003 (Preliminary data from WHS).
WHS: Philippines, 2003.

WBAHSD Demographic and Health Survey: Benin 2001

WHS: Russian Federation, 2003.

WBAHSD Enqute Camerounaise Auprs des Mnage (ECAM).

WHS: Senegal, 2003 (Preliminary data from WHS).

WBAHSD Enqute Prioritaire (EP).

WHS: Slovakia, 2003 (Preliminary data from WHS).

WBAHSD Enqute Prioritaire auprs des Mnages (EP).

WBAHSD Demographic and Health Survey (DHS)

WHS: Slovenia, 2003 (Preliminary data from WHS).

World Bank. World Bank Agricultural Census, http://www.bps.go.id/index.shtml, accessed

2004 [WBAC].

WHS: South Africa, 2003.

WHS: Spain, 2003 (Preliminary data from WHS).

World Bank. World Bank Databank Demographic and Health Survey (DHS),
http://www.measuredhs.com/countries/country.cfm, accessed 2004 [hereafter WBDDHS].

WHS: Sri Lanka, 2003.

World Bank. World Bank Family Budget Survey (FBS), 2004 [WBFBS, 2004].

WHS: Tunisia, 2003 (Preliminary data from WHS).

WHS: Swaziland, 2003 (Preliminary data from WHS).

World Bank. World Bank Household Budget Survey (HBS), 2004 [WBHBS, 2004].

WHS: Ukraine, 2003 (Preliminary data from WHS).

World Bank. World Bank Living Standards Measurement Study (LSMS),

http://www.worldbank.org/html/prdph/lsms/, accessed 2004 [WBLSMS].

WHS: United Arab Emirates, 2003 (Preliminary data from WHS).

WHS: Uruguay, 2003.
WHS: Vietnam, 2003.
WHS: Zimbabwe, 2003 (Preliminary data from WHS).
World Resources Institute (WRI), 2003. Earth-Trends: The Environmental Information Portal,
January.

WBAHSD National Household Poverty Survey, Gambia 1998, 2000

WBAD, 2000 citing Household Priority Survey, Enqute permanente de conjuncture
conomique et sociale, 1995, 2000.
WBAD, 2000 citing Household Priority Survey, 1993, 2000.

44

Energy for Sustainable Development

Volume X No. 3

September 2006

Articles

Appendix A. World Health Organization epidemiologic regions

Region

Sub-region

Mortality stratum
Child

Adult

Africa

AfrD

High

High

Africa

AfrE

High

Very high

The Americas

AmrA

Very low

Very low

The Americas

AmrB

Low

Low

The Americas

AmrD

High

High

Eastern Mediterranean

EmrB

Low

Low

Eastern Mediterranean

EmrD

High

High

Europe

EurA

Very low

Very low

Europe

EurB

Low

Low

Europe

EurC

Low

High

South-East Asia

SearB

Low

Low

South-East Asia

SearD

High

High

Western Pacific

WprA

Very low

Very low

Western Pacific

WprB

Low

Low

Energy for Sustainable Development and

its role in IEIs mission
IEIs journal Energy for Sustainable Development (ESD) is meant to be a vehicle of communication
between Northern and Southern energy actors on the energy goals and policies as well as energy
plans and projects in developing countries. It also aims to promote South-South interactions to tackle
these problems.
ESD attempts to provide a balanced treatment of:
renewables and end-use efficiency (in the industrial, transport, domestic, commercial and agricultural sectors);
generation (including development) of energy technologies and their dissemination;
hardware (for the conversion of primary energy as found in nature into convenient secondary
energy, for transmission/transport and distribution and for utilisation in end-use devices); and
software (relating to energy policies, institutions, management and financing).
Articles and short articles published are subject to a formal process of peer review.
Find out more about ESD on the Internet at www.ieiglobal.org/esd.html.

Energy for Sustainable Development

Volume X No. 3

September 2006

45