A Comparison of Measures by Consumption and Supply Surveys
June 15, 1988
Energy policy makers-be they industry executives, members of utility commissions, or government officials require reliable data to make informed decisions. The need is particularly acute when the task is to forecast the future of energy markets. Making reasonable estimates of possible future prices, supply, and demand for energy requires accurate and detailed data about the current and historical situation of energy markets.
The Energy Information Administration (EIA) was mandated by Congress to be the agency within the Department of Energy that collects, analyzes and disseminates impartial, comprehensive data about energy-how much energy is produced, who uses it, and the purposes for which it is used. Energy data are also collected by a wide range of other groups; including other governmental agencies Federal, State and local; individual utility companies; and members of the energy industry, including individual companies and industry associations, such as the Edison Electric Institute.
To forecast energy market behavior, it is useful to bring together as much data as possible about the supply and demand of energy. The integration of data on the same subject from different sources, however, usually entails some difficulty since data collected by different groups and/or by different means are seldom completely compatible. Some effort must therefore be undertaken to reconcile the differences in data sets before the combined data can be used in an analysis.
EIA collects data from two distinct sources that, in their entirety, provide a comprehensive picture of energy production, marketing and use in the United States. The first set of surveys- termed "supply" surveys in this report-are directed to the suppliers or marketers of specific fuels (including electricity). These surveys measure the quantity of the specific fuel produced and/or supplied to the market, along with other information related to the fuel's production and supply. Supply surveys are conducted for petroleum, natural gas, electricity, and coal. (The "supply" surveys for coal include surveys of consumers, which provide information on coal consumption; these surveys are discussed in the chapter on the industrial sector.)
The second group of surveys-termed "consumption" surveys in this report-collect data from samples of end-use consumers. These consumers use one, several, or all kinds of fuels. These surveys gather information on the types of fuels used by the consumer, the purposes for which each fuel is used, and the characteristics of the users. The end-use consumption surveys are conducted for the residential sector, commercial buildings, manufacturing establishments, and personal transportation.
Together, the uata from the supply and consumption surveys provide a ncn, aeianeu source or data on energy supply and use. The supply surveys have been conducted for many years, starting with predecessor agencies to EIA, The consumption surveys were started during the past decade under the auspices of EIA.
It is tempting to merge the time series data from the supply surveys with the detailed cross- sectional data from the consumption surveys. However, there are important differences between the supply and consumption surveys which need to be taken into account in any analysis that uses both data sources.
This report discusses the relationship of EIA data from the supply surveys and from the consumption surveys. This comparison is made for three major end-use sectors: residential, commercial, and industrial. The discussion of industry is limited because the consumption survey covers manufacturing industries while the supply surveys (with the exception of coal) cover the entire industrial sector. No discussion is given of the transportation sector because the end-use consumption survey collects data only for personal transportation, and there are no comparable supply data available. Estimates of energy supply and consumption in each sector are compared for electricity, natural gas, petroleum fuels, and (for manufacturing industries) coal.
It is tempting to merge the time series data from the supply surveys with the detailed cross- sectional data from the consumption surveys. However, there are important differences between the supply and consumption surveys which need to be taken into account in any analysis that uses both data sources. This report discusses the relationship of EIA data from the supply surveys and from the consumption surveys. This comparison is made for three major end-use sectors: residential, commercial, and industrial. The discussion of industry is limited because the consumption survey covers manufacturing industries while the supply surveys (with the exception of coal) cover the entire industrial sector. No discussion is given of the transportation sector because the end-use consumption survey collects data only for personal transportation, and there are no comparable supply data available. Estimates of energy supply and consumption in each sector are compared for electricity, natural gas, petroleum fuels, and (for manufacturing industries) coal. There are three important differences between the supply-based and consumption-based surveys:
- The
supply
surveys
measure
product
supplied
at
some
point
in
the
supply
chain,
while
the
consumption
surveys
measure
the
actual
quantity
of
fuel
consumed
(except
for
petroleum
fuels
for
which
the
quantity
delivered
to
the
consumer
is
measured).
Since
fuels
(except
electricity)
can
be
stored,
the
amount
of
product
supplied
to
a
sector
in
a
given
period
is
not
necessarily
equal
to
the
amount
consumed.
- It
is
tempting
to
merge
the
time
series
data
from
the
supply
surveys
with
the
detailed
cross-
sectional
data
from
the
consumption
surveys.
However,
there
are
important
differences
between
the
supply
and
consumption
surveys
which
need
to
be
taken
into
account
in
any
analysis
that
uses
both
data
sources.
This
report
discusses
the
relationship
of
EIA
data
from
the
supply
surveys
and
from
the
consumption
surveys.
This
comparison
is
made
for
three
major
end-use
sectors:
residential,
commercial,
and
industrial.
The
discussion
of
industry
is
limited
because
the
consumption
survey
covers
manufacturing
industries
while
the
supply
surveys
(with
the
exception
of
coal)
cover
the
entire
industrial
sector.
No
discussion
is
given
of
the
transportation
sector
because
the
end-use
consumption
survey
collects
data
only
for
personal
transportation,
and
there
are
no
comparable
supply
data
available.
Estimates
of
energy
supply
and
consumption
in
each
sector
are
compared
for
electricity,
natural
gas,
petroleum
fuels,
and
(for
manufacturing
industries)
coal.
There
are
three
important
differences
between
the
supply-based
and
consumption-based
surveys:
The
supply
surveys
measure
product
supplied
at
some
point
in
the
supply
chain,
while
the
consumption
surveys
measure
the
actual
quantity
of
fuel
consumed
(except
for
petroleum
fuels
for
which
the
quantity
delivered
to
the
consumer
is
measured).
Since
fuels
(except
electricity)
can
be
stored,
the
amount
of
product
supplied
to
a
sector
in
a
given
period
is
not
necessarily
equal
to
the
amount
consumed.
The
consumption
surveys
are
based
on
a sample
of
the
consumers
in
each
economic
sector,
so
there
is
a
sampling
error
associated
with
the
data
from
these
surveys
in
addition
to
the
nonsampling
errors
that
are
common
to
both
sample
and
census
surveys.
In
many
cases,
the
supply
surveys
are
a
census
of
suppliers
and
so
have
no
sampling
error.
- There are differences between the consumption and supply surveys in the coverage of each sector. The residential consumption survey, for example, does not cover the same set of residences as do the supply surveys. Further, there are differences among the various supply surveys in the coverage of each sector.
These differences are discussed further in the next chapter. It is important to note that the existence of differences between the two types of surveys does not mean that one or the other is incorrect or improperly conducted. In both types of survey, the decisions on the kinds of data to gather and the means to collect them have been based on sound statistical criteria and a full understanding of the types of data that could be collected without placing an undue burden on the respondent. For the most part, the differences between the two types of surveys reflect limitations imposed by efforts to minimize the burden of responding to the surveys and the need to limit the costs of data collection.
The result of the comparison in this report shows that, taking into account their differences, the two types of surveys present a consistent picture of energy consumption in the United States. Neither type of survey alone provides a complete description of energy consumption by economic sector. A fuller understanding of energy use by sector can be obtained by using data from the two sources, allowing for their differences.