Energy consumption in households

Name: Energy consumption in households
Topic: Energy and manufacturing

Division for Energy, Environmental and Transport Statistics

The results are mainly on national level. In addition, tables are available on regional level.

The survey has been carried out for the years 1993-1995, 2001, 2004, 2006 and 2009. The next survey will be conducted for 2012. The survey is usually carried out every third year.

Not applicable

Raw data and results from the processing routine will be stored on household level in SAS / Unix data banks.

The purpose of the statistics is to present figures for average energy consumption per household for different household types and house types, and to study the development over time. This shows how the energy consumption varies between different types of households and indicates what kind of factors influence the households' energy consumption. The statistics also comprise combinations of heating equipment in dwellings, which give an indication of the energy flexibility in households or the households' possibility to switch between different energy sources. Using the survey of consumer expenditure as a basis, we have access to a large amount of relevant background information. This facilitates studies of correlations between energy consumption and relevant factors such as house types, dwelling area, number of persons in household, income etc. The survey can also be used to study the effect on energy consumption of heat pumps by comparing similar households with and without heat pumps. In addition, questions about other energy efficiency efforts (in 2001, 2006 and 2009) enable the study of the extent and effects of such efforts in the households.

The survey was first time carried out for 1993 and then repeated for 1994, 1995, 2001, 2004 2006 and 2009. A similar study of energy consumption in households was conducted by Statistics Norway for 1990, but other sources and methods were used in that survey. The results for 1990 are therefore not directly comparable with the results for the other years. For 1990, 2001 and 2006 the data were also used in analyses of the average share of electricity consumption for different purposes such as cooking, space heating, water heating etc. The research department in Statistics Norway was responsible for these analyses.

The statistics are among other things used as background information in energy reviews/reports in ministries, in various types of research, in the Norwegian Water Resources and Energy Directorate, and in Enova (a public enterprise which works towards more efficient and environmentally friendly energy consumption). The data and results are used for research purposes on the Research Department in Statistics Norway, and in work with energy indicators for households. There is also a lot of public interest in the results.

Figures for all households total electricity consumption are published in the annual electricity statistics that are available on the following web site:

http://www.ssb.no/english/subjects/10/08/10/elektrisitetaar_en/

Figures for the households total energy consumption are also a part of the energy balance that are published annually, usually in October / November, see:

http://www.ssb.no/english/subjects/01/03/10/energiregn_en/

The electricity consumption for households in the annual electricity statistics is used as a source for this figure in the energy balance.

By multiplying results for energy consumption per household from the household survey with the number of all households (from the family and housing statistics) we use to get similar figures as in the energy balance. There are some differences because the data sources and methods for calculating these figures not are the same in these statistics.

The survey is voluntary for households. The data for electricity consumption are collected from the households' electricity supplier pursuant to §2-2 and 2-3 of the Statistics Act.

Not applicable

The statistics are based on a sample survey where the household is the statistical unit.

The source is the annual survey of consumer expenditure and related information from registers on income. For years when energy surveys have been conducted, the survey of consumer expenditure has been supplemented with additional questions on energy consumption and heating equipment in order to give a better basis for statistics and analyses of the energy consumption in households. Data for electricity consumption are collected from the households' electricity supplier. In the 2009 survey, it was collected additional data from households that had a heat pump. In addition to electricity consumption for 2009, we collected consumption data for the year when the heatpump was installed, and the consumption before and after the installation year. The purpose was to compare the electricity consumption before and after the installation of a heet pump.

The survey of consumer expenditure is based on personal interviews and detailed accounting in a representative sample of 2 200 households. For 2004, 48 per cent of the households did not respond, and the net sample became 1 148 households. The tables for combination of heating equipments in dwellings are based on information from all 1 148 households, while the tables for energy consumption are based on information from a somewhat smaller sample of 1 091 households. The smaller sample is due to that households are excluded from energy calculations if important data for energy consumption or background information about dwelling area, year of construction etc. are missing.

For 2006, tables for heating equipments are based on answers from 1027 households, while results for energy consumption are based on answers from 1010 households. This accounts for 47 and 46 per cent of the gross sample respectively.

For 2009, tables for heating equipment are based on data from 1155 households, while the figures for energy consumption is based on answers from 1131 households. This corresponds to 53 and 51 per cent of the gross sample respectively.

For the next survey in 2012, the sample will be three times higher than earlier, so we can assume that the net sample becomes about 3000 households.

See also "About the statistics" for the consumption survey:

http://www.ssb.no/english/subjects/05/02/fbu_en/

Computer Assisted Personal Interview (CAPI) and detailed accounting for a period of 14 days. Data from the electricity plants are collected in questionnaires by post, or electronically in excel sheets. The survey of consumer expenditure is carried out once a year, while the questionnaire on energy and heating equipment only is added if a separate energy consumption survey for households will be carried out . We have not measured how long it takes to answer the additional energy questions or how long it takes the electricity plants to fill in sales figures (in kWh) for electricity to the households. The response burden for individual electricity plants depends on how many households they have to provide information on. In general, the burden becomes larger for large electricity plants with many subscribers.

Revisions and controls of data from the additional questionnaire on energy are done in the data processing programs. The data processing and calculations are done in SAS, a software specially developed for preparing statistics. The software contains some automatic controls that reveal extreme figures. Some obvious errors are corrected automatically in the software, while others require specific treatment. If the energy data for a particular household are not sufficient to calculate the energy consumption, the household is excluded from the energy calculations.

See also "About the statistics" for the survey of consumer expenditure: http://www.ssb.no/english/subjects/05/02/fbu_en/

For 2004, 92 per cent of the respondents gave Statistics Norway permission to collect data on their electricity consumption from the electricity plant. For the 2004 survey, Statistics Norway collected the electricity consumption from the grid supplier, while information about the type of electricity contract are collected from the power supplier. Information about electricity contract type is collected for research purposes only, for instance price elasticity on electricity. Data from the electricity plants are collected pursuant to the Statistics Act and thus the response rate is very high. However, in some cases the electricity plant is unable to find the household in their register, for instance because the subscriber's name is incorrect or the questions are sent to the wrong electricity plant. As a result, we do not receive data for all households who have given StatistiCS Norway permission to collect data from their plant.

For 2004, the electricity consumption is based on information from electricity plants covering 90 per cent of the households in the sample. However, for various reasons some of the plants have only supplied the consumption for parts of the year, for instance for ten months. In such cases, the consumption is converted to annual consumption on the basis of a profile that shows the percentage distribution of average electricity consumption by month. For 81 per cent of the households, the electricity consumption provided by the plant covered the whole year (for 2004).

For 2006, about 90 percent of the households gave their permission to Statistics Norway to collect electricity consumption from their electricity plant. We received data from the electricity plant for most of these households. However, also for 2006, we had to make some adjustments of these data, in cases where the plant reported data for another period than the calendar year 2006. In cases where data for the electricity plant were totally absent, the electricity expenses, reported by the households, are used as a source for the electricity consumption

For 2009, 1000 households, corresponding to 88 per cent of the households, permitted Statistics Norway to collect electricity consumption from their electricity plant. For the remaining households, and for those where we of other reasons not have received information from the electricity plant, we have used data for the households expenditures to electricity, if it appeared to be reasonable.

The survey of consumption expenditure contains information on household's expenses and demand for (in tonnes or m 3 ) fuel wood, heating oils and kerosene for the last 12 months for dwellings and cottages/holiday houses. We only ask about electricity expenses, not the quantity in kWh. For households where the consumption data from the electricity plant are missing, the electricity consumption is calculated on basis of the households' expenses for electricity in the last 12 months. The household's own reporting of expenses is less reliable than data from the electricity plant. In cases where this figure seems to be wrong and other information about the household's energy consumption is missing, the household is excluded from the energy calculations. Because the interviews of the households takes place throughout the year, the consumption refers to parts of two calendar years (partly 2008 and partly 2009 for the 2009 survey). This is also the case for consumption of fuel wood, oil and kerosene that are calculated on basis of purchases, in physical or monetary units, to these energy sources the last 12 months.

In cases where the consumption is calculated from the expenses, the expenses are divided by an average price for the energy source for the relevant period (average price for 2008-2009 for the 2009-survey).

In the survey of consumer expenditure, data on expenses for energy sources for dwellings and cottages/holiday houses are collected gathered. The consumption in cottages/holiday houses is collected in the additional questionnaire on energy and then subtracted from the energy consumption, as the survey only comprises energy consumption in permanent houses.

The statistics for combinations of heating equipment are calculated on the basis of information that is collected in the additional questionnaire on energy. For some households, the information about energy consumption and heating equipment does not correspond to each other. For example, some households report consumption of wood, but not wood stove. In such cases we have assumed that the household has forgotten to report all types of heating equipment, and have therefore corrected the heating equipment accordingly.

All data for the households are treated confidential

In connection with a report which where written in 2005 about energy consumption in household, it was undertaken a retrograde revision of the results from the household surveys, where the same methods were used to correct the data for biases in the sample. The purpose was to obtain comparable results over time.

For the years 1993-1994 and 1995 it was originally used other methods and data sources to correct for biases in the sample. The results were presented in a report which was published in 1999. This report shows somewhat different results compared to results that were presented later on. For 2001, it was originally corrected for size of dwelling area, in addition to household size and house type. However, this was not done in the revised figures, because it came to light that the definition of dwelling area was somewhat different in the consumer expenditure survey and the Population and housing census, which is used as a key for the right distribution.

For 2006, the consumption of fuel wood and kerosene in cottages /holiday houses are included in the households energy consumption, while this is deducted in consumption figures for earlier years. To obtain a comparable development between 2004 and 2006, the 2004-figures were calculated again, including consumption of fuel wood and kerosene in cottages. The timeserie tables in the publication of results for 2006 show by this two sets of consumption figures for 2004, both including and not including consumption of fuel wood and kerosene in cottages.

In the questionnaire used in the survey of consumer expenditure, we ask about expenses for energy in the last 12 months, not the actual consumption. Wood, oil and kerosene can be stored and are not always consumed the same year as they are purchased. The calculated energy consumption may therefore differ from the actual consumption. We have assumed that this evens out between the households, and that it has no consequence for the calculated total energy consumption on average. However, this might lead to errors in detailed tables where the energy consumption for certain groups are based on a small sample of households. Errors might also occur if the households not remember, or do not have any information on how much they have purchased of the different energy sources during the last 12 months. Furthermore, errors can occur from incorrect recording of household answers.

Households that have common central heating and pay for their use of energy through the rent, normally have limited information about their energy consumption. This is because many of them do not have an electricity meter, or they have a meter that only covers parts of the consumption. The electricity plants are usually unable to provide information about the electricity consumption in these cases. Thus special calculations are made for these households, among other things based on expenses for energy consumption in the rent if available. The consumption is divided between electricity and oil.

For households that use district heating, the energy expenses in the rent are converted to district heating consumption, or special calculations are made based on other information. Consumption of district heating is not given as a separate figure in the tables, because few households in the sample use district heating, and they seldom know how much they have used. The district heating plant usually cannot provide this information for individual households because they make deliveries to a central heating system in a housing cooperative. The calculations for district heating consumption are therefore uncertain. However, calculated consumption of district heating is included in the figures for total energy consumption per household. District heating makes up only about 1 per cent of the energy consumption in households, and is not a big source of error.

For 2004, 48 per cent of the households did not respond. For 2006, the non-response-rate was 53 per cent while the non-response-rate for 2009 was about 47 per cent.This also includes questionnaires that were too insufficient filled in to be used for statistics purposes.

The non-respondents are not evenly distributed among all household groups, and this implies that the distribution by household category does not become representative for the total population. In order to correct sample bias, corrections have been made for non-response. Household groups with a high non-response rate get a high weight when estimating average figures. The non-response rate is particularly high among one-person households. As a result of this, they get a higher weight than the other households in the sample.

Variables that have been used in the correction of bias in the sample are household size (1,2,3,4 and more than 4 persons) and three house types (detached house/farming house, row house/semi-detached house and flat). The source for the correct distribution of household size and house types is the Population and Housing census for 1990 and 2001 and the family and housing statistics combined with the housing register, that started in 2005. The distribution for 2004 is interpolated on basis of data for 2001 and 2005, while the family and housing statistics is used directly as a source for right distribution for the 2006- and 2009-survey.

Even if the data are weighted, there may still be biases in distributions of certain variables, as age, region etc. that influence the results. Of tehnical reasons, it is not possible to correct for all kinds of biases. We have chosen to weight the data on basis of household size and house type, because these variables are some of the most important for the energy consumption. However, the correction that is undertaken can also correct for other biases. It is for instance a connection between age and household size, and correction of household size can by this also make the agemix in the sample more representative.

Other errors that can occur is for instance if the electricity plants not have real meter readings from the households, and report stipulated figures instead of. Normally, data from the electricity plants is much better than the households' own reporting, even if it is stipulated.