Australian Bureau of Statistics
1504.0 - Methodological News, Mar 2012
Previous ISSUE Released at 11:30 AM (CANBERRA TIME) 22/03/2012
|Page tools: Print Page Print All RSS Search this Product|
An Update on the Methodologies for Estimating Energy Consumption
The new data source for the Australian Energy Statistics – the National Greenhouse and Energy Reporting System (NGER) – does not require businesses using less than 200 terajoules to report their energy consumption. This results in a data gap in the estimate of total industry energy consumption by businesses. The Analytical Services Branch is currently working with the ABS Centre of Environment Statistics on a project aiming to develop a methodology for estimating statistics for this under-coverage.
The energy modelling literature review identified several possible modelling approaches. Of these, the most appropriate approach based on the available data, was to estimate business energy consumption using the static single equation approach with turnover and industry dummy variables (at 3-digit ANZSIC06 level) as the main explanatory variables. Data were sourced from the ABS 2008-09 Energy, Water and Environment Survey (EWES) and the Business activity statement Unit Record Estimates (BURE).
Different model specifications were tested including: (1) log-linear models, (2) a model with interaction dummy variables, (3) a model using derived price measure ($ expenditure per gigajoule), (4) a model with ANZSIC06 2-digit industry dummy variables, (5) models at division level, and (6) a model using energy expenditure data.
Several methodological issues were also highlighted during the estimation process. These included transforming data, dealing with zero observations of turnover and energy consumption volume, and correcting for log transformation bias.
These documents will be presented in a new window.
This page last updated 11 July 2012