The Australian Bureau of Statistics will explore the privacy implications of using electricity usage data to improve the accuracy of the next census.
Subscribe now for unlimited access.
$0/
(min cost $0)
or signup to continue reading
The bureau has been researching how it can use administrative datasets - including de-identified utility data and government administration records - as part of the 2021 national survey.
It comes after a post-mortem of the last census found the bureau overestimated the number of occupied houses in Australia, and counted people it shouldn't have.
It also found a problem with the way the bureau was using "donor houses" to estimate age, sex and marital status for houses where no census form was returned.
The bureau "borrows" census counts from a similar nearby house where a form was returned. They do this so the dataset is "more representative" of the true number of people in the country and in each region on census night.
However the team assessing the 2016 census found this approach was problematic, as "donor houses" tended to overrepresent older Australians, who were more likely to respond to the Census.
"Administrative data, such as counts of people from de-identified Medicare and Centrelink data, is showing us that people in the houses where forms aren't provided tend to have a younger age profile on average," the Bureau said.
"Our research is showing that this administrative data can help us choose donor houses from the census that more truly reflect the ages of people in houses that don't provide forms."
An Australian Bureau of Statistics spokeswoman said information like electricity usage could help determine whether a house was occupied or not on census night.
READ MORE:
"The Census provides useful data on residential occupation rates. However, in some cases it can be difficult to determine if a house was unoccupied on Census night, particularly in secure apartment blocks," she said.
"In addition to our observations during field collection, electricity readings at Census time can help indicate the likelihood of a house being unoccupied and improve the quality of the data available for governments, businesses and the community."
The bureau spokeswoman said the research was being done "within our commitments to privacy and security".
It has hired privacy specialist Information Integrity Solution to due a privacy impact assessment on the use of the data for the census, to the tune of nearly $30,000.
The assessment will look at the possible privacy risks from using the data and how to mitigate them. It will be published next year.
The Australian Bureau of Statistics already draws on other sorts of administrative data collected by governments to produce its national statistics.
Countries like Canada, the United Kingdom and New Zealand have used administrative data to help with their censuses.
"International experience suggests that administrative data has the potential to enhance the operational efficiency and value of the Census for the Australian community," the bureau said.
Using administrative data may reduce the cost of running the census, allow new information to be collated and provide a contingency for areas that are affected by a natural disaster or other emergency during census time.
