4661.0 - Using electricity data to understand COVID-19 impacts , 2020  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 02/10/2020  First Issue
   Page tools: Print Print Page Print all pages in this productPrint All


In aggregate, business purchasing of electricity began a declining trend around 16 March 2020. Examination of unit level data revealed that businesses exhibited a range of responses after government restrictions were announced: some markedly reduced electricity purchases, others had some reduction, while others showed no discernible change.

We linked the ABS Address Register and the ABS Business Register to the business electricity data to enable us to examine trends across different industries (ANZSIC classification), and by the number of employees in a business.

Image: Line graph showing that electricity usage for Accommodation and Food Services, and Retail Trade tracked similarly to All Industries in 2019 and 2020 until March, when these two industries began to reduce usage.
Figure 5: Electricity usage by the Accommodation and Food Services, and Retail Trade Divisions, indexed to March 2–15, 2020, rolling 7-day average.

Some industries appeared to be more affected than others by the progressive restrictions starting in mid-March 2020. The Accommodation and Food Services industry electricity purchases declined more over the following month than the overall industry average, consistent with this industry being more affected than average by lockdowns. Detailed data analysis revealed that the aggregate trends were heavily affected by a small number of high-intensity users, and examination of unit level trends yields further insights.

We examined changes in daily electricity purchases by businesses between March 2–15 and March 16–29. We found that certain industry groups within the Accommodation and Food Services, the Retail, and the Other Services industry divisions had the highest proportion of businesses with a reduction of over 30% in electricity purchases between the two periods. These results are broadly consistent with findings from the ABS Business Impacts of COVID-19 Survey.

Image: Bar graph showing the ANZSIC groups with the highest proportion of businesses substantially reducing electricity usage. The top is "Pubs,  Taverns and Bars" by a substantial margin.
Figure 6: The 10 industry groups (ANZSIC) with the highest proportion of businesses reducing electricity purchase by more than 30%. Groups with fewer than 20 businesses in the dataset were excluded.

The industry groups that demonstrated the highest proportion of substantial (30%) reduction in electricity usage are largely those groups that have been particularly affected by COVID-19 and associated government lockdowns. Businesses that rely on personal interaction or colocation feature prominently in Figure 6, e.g. bars, clubs, gyms and cafes.

Map: For inner Melbourne, areas are coloured by what proportion of businesses have substantial reduction in electricity usage from the start to the end of March. The CBD is most affected, as are some inner South areas. Most areas show between 5 - 15 %.
Map 3: Inner Melbourne, proportion of businesses reducing daily electricity usage by 30%, comparing the fortnight March 2–15 to March 16–29. Darker shading indicates a greater proportion of businesses with reduced usage.

Analysing geospatial trends, we found that areas close to the city centre had a greater proportion of businesses with a substantial reduction in electricity purchasing.

Business Conclusion

We found that the smart meter data can be used to shed light on the impact of COVID-19 and government lockdowns on businesses, and how this varies through time, by industry, and geospatially. The advantages of this data source are its granularity and timeliness, enabling us to drill down and investigate trends. Integrating this new data with ABS assets such as the ABS Business Register greatly enhances its value and the discoverability of new insights. Some of the challenges include the limits to the geographic coverage, the extent to which the data can be accurately geo-coded and linked to the ABS Business Register, and interpretability of the data. The availability of a greater scope of data in time and geography, and the ability to compare unit level data with other data relating to business activity would enable a more comprehensive and revealing analysis.