The story in one line

Victoria just recorded its first annual fall in crime in three years. It will be a talking point before the November election. But the fall is small, and it comes after three years of sharp increases. The bigger question is what actually went up.

The answer is theft. Property crime fell sharply during COVID, then rose about 50% in three years. Burglary fell over the decade, but theft surged. Four of the five biggest risers are theft: shoplifting, car theft, stealing from cars and other theft. The fifth is breaches of family violence orders, which says more about policing and reporting than about new offending.

Two things make this worth a second look. Most of the fastest-growing councils did not see property crime outpace their population. And most property offences never get solved.

1. Property crime fell in COVID, then bounced back

Source: Crime Statistics Agency (2026a), Recorded offences, Table 01. Division F (other/regulatory) excluded, see notes.

Hover to compare all five divisions in a given year. Click a legend entry to hide it, double-click to isolate it.

2. Theft and family violence order breaches drove the rise

Source: Crime Statistics Agency (2026a), Recorded offences, Table 01. Offence types under 100 per 100,000 in both years excluded.

3. Population growth does not explain it

Sources: Crime Statistics Agency (2026b), LGA recorded offences, Tables 01 and 02; Australian Bureau of Statistics (2026), Regional population.

Filter by police region or population growth. Click a point to highlight it, double-click to reset. Any council in the top right grew fast and saw its property offence rate rise faster than the population did.

4. The theft surge started in 2023

Source: Crime Statistics Agency (2026a), Recorded offences, Table 01. Colour scale capped at 50 and 200.

5. Most property crime goes unsolved

Source: Crime Statistics Agency (2026a), Recorded offences, Table 04. Recent years show more unsolved offences partly because investigations are still open.

What this means

This year’s fall in crime is real, but small. It does not undo three years of sharp growth.

That growth is mostly theft. Shoplifting and car theft rose. Violent crime per person rose much less. Family violence order breaches also rose, but that likely reflects more reporting and enforcement.

Population growth is not the answer either. Rates already account for it, and most fast-growing councils did not see property crime climb faster.

The bigger problem is what happens next. Most property offences are never solved. For victims of theft, the chance of anyone being caught is low, and it has not improved in ten years.

So when crime comes up before the election, the useful question is not whether it went up or down. It is what kind of crime, and whether police can solve it.

Notes on the data

  • Rates use CSA’s population denominators (ABS estimated resident population). Where a subgroup rate is summed to a division, this is valid because rates share the same denominator within a year.
  • Offences recorded from March 2020 to December 2021 include breaches of Chief Health Officer directions. CSA codes these under Division F (other offences), which is otherwise a small regulatory category. This produced a spike of roughly twenty times the 2017 level, so Division F is excluded from Chart 1. The restriction period is shaded so the reader can see its effect on the other divisions.
  • ABS population estimates run to June 2025, one year short of the offence data. Population growth is therefore measured over eight years and offence-rate change over nine. LGA names were matched after removing the ABS council-type suffix; unmatched councils, if any, are listed here: none.
  • Drunk and disorderly falls to zero because Victoria decriminalised public drunkenness in November 2023. It is a law change, not a fall in offending.
  • Investigation status is a snapshot at the time of extract. Recent years will show more “unsolved” simply because investigations are still open.
  • Colours for the offence divisions use the Okabe-Ito palette, chosen to stay distinguishable for readers with colour blindness.
  • Preprocessing was done entirely in R within this document: renaming columns, summing subgroup rows to division level, indexing to 2017, and grouping investigation statuses.

Acknowledgements

I acknowledge the use of Claude (Anthropic, 2026) to hwlp debug the R code. The choice of story, the data selection and the interpretation are my own.

References

Anthropic. (2026). Claude [Large language model]. https://claude.ai

Australian Bureau of Statistics. (2026). Regional population, 2024-25 financial year [Data set]. https://www.abs.gov.au/statistics/people/population/regional-population/latest-release

Crime Statistics Agency. (2026a). Data tables: Recorded offences, year ending June 2026 [Data set]. State Government of Victoria. https://www.crimestatistics.vic.gov.au/crime-statistics/latest-victorian-crime-data/download-data

Crime Statistics Agency. (2026b). Data tables: LGA recorded offences, year ending June 2026 [Data set]. State Government of Victoria. https://www.crimestatistics.vic.gov.au/crime-statistics/latest-victorian-crime-data/download-data

Crime Statistics Agency. (2026c). Explanatory notes. State Government of Victoria. https://www.crimestatistics.vic.gov.au/about-the-data/explanatory-notes

Sievert, C. (2020). Interactive web-based data visualization with R, plotly, and shiny. Chapman and Hall/CRC. https://plotly-r.com

Wickham, H. (2016). ggplot2: Elegant graphics for data analysis (2nd ed.). Springer.

Data licensed under Creative Commons Attribution 4.0 International. Files downloaded 28 September 2026.