Local Government Productivity

The Challenge: Benchmarking Diverse Councils

Local governments across Victoria vary widely in size, service mix, and financial base. The government needed an efficiency measurement framework to guide policy on rate setting and resource allocation. The analysis needed to capture the full scale of activities-from waste collection to road maintenance-and fairly compare vastly different council types, from metropolitan hubs to small rural shires, despite potentially sparse or inconsistent data.

The goal was to provide efficiency scores that were credible, fair, and could withstand public and legal scrutiny, forming a sound basis for government policy.

Our FDE Approach: Data Envelopment Analysis

Working with the government, our Forward Deployed Engineers (FDEs) applied advanced Data Envelopment Analysis (DEA) modelling. We built a comprehensive framework by exploring multiple modelling avenues to ensure the final results were stable and defensible.

• Total Factor Productivity Models: Our primary model was built from inputs like staff, capital, and operational expenditure, and outputs such as the number of households, businesses, road length, and waste collected.
• Single and Multiple Group Analysis: Councils were first assessed as one group to establish a sector-wide benchmark, then segmented into five distinct peer groups (Metropolitan, Regional Centre, Interface, Large Rural, and Small Rural) for fairer, like-for-like comparisons.
• Alternative Model Testing: We developed and tested several model variants-some excluding depreciation, others incorporating community satisfaction data-to examine how different assumptions affected the outcomes and to ensure the core findings were robust.
• Scenario Analysis: We compared alternative specifications and frontier assumptions to pressure-test the model, ensuring our final recommendations were stable and defensible from every angle.

Making the Decision-Maker's Life Easier

Our framework was designed not just for analytical purity, but for practical application. It translates vast, complex datasets into a clear, actionable tool that empowers government leaders to make informed and confident policy decisions.

• From Data Chaos to a Clear Narrative: The AI model distils millions of data points from 79 unique councils into a single, intuitive efficiency score. This allows leaders to quickly grasp the performance landscape without getting lost in spreadsheets.
• Politically and Legally Defensible Evidence: When setting rate caps or allocating resources, decisions face intense scrutiny. Our multi-model, peer-reviewed approach provides a robust, evidence-based foundation that can be confidently defended in public forums and legal challenges.
• Shifting the Conversation to Solutions: By establishing a fair and objective benchmark, the framework moves the debate away from arguments over data and towards collaborative problem-solving. The focus becomes "how can we improve?" rather than "is the comparison fair?".
• Simulating Policy Before Implementation: The model allows leaders to test different scenarios and assumptions, providing a clear forecast of how policy changes might impact various council groups. This reduces risk and leads to better-designed, more effective policy.

The Outcome: A Defensible Benchmarking Tool

Our analysis revealed that most councils had mean technical efficiencies between 70% and 83%, with Interface and Metropolitan councils tending towards higher efficiency. By confirming the robustness of these findings across multiple models, we provided the government with a transparent, quantitative basis for determining efficiency factors.

• Policy Relevance: Provided government with a transparent, quantitative basis for determining efficiency factors under rate-setting frameworks.
• Flexibility: The framework was designed to be easily recalibrated with new inputs, outputs, or service-specific data as government priorities evolve.
• Credibility: Using a peer-based DEA benchmarking approach created a fair and defensible efficiency measure for all 79 Victorian councils.