Research on Innovation Ecosystems: Getting the Unit of Analysis Right

John H Howard, 27 January 2026

Innovation policy increasingly targets specific places: precincts, districts, clusters, corridors and hubs to achieve economic, business, and socio-cultural outcomes through investments in physical, knowledge, financial and social capital.

At the same time, much of the evidence used to justify these investments relies on metropolitan or regional analysis and indicators. This mismatch creates a recurring policy risk. Decisions taken at precinct scale are often informed by analysis conducted at metropolitan scale, perhaps obscuring the mechanisms that determine whether place‑based investments deliver productivity and innovation outcomes.

This Innovation Insight argues that innovation ecosystems cannot be understood through a single spatial lens. Metropolitan regions and innovation districts perform different functions. Innovation outcomes emerge from their interaction. Treating them as separate units of analysis can mislead policy. Treating them as nested and complementary improves both explanation and intervention.

The policy problem

Governments across OECD economies have turned to place‑based innovation strategies in response to weak productivity growth, uneven urban and regional development, and the diffusion challenges of general‑purpose technologies such as information technology, biotechnology, and artificial intelligence. These strategies typically involve concentrated investment in designated locations, supported by planning, infrastructure, and governance reform.

However, evaluation frameworks often rely on available urban and region level indicators such as demographics, education, incomes, infrastructure investment and assets, employment growth, risk capital, and new business creation. These indicators describe enabling conditions, but they say little about how innovation actually occurs within specific sites. The result is a persistent gap between where policy acts and how performance is measured.

Metropolitan regions as enabling systems

Metropolitan regions remain a necessary unit of analysis. They approximate functional economic systems in which labour markets, transport networks, housing supply, and institutional arrangements operate across municipal boundaries. Knowledge diffusion through labour mobility, professional networks, and supplier relationships largely occurs within metropolitan catchments. This is the logic underpinning Edward Glaeser’s The Triumph of the City: cities as dense, metropolitan-scale systems of interaction that accelerate learning, matching, and innovation.

From an analytical perspective, metropolitan regions offer practical advantages. Data availability is stronger, indicators are more standardised, and international comparison is feasible. These characteristics explain why metropolitan regions have become the default unit in much comparative innovation research, and urban studies.

Crucially, metropolitan regions establish enabling conditions. Skills depth, research capacity, capital availability, digital infrastructure, and regulatory stability is difficult to generate efficiently or effectively at precinct scale. A region either possesses these foundations or it does not. Precinct‑level effort cannot substitute for their absence.

What metropolitan analysis cannot see

Despite these strengths, metropolitan analysis systematically masks variation within regions. Innovation outcomes differ markedly across districts that share the same labour market, regulatory regime, and infrastructure base. Aggregated indicators cannot explain why some precincts thrive while others stagnate within the same city.

The mechanisms that drive this divergence operate at finer spatial scales. Face‑to‑face interaction, informal learning, and observational imitation depend on proximity, and even serendipitous encounters. These processes typically occur within walkable distances rather than across metropolitan space. Governance arrangements, anchor institution strategies, and place‑specific coordination also function at precinct scale.

For policy, this blind spot matters. Most place‑based interventions target specific locations. If analysis does not engage at the scale of intervention, it cannot explain success or failure.

Districts and precincts as combinatorial spaces

Innovation districts and precincts function as combinatorial spaces. They bring together people, organisations, and assets in ways that enable recombination of knowledge and capability. Translation between research and application, cross‑sector collaboration, demonstration effects, and talent circulation occur in particular places rather than across abstract regions.

Design choices matter at this scale. Built form, land‑use mix, governance structures, and programming shape interaction patterns. These factors are resistant to aggregation and are largely invisible in metropolitan statistics.

Table 1. Metropolitan regions and districts as units of analysis

Dimension	Metropolitan region	District or precinct
Primary function	Establishes enabling conditions	Enables interaction and recombination
Spatial logic	Labour markets and infrastructure catchments	Walkable proximity and shared spaces
Typical scale	City‑region	1–2 km radius
Dominant mechanisms	Skills supply, capital access, institutions	Collaboration, translation, demonstration
Data characteristics	Standardised, comparable	Contextual, often qualitative
Policy focus	System capacity	System performance

Why single‑scale analysis fails

Selecting a single unit of analysis forecloses key questions. As mentioned, metropolitan analysis cannot explain why particular districts outperform others within the same region. District analysis cannot explain why similar precinct models succeed in some cities and fail in others. Each scale answers different questions.

This limitation becomes acute in comparative research. Without metropolitan context, district case studies risk becoming anecdotal. But without district analysis, metropolitan comparisons risk attributing outcomes to factors that never directly influence innovation behaviour.

A nested analytical framework

A more effective approach treats metropolitan districts, precincts and clusters as nested components of a single system. Metropolitan regions define the possibility space for innovation. Districts determine which possibilities are realised. Outcomes emerge from their interaction.

This framing clarifies analytical responsibility. Metropolitan analysis identifies whether foundational conditions exist. District analysis explains how those conditions are mobilised. Neither is sufficient on its own.

Table 2. What each scale can explain

Question	Metropolitan scale	District scale
Cross‑city performance differences	Strong	Weak
Within‑city variation	Weak	Strong
Readiness for AI adoption	Strong	Partial
Realisation of productivity gains	Limited	Strong
Returns to place‑based investment	Indirect	Direct

Implications for AI and productivity research

This nested perspective is particularly relevant for research on AI and productivity. Skills availability, digital infrastructure, and institutional readiness vary across metropolitan regions and shape the feasible frontier of adoption. These are metropolitan attributes.

Productivity gains, however, depend on how organisations combine AI with complementary assets such as domain expertise, workflow redesign, governance arrangements, and inter‑organisational learning. These combinations occur locally, within districts and precincts. Ignoring either scale misrepresents the drivers of performance.

Policy leverage by scale

Different actors possess leverage at different spatial levels. Confusion arises when responsibilities are blurred or when precinct interventions are expected to compensate for metropolitan deficits. Table 3 endeavours to provide some clarity.

Table 3. Policy roles across scales

Actor	Metropolitan role	District role
National government	Skills, research funding, digital infrastructure	Limited
State government	Transport, universities, system coordination	Precinct designation, anchor investment
Local government	Marginal	Planning, public realm, coordination
Universities	Workforce and research capacity	Anchor presence and engagement
Firms	Regional location choice	Site‑specific collaboration

Governance structures and why the Australian case differs

The nested framework operates differently depending on how metropolitan governance is organised. In North America and much of Europe, metropolitan regions are not only analytical constructs but political or administrative realities. Formal metropolitan governments or planning authorities provide a platform for coordination across municipalities, aligning land use, transport, infrastructure, and economic development. They also tend to collect useful administrative data.

In the United States, metropolitan coordination often occurs through counties or metropolitan planning organisations. These bodies do not eliminate local autonomy, but they provide a forum for strategic alignment. Innovation districts therefore operate within a broader metropolitan policy envelope rather than compensating for its absence.

Canadian cities provide an even clearer contrast. Large, consolidated municipalities and recognised metropolitan regions allow city-wide strategies to coexist with district-level initiatives. This reduces duplication, moderates zero-sum competition between local governments, and supports coherent investment in skills, transport, and innovation infrastructure.

German metropolitan regions operate through formalised regional planning associations embedded in federal systems. State governments play a strong role, but they do so through metropolitan institutions rather than bypassing them. This allows differentiated district development within a shared regional strategy.

The United Kingdom occupies an intermediate position. London’s mayoral system and the Greater London Authority provide strategic oversight across boroughs, while preserving local planning control. Innovation precincts therefore sit within a negotiated but explicit metropolitan framework.

Australian cities lack an equivalent layer.

Metropolitan regions such as Sydney and Melbourne comprise dozens of local governments with no metropolitan authority. State governments retain control over strategic planning, major infrastructure, transport, and urban development, often through central departments or public corporations. In political terms, strong metropolitan governments have been viewed as competitors to state authority and have not been sustained. In Sydney multiple forms of Development Corporations have been tried, and discontinued.

A current exception is the Bradfield Development Authority, responsible for delivering the new Bradfield City Centre in Western Sydney. It is also mandated to attract investment, drive industry growth (especially in advanced manufacturing), and develop a 24/7 city adjacent to the new Western Sydney International Airport. 

Currently metropolitan Sydney might include at least eight major innovation districts, precincts or hubs that are recognised as distinct spatial or functional nodes in the innovation ecosystem, with additional smaller co-located activity clusters. This inventory will evolve as precinct plans and ecosystem strategies mature. A critical issue concerns how well these places complement each other through specialisations rather than building duplication and overlaps. In almost all multi-nodal metropolitan innovation ecosystems, rapid public transport is a key enabler for building collaboration and connectivity.

At the same time, Australian local governments retain responsibility for a wide range of social, cultural, public health, and environmental services. Many have developed capable economic development functions, including business engagement, cultural activation, and place promotion. However, these capabilities can operate in competition rather than coordination across metropolitan areas, reflecting the absence of a shared metropolitan mandate similar to the innovation imperative outlined above.

The absence of strong metropolitan governance in Australian cities has two implications for innovation ecosystems.

• Metropolitan enabling conditions tend to be managed vertically by state governments rather than horizontally through metropolitan institutions.

• Paradoxically, district and precinct governance arrangements are expected to integrate economic, spatial, and institutional functions that elsewhere are distributed across metropolitan systems.

The consequence is a heavier policy burden placed on precincts. Expectations of impact may exceed what district-level mechanisms can deliver when metropolitan foundations are weak or misaligned. The nested framework remains valid, but in Australia the interaction between scales is mediated primarily through the state rather than through metropolitan governance. This increases the importance of clarity about the policy instruments available to each layer of governance, what they are intended to influence, and how.

A major exception to this narrative is Canberra, which is a city-state with an integrated state and local level governance. It is characterised by a network of districts and precincts across the city, with the Canberra Innovation Network (CBRIN) taking a strong leadership role.

Learning from comparative cases

Comparative cases illustrate the value of this framework. Metropolitan Munich provides strong enabling conditions, but distinct districts have developed different innovation trajectories based on anchor institutions and local coordination. Berlin illustrates how district‑level effort can partially compensate for weaker metropolitan conditions, albeit with sustained public support.

Australian precincts show a similar pattern. State‑led designation and investment can accelerate concentration effects when metropolitan foundations exist. Where they do not, expectations should be tempered.

Why Australian precincts carry more policy weight than their overseas counterparts

In many OECD countries, innovation precincts operate within a mature metropolitan governance framework. Strategic planning, transport coordination, skills pipelines, and In In many OECD countries, innovation precincts operate within a mature metropolitan governance framework. Strategic planning, transport coordination, skills pipelines, and regional economic development are handled at metropolitan or city-wide scale. Precincts focus on local interaction, translation, and place-specific coordination.

In Australia, the absence of metropolitan government shifts this balance. State governments manage metropolitan-scale functions through central departments and public corporations, while local governments retain responsibility for social, cultural, environmental, and community services. Economic development capability exists at local level, but it is fragmented and competitive across metropolitan areas.

As a result, innovation precincts are asked to perform integrative roles that elsewhere sit above them. They become sites where economic development, spatial planning, institutional coordination, and policy experimentation are bundled together. This inflates expectations of what precinct-level governance can achieve and increases delivery risk when metropolitan enabling conditions are weak or misaligned.

Understanding this structural difference is essential when importing international models. Precinct success in Australia depends as much on how state-level systems support metropolitan coordination as on local design and governance quality. 

Implications for evaluation and research design

Evaluation frameworks that rely on single‑scale indicators risk false positives and false negatives. Strong metropolitan performance may mask underperforming districts. Successful precincts may conceal regional weaknesses. Robust evaluation requires indicators aligned to both scales and explicit assessment of their interaction.

For researchers, this implies mixed‑method designs that link metropolitan datasets with district‑level qualitative and organisational analysis. For policymakers, it implies discipline in aligning instruments with the scale at which mechanisms operate.

Conclusion

Getting the unit of analysis right is not a technical detail. It shapes how innovation systems are understood, how policy is designed, and how success is judged. Treating metropolitan regions and districts as nested and complementary provides a clearer account of how innovation actually occurs. It also offers a more reliable foundation for place‑based policy in an era of constrained public resources and heightened expectations of impact.

---

Dr John H Howard is the Founder of the Acton Institute for Policy Research and Innovation. He is a management consultant, public policy adviser and economist with more than four decades of experience in innovation policy, governance, and digital transformation across government, universities, and industry.

This Insight draws on the research and analysis for the Handbook of Innovation Ecosystems: Placemaking. Economics. Business. Governance, published by the Acton Institute for Policy Research and Innovation, published in October 2025. The Handbook is available in paperback from Amazon Publishing and as an ebook from Kindle

For further information, advice and guidance about this Insight, contact john@actoninstitute.au.

© 2026 John H Howard.All rights reserved. No part of this publication may be reproduced, stored, or transmitted in any form or by any means without prior written permission of the author, except for brief quotations used for review or scholarly purposes.