The Complementarity Thesis and Place-Based Innovation: Why Technology Alone Is Never Enough

John H Howard, 10 February 2026

Innovation policy has long been preoccupied with the supply side: funding research, building laboratories, commercialising discoveries, and attracting talent. These investments matter. But they leave unanswered a persistent question: why do some places convert the same investments into measurable productivity gains while others, with comparable resources, do not?

The complementarity thesis offers an explanation. It holds that the effects of any technology depend on what it is combined with, and that the rate-limiting factors are usually the complements rather than the technology itself. The technology does not determine the result. The system in which it is embedded determines the outcome.

This Insight applies the complementarity thesis to place-based innovation, drawing on the four-domain framework developed in The Handbook of Innovation Ecosystems: Placemaking, Economics, Business, and Governance (Howard, 2025).

The argument is that successful innovation ecosystems, indicated by outcomes and results, are those that systematically provide complements across all four domains, underpinned by infrastructure investment. And it extends the governance domain beyond its conventional structural emphasis to examine the human dimensions: leadership, collaboration, and the growth mindsets that enable ecosystems to learn and adapt.

The Complementarity Thesis: Origins and Application

The concept of complementarity has deep foundations in economics. Milgrom and Roberts (1990, 1995) formalised the idea that activities within firms are complementary when doing more of one increases the returns from doing more of another. Their work on modern manufacturing demonstrated that technology investments pay off only when accompanied by changes in human resource practices, decision systems, product development processes, and organisational routines.

Furthermore, piecemeal adoption of complementary practices may actually worsen performance. In the 1980s, General Motors spent an estimated US$80 billion on robotics and capital equipment, without making corresponding changes to its human resources policies or manufacturing procedures. The investments disappointed because the complements were absent.

Brynjolfsson and Milgrom (2012) extended this analysis to information technology more broadly, showing that firms combining IT investment with organisational redesign, workforce training, and new business processes achieved substantially higher returns than those investing in technology alone. The same pattern appears in contemporary AI adoption. Firms that purchase AI software without investing in complementary technologies, data infrastructure, management capability, and workforce skills see far smaller gains, if any.

Robert Solow observed in 1987 that computers could be seen everywhere except in the productivity statistics. That paradox eventually resolved, but only after roughly fifteen years of complementary investment in software, training, and business process redesign. The current lag between AI capability and measured productivity may reflect a similar dynamic: the technology is available, but the complements have not yet matured.

From Firms to Places: Complementarity at Ecosystem Level

The complementarity thesis was originally developed at firm level. Extending it to places requires an additional analytical step. Individual firms need complements they often cannot build alone. Data infrastructure, skilled labour markets, regulatory clarity, shared research facilities, and business networks are collective goods that operate at ecosystem level. Places that provide these shared complements are able to lower the threshold for technology adoption across many firms simultaneously.

This is why innovation districts, precincts, and hubs matter for national productivity. They are the sites where complementary assets combine in observable ways. A well-functioning district does not merely co-locate organisations. It creates the conditions under which firms, universities, government agencies, and intermediaries can access the complements they need to convert new technologies into productive use.

The analytical challenge is to specify what those place-level complements are and how they interact. The four-domain framework provides a structure for doing so.

The Four-Domain Framework

The Handbook of Innovation Ecosystems identifies four interdependent domains that shape ecosystem performance: placemaking, economics, business, and governance. The integration of these domains is a sufficient condition for ecosystem success. But their effective operation depends on a necessary condition: sustained investment in supporting infrastructure, including physical, knowledge, financial, and digital assets.

Placemaking

Placemaking addresses the spatial, physical, and cultural conditions that attract talent and enable interaction. It encompasses urban design, mixed-use development, public realm quality, and the deliberate creation of environments that encourage serendipitous encounter. Innovation requires both planned collaboration and unexpected connections. Smart precinct design maximises opportunities for cross-pollination between organisations and disciplines without forcing artificial interactions.

The shift from traditional science park models to contemporary innovation districts reflects a move from efficiency-driven design to what might be called productive proximity. Leading precincts integrate housing, amenities, transport, and workspace in ways that compress the distance between the people and organisations whose interaction generates knowledge spillovers.

Placemaking serves as a complement to the other three domains. Without quality places, economic activity fragments, business formation slows, and governance loses its spatial anchor. The physical environment is the platform on which the other domains operate.

Economics

The economics domain concerns economic and industry development, job creation, market dynamics, and the economic logic of agglomeration. Innovation ecosystems generate benefits that individual actors cannot fully capture: knowledge spillovers, talent mobility, and shared infrastructure create positive externalities that benefit the entire system.

Economic thinking within innovation ecosystems has moved from linear models, where R&D inputs generate outputs that create growth, to a systems perspective that recognises increasing returns to scale, path dependence, and the role of social capital in reducing transaction costs. Marshall's insights about industrial districts, first articulated in the 1890s, have gained renewed relevance as policymakers seek to understand why some places develop self-reinforcing cycles of innovation while others stagnate.

The economics domain provides the market signals and resource flows that sustain ecosystem activity over time. It determines whether the value created through innovation is captured locally or dissipates, and whether the gains are distributed in ways that maintain social licence for continued investment.

Business

The business domain addresses commercial logic, investment flows, entrepreneurship, and the translation of research into marketable goods and services. It includes venture capital, corporate R&D, startup formation, and the intermediary organisations that connect researchers with commercial partners.

One of the most persistent challenges in innovation policy is what has been described as the management chasm: the gap between research capability and the commercial and operational skills required to bring discoveries to market. Promising ventures continue to fail despite increased financial support because the missing complement is often managerial rather than financial. One of the most effective models pairs lead researchers with experienced general managers who bring financial modelling, operational capability, and market knowledge, building on a foundation of trust developed over extended collaboration.

The business domain functions as a complement to placemaking (which provides the physical environment for commercial activity), economics (which provides the market context), and governance (which provides the institutional framework). Weakness in any one domain constrains the performance of the others.

Governance

Governance is the domain most frequently underestimated in innovation policy, and the one where the complementarity thesis has the sharpest implications. If ecosystem performance depends on the alignment of multiple domains, then someone or something must orchestrate that alignment. Markets will not do it. Governance provides the coordinating mechanism.

The governance challenge in innovation ecosystems is distinctive. Innovation districts typically involve multiple landholders with different missions and performance indicators, several layers of government with potentially conflicting policy objectives, and diverse institutional actors including universities, hospitals, cultural organisations, and private developers. The number and diversity of stakeholders make governance inherently complex.

Conventional treatments of governance tend to focus on structural arrangements: boards, committees, partnership agreements, and regulatory frameworks. These structural elements matter, but they are insufficient. The complementarity thesis suggests that governance, like technology, depends on what it is combined with. Governance structures without capable people, collaborative relationships, and adaptive mindsets will underperform, regardless of how well the organisational charts are drawn.

The Human Dimensions of Governance

Extending the governance domain to encompass its human dimensions adds three elements to the conventional structural analysis: leadership, collaboration, and growth mindsets. Each serves as a complement to the others and to the structural arrangements within which they operate.

Leadership

Innovation ecosystems require a distinctive form of leadership. Unlike hierarchical organisations where authority flows from position, ecosystems are networks of autonomous actors who cannot be directed by any single institution. Leadership in this context is less about command and more about convening, brokering, and sustaining shared purpose across institutional boundaries.

The most effective ecosystem leaders are what the Handbook describes as system orchestrators and integrators: individuals and organisations that connect actors, align resources, and maintain strategic coherence over time. They operate at the boundaries between institutions, translating between the different languages and incentive structures of universities, government agencies, and businesses.

This form of leadership is scarce. It requires deep knowledge of multiple sectors, credibility across institutional boundaries, and patience measured in decades rather than electoral or budget cycles. Many Australian innovation precincts suffer from rapid leadership turnover and a revolving door of governance arrangements, which undermines the trust and institutional memory on which effective ecosystem leadership depends.

Collaboration

Collaboration in innovation ecosystems differs from collaboration within organisations. It occurs between entities with different missions, incentive structures, and time horizons. Universities prioritise publications and teaching quality. Governments respond to electoral cycles and ministerial priorities. Businesses pursue commercial returns within competitive markets. Aligning these divergent logics is the central challenge of ecosystem collaboration.

Effective collaboration rests on trust, and trust develops through sustained interaction rather than formal agreements. The pairing model observed in research commercialisation, where academic founders and experienced general managers build working relationships over extended periods, illustrates a broader principle. Collaboration across institutional boundaries requires investment in relationship infrastructure: neutral convening spaces, boundary-spanning roles, and long-term commitment to joint endeavour.

The complementarity thesis explains why collaboration cannot be mandated. Collaboration is itself a complement to the structural governance arrangements, the leadership capabilities, and the shared understanding that enables diverse actors to work together productively. Absent these complements, collaboration mandates produce meetings without outcomes.

Growth Mindsets

Carol Dweck's research distinguishes between fixed mindsets, which treat capabilities as innate and unchangeable, and growth mindsets, which see potential as expandable through effort, learning, and persistence. Applied to innovation ecosystems, this distinction has practical implications for how governance bodies, institutions, and entire districts approach their development.

An ecosystem exhibiting a fixed mindset tends to defend existing positions, avoid risks, and treat setbacks as confirmation that change is unwarranted. Decision-makers in such environments are more likely to replicate familiar models, resist experimentation, and view success as a function of inherent advantages rather than adaptive effort. The result is institutional inertia that prevents the ecosystem from responding to new opportunities or technologies.

By contrast, an ecosystem with a growth mindset embraces challenges as learning opportunities, invests in building new capabilities, and treats failure as information rather than as a verdict. Leaders in growth-oriented ecosystems seek feedback, invest in capability development, and are willing to revise strategies when evidence warrants.

The growth mindset concept connects to the complementarity thesis in a specific way. Technology adoption requires organisations and places to develop new capabilities: data management, AI literacy, redesigned business processes, and updated governance arrangements. A fixed mindset treats these requirements as barriers. A growth mindset treats them as the necessary complementary investments that will determine whether technology investment pays off.

There is a further dimension. Developmental psychology, which describes how individuals progress through stages of meaning-making complexity, may offer a useful lens for understanding place-based innovation cultures. The framework could be extended to explore the extent to which cities and regions exhibit collective mindset characteristics that shape their approach to innovation.

A city whose civic culture prioritises consensus and deference to established frameworks may approach innovation differently from one that pursues autonomous, vision-driven strategies. This parallel suggests that the developmental maturity of a place's institutional culture could influence its capacity to absorb and deploy new technologies such as AI.

These collective dispositions could further inform how ecosystems balance the four domains and whether governance arrangements are genuinely adaptive or merely performative.

Infrastructure: The Necessary Condition

The four-domain framework identifies the integration of placemaking, economics, business, and governance as a sufficient condition for ecosystem success. But a necessary condition sits beneath all four domains: sustained investment in supporting infrastructure.

Infrastructure in this context extends well beyond its conventional definition. It encompasses physical infrastructure (transport, buildings, public spaces), knowledge infrastructure (universities, research facilities, data centres), financial infrastructure (venture capital, development finance, patient capital instruments), digital infrastructure (broadband, undersea cables, cloud computing platforms), and social infrastructure (networks, intermediaries, trust relationships).

What is not always recognised is that much of this investment is invisible to standard innovation metrics. R&D expenditure captures annual spending on research. It does not capture the accumulated stock of infrastructure capital (or its adaptive reuse) that enables research to be productive. A city with reliable metro systems, high-speed fibre networks, quality public spaces, and deep social networks has built up innovation capital that does not appear in national innovation statistics.

This measurement gap has policy consequences. It understates the total investment in innovation capacity and distorts comparisons between places. More importantly, it directs policy attention toward R&D funding (which is measured) and away from infrastructure investment (which largely is not), even though the latter may be the binding constraint on ecosystem performance.

Implications for Policy and Practice

The complementarity thesis, applied through the four-domain framework, yields several implications for innovation policy and practice.

• Technology-focused policies will disappoint if they neglect the complements. Funding research, subsidising AI adoption, or building new laboratories will produce smaller returns than expected unless accompanied by investment in management capability, workforce skills, governance capacity, and supporting infrastructure. This is the lesson of General Motors in the 1980s, the Solow paradox of the 1990s, and the current AI productivity lag.

• Place-based policy has a distinctive role. Many complements are collective goods that operate at ecosystem level. Skilled labour markets, shared research facilities, regulatory frameworks, and trust networks cannot be built by individual firms. Innovation districts, precincts, and hubs are the organisational forms through which these collective complements are assembled and maintained.

• Governance requires investment in people as well as structures. The human dimensions of governance, including leadership, collaboration, and growth mindsets, are themselves complements to the formal arrangements. Appointing boards and signing partnership agreements achieves little without the capable individuals, trusted relationships, and adaptive dispositions that make governance effective. This means investing in long-tenure leadership, relationship infrastructure, and cultures of learning within governance bodies.

• Infrastructure investment should be recognised as innovation investment. The conventional separation between infrastructure policy and innovation policy is analytically unjustified and practically harmful. Transport, digital networks, public spaces, and social infrastructure are enabling conditions for innovation. Including them in the innovation investment calculus would provide a more accurate picture of national innovation effort and direct policy attention toward binding constraints.

• The nested geography of innovation matters. The conditions that enable technology adoption, such as workforce skills, institutional quality, and broadband infrastructure, largely operate at metropolitan or regional scales. The combinations that produce distinctive outcomes, including anchor institution strategies, precinct governance, and cross-sector collaboration, manifest at district level. Effective policy operates across both scales simultaneously.

Conclusion

The complementarity thesis reframes a familiar policy question. Rather than asking how to develop better technology, it asks what technology must be combined with to produce value. The answer, consistently, is a set of complements that spans management capability, workforce skills, governance capacity, infrastructure, and the human qualities of leadership, collaboration, and adaptive learning.

Innovation districts, precincts, and hubs are the places where these complements converge. Their performance depends on how well they align placemaking, economics, business, and governance, and on whether the underlying infrastructure investment is sufficient to support productive activity across all four domains.

Technology alone is never enough. It is the combinations that count.

---

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.

References

Brynjolfsson, E., & Milgrom, P. (2012). Complementarity in organizations. In R. Gibbons & J. Roberts (Eds.), The handbook of organizational economics (pp. 11–55). Princeton University Press.

Brynjolfsson, E., Rock, D., & Syverson, C. (2021). The productivity J-curve: How intangibles complement general purpose technologies. American Economic Journal: Macroeconomics, 13(1), 333–372.

Dweck, C. S. (2006). Mindset: The new psychology of success. Random House.

Dweck, C. S., & Yeager, D. S. (2019). Mindsets: A view from two eras. Perspectives on Psychological Science, 14(3), 481–496.

Howard, J. H. (2025). The handbook of innovation ecosystems: Placemaking, economics, business, and governance. Acton Institute for Policy Research and Innovation.

Jones, C. I. (2024). The AI dilemma: Growth versus existential risk (NBER Working Paper No. 32763). National Bureau of Economic Research.

Kegan, R. (1994). In over our heads: The mental demands of modern life. Harvard University Press.

Marshall, A. (1890). Principles of economics. Macmillan.

Milgrom, P., & Roberts, J. (1990). The economics of modern manufacturing: Technology, strategy, and organization. American Economic Review, 80(3), 511–528.

Milgrom, P., & Roberts, J. (1995). Complementarities and fit: Strategy, structure, and organizational change in manufacturing. Journal of Accounting and Economics, 19(2–3), 179–208.

Mohnen, P., & Röller, L.-H. (2005). Complementarities in innovation policy. European Economic Review, 49(6), 1431–1450.

Solow, R. M. (1987, July 12). We’d better watch out. New York Times Book Review, 36.

About the Author

John H. Howard is the Executive Director of the Acton Institute for Policy Research and Innovation. He has more than four decades of cross-sector experience spanning universities, business, and government, including management consulting partnerships at Ernst & Young and Coopers & Lybrand, and service as Pro Vice-Chancellor at the University of Canberra. He is currently managing a UTS/Google.org research project examining how innovation ecosystem quality explains AI productivity outcomes across global innovation districts.