"Innovation as a Systemic Phenomenon : Rethinking the Role of Policy" by Keith Smith. A summary.

The following is a discussion of Keith Smith’s 2000 paper on “Innovation as a Systemic Phenomenon : Rethinking the Role of Policy”.    The purpose of these LinkedIn posts is to comment on key innovation themes from academic papers, thereby introducing these concepts to a broader audience.  There are three topics in the paper 1) conceptual approaches to innovation systems 2) learning and technological knowledge and the firm level and 3) policy implications.   

The key takeaway I take from this paper is that what determines performance in a system of innovation is not knowledge creation per se as the ability of the system to provide innovators to timely access to relevant stocks of knowledge.  The role of policy makers is to appreciate the nuances of their regional innovation system and implement policies that support learning and knowledge base evolution.

Systems Approaches to Innovation

The ability of a firm cannot be understood in terms of independent decision-making by the firm, rather it must be understood as a series of complex interactions between the firm and its environment (e.g. Customers, suppliers).   These linkages are much more than market transactions and can be quasi-cooperative relationships that shape the firms technological and learning capabilities.  The environment can also influence the firms die to cultural context, institutions, local infrastructure, global knowledge base etc.

Systematic approach to innovation is important as it impacts the competitiveness amongst firms, however it is also pervasive and underpins growth at the regional level.   System of innovation theories suggest that a regions macroeconomic performance can be traced back to underlying differences in system structure.  At the national level the structural differences emerge in two ways 1) the systems of governance (I.e. Regulations for corporate behaviour and informal rules of the game for corporate behaviour) and 2) regional technological capabilities (reflected in patterns of R&D expenditure, patents, scientific publications etc).  Even in the context of increased globalization a national government will still play an important role in setting the context for economic behaviour through its policy interventions.

The innovation systems literature aim to develop an understanding of the trajectories and processes of innovation and how these processes fundamentally shape economic growth.  This approach stands in stark contrast to neoclassical (mainstream) economic theory which still does not incorporate institutional issues (including the role of government), has a limited concept of ‘uncertainty’ and does not have adequate approaches to explain the creation of technological knowledge and the technological interdependence between firms.   Generally speaking, the innovation systems literature emerged from two approaches the first was examining the evolution of specialization and the associated patterns of learning and interaction, and the second approach of examining an economy-wide patterns of corporate behaviour, policy and support (e.g. Education).  These approaches are crudely conceptualized as being narrow (including a limited number of elements) or broad (holistic).   

Regardless of the conceptualization, the process of learning and knowledge creation are central to understanding the innovation capabilities of a system.   This is a topic that is well beyond my ability to succinctly summarize but the key point related to firm level economic performance with a focus on scientific and technological knowledge (leaving aside the knowledge of finance, marketing, design etc) is that what determines performance is not knowledge creation as the ability of the system to provide innovators to timely access to relevant stocks of knowledge. There are five processes of knowledge distribution within a regional system of innovation: 

• The distribution of knowledge among universities, research institutions and industry; 
• The distribution of knowledge within a market, and between suppliers and users;
• The re-use and combination of knowledge;
• The distribution of knowledge among decentralized R&D projects; and
• Dual technological developments (especially civil and military).

It is easy to see how government policies can affect the practise of many of these organizations and therefore affect their ability to share knowledge with the innovation system.

Other academically approaches to exploring the phenomenon of innovation from the level of the firm, industry, region or national context include:

The ‘History of Technology’ approach : Conceptualizes technologies not as artefacts, but rather as integrated systems of components and supporting managerial and societal arrangements.

The ‘Science and Technology Studies’ approach : A sociological research approach that places the development and use of technology within a social framework, the difference in technological performance in societies will have their origins in the differences in the societal culture.

‘Business Organization Theories’ and ‘The Theory of the Firm’ approaches : An approach that focuses upon the evolution of the firm (particularly in terms of vertical integration and systematic integration of inter-firm functions).  Research in this area relates system differences at the firm level and differences in macroeconomic trajectories.

The ‘Regulation School’ approach : Based upon the concept of a ‘filière’ which is a specific set of infrastructures, technologies, institutions, practices and actors.  This perspective also considers technologies as integrated systems, not as individual techniques and thus technology firms are strongly interdependent upon other firms for technological knowledge.

The ‘Industrial Cluster’ approach : An analysis approach with focuses upon the performance of industrial sectors in terms of the integration of firms (often around key technologies) with an appreciation of environmental conditions.

The author argues that these systems approaches have three conceptual underpinnings:

• Economic behaviour rests upon institutional underpinnings (legal or informal ‘rules of the game’) that function to reduce uncertainty in the system.  Different institutional configurations lead to different economic behaviour and outcomes.
• Competitive advantage results from variety and specialization.   These arise from economic choices made in the past (they are path dependant) and create path dependancies going forward (they shape future choices).   Successful specializations are self replicating and a driver of regional industrial structures.
• Technological knowledge is generated by interactive learning between different types of economic agents.  An important point is that the learning is driven by interaction.

The Nature of Technical Knowledge

Since innovation is conceptualized with effective learning, then policy should concern itself with the nature of learning and the knowledge that results from learning.  This section discusses the characteristics of knowledge which imply 1) the existence of systems and 2) the need for new innovation policy approaches.

Much of the existing theoretical rationale for innovation policy have been derived from neo-classical economic ideals.  The essence of these neo-classical ideas are that firms face decisions about what to produce (based upon estimation of the highest rates of return) and how to produce it (again using a profit maximizing rationale).   Changes in the ‘choice set’ (e.g. Technologies) are assumed to be optimally and instantaneously adopted by firms.  Under these assumptions, economic efficiency relies upon economy-wide flexibility and firm level flexibility.  Under this paradigm, policies should focus upon freeing markets, removing barriers to entry (and not worrying about exit), removing barriers to change within the firm, and increasing competitive pressure thereby increasing incentives to optimize.  All in the name of allocative and technological efficiency.   However, what assumptions are made about the nature of technological knowledge?   The author argues that for the neo-classical production theory to work the following are the [questionable] attributes of knowledge:

• Knowledge is generic and can be applied widely between firms and industries
• Knowledge is codified and recorded in a fairly compete and usable form
• Knowledge is costlessly accessible and can be brought into use for production with negligible expense
• Knowledge is context independent and every firm has equal capabilities for transforming knowledge into production capabilities.

With these [questionable] assumptions in place a firm can make optimal profit maximizing choices relying upon a straightforward calculation rather than its technological capabilities and organization.  These assumptions lead to problems when it comes to the development of technology.  The first is uncertainty, knowledge outputs are not predictable from knowledge inputs and any producer must commit resources without knowing the results with any accuracy.  The second problem is appropriability, the producers of knowledge cannot appropriate the full benefits of knowledge (anyone who learns the new knowledge does so at a much lower cost) and thus creating an efficient market economy for knowledge is challenging as the producers will have poor incentivization.  Another aspect of technological knowledge is indivisibility.  Like a railroad a body of knowledge must be ‘complete’ before it can be useful for production.    So with the above (neoclassical) assumptions about knowledge The problems of risk, inappropriability and indivisibility suggest that market economies will underinvest in knowledge.  However, what kinds of knowledge have these characteristics?  Uncertainty and appropriability certainly apply to knowledge generated by fundamental scientific research, however can’t be extended to other forms of knowledge important to innovative activity.  This ‘market-failure’ approach to understanding technological knowledge leads to straightforward policy proposals, encourage discovery-orientated activities, and then protect the use of the results.  The risk and indivisibility issues lead to the public sector subsidizing the production of knowledge and the appropriability problem suggests policies of subsidies and/or intellectual property rights.  The author suggests that the main problem with this approach is understanding how to identify market failures and what is the adequate amount to address the failure.  The author also notes that this approach to innovation policy compliments the ‘linear-model’ of innovation which recent innovation theory and research are not a good guide to understanding the nature of technological approaches.

Another approach to explaining the dynamics of technological knowledge can be found with the so-called ‘new-growth theory’.   In these models the process for explaining economic growth is increasing returns to scale that arise from the production of knowledge.   The increasing returns exist because knowledge flows to multiple users without being traded.  Knowledge is partly appropriable by the firm, but also provides value to other firms through spillovers.  Thus, under this model a firms final output is not simply dependant upon its own stock of knowledge it also includes the accumulated spillovers from knowledge production in the economy.  The author argues that this model (and the other neoclassical models of growth) have very little consideration of the reality of knowledge creation as a systemic process. 

Learning and Technical Knowledge in Modern Innovation Theory

This section discusses the nature of knowledge bases that are generated by learning.  The author takes the technological knowledge base of the firm and divides it into three types of production relevant knowledge, firm-specific knowledge, sector product-field specific knowledge and generally applicable knowledge.   Against this background the author also makes three further distinctions between the form of knowledge, the object of knowledge and the institutional structure of production of knowledge.

Firm specific knowledge is highly localized and specific to product characteristics.  The character of this knowledge base is not simply technical, but includes social dimensions (e.g. How technical processes can integrate with firm routines, management processes etc.).  A key characteristic of knowledge at this level it that it is bounded, a firm may run into innovation challenges beyond their expertise and their ability to search for the needed knowledge may also be limited.   Firms have a limited knowledge-base and capabilities and their capabilities are shaped by their history and the niches that they have been able to occupy.

At the level of the product-field or industry knowledge is often a shared scientific or technological parameters and shared understanding of technical functions, performance characteristics, use of materials in products.   This part of the industrial knowledge base is public and, in principle, more or less available to any firm with the ability to utilize it.  This knowledge base thus shapes the performance of all firms in an industry and requires significant investment to be developed, maintained and disseminated.

Finally there is generally applicable knowledge, such as the general scientific knowledge base.  This type of knowledge is extremely variable and of varying relevance for industrial production.

These three types of knowledge are not separate but are intertwined and evolve over time.  The resulting knowledge structures are cumulatively built over time.   Now the discussions turn to three forms of innovation and learning 

1) The types of change that are sought : Firms engage in highly differentiated search processes for new knowledge under conditions of high technological and economic uncertainty.  The selection process of the firm narrows down the technological variety and leads to the cumulative development of technological knowledge along discernible trajectories.  The knowledge that results from searching are usually specialized around areas of areas of previous experience and often are involve substantial tacit dimensions (which are embodied in individuals).  Thus, firms innovation strategies typically rely on the knowledge portfolio that they have built over time. 

Economically relevant information can be of four kinds, specific factual information, knowledge of scientific principles, social/network knowledge, and practical skills and capabilities. 

2) The degree of change in the underlying knowledge base : Learning implies changes in the underlying knowledge base, the change may be incremental or quite radical (paradigm shifts).   A firms product is not itself an advantage, but rather is a manifestation of the firms internal reality.  These technological capabilities can also be reflected in the region itself or in a ‘technological paradigm’ (the common features of a technology shared by an industry or groups of industries).  The author also describes a ‘technological regime’ which includes both the manifestation of the technology and the processes by which it is created.  When it comes to the dynamics of the knowledge base shifts in a technological regime involve major shifts in the underlying technical knowledge and and a wide range of managerial, organizational and social changes.  Paradigm shifts are not rare, but are often much more gradual than are commonly represented. 

3) The modes of learning through which innovation occurs : Innovation is best thought of as an interactive process where engagement creates knowledge that flow’s into the firms (product and operations) development process.  Innovation is far from being a simple ’transfer’ into the firm  knowledge developed elsewhere.  The successful innovative firms are deliberately open to their environments and engage in interactive learning with a range of institutions, that is to say they increase the ‘boundedness’ of their knowledge bases.  These ability of firms to engage in interactive learning are shaped by the structure, type and scale of communication processes between relevant knowledge institutions.   It is important to note that these institutions have a tangible geographic dimension and are thus shaped by regional institutions and culture.

The author summarizes the learning processes and knowledge bases of industrial firms as having the following characteristics:

• Knowledge uses are differentiated, multi-lated and involve the systematic integration of many different types of knowledge
• A firms knowledge base is highly specific, organized around a relatively limited set of functions which the firm understands well.  This leads to a characteristic boundedness, bounded rationality and ‘bounded vision’
• The knowledge base involves significant tacit components, embodied in the skills of engineers, R&D staff, workers and managers.
• The knowledge bases are developed cumulatively, developing through time as firms build up experience with particular technologies.   This reinforces the idea that technological knowledge is path dependant.
• Knowledge bases are developed through costly processes of search, through processes of learning and adaptation.
• The knowledge bases are part of an internal to the firm system that has many components (e.g. Production, marketing)
• The knowledge bases are part of an external to the firm system that relies on interactions between firms and other agents, supported by regional institutions and infrastructure.

Policy Issues  

There are two main policy issues that the author discusses in the context of supporting innovation and the underlying learning processes.  1) What is the rationale for policy action? 2) What are the policy capabilities do the policy makers have (and what capabilities do they need to develop).

The standard rationale for innovation policy usually argues that a competitive, decentralized market system will provide sub-optimal levels of knowledge and thus leads to a case for public subsidies for knowledge creation and or intellectual property rights.

A systems approach would not disagree with these policies per se.  However, a systems approach offers policymakers a greater range of potential options to where public support may go.   Furthermore, the authors argue that a market-based approach to innovation policy may lead to areas of systemic failure which can include:

Failures in infrastructure provision and investment : Under a neo-classical model the returns on investment for knowledge infrastructure (universities, databases, government ministries, etc.) are difficult to assess due to very large scale, indivisibilities and long time horizons.  This can lead to difficulties justification of investment if standard financing assessment techniques are used.

Transition failures : Adjusting to technological change is an important consideration for firms and is challenging since firms (especially small ones) have limited technological horizons.  Also, firms will focus on what they know best (products that compliment their skills and capabilities) and thus will have relatively limited capabilities in other areas (even if they are relatively adjacent).  This can lead to situations where challenges arise that are beyond the firms technological capabilities.   There may environmental shifts in demand that lead to discontinues shifts in technology, which can be very challenging for firms to adopt to (even if the technological shifts are minor, they can cause firms serious problems).  Firms may also face major technological paradigm shifts which may present a challenge to adoption of completely new generic technologies (the required capabilities for the firm may actually be more organizational than technological).

Lock-in failures : Technologies often have a close link to their social and economic environment and can lead to technologies becoming locked-in.   This means that new alternative technologies need to overcome the existing technology AND the elements of the system that the technology is embedded in.  Both firms and societies can be locked into a technological paradigm.   To transform a technological paradigm would require deliberate changes in science, engineering practise, physical infrastructure, social organization etc.  The author presents the example of the hydrocarbon fuel technological paradigm as an example to illustrate key challenges.

Institutional failures : The integrated set of public and private institutions (e.g. Contract laws, technical standards, political culture, etc) make up the overall context of economic and technical behaviour which shapes the opportunity set available to firms as well as their capabilities.   Policy agencies are largely responsible for the evolution of theses institutions (deliberately or otherwise)  and thus assessing and adjusting institutions is key area for public action.

Policy capabilities and Knowledge bases : Policy makers need to be able to assess system specificities and be able to implement innovation policies with regional context in mind (this requires significant analytical capabilities).  They also need to be able to understand the knowledge base required by the regional innovation system and provide support for their development.

 A key policy issue is the need to identify and support ‘nodes’ in the knowledge creation and distribution system.   The author suggests that a key task is the identification of key points and functions where public support could improve the overall knowledge distribution capability.  This is certainly challenging since knowledge systems are complex and managed by multiple institutions.  The author suggests that developing information systems for policy coordination is core priority.

Conclusion  

Systems approaches offer policy makers a powerful approach for considering the core elements of economic performance in the context of their region.   Policy makers have a key role in the evolution of innovation systems, both in terms of the development of knowledge bases and the provision of physical and knowledge infrastructures upon which the technological regimes (of today and tomorrow) rest.   It is also clear that the observation and analysis of the dynamics of a regional innovation system is necessary for astute policy intervention.

Reference

Smith, K. (2000) Innovation as a Systemic Phenomenon : Rethinking the Role of Policy.  Enterprise & Innovation Management Studies., 1, 1, 73-102