Chapter 2 : Systems of Innovation Literature Review

The following is the latest draft of Chapter 2 from my PhD dissertation on innovation agencies.  The goal of this chapter is to discuss the relevant literature to provide an regional and systemic  perspective. Comments and feedback are welcome.

 

The Oslo Manual (2005) defines innovation as “the implementation of a new or significantly improved product (good or service), or process, a new marketing method, or a new organizational method in business practices, workplace organization or external relations”. This dissertation takes a broad view of innovation as a phenomenon that can result in improved socioeconomic performance and can be expressed in a range of mediums (i.e. technology, governance, institutions, organizational form, household activity, etc.).

The essence of innovation is that people (individually and collectively) have the ability to recognize and propose new and valuable combinations of socioeconomic elements.   Additionally, in this perspective there is a protagonist role for individuals and organizations (Malerba, 2002).

This chapter will begin with a discussion of some of the early economic thinkers who developed approaches to conceptualize the economy using the concepts of technical change, learning and entrepreneurship.

1.1      Background

1.1.1     Schumpeter

Joseph Schumpeter developed a perspective suggesting that capitalist development was the result of the long run co-evolution of elements in the economy. Innovation is the result when the elements are recombined in qualitatively different ways that add positive socioeconomic value (Schumpeter, 1942).  

Joseph Schumpeter emphasized the co-evolution of economic elements and the importance of integrating theoretical work with historical analysis (Fagerberg, 2003). Schumpeter was inspired by many economists with views as diverse as Karl Marx (who believed that technological competition between organizations drove capitalist economic evolution) and Léon Walras (who begat the 'equilibrium' principle at the core of neoclassical economic analysis).

Schumpeter appreciated Marx’s perspective that technological change is the main source of industrial dynamism, suggesting that new markets, new sources of supply and new organizational structures are all methods by which industrial organizations can compete in a capitalist economy (Fagerberg, 2003; Hospers, 2005). Schumpeter’s view was that qualitative change (aka creative destruction) was a crucial force that arose from individual entrepreneurs (later, he included organizations as a source of this entrepreneurial function). His context rich approach to economic assessment differed from neoclassical models that underemphasized the evolutionary elements to economic growth. A neoclassical economic picture is what an economy would look like without dynamic forces like entrepreneurship causing disequilibria (Fagerberg, 2003). Schumpeter believed that some neoclassical models were useful as they described how the economy behaved (i.e. returning to a state of equilibrium). Schumpeter believed that competing (industrial) organizations would imitate and adapt to innovation thus creating new investment that would allow the economy to grow; this attempt to catch up would erode the advantage of the original innovator. (Interestingly, Schumpeter didn’t have much to say in regards to why certain firms would be better able to innovate that others (Magnusson,1994)).

Schumpeter applied this concept of imitation to government policy when one jurisdiction would attempt to apply policy that had been successful in another region. Schumpeter believed that this approach of copycat policy design needs to incorporate regional context in order to be successful (Hospers, 2005).

1.1.2     Neoclassical economics

After World War II, economic science embraced a neoclassical economic paradigm (Nelson and Winter, 1973; Nelson and Winter, 2002). Neoclassical economics is characterized by analyzing an instant of economic time (i.e. the subject of analysis is static) with a relatively low amount of contextual analysis. For example, technological change is widely accepted in neoclassical thought as a driver of economic growth but is treated as external (i.e. exogenous) to the neoclassical models (Abernathy and Clark, 1985; Freeman, 1994). Neoclassical economics is also characterized by economic actors all having 'friction free' and immediate access to knowledge (Stoneman, 2002).  

When it comes to appreciating the role of technology, neoclassical models require that the rate of technological growth increase deliberately in order for per capita growth to occur. However, the models do not include technological growth in their design which results in technical change being exogenous to the model. Technological change might be unexplainable, or explainable after the fact, however, the neoclassical model does not have much useful insight on the process that can be incorporated into a model of aggregate growth (Solow, 1994). Neoclassical models admit that the rate of per capita growth requires a deliberate investment in technology. This technological dynamic is exogenous to the neoclassical model and addressing this exclusion is extremely challenging due in part to the limitations of modeling knowledge development (e.g. serendipitous and unexpected research insights, R&D knowledge being only one type of innovation affecting productivity, etc.) (Solow, 1994).  

1.1.3     Neo-Schumpeterian and evolutionary economics

Schumpeter’s contributions; insights on entrepreneurship and innovation and the process of economic growth, provided a different conceptual starting point for economic analysis. Economists who embrace this model emphasize the dynamic ways that socio-economic systems qualitatively evolve in an open-ended process (Fagerberg 2003; Hospers, 2005). The central role of technical change in economic growth also became widespread in these economic models. There were now a range of economic approaches incorporating technological competition and evolutionary dynamics.

 Widespread interest in evolutionary economics had faded after Schumpeter’s death in 1950. Then, gradually, economists began to propose economic models that incorporated innovation as a factor explaining differences in international trade that could not be accounted for by contemporary equilibrium models (Fagerberg, 2003). In the 1980’s, academics such as Freeman, Dosi, Pavitt, and Fagerberg proposed models that had innovation as a primary driver of economic growth.

1.2      Learning and skill development

Learning and skill development impact the performance of an economy because ‘intangible’ resources (i.e. knowledge, routines, and skills) are not evenly distributed throughout a socioeconomic system (i.e. individuals, organizations and regions) (Nelson and Winter, 1982; Lundvall, 1998; Nelson and Winter 2002; Lundvall, 2007).

The core concepts of learning and skill development are particularly relevant for innovation agencies for two reasons. First, these concepts suggest that innovation agencies, like other organizations, learn and grow their capabilities within their organizational boundaries. Second, they suggest that the backers of innovation agencies in the government can learn to launch new innovation agencies more effectively.

Should innovation agencies focus on improving the overall balance of targeted skills in the region, they may improve the depth of skill by supporting education programs and/or focusing upon attracting people to the region. The next chapter will discuss how an innovation agency uses the aggregate skills of its people to design and implement instruments to affect the behaviour of universities and industry via investments in initiatives led by scientists, engineers and doctors.

A skill is the simplest piece in the socioeconomic framework of this dissertation. A skill is a capability for a smooth sequence of coordinated behaviour that is ordinarily effective relative to its objectives, given the context in which it normally occurs. Skills are described as programmatic (i.e. involving a series of steps) and become more unconscious as one becomes more skilled. Skills can be aggregated into larger processes called routines. The aggregate skills of the individuals manifest in the capabilities of an organization or a regional economy (Nelson and Winter, 1982). 

Many elements of skills are tacit (i.e. gained through experience) and thus difficult to develop or transfer without learning by doing, using and interacting. A particular skill (and its aggregate network of skills) may be difficult to transfer because it is challenging to communicate either as a result of the inherent ineffable nature of the skill or due to a poor vocabulary that prevents the people from articulating the desired skill. Repeated experience is a primary method for individual development of skills with an inherently tacit nature (Lundvall, 1998). 

Routines are the focal point for a learning-centered approach to the questions relating to organizational behaviour. Like the concept of tacit knowledge becoming embedded within an individual, routines are repeated organizational processes that become ingrained and efficient. Individuals or organizations will make decisions about their next course of action, generally, by sticking with a routine as long as it provides a satisfactory outcome. These routines will determine behaviour, are inherited (through organizational culture) and evolve (since some routines may lead to organizational success or failure) (Nelson and Winter, 1982). Schumpeter (1934) discusses how difficult it can be for individuals and organizations to overcome their ordinary routines in search of the innovative. He attributes this difficulty to uncertainty about the implications of the new routines, the limited time that the new opportunity may be available and the inherent biases towards the status quo that arise from existing routines. This means that routines can continue to be utilized even when they lead to a sub-optimal outcome due to outright irrationality or because developing a new routine can be expensive and can disrupt existing relationships and organizational culture (Nelson and Winter, 1982). This inertia that resists evolution in routines is a source of path dependency in a socioeconomic system.

On a macroeconomic scale, learning and skill development affect the performance of an economy because ‘intangible’ resources (i.e. knowledge, routines, and skills) are not evenly distributed throughout the economic system (Nelson and Winter, 1982; Lundvall, 1998; Nelson and Winter 2002; Lundvall, 2007). Individuals, organizations and regions are heterogeneous and possess bounded rationality about their environment and the opportunities available. Knowledge matters to actors because it enables agents an understanding of their ideal ways to optimize their economic outcomes (e.g. actors have bounded rationality) (Nelson and Winter, 2002).  

The creator of knowledge generally captures a portion of the possible benefit, however, the benefit can ‘spillover’ within and beyond the home industry of the firm. The knowledge spillover can significantly impact the research activities of others; a firm from industry sector ‘A’ can benefit from the research that is conducted in industry sector ‘B’. The productivity of one firm is not limited by their own research efforts but by the general level of knowledge that is accessible to it (Griliches, 1998).

Conceptually, messages are a wide range of information about the environment that an individual or organization can discover and interpret through a process called search routines. The skill of an agent comes from its ability to understand the potential of messages in conjunction with the ability to choose appropriate routines from their repository to exploit situations (Nelson and Winter, 1982).

1.3      Institutions and institutional dynamics

While it is valid to conceptualize organizations as institutions aggregated at a different scale, this dissertation presents institutions and organizations as separate concepts (Hollingsworth, 2000). Organizations are coalitions of elements with distinct boundaries meant to leverage economies of scale (Tirole, 1988). Institutions are defined as the rules of the game in a society and act to constrain and shape human interaction with an inherently crucial impact upon the way that societies evolve over time (North 1990). Institutions are conceptually crucial in this dissertation paper as they shape the interaction between regional systems of innovation elements (e.g. organizations, natural resources, and laws). Equally important to this paper is the impact that organizations and natural resources will have on the dynamic evolution of institutions (Edquist, 2001).

Institutions are a source of inertia in a path dependent system. Evolutions in institutions may also provide the catalyst for economic and technological change (Johnson, 1992; Freeman, 1995). The coevolution of institutions and other system elements results in fairly unique systems to emerge over time. Institutions in the form of rules and norms are generally durable[1]. This institutional durability contributes to longitudinal path dependency within a region (Hollingsworth, 2000). 

Organizations that are best able to adapt their activities to their institutional environment will be better positioned for success and innovation (Abernathy and Clark 1985, Hollingsworth, 2000). Interaction amongst organizations and their institutional environment is a multifaceted process. If organizations are empowered, they cannot only respond to their institutional environment but can attempt to modify it (Hollingsworth, 2000).

1.4      Systems of innovation

The system of innovation (SOI) approach has emerged from the field of evolutionary economics as a field of study for exploring the determinants of innovation (Edquist, 1999; Edquist, 2001; Geels, 2004).

 In his paper examining innovation as a systemic phenomenon Smith (2000) describes systems as possessing the following basic underpinnings:

• Economic behaviour rests on institutional foundations that afford individuals and organizations reduced uncertainty.
• Differences in institutional arrangements are critical in understanding differences in socioeconomic behaviour and outcomes.
• Competitive advantage results from variety and specialization.
• Institutional evolutional processes are self-reinforcing allowing path dependent specialization in socioeconomic structure.
• Technological knowledge is distributed amongst the individuals and organizations within the system. 

Much like the evolutionary economic approaches reviewed in the previous section, the systems of innovation approach emphasizes:

1. Putting innovation at the centre of the approach.
2. Including all the elements (i.e. organizational, political, social, natural resources, etc.) that are relevant to innovation in the model.
3. Exploring a historical perspective where system elements co-evolve and require historical context to understand how they have emerged.
4. There is no optimal system of innovation and that learning comes from comparison between idiosyncratic systems.
5. System elements (e.g. industrial firms) never innovate in isolation but that innovation occurs due to interplay with other system elements in a process that is guided by institutions (e.g. laws, regulations, habits etc.).
6. Innovation must be conceptualized beyond technological products and services in order to understand the relationship between innovation and economic growth.

 The systems of innovation approach suggests that innovation and economic performance are driven by the configuration of the elements of the system (see Table 1), how optimal these elements are relative to the demands upon the system and how effectively the system can evolve in response to demands. Knowledge is generated and applied by interactive learning between individuals and organizations within the system.

 

(Edquist, 2001; Geels, 2004)

Another approach to analyzing a system of innovation is presented in Table 2. On one axis, resources are characterized by their ease of replicability and thus regionally bounding more difficult to replicate resources. On the other axis, the resources are characterized by their physical attributes.  

 

(Lundvall, 2007)

Feedback in the system of innovation provides instability that will, over time, lead to qualitative changes in the economic structure. This evolutionary feedback affects institutions at micro, meso, and macro levels; the economy of today is resultant of historical evolution having played itself out (Smith, 2000; Aghion & David, 2009).

It is important when discussing a system of innovation that one is as specific as possible about the ‘level of analysis’ under discussion (Carlsson et al. 2002, Fagerberg et al., 2005). The regional perspective is the default system of innovation concept in this dissertation due to the sub-national context. Systems of innovation can generally be conceptualized at multiple levels, as described in the following table : 

(Breschi and Malerba, 1997; Cooke, P. and Uranga, M. and Etxebarrie, G.,1997; Smith, 2000; Edquist, 2001; Doloreaux, 2002; Geels, 2004; Doloreaux and Pareto, 2005). 

1.4.1     National systems of innovation

 The idea of a system of elements embedded in a national system of relationships is actually quite old with Freeman (1995) and Lundvall (1992) suggesting that the idea can be traced back to Friedrich List’s The National System of Political Economy. List’s work advocated policy ideas to protect emerging industries and encourage industrialization to catch up to the rapidly developing British economy. List recognized that investment in knowledge accumulation (versus physical capital investment) was a decisive factor in economic development (Freeman, 1995). List (1841) also foresaw the value of linking industry and university organizations: 

There scarcely exists a manufacturing business which has no relation to physics, mechanics, chemistry, mathematics or to the art of design, etc. No progress, no new discoveries and inventions can be made in these sciences by which a hundred industries and processes could not be improved or altered. In the manufacturing state, therefore, sciences and arts must necessarily be popular. 

List understood the importance of many of the elements discussed in contemporary national system of innovation studies (i.e. education, science, interactive learning between producer and user, integration of imported knowledge, technical institutions, etc.) and also emphasized the role of the state in coordinating activity and long-term economic development policies (Freeman, 1995). The advocacy of List and other economists, combined with the influence of Prussian institutions, induced Germany to develop one of the best technical training and educational systems in the world. List also observed the emergence of the unique American system which promoted knowledge based initiatives, enjoyed abundant natural resources and hosted an institutional environment that encouraged development and waves of immigration (Wright 1990; Freeman, 1995).

Fast forward to the 1980’s and research was revealing that the systematic aspects of innovation were crucial determinants of the efficacy of knowledge diffusion and associated productivity gains. Researchers and policy makers noted the astounding performance of Japan and South Korea as examples of this phenomenon (Freeman, 1985). It was hypothesized that national institutions were extremely influential in the rates of technological change and economic growth and, in the 1980’s, the National System of Innovation approach emerged as a discrete concept.

While many of the factors affecting a system of innovation will span the globe (i.e. social, technological and institutional), the majority of these factors are geographically linked to a nation. Lundvall (1998) suggests there are there are strong reasons to focus on the national level when examining or comparing SOIs:

• When examining a nation, there is often reduced variation in culture, institutions and language.
• Systems of innovation can vary dramatically due to institutional differences even when history, geography, culture etc. are reasonably similar (e.g. Canada and the United States).
• The vast majority of economic data is very national in its focus.
• The focus of much economic policy is directed at the national level.

 

1.4.2     Regional systems of innovation

A regional system of innovation derives much of its character from local knowledge networks (i.e. geographic proximity to advanced users, institutionalized user-producer relationships, local supplies of talent) and geographical features. This dissertation purports that regional systems of innovation are largely an institutionally defined concept. Institutions that are manifested in government policies (e.g. taxes, subsidies, R&D organizations, innovation infrastructure, financial support, regulation, procurement, etc.) define national (e.g. Canada) and sub-national (e.g. Alberta) regional systems of innovation (Doloreaux, 2002). Thus, a regional innovation system has much of its form defined by innovation policy instruments projecting from both national and sub-national levels of government. 

Beginning in the 1970’s, regional institutions (e.g. sub-national R&D subsidies to sectors) and regional organizations (e.g. sub-national research organizations and universities) began to play an increasing role in the evolution of their regional system of innovation (Cooke et al., 1997). In theory, national and regional governments should act, in informal and formal ways, to coordinate inputs from government, industry and academia in order to achieve regional innovation outcomes (Nelson, 1993; Hawkins, 2012; Freeman, 2004).   Collaboration is impacted by provincial institutions that demarcate the flow of economic factors in a manner similar to the function of national borders (Niosi, 2005). Hence, it is relevant to understand the idiosyncrasies of state structures and multi-level divisions of power particularly when studying the way in which policy instruments have been designed and developed. (Salazar and Holbrook, 2007; Borrass and Edquist, 2013).

This dissertation will return to the literature regarding the particular importance of natural resources as both cause and consequence of innovation (David and Wright, 1997). Localized natural resources that are valuable and enigmatic can have a primary evolutionary impact on a region’s institutions and catalyze the development of a regional technical and scientific knowledge base. These specialized knowledge bases emerge in a region based upon the needs of one sector, however, the tendency for knowledge to ‘spill-over’ from one specialty to another (Griliches, 1998) is an important evolutionary economic dynamic.

1.4.3     Sectoral system of innovation

Sectors are often a primary consideration when governments create policy instruments. Nelson and Winter (1997) posit that “policies need to be designed to influence particular economic sectors and activities.”. Malerba (2002) defines a sectoral system of innovation as “a set of new and established products for specific uses and the set of agents carrying out market and non-market interactions for the creation, production and sale of those products. Sectoral systems of innovation have a [sector specific] knowledge base, technologies, input and demand.”

A sectoral perspective on a system of innovation analyzes the structure of a sector, its boundaries, the agents within the sector and their interactions, learning and innovation processes, production processes, how the sector evolves and any factors that impact differential organizational and regional performance (Malerba, 2002).   

A crucial aspect of the sectoral system of innovation is that the knowledge, actors and institutions will vary significantly between sectors (Malerba and Vonortas, 2009). It is important to note that the geographic dimensions of a system of innovation may differ from one sector to another (Carlsson and Stankiewicz, 1991). Griliches (1998) observed that technological capabilities spill over from one industry sector to another and thus the knowledge base of adjacent sectoral systems influences the general knowledge base of the primary sector.

1.4.4     Technological system perspective

There is a perspective that takes technology not as an individual artifact but as an integrated technological system supported by managerial and societal elements (Smith, 2000). Technological systems focus upon knowledge competence flows (Carlsson and Stankiewicz, 1999). Technological systems of innovation develop and deploy technologies, clustering of resources and institutional infrastructure into new business opportunities. The technological system may be working to integrate diverse inputs into a product (Smith, 2000). 

1.5      Natural resources

Conventional theories of innovation have very little to say about the role that natural resources play in the evolution of innovation systems. This presents a challenge for understanding the development of systems of innovation in resource rich regions (e.g. Alberta).

Natural resource development can create economic impact directly through extraction and use. The development of natural resources can also create valuable externalities that impact the development of the regional economy (i.e. affect the capabilities of organizational supply chains and technological capabilities in the region) (Hawkins, 2011). Knowledge based natural resource development is an example of the spillovers that Griliches (1998) describes. Perceived natural resource opportunities directly influence the institutions and policies that will be introduced by the regional government (Boothe and Edwards, 2003). 

The existence of resources is certainly no promise of a well functioning socioeconomic system (Wright and Czelusta, 2004). The socially constructed elements of an economy have as much impact upon economic development as the resources themselves. Examples of social constructed elements include:

• The intensity applied by industry, government and academia in the search for natural resources.
• The emergence of new innovations for extraction, refining, utilization, etc.
• The emergence of substitutions for local natural resources.
• The existence of an accessible market for the natural resources.
• The legal, institutional and political structures that govern all of the above.

(David and Wright, 1997; Sachs and Warner, 1997; Sachs and Warner, 2001; Lundvall, 2007). 

Industrial organizations that directly develop natural resources are embedded within a system of suppliers. Thus, natural resource development can have a significant impact on supply chains related to resource extraction. The supply chain impact can be long, drawing upon an enormous range of inputs from a wide variety of other sectors, and can also be deep, drawing inputs that are crude and unrefined as well as inputs that represent the edge of human technological capability (Hawkins, 2011). The natural resource sector is strongly linked to innovations in exploration, extraction and substitution which are often driven by fears of impending scarcity since natural capital cannot always be reproduced (Wright and Czelusta, 2004; Lundvall 2007). The natural resource opportunities of a region and its development demands will influence the path of the region’s socioeconomic character. The emergent network of industrial organizations, institutions and ethos provides layers of ‘character’ to the region (David and Wright, 1997).  

David and Wright (1997) describe how the perception of natural resources as being nearly economically viable can create a type of innovation feedback loop. There can be positive feedback between resource development ‘savvy-ness’ (causation) and the actual amount of resources in a region (consequence). The creation of new knowledge (e.g. location of deposits, development techniques and technologies) can lead to development of widespread resource management routines which may create demand for further knowledge resources (e.g. scientists) (David and Wright, 1997).  Since natural resource abundance can be both a cause and a consequence of the deployment of knowledge resources (i.e. social, knowledge, infrastructure, etc.), it creates what David and Wright (1997) purport is a positive feedback loop; “The more [natural resources] you find, the closer you look, and the closer you look the more you find.”. Again, note the further benefit of these knowledge resources establishing within a region - they may spill over into non-resource related sectors in the form of new initiatives (Hawkins, 2011).

Developing institutions and policies for the development of the resource-focused economy is often a central economic goal of regional governments. A government can seek to improve the development of natural resources by making deliberate investments in knowledge capabilities such as an educated workforce, geological surveys and extraction technologies that can lead to a comparative advantage. A common natural resource ‘bottleneck’ is the lack of accurate knowledge about the extent and distribution of the potential natural resource deposit (Wright, 1990; Edquist, 1999; Wright and Czelusta, 2004)[4]. Expectation of a poor outcome from exploration or development is another sociological factor that can cause underperformance. A lack of expectation of new discoveries (or the perceived potential of existing natural resources) may be a more potent source of resource underdevelopment than many of the more common explanations (i.e. small population, distance to market, export factors, geological challenges) (David and Wright, 1997).

The perceived natural resources opportunities within a region guide the policies of the regional government. The type of natural resource, its location, the ability to economically harvest and distribute and the demand of the market are all variables that affect the impact of the natural resource and the concurrent evolution of government institutions (Boothe and Edwards, 2003). Regional governments and their institutions within the region and beyond can affect innovation by creating new demands that require technological solutions (e.g. stricter environmental regulations) and/or affect the economic viability of a resource (e.g. putting a resource in strategic reserve). A weak institutional environment will challenge the efficient, economic and orderly development of natural resources.   

Resource abundance (whether catalyzed by innovation or not) hardly ensures prosperity. Resource abundance can lead to economic challenges such as:

• An industrial structure focused upon exporting resources or providing inputs to the resource sector creating economic linkages to other sectors that can magnify economic growth or shrinkage from the resource sector.  
• Increased economic volatility arising from the interdependent nature of economic sectors.
• A regional government becoming dependent upon resource rents to finance ongoing operations leading to operating challenges during lean time.
• The growth of a resource sector challenging overall economic growth in a region if the growing sector has less externalities than the compromised sector(s).
• The growing sector challenging the growth of other sectors by raising the cost of inputs such as labour or capital.

(Wright, 1990; Mansell and Percy, 1990; Sach and Warner, 1997; Sachs and Warner, 2001; Wright and Czelusta, 2004; Alberta Chamber of Resources, 2011; Borrass and Edquist, 2013).

1.5.1     Transforming natural resources into knowledge resources

Policy makers in government often appreciate the knowledge based character of resource development. In such cases there may be a policy practice of taking non-renewable resource revenues and large saving and investing funds) (Edquist, 2001; Lundvall, 2007; VanSlyke, 2007). Knowledge resources are crucial for regional resource focused economies that want to manage production of natural resources in a forward looking way (Wright and Czelusta, 2004). A noteworthy policy instrument that a government may use to manage a resource bonanza is the creation of a savings fund financed from resource rents; this instrument of public policy can navigate multiple mandates such as helping to smooth government expenditures and support investments in economic diversification or quality of life improvements in the region (Magud and Sosa, 2011).