Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
Substantiating endogenous models on induced
technology change
Silvia Ulli-Beer, Alexander Wokaun
silvia.ulli-beer@ psi.ch, alexander.W okaun@ psi.ch
Paul Scherrer Institut, Villigen, Switzerland
Abstract
In this paper we reflect on effective research strategies for building helpful system
dynamics models on induced technology change that are substantiated in the relevant
literature and empirical data sources. The paper positions the innovation system literature
within the overall field of induced technology change as a distinct systemic approach that
offers relevant conceptual starting points for a system dynamics modeling experiment on
induced technology change analyses. Innovation system research is interested in
identifying the processes underlying innovation, industrial transformation and economic
growth. Also, the interest in the functional dynamics of innovation systems creates an
opportunity for system dynamics researchers that are applying a scholarly developed
modeling approach aiming to identify the structure and processes that explain behavior
patterns of induced technology change.
Based on literature research the paper summarizes the main modeling steps applied by
system dynamics scholars and compare it with research approaches of innovation system
scholars. An unifying research strategy framework for a scientific modeling approach is
introduced that highlights the main necessary requirements in order to be most useful for
areal world problem situation and for theory building and refinement in general, and
specifically for system approaches on induced technology change.
6/17/2010
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
1 Introduction
The “need to innovate” is growing not only in order to survive and prosper in highly
competitive and globalised economies but also to induce technology change as a response
to critical societal challenges such as climate change and security of energy supply.
In the last ten years system approaches to innovation and induced technology change
have been increasingly adopted by different agencies at different levels (e.g. the OECD,
the European Commission, or the European Environment A gency). Scholars of these
approaches (e.g. 1988:87; Kemp, Loorbach et al. 2007; Bergek, Jacobsson et al. 2008;
Geels, Hekkert et al. 2008; Jacobsson 2008; Fagerberg 2009) aim at offering a
complementary analysis framework to economic, mainly market failure based studies on
innovation and technology change.
These innovation system researchers from different scientific backgrounds are interested
in identifying the processes underlying innovation, industrial transformation, and
economic growth. Their ambition is to come up with a practically useful analytical
framework that allows for the assessment of system performance as well as the
identification of decisive performance factors. Y et, criticism concerns lack of a unifying
theoretical framework and therefore the comparability between single studies (Edquist
2004).
The interest in the feedback mechanisms and the functional dynamics of innovation
system creates an opportunity for system dynamics researchers applying a scholarly
developed modeling approach that aims at identifying the structure and processes
underlying the behavior patterns of dynamical complex systems (A ghion, David et al.
2009). Simulation models as the third branch of science (Pool 1992) allows for building
and testing theories that explain longitudinal patterns of system performance and for
conducting comparative policy and scenario analysis (Forrester 1971; Morecroft 1988;
Sterman 2000). Applying such computer based analysis frameworks would enhance the
progress of system approaches to induced technology change in theory building and
practical decision support. Modeling exercises would help to explore how sensitive
6/17/2010
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
coupled innovation systems are concerning policy action or inaction as well as to identify
most effective policy packages.
In this paper we reflect on effective research strategies for building helpful system
dynamics models on induced technology change that are substantiated in relevant
literature and empirical data sources. The aim of the paper is to identify opportunities and
challenges for system dynamics modelers and innovation systems researchers on how
they can effectively contribute to computer assisted development of carefully grounded
behavioral theories on induced technology change. Our experience is, that building a
carefully grounded model often requires both a broad literature research in the relevant
often interdisciplinary fields and experimenting with different social science methods
(Aghion, David et al. 2009). Unstructured approaches become overwhelming and may
consume too much resources delaying the overall modeling progress as well as the
development of an unifying body of theory and useful policy guidance. Hence, the
challenge is of knowing how to integrate and organize fragmented, scattered and thinly
spread innovation knowledge and data.
Inspired by Beer’s methodology of topological maps and scientific modeling (Beer 1984)
we suggest an elaborated research strategy framework for system dynamics scholars in
the field of induced technology change. On the one hand SD literature is guiding the
modeler by highlighting important modeling steps (e.g. Sterman 2000), by providing best
practice approach on group model building (e.g. Vennix 1996; Andersen and Richardson
1997) and on grounding and testing the model in the data and previous research (e.g.
Barlas 1992; Luna-Reyes and Andersen 2003). On the other hand we miss an ongoing
discussion about an efficient and methodological approach for both substantiating system
dynamics models against the background of relevant literature and empirical data
sources. A well grounded methodology as discussed in this paper is seen as important and
helpful since it will facilitate the confidence building process between researchers from
different communities (e.g. innovation system researchers and system dynamicists) as
well as interdisciplinary communication.
6/17/2010
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
2 Defining the starting position: system approaches to induced technology change
2.1 The field of induced technology change and innovation system research
Before providing an overview on main research streams on induced technology change
we will define the term induced technology change.
Dowlatabadi and Oravetz (2006) showed in their study that when energy efficiency has
not been perceived as a problem in the USA economy, the forces of change were not
efficiency enhancing. For example in the period 1954-74 structural change, innovation
and diffusion did not target energy conservation. However after the energy crisis, energy
efficiency became a target of change, e.g. energy efficiency was identified as a goal in
technological progress.
In the modeling literature in general, modeling frameworks that address “induced
technology change (ITC)” change (i.e. endogenous technical change models) have an
active technology equation that adjusts to policy (Goulder and Schneider 1999; Gerlagh
2007). Reflecting a similar understanding, the new economics of technology policy
applies the term “technological policies” for “all public interventions intended to
influence the intensity, composition and direction of technological innovations within a
given entity (region, country or group of countries)“ Foray (2009:1). While this
understanding highlights public policy initiatives, induced technology change refers to a
broader conception that include every realm of economic activity as a response to societal
challenges including innovation policy and entrepreneurship (Trajtenberg 2009). The
multi-level analysis of the transition pathway from horse-drawn carriages to automobiles
by Geels (Geels 2005) illustrates a historical successful regime shift supporting
technology change. The term regime refers rules and its embodiment in practices and
products. A broader sociotechnical regime definition is applied that refer to a regime rule-
set that binds producers, users and regulators together. Examples are shared engineering
search heuristics, ways of defining problems, user preferences, expectations, product
characteristics, skills, standards and regulatory frameworks (Schot and Geels 2007).
Based on the illustrative cases and the implicit understanding in the scientific literature
(Nordhaus 1997; Pizer and Popp 2008; Gancia and Zilibotti 2009; Nemet 2009),
6/17/2010
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
we define induced technology change as a regime shift being induced by both major
challenges of an economy or societal concerns (e.g. energy crises, climate change
threats) and new improvement paradigms being supported by policy initiatives that
are affecting the rate and direction of technological change.
In the following, main research arguments are highlighted that are still unsettled in the
field of induced technology change. Induced technology change is often traced back to
two different sources of change that are called technology (or supply) push and demand
pull, respectively. However, the economics of innovation literature and the studies on
energy technology policy have come to a consensus that both technology push as well as
demand pull policies are necessary to induce non-incremental technological change and
that this distinction may be artificial (Kline and Rosenberg 1986; Nemet 2009).
Nonetheless, the ongoing debate between target-focused and technology-led approaches
to climate policy (Galiana and Green 2009) still indicate the need to an improved
understanding of the mechanisms that are linking public policies and incentives which
innovators face as well as purchase behaviors. Such a coupled view is seen as necessary
for informed policy decisions and effective allocation of public funds (Nemet 2009).
Likewise linear innovation models that generally envision a progression from basic
research to applied research to product development and commercial products have been
challenged by recursive models in which cumulativeness, networks, interaction and
feedback effects between the different phases are accounted for’. For example, Kline and
Rosenberg (1986) suggests a “Chain-linked model” showing flow paths of information
among invention, innovation, and production. A still broader perspective is represented
by system approaches to innovation as initiated by Freeman (1987; 1988). In these
studies he coined and defined the term national system of innovation as “the network of
institutions in the public and private sectors whose activities and interactions initiate,
import, and diffuse new technologies” (1987:1). Since then different perspectives and
' However, improved models have not yet come into widespread use and consequently, the linear model is
still often invoked in current political discussions. This is not surprisingly as Thomas Kuhn (1967) has
argued, we do not abandon a model for thinking about a complex situation until we have a better model to
put in its place Kline, S. J. and N. Rosenberg (1986). An overview of innovation. The positive sum
strategy: Hamessing technology for economic growth. R. Landau and N. Rosenberg. Washington D.C. ,
National Academy Press.
6/17/2010
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
specifications of systems of innovation can be found in the literature ranging from
theoretically oriented work (e.g. Lundvall 1992) to case study approaches with a more
narrow focus on national innovation systems (e.g. Nelson 1993), to technologically
specific innovation systems (Carlsson and Stankiewicz 1991) or to sectoral (Malerba
2002) as well as regional innovation systems (Cooke, Uranga et al. 1997). The different
geographical foci are seen as variants of a single generic “system of innovation”
approach. According Edquist (2004) for each investigation the respective research
question has to guide the choice of the right conception.
Expanding the research boundary further allows connecting innovation research with
economic growth deliberations. For example, Abernathy and Utterback, or Christensen
(1978; 1996; 1997) explicitly focus on the dynamics of innovation in the sense of typical
behavioral phenomena of firms and industries and their structural explanations, at the
industrial level. Their understanding highlights the role of frame conditions for
newcomers and incumbents that can foster or hinder technology change and economic
prosperity. A further distinct perspective offers the economy of the knowledge society
with its innovation model by David and Foray (David and Foray 2002; Foray 2009): The
term “knowledge-based economy” has been recently coined and stands for a new kind of
organization that is fueling development. At the core are
“knowledge-based communities, i.e. networks of individuals striving, first and
foremost, to produce and circulate new knowledge and working for different, even
rival, organization (David and Foray 2002:9).”
Typically, a larger community creates a public space for exchanging and circulating the
knowledge with adequate communication and information technologies to codify and
transmit the new knowledge. Such boundary spanning communities are seen as the agents
of economic change. This approach also emphasis “research-without-frontiers”:
Therefore territorial limitation of funding of research and innovation activities appears
contrary to the need for global solutions. Their argument is that achieving technological
international competitiveness might well have become today a dangerous European
obsession, certainly when viewed against the global challenge and threats to national
welfare (Krugman 1994; Soete 2009).
6/17/2010
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
Freeman and Soete (2009) and Trajtenberg (2009) describe the two main shifts regarding
the boundary of science-technology-innovation (STI) research towards user driven
innovation and open innovation collaboration as an answer for reaching a critical mass in
research while keeping the escalating costs in balance with the involved risks and
benefits”. However, they also acknowledge that in reality different innovation models
prevail (e.g. the more linear as well as the open innovation conception). For example in
industrial sectors ranging from motor vehicles, semiconductors or electronic consumer
goods, knowledge-generation still follows clearly agreed on criteria of the more linear
R&D-system concept whereas in the open-source software development community, the
technological progress is less predictable.
The OECD (1981; 1997; 2005) undertakes the journey of offering guidelines for
collecting and interpreting innovation data that should help to better understand
innovation and its relation to economic growth, and to provide indicators for
benchmarking. Homogenous data sets should support diagnosis of performance gaps and
inform policy making as nicely illustrated by Ecaoua (2009). Catching up with new
insights in innovation research, the OECD have regularly adjusted their theory based
measurement frameworks, starting from the original R&D model in the so called Frascati
Manual (1981), to the firm model “innovation dynamo” in the 2"™ Oslo Manuel (1997),
and finally to the “innovation system” model accounting for innovation within the firm,
linkages with other firms and public research institutions, and the institutional framework
in which firms operate, as well as the role of demand. It is documented within the 3
Oslo Manual (2005).
This broad overview of the development of system approaches of innovation research and
induced technological change illustrates the wealth and heterogeneity of available
? Recent discussions on the multiple connections of technology change and economic prosperity refer to the
challenge of conceptualizing so called ,,science, technology, innovation and growth systems” (STIGS). See
for example: Aghion, P., P. A. David, et al. (2009). Can we link policy practice with research on "STIG
systems"? Toward connecting the analysis of science, technology and innovation policy with realistic
programs for economic development and growth. The new economics of technology policy. D. Foray.
Cheltenhem UK, Edward Elgar Publishing Ltd, Aghion, P., P. A. David, et al. (2009). "Science, technology
and innovation for economic growth: Linking policy research and practice in 'STIG Systems'." Research
policy 38: 681-693.
6/17/2010
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
theoretical information and data that inform innovation system research and technology
policy analysis. In sum, three developments trends have been evidenced:
First, the width of the innovation system boundary has been increased in order to
improve the understanding of the mechanisms linking public policy, entrepreneurial
incentives, and demand determinants. Second, researchers of the innovation policy field
are striving for a better understanding and representation of the dynamical complexity of
innovation processes. Finally, the variety of helpful innovation models as a basis of
policy formation is increasing.
Subsequently we will elaborate on innovation system approaches that have been applied
for identifying processes, feedback mechanisms, and functions of innovations system as a
basis for policy analysis. These approaches offer interesting starting points for system
dynamics modeling approaches since they intend to go beyond the meticulous
econometric analysis of causal relationships of picked factors that does not allow for a
better understanding of its role in the functioning of complex causal loop interactions
within innovation systems (see also Foray 2009).
2.1.1 Generics of systems of innovation
In order to elaborate more specifically on the contribution system dynamics modeling
can make to innovation system research, we will focus on generic properties of the
“systems of innovation” approach as described by Edquist (2004) which may build
bridges for system dynamics modelers interested in innovation processes. The generic
approach is characterized by common strengths and limitations. The common strengths
include its conceptualization as a complex system being made up of components,
relationships, and attributes (Carlsson, Jacobsson et al. 2002:234). Likewise system
dynamics terminology is used which emphasize characteristics such as feedback
processes, interdependences and non-linearities’ or accumulation, path dependence and
lock-in effects.
3 A typical study addressing non-linearity in innovation system research can be found by Geels, F. W. and
R. P. J. M. Raven (2006). "Non-linearities and expectations in niche-development trajectories: ups and
downs in Dutch biogas development (1973-2003)." Technological Analysis & Strategic Management
18(3/4): 375-92.
6/17/2010
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
For example Edquist argues that the generic characteristics are “ based on the
understanding that firms normally do not innovate in isolation but interact with
other organizations through complex relations that are often characterized by
reciprocity and feedback mechanisms in several loops. Innovation processes are
not only influenced by the components of the system, but also by the relations
between them. This captures the non-linear features of innovation processes and
is one of the most important characteristics of the SI approach” (Edquist
2004:185).
Recently, the identification and assessment of innovation system functions’ has gained
increasing attention concerning conceptual and practical guidance (Bergek, Jacobsson et
al. 2008; Negro and Hekkert 2008; Hekkert and Negro 2009). It serves as an heuristic
analysis device that goes beyond of previous system failure concepts with a focus on
infrastructure failure, institutional failures, interaction failures or capability failure (e.g.
Woolthuis, Lankhuizen et al. 2005). Bergek et al. (2008) argues that the functions
approach to innovation systems implies “a focus on the dynamics of what is actually
‘achieved’ in the system rather than on the dynamics in terms of structural components
only” (409).
A further acknowledged strength of generic innovation system approaches is its ability of
relating micro level processes to meso- or macro- level dynamics. Specifically for this
purpose, the multi-level perspective on transitions offers an explanatory framework
(Berkhout, Smith et al. 2004; Geels 2005; Geels and Raven 2006; Geels, Hekkert et al.
2008; Mahapatra, Gustavsson et al. 2008). This framework integrates conceptual findings
of strategic niche management as well as transition management approaches (Kemp,
Loorbach et al. 2007; Schot and Geels 2008). The strategic niche management approach
focuses primarily on endogenous steering by a collective that applies distinct selection
criteria in a niche and triggers a divergent evolutionary path (see also Levinthal 1998). It
is seen as a form of reflexive governance with collaborative learning processes as key
drivers for the emergence of a sustainable development path. However, since niche
internal processes alone may not be strong enough for an overall regime shift inducing
* Bergek et al (2008) suggests the following seven main functions of a technological innovation system:
knowledge development, resource mobilization, market formation, influence on the direction of search,
legitimation, entrepreneurial experimentation and development of external economies.
6/17/2010
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
technology change, the multi-level perspective aims at identifying the decisive forces and
processes at the niche, regime and landscape level which collude and finally foster
technological change.
In addition to this outline of the characteristics and strength of generic innovation system
research, its identified limitations highlight further improvement potential and
opportunities also for system dynamics modelers. Concretely the critiques give hints,
how system dynamics research could contribute to this scientific field. According Edquist
(2004) the main weaknesses of generic innovation system research is the conceptual
diffuseness conceming unclear defined terms, and its theory status, as well as the
definition of system boundaries.
“The SI approach has often been used more as a label than as an analytical tool.
It has not influenced the empirical studies in depth; for example it has not been
used to formulate hypothesis to be confronted to empirical observations” ... but he
also suggest that “SI approach should be used as a conceptual framework in
specific empirical analyses of concrete conditions. Testable statements or
hypotheses should be formulated on the basis of the approach and these should be
investigated empirically, by using qualitative as well as quantitative observations.
Theoretically based empirical work is the best way to straighten up the SI
approach; the empirical work will, in this way, serve as a “ disciplining” device in
an effort to develop the conceptual and theoretical framework” (Edquist
2004:202).
Although Edquist (2004) envisions the development of a theoretical framework for the
field, he also acknowledges the ongoing debates addressing the questions if the
innovation system concept should and could be advanced towards a formal theory and if
a deductive closed boundary definition could be elaborated.
This very focused outline have illustrated the meta assumptions guiding generic
innovation system research in order to identify opportunities for system dynamics
6/17/2010
10
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
research to contribute to this field. The following paragraph provides an overview of the
field’s practice of analysis.
2.1.2. Research steps in innovation system research
Innovation system investigations are typically based on case studies and the analyses of
time series describing the innovation system performance. The case study analysis
provides identification and verbal description of interaction between different actors,
networks and institutions and an assessment of how well specific functions are fulfilled.
The field offers some analysis scheme that give guidance for researchers or policy
makers on how to conduct a disciplined innovation system analysis. The aim of such an
analysis is primarily seen as providing practical guidelines for policy makers.
Bergek, Jacobsson et al. (2008) identifies in their scheme of analysis of technological
innovation systems six steps (see figure 1).
In the first step the context specific innovation system under focus will be defined. This
step is seen as most critical since the choice of the unit of analysis or the focus of the
study will strongly influence the outcome of the analysis. Researchers need to choose (1)
between the knowledge field or the product as a focusing device, (2) between breadth and
depth of the investigation, and (3) choose the spatial domain. In the second step the
structural components are identified (i.e. actors, networks and institutions). In the third
step, the conceptual framework of innovation functions is applied to describe the actual
occurrence of “functional dynamics” (i.e. how strong they may influence the innovation
system performance)’. The forth step is a relative assessment of how well the innovation
system is developed and is functioning in terms of the identified key processes. The
analysis often also includes a comparison between the focal innovation system and others
across regions or nations. Based on the analysis process goals are articulated which
describe how functional pattems (e.g. developing the knowledge base) should develop in
5 For example in the German wind turbine case, the development of positive extemalities as an innovation
function would be described as followed: “First, new entrants into the wind turbine industry, as well as into
wind power production, increased the political power of the advocates of wind energy so that they could
win against opposing utilities in several courts and defend a favorable institutional framework” Bergek, A.,
S. Jacobson, et al. (2008). "Analyzing the functional dynamics of technological innovation systems: A
scheme of analysis." Research Policy 37: 407-429.
6/17/2010
11
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
order to reach higher functionality. In the fifth phase exogenous factors influencing the
internal dynamics either as inducement mechanisms or as blocking mechanisms are
identified. Finally in the sixth phase key policy issues are specified that block or induce
the development of the above defined process goals.
5. Inducement &
blocking mechanisms
6. Key policy issues
4. Assessing
Functionality &
setting process goals
3a. Functions
2. Structural devcopment poblieson
Components
Market Influence on the
Actors formation ‘rection of search
>
Networks Entreprencurial
Legitimation experimentation
Institutions
Development of
external economies
N
3b. Achieved
Functional
Pattern
4. Starting-point;
defining the TIS in
focus
Figure 1: Innovation system analysis scheme (Bergek, Jacobsson et al. 2008)
In order to identify further similarities or distinctive characteristics of innovation system
and system dynamics modeling approaches on induced technology change relevant
characteristics of system dynamics modeling studies are highlighted in the following.
2.2 The field of system dynamics and typical studies on induced technology change
System dynamics embraces a worldview in which system behavior arises endogenously
from feedback loop structures. The field has been inspired by feedback thinking and
system theory and is basically grounded in control theory and nonlinear dynamics
(Forrester 1968; Richardson 1991). System dynamics has been applied to many real
world challenges such as industrial dynamics, urban dynamics (Forrester 1961; Forrester
1969) or climate change issues (Sterman and Sweeny 2002; Fiddaman 2009).
6/17/2010
12
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
Characteristic concepts of the system dynamics approach are accumulations and rates of
flow interconnected by information feedback loops with delays, counterintuitive
behavior, limits to growth, boom and bust patterns, nonlinearity, or tipping points as
typical behavioral characteristic of social systems. Policy studies often address
unintended consequences, policy resistance and compliance, or positive feedbacks
dominating undesired system development such as global warming (Sterman 2007).
System dynamics methodological aspiration goes beyond the analysis of dynamical
complex (policy / strategy / management) issues. Methodology strongly strives to
overcome impediments to learning in and about complex systems (Sterman 1994) and to
increase actors intuition conceming how system structure is linked to its behavior
(Forrester 1961; Repenning 2003). Therefore system dynamics methodology mainly
focuses on mapping and simulation modeling principles based on differential equation
models in the form of
r=dx/dt = f(xu)
where r is a vector of rates associated with x, a vector of states. The rates are some
nonlinear function f() of the state-vector x and u, a vector of exogenous inputs that
includes the parameterization of table functions. The resulting maps or simulation model
can be used as experimental laboratories both for research and in support of management
education.
Meadows and Robinson (1985; 2002) suggest that the models “ could be used to
search for imprecise policies that are robust against uncertainties rather than
precise policies that try to optimize something not understood. Perhaps most
important, they could simply serve as communication devices in which different,
partial, mental models of the social system could be expressed and integrated” (p.
300).
The double objectives of system dynamics research to understand complex feedback
systems, and to enhance managers’ or policy makers’ intuition on steering complex
system have coined many investigations in a broad array of societal and corporate
6/17/2010
13
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
challenges including transition and innovation studies (e.g. Maier 1987; Davidsen,
Sterman et al. 1991; Morecroft, Lane et al. 1991; Milling 1996; Pardue, Jr. et al. 1999;
Weil 2007; Rahmandad and Weiss 2009).
Milling’s outstanding work on dynamics of innovation diffusion can be used to illustrate
the conception and implementation of a system dynamics modeling approach (Milling
1996; Milling 2002; Sterman 2002). This work goes beyond the usual focus on the
lifecycle of a single innovation of new product and “S-curve” pattem. Endogenous to the
boundaries of his models are physical and managerial structures, advertising and word of
mouth effects, production capacity and cost (including learning curves) as well as
competition. He also has extended his models to include the processes of research and
development and their relationships to resource allocation and market success.
Exemplarily for the system dynamics quest to enhance managers’ intuition on managing
complex systems, he has adjusted his innovation management models into an “Innovation
Management Simulator”. The simulator can be applied by managers as a virtual
management laboratory that allows them to test different courses of action, also ina
competitive game setting but without its costly real-life consequences.
Another work described by Morecroft, Lane et al. (1991) illustrates a typical system
dynamics modeling approach which emphasizes “how project team members participate,
how their ideas are captured and mapped, and how simulations are used to challenge the
team’s intuition about policy options and consequences” (93).
In contrast to the above mentioned work, Rahmandad and Weiss (2009) describe a
system dynamics modeling study that mainly fosters computer-assisted theory
advancement based on rich case study data that aims to come up with conclusive
statements and practical recommendations. It nicely illustrates how boundary selection
influences the identification of critical reinforcing feedback processes that may lead to
tipping dynamics creating the challenges and opportunities in management tasks.
This short selective overview of some typical system dynamics studies with a focus on the
management of technology and innovation demonstrate the similar interest in a better
understanding and management on a tailored dynamical complex innovation system
6/17/2010
14
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
excerpt. The following presentation of the system dynamics scheme of analysis provides a
further basis for elaborating on main similarities and differences.
2.2.1 Research steps in system dynamics research
The design of system dynamics research is strongly guided by best practice approaches to
modeling that aims at developing helpful models for dealing with a complex issue.
Sterman (2000) describes a “disciplined process” with five iterative steps that successful
modelers normally follow (see figure 2). In the first step, the dynamic problem situation
and the system boundary is specified with crucial time series characterizing the observed
dynamical phenomena under focus. In the second step prevailing theoretical explanations
of the problematic behavior are challenged by a dynamical hypothesis. The new
formulated dynamic hypothesis explains the dynamics as an endogenous consequence of
the feedback structure. It should provide a more accurate picture of the problem situation
than previous theoretical explanations. In step three and four the dynamical hypothesis is
operationalized within the simulation model. The modeling activities include rigorous
specification, parameterization as well as structure and behavior testing. Step five is
dedicated to policy development and analysis. One important objective is to identify
robust policy recommendations under different scenarios and given uncertainties.
According Luna-Reyes and Andersen (2003) each modeling step can profit from social
science methods for collecting and analysis of qualitative data. Hence scientific rigor of
the system dynamics analysis originates from both proper data collection and analysis as
well as from the precise mapping and mathematical formalization and testing of clearly
operationalized concepts and feedback process. However, system dynamics researchers
have still a high degree of freedom on how they specifically conceptualize and term the
identified feedback structures depending on scientific theoretical frameworks or theories
in use they are applying (Andersen 1980).
6/17/2010
15
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
1. Problem Articulation
(a (Boundary Selection) \
5. Policy
Formulation
& Evaluation
2. Dynamic
Hypothesis
3. Formulation
Figure 2: Iterative research steps of the system dynamics modeling approach (Sterman
2000:87).
In sum, the result of a system dynamics analysis would not only be to develop a useful
simulation model addressing a specific issue but also to offer a dynamic theory for “the
family of systems to which the specific one belongs” (Forrester 2003:4).
2.3 Challenges and opportunities for system dynamics researchers
This section discusses the main similarities and differences between the two approaches
and draws research implications for system dynamics researchers. It specifically should
help to identify the challenges and opportunities for system dynamics modelers to
contribute to the field of innovation system research.
Our discussion begins with a short description of the nature of the scientific field on
innovation studies and departs from the following statement made in the field of
organizational science by Pfeffer (1993).
“Disagreement in theoretical approaches and even in methodology will not prove
detrimental as long as there is some agreement about what the fundamental questions
or issues are and as long as there are some agreed upon ways of resolving theoretical
and methodological disputes” (617).
In general the field of innovation studies is a thematically focused research community
that is organized in 136 research units world wide but not (yet) organized as a scientific
discipline with departments, undergraduate, graduate and post-graduate teaching,
6/17/2010
16
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
curricula, textbooks (Fagerberg 2009). Fagerberg (2009) identified in his network
analysis that the field consists of a large number of small groups of strongly interacting
scholars and that weaker ties to shared sources of scholarly inspiration, meeting places
and journals identifies them as a cognitive community. He identifies a mainstream
clusters of researchers called “Schumpeter crowd" since it is the largest community and
has the closest association with the core innovation literature®, and with the main meeting
places (e.g. The International Schumpeter Society ISS, or the Danish Research Unit for
Industrial Dynamics DRUID) as well as with the most important journals (e.g. Research
Policy)’.
In contrast, the field of system dynamics is a world view and a methodological focused
research community that is organized in research units with a comparable magnitude to
innovation studies. In some universities there are installed structures of a scientific
discipline (e.g at the Massachusetts Institute of Technology, USA or at the University of
Bergen, Norway). However, Fagerbergs (2009) analyses show that system dynamics
innovation modelers cannot be identified as a separate cluster contributing to the
innovation research community. Also, a short inquiry within the most important journal
on innovation studies (i.e. Research Policy) illustrates’ that system dynamics modeling of
innovation systems hardly contribute to the cognitive community of innovation studies.
While these findings show that system dynamicists and innovation system researchers do
not have an active common cognitive community, the generic properties as well as the
fundamental research questions of both fields as described in the previous chapters are
pointing to essential agreements on important research paradigms and questions. Also,
the methodology of both fields does not basically differ regarding their problem
© Authors of the most influential publications are Freemann, Schumpeter, Arrow, Schmookler, Nelson and
Winter, Rosenberg, Pavitt, Teece, Cohen and Levinthal and Lundvall (compare Fagerberg, J. (2009).
"Innovation studies - The emerging structure of a new scientific field." Research policy 38: 218-233.)
7 Based on an alternative classification scheme Monlacchi and Martin (2009) distinguishes three sub
communities within the innovation field that have among other things distinctive preferences for the unit of
analysis: the ,, science, technology and innovation” (STI) group focuses on interconnections of the
entrepreneurial, the industrial and national level and is dominated by economists; the “science and
technology studies” (STS) group emphasizes knowledge production processes and is dominated by
sociologists; finally the “technology and innovation management” (TIM) group is populated by researchers
with a business perspective at the firm level.
8 In the Research Policy inquiry 995 papers show up that apply the term “system AND dynamics”, out of
this sample only 140 papers refer to “simulation” and of them only 3 papers to “simulation AND
Forrester’.
6/17/2010
17
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
orientation and empirical data gathering. However, the analysis schemes of both fields
show stronger disparities. The system dynamics modeling approach goes beyond the
verbal description and assessment of observed causal structures and behavior patterns as
applied by the innovation system approaches. System dynamics modeling aims at
developing a virtual laboratory that allows theory and policy testing. In addition, system
dynamics approaches are in general problem focused and not bounded to a priory theory
and evaluation frames such as suggested by Bergek, Jacobsson et al (2008) (compare
chapter 2.1.2 and footnote 8). In addition, the reflexive overview on science, technology
and innovation (STI) policy research by Morlachi and Martin (2009) point to skepticisms
of this field regarding computerized model-building in the social science. They
specifically refer to the discussions around ‘limits to growth’ in the 1970s (see Meadows,
Meadows et al. 1972). The Science and Technology Policy Research (SPRU) group lead
by Freeman and Cohen’ was intensively involved in a sustained debate with the MIT
system dynamics group at that time (e.g. Meadows, Meadows et al. 1973; Streatfeild
1973). Some more up to data cautionary words regarding the use of simulation models in
the innovation policy research field are expressed by A ghion, David et al. (2009). For
example, they highlight the role of simulation models to leam about certain qualitative
dynamic properties but that these may be counteracted at the same time by change agent.
“A further complicating factor is that policy-decision makers and implementation
agents are themselves part of the interdependent processes and may contribute to
the creation of destabilizing positive feedback dynamics” (692).
These issues create the main challenges but also promising opportunities for system
dynamics researchers as well as for the field of innovation studies. The main challenge
for system dynamicists arises from identifying the most promising theoretical starting
point for their study that allows connecting to ongoing theorizing in the field’. In
addition to providing a coherent understanding of the research objectives, stated
assumptions, empirical data sources, findings and limitations, the communication of
° SPRU group is based at the University of Sussex, UK
‘© The STI policy research field defines itself as a problem-oriented field that focuses on practical issues but
positions its theorizing activities as mostly inductive reflecting on what empirical records appears to show.
Morlacchi, P. and B. R. Martin (2009). "Emerging challenges for science, technology and innovation policy
research: A reflexive overview " Research policy 38: 571-582.
6/17/2010
18
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
system dynamics research should refer to the term and explanation frames used in
innovation studies in general and innovation system approach specifically. They need to
illustrate how their findings may provide increased understanding and practical guidance.
This step however increases the cost for SD modelers since they must become familiar
with the literature of this fragmented field. In addition, it involves the risk of referring to
premature term frames and concepts from different innovation system communities and
their struggle on the accurate definition of system boundaries. Also the reflections made
by Repenning (2003) highlights relevant stumbling blocks for system dynamicists that try
to enter in a thematically focused research field or to enlarge their impact in general.
1. The failure to ground the work in the language and literature of the field.
2. Developing and communicating models that were too large and too complex for the
non-system dynamicist to absorb the embedded (new) insights.
3. The use of inadequate methods to build intuition concerning the link between a
model's structure and its behavior.
4. To target scholarly communities interested in modeling rather than those interested in
understanding complex social phenomena.
The benefits of sharing system dynamics modeling approaches with the scientific
innovation field are promising: It means to enter in a fresh dialogue with the existing
cognitive community and the existing market place of innovation research that could help
to increase the absorption of computer assisted theory development, and subsequently
their practical relevance and impact. This benefit would reinforce the opportunity for an
iterative improvement of an interdisciplinary understood term and conceptual frame that
may guide further empirical investigations and simulation experiments. Such a strategy
would also provide the innovation system field with an additional analytical tool
departing from clearly stated dynamic hypotheses that can be confronted with empirical
observations and virtual simulation experiments. The common meeting places as well as
publication channels could be used for resolving the theoretical and methodological
dispute emerging from system dynamics modeling approaches in innovation studies.
By tapping these synergies the acceptance of computer assisted theory refinement on
induced technology change in general may increase since
6/17/2010
19
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
“the acceptance of scientific theories is a function of both their truth value and
the political and rhetorical skill and power of their proponents and opponents”
(Ferraro, Pfeffer et al. 2009:670).
3 Towards a research strategy framework
In the previous chapter we have analyzed the challenges and opportunities of system
dynamics modeling of induced technology change. In this chapter we suggest a research
strategy framework that highlights the main requirements of a scientific modeling
approach that is both most useful for the real world problem situation and contributes to
theory building and refinement of complex models on induced technology change.
The suggested research strategy framework emphasizes the double objectives of system
dynamics research that is to contribute to theory building and to enhance the management
of dynamical complex systems. While the suggested research strategy framework has
been developed and applied in different studies on dynamics of innovative systems (e.g.
Ulli-Beer, Bruppacher et al. 2006) it has been inspired by Beer’s methodology of
topological maps and scientific modeling (Beer 1984).
3.1 Stafford Beer’s framework on scientific modeling
Stafford Beer distinguishes in his framework on scientific modeling the managerial
situation and the scientific situation (see figure 3). By definition the perception and data
reduction process in the managerial situation is systematically unrecognized whereas in
the scientific situation the reduction process is deliberate and creates a well understood
detail of a scientific view. He describes the process of systemic modeling as mapping of
the managerial concept model into a scientific concept model in such a way that the
structure is preserved. The mapping is a many to one reduction (homomorphism) that
needs to be made in a deliberate way. That means that the rigorous formulation of the
scientific interpretation of the managerial situation should preserve the discriminating
structures (isomorphism) of the managerial challenge. He postulates that
“if we find invariances between two systems, then these are isomorphic mappings,
one-to-one in the elements selected as typifying systematic behavior in some selected
6/17/2010
20
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
but important way. The generalized system that comes out of this process, which
applies to all systems of a particular class, is a scientific model” (p).
This view is in line with the qualification of system dynamics modeling as a structural
theory of system behavior (Lane 2001a) and with Forresters
“assertions of the ability of system dynamics to understand surprising behavior in
terms of a formal, analytical framework which allows insights learned in one domain
to be transferred to another” (Lane 2007:106).
Beer also emphasizes four challenges of scientific modeling:
First the research team must have an adequate understanding of relevant scientific
domains in order to relate the appropriate scientific concepts to managerial problem
situations. Second, intensive empirical investigation into the problem situation is
necessary to effectively apply a helpful scientific analogy. Third, the modeling process
that is the many to one simplification must be guided by empirically valid explanations of
the discriminating phenomena in order to come up with a useful model outcome. And
finally its fourth challenge refers to drawing an adequate model boundary in such a way
that control mechanisms of the managerial challenge can be identified. Beer concludes
that these four challenges of scientific modeling can only be met by grounding the
mapping in empirical science (Beer 1984).
6/17/2010
21
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
INSIGHT
Scientific
situation
Managerial
situation
Extra samples
of the class
‘4
ANALOGY
Scientific
model
Figure 3: Beer’s methodology of topological maps and scientific modeling (Beer 1984).
While the analysis scheme suggested by innovation system researchers (figure 1) as well
as system dynamics theorists (2) is highlighting the iterative character of the system
analysis, Beer’s methodology emphasis progression in the analysis process towards a
final product and indicates when the specific research journey can reach closure! It
actually addresses what steps within a scientific modeling - project of individual or group
of researchers are seen as necessary. It also highlights the mutual dependence of
empirical inquiry and scientific modeling.
In addition, it triggers different questions in the research progress that helps firstly to
position the research endeavor and to identify in advance the relevant research
community where from existing knowledge can be drawn.
6/17/2010
22
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
Last but not least it emphasizes the generalization from the specific case in such a way
that the generic properties of the case are guiding the development of a simplified model
with sufficient detail to guide effective decisions. Aghion (2009) illustrates this
requirement by his words of caution for using a simulation model to learn about certain
qualitative dynamic property of a complex system.
“The task of navigation in the terrain of ‘political economics’ will not be advanced by
furnishing either researchers or policy-makers with ‘a map that is as big as the
territory’” (692).
3.2 Theses on critical requirements for efficient scientific innovation system
modeling
System dynamics modeling can be used to accomplish various aims such as problem
solving in practice, or measuring and enhancing the ability of people to deal with
complex systems or contributing to theory building and refinement. The following
discussion on substantiating endogenous models on induced technology change
specifically refers to the scientific task of theory building and refinement based on case
study research. It draws on the account of innovation system research, system dynamics
modeling and combines Beer's deliberations on scientific modeling and case study
research (e.g. Eisenhardt 1989; Y in 2003). More specifically we derive proposition,
conceming what modeling tactics of system dynamicists are most fruitful to contribute to
theory building and refinement within the field of innovations system research and to
overcome the identified challenges.
Specifying the problem situation
Eisenhardt (1989) points out that theory-building research should begin as close as
possible to the ideal of no theory under consideration and no hypothesis to test.
“Attempting to approach this ideal is important because preordained theoretical
perspectives or propositions may bias and limit the findings” 536. Nevertheless in
innovation research the empirical phenomena in focus can often easily be related to
existing innovation concepts. But, different cases have their anomalies therefore each
case needs to be carefully described if it should enhance theory building. Eisenhardt
6/17/2010
23
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
(1989) emphasizes that in this phase the data gathering process may be very flexible, and
that it may combine data collecting and analysis at the same time. She also suggests to
take field notes and to trigger thinking by asking question such as “How does this case
differ from the last?” (539). At this stage, also “personal theories” or practitioner’ s
accounts of the problem situation are an important data source. In a messy problem
situation collaborative research methods may be most useful in order to define what
exactly the relevant problem is for the involved practitioners (c.p. Mueller, Ulli-Beer et
al. 2010). Without a clear description of the phenomena under study, the subsequent
theory building process may be difficult to follow by other researchers working on
innovation policy issues (c.p. Carlile and Christensen 2005). These researchers are used
to work with rich empirical case descriptions and are therefore open minded on similar
accounts of a new case. It provides the common starting ground for a dialogue with the
thematically interested scientific community. System dynamics has developed helpful
methods or scripts that gear the focus towards dynamical characteristics of the
phenomenon. The use of longitudinal graphs as reference modes of a dynamical problem
is a typical example (e.g. Andersen and Richardson 1997; Sterman 2000). Beer’s
framework points out, that this first empirical step “defining the problem situation” builds
the basis for choosing the most adequate scientific framework. It actually answer the
question “What is the object of research or the unit of analysis?” and does not yet
contribute to answering questions of theorizing such as how, when and why (c.p.
Bacharach 1989).
6/17/2010
24
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
1. Proposition - Specifying the problem situation: The problem definition is mainly
an empirical task and serves as basis for the dialogue with thematically interested
innovation policy researchers. It requires a detailed account on the empirical
phenomena under focus that tries to push thinking beyond theoretical lenses.
The outcome would be an empirical case description.
Selecting the thematic reference-authorities
Acknowledging the assumption made by Ferraro, Pfeffer et al. (2009) that the acceptance
of scientific theories does also depend on the political and rhetorical skill, and power of
the proponents and opponents it becomes important to identify the relevant cognitive
community in the field of innovation research. Since the innovation policy research field
has developed its own evaluation communities based on a merit-based market system, it
becomes crucial to identify early on the appropriate thematic audience and their
communication system (for example journals and conferences). For this purpose
Fagerberg (2009) provides a very helpful overview on the field of innovation studies.
Different research objects and levels of analysis may point to different target audiences
and publication outlets. Identifying the thematic reference-authorities also helps to
identify corresponding and helpful research perspectives and approaches and
frameworks. Finally it may become crucial for identifying the dialogue partners and for
legitimating the own ongoing research.
2. Proposition - Selecting the thematic reference-authorities: Increasing expertise
in innovation system modeling is an applied modeling discipline that should be
grounded in a thematically field. It requires that its contributions to theory
development and refinement can be discussed and judged within innovation system
communities and researchers of the broader STI policy research field.
The outcome would be a tentative communication and publication plan
6/17/2010
25
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
Discussing tentative reference frameworks
The empirical case description and first speculations on a conceptual model from a
practitioners view guides the identification of helpful scientific perspectives and
frameworks from previous research on similar cases. The comparison of emergent
concepts, hypotheses or theory with the extant literature is an essential feature of theory
building. An important aspect in this step is to discuss case selection based on the concept
of theoretical sampling. How adequate is the case in order to replicate or extend an
emergent theory. This first evaluation helps to define the limits early on for generalizing
the findings and to identify its relevance. The theoretical positioning results in the
identification of the research gap based on an evaluation of previous explanation frames.
The evaluation process helps to answer the question how adequately previous theorizing
may explain the observed phenomena and where the limitations are. Subsequently, it can
be speculated where the new case may contribute to an increased understanding and to
informed policy making. Also, this evaluation is guiding the formulation of alternative
(dynamic) hypotheses and the actual research questions.
3. Proposition - Discussing tentative reference frameworks: System dynamics
modeling can enhance innovation system research by providing scientific tools for
virtual experimentation and computer assisted theory building that would be
otherwise impossible. It requires the identification of corresponding theorizing
starting points within the broader innovation policy research field in an early
conceptualization phase. It requires further the theoretical discussion of case
sampling and the dynamic hypothesis regarding its likelihood to replicate or extend
emergent theory within the conceptualization phase.
The outcome would be a clear definition of the scientific modeling task with
formulated research questions and scientific sound dynamic hypotheses including
the definitions on time scale, boundary and levels of analysis.
6/17/2010
26
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
Developing expertise in innovation system modeling
Corresponding with Beer's call for grounding computer modeling in empirical science,
ongoing research on system dynamics modeling is highlighting the importance of
empirical rigor and is pointing to traditional theory building research such as grounded
theory approaches or case study research that are most helpful for supporting model
conceptualization (Kopainsky and Luna-Reyes 2008). It points out the tension between
empirical rigor and parsimony of the resulting theory:
“The potential danger in the application of grounded theory and case study research
is that the intensive use of empirical evidence can yield theory which is overly
complex” (483).
In addition, a second challenge of empiricism and system dynamics modeling needs to be
addressed. The efforts and cost of computer assisted theory building can become
overwhelming if the empirical data collection and analysis methods are applied as
comprehensively as in traditional theory building approaches and if modeling manly
becomes an additional specification. In the traditional theory building literature different
tactics and methods are highlighted that provide stronger substantive of hypotheses and
their internal validity: For example case study write-ups are suggested as simply pure
narrative descriptions with the use of longitudinal graphs, because they would help
researchers to cope early in the analysis process with the often enormous volume of data
(Eisenhardt 1989). Also triangulation by multiple data collection and analysis methods or
relying on multiple investigators should help to increase validity (Y in 2003). However,
the challenge is not to become highly competent as social scientist in the first phase and
then as system dynamics modeler in the second, but to develop expertise! in computer
assisted theory building on innovation research that is optimally effective and efficient.
" According Herling competence is characterized by a minimal efficient and effective
demonstrated behavior whereas optimally efficient and effective behaviors characterize
expertise; Herling, R. W. (2000). "Operational definitions of expertise and competence."
Advances in developing human resources 2(1): 8-21. Hence an expert is seen as an
problem solver geared to performance; Breiter, C. and M. Scardmalia (1993). Surpassing
ourselves: An inquiry into the nature and implications of expertise. Chicago, Open Court,
Swanson, R. A. (2003). "A disservice to the ideas of theory, research and expertise."
human resource development review 2: 206-2010..
6/17/2010
27
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
Therefore tailored empiricism for computer assisted theory development needs to be
applied that meets the critical requirement of scientific modeling and scientific theory
building. Such ambitions become an important research avenue for applied modeling
projects in general. Research into synergies and redundancies between traditional social
science approaches and computer assisted approaches for theory building needs to be
identified. In the field of system dynamics the ongoing discussion on empiricism and
system conceptualization, participatory modeling or validation is contributing to this task.
Further work may be required that compare quality standards of empirical data collection
and analysis for traditional theorizing (e.g. Bacharach 1989) and for scientific modeling
that can add scientific rigor through simulation in further steps. But also the first
challenge that is addressing homomorphism and parsimony of scientific modeling
remains an important task of system innovation modeling. So called “small models”
(Ghaffarzadegan, Lyneis et al. 2009) may become most helpful in complex innovation
policy tasks. However, such expertise needs to be further developed in order to exploit
the opportunities for system dynamics research to enhance the field of innovation policy
research.
4. Proposition - Developing expertise in innovation system modeling: Innovation
system modeling expertise depends on tailored methods of computer assisted theory
building. It requires a methodical synthesis of innovation system empiricism and
modeling that is guiding the development of small models for innovation policy.
The outcome would be a parsimonious model with high explanatory power on
innovation dynamics that allows conducting virtual policy experiments.
6/17/2010
28
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
4 Discussion and conclusions
The given account on effective research strategies for building helpful system dynamics
models on induced technology change makes three main contributions. In a first step, we
have identified three development trends in the field of innovation research: First, the
boundary of coherent theorizing on science, technology and innovation policy is
expanding. Second, theorizing efforts are increasingly addressing the dynamical
complexity of innovation processes. Third, alternative frameworks and models are
emerging as a basis of policy formation.
Ina second step we have identified challenges and opportunities for system dynamics
modelers to contribute to the field of innovation system research. One important
challenge for both communities is to establish a fruitful dialogue. A second challenge is
to overcome the methodological reservation of innovation policy researcher regarding the
use of simulation models for policy making. The main opportunity is to provide the
innovation system field with an additional tool that is most useful in enhancing theory
developing conceming the dynamical complexity of innovation processes.
Finally, we have formulated four propositions for innovation system modeling
approaches on how to overcome the challenges and to deploy the opportunity. The
propositions point to four critical leverage points of innovation system modeling tasks:
(1) the empirical description of the problem situation, (2) the selection of thematic
reference authorities, (3) the adequate discussion of tentative reference frameworks and
(4) the development of tailored expertise in innovation system modeling.
The findings are not contradicting any theorizing in the field of system dynamics on how
the concepts of feedback loops, rates and stock variables should be used to construct
models on social phenomena. But it reflects on critical requirements for positioning this
scholarship as an applied modeling discipline in the field of innovation policy research.
The main argument is that for the scientific legitimation of endogenous simulation
models on induced technology change, the thematically oriented community of the
innovation research field becomes as important as the methodological focused system
dynamics community. In addition, it highlights that expertise in innovation system
6/17/2010
29
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
modeling requires a tailored methodological synthesis that supports both the empirical
identification and the scientific mapping of the discriminating structures of the specific
behavioral phenomena of an innovation system. Empirical research methods should
facilitate the exploration of the link between practice, theory and modeling.
At the best such a methodical synthesis may be guided by general innovation frameworks
(for example the multilevel perspective on transition) or core concepts of innovation
system analysis as suggested by Bergek, Jacobsson et al. (2008). A similar synthesis can
be found in the field of strategy making that applies theoretical concepts and methods
from strategy and policy making and combines them with the strength of the system
dynamics method (e.g. Howick, Ackermann et al. 2006; Eden, Ackermann et al. 2009).
Also in innovation system modeling approaches the aim of such a synthesis should be to
get the best out of traditional innovation system analysis and the additional analytic
power that system dynamics modeling can offer in respect to effectiveness and
efficiency. This implies that research also should reflect on (less) successful method
mixing experiences in order to improve expertise on innovation system modeling.
In addition to the above discussed implications, the convergent view on computer
assisted theory development as offered by Beer’s methodology is providing an useful
template for a research project road map. It allows for illustrating research progress by
identifying important milestones. In addition it highlights how theory, empiricism (or
mutual learning) is logical embedded in the project design of innovation system modeling
and policy. Figure 4 provides an illustrative example of its implementation in a successful
project proposal for the Swiss National Science Foundation”.
2 Project No. 405440-107211/1 “Diffusion dynamics of energy efficient buildings DEEB” lodged with the
Interfacultary Center for General Ecology in Berne, Switzerland.
6/17/2010
30
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- | uly 25 - 29, 2010.
Project road map
Identification of
important stakeholders,
networks
Conceptual research
heuristics,
A Decision-model__[
B Simulation model
processes and
decision rul
ransformation support |}
tool with robust policies 7
and policy implication:
Figure 4: Project road map - integrating system experts into the system dynamics
analysis in order to foster a mutual inquiry attitude during the investigation (see also
(Ulli-Beer, Bruppacher et al. 2006)
In addition to this communicative advantage, Beer’ s mapping framework, the discussion
of requirements and the propositions explicitly disentangle the logic of the interplay
between empiricism and scientific modeling for computer assisted theory building. Most
statements and propositions are well known in traditional theory building research and
also in the tradition of system dynamics modeling. They concur with good research
practice in general, - but the paper emphasizes the critical leverage points for establishing
a fruitful research dialogue between a thematically focused research community and a
modeling based community.
Finally the methodological deliberations also point to limits how far system innovation
modeling may enhance formal theory building on innovation issues and its adequate
6/17/2010
31
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
boundary. System dynamics as a problem oriented approach may finally contribute to the
formalization of behavioral theories on invariant innovation phenomena that can be
replicated or refined by cross case analysis. It can enhance concept development and
operationalization for a particular class of systems with its specific boundaries. However,
further research resulting in classification schemes of different classes of systems and the
development of general innovation frameworks remains beyond computer assisted theory
building. For such an endeavor an holistic theory framework for applied disciplines as
suggested by Swanson (2007) may provide an interesting starting point.
6/17/2010
32
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
Literature
Abernathy, W. J. and J. M. Utterback (1978). "Patterns of Innovation in Industry."
Technology Review 80(7): 40-47.
Aghion, P., P. A. David, et al. (2009). Can we link policy practice with research on
"STIG systems"? Toward connecting the analysis of science, technology and
innovation policy with realistic programs for economic development and growth.
The new economics of technology policy. D. Foray. Cheltenhem UK, Edward
Elgar Publishing Ltd.
Aghion, P., P. A. David, et al. (2009). "Science, technology and innovation for economic
growth: Linking policy research and practice in ‘STIG Systems'." Research policy
38: 681-693.
Andersen, D. F. (1980). How differences in analytic paradigms can lead to differences in
policy conclusions. Elements of the System Dynamics Method. J. Randers.
Massachusetts, The MIT Press: 61-75.
Andersen, D. F. and G. P. Richardson (1997). "Scripts for group model building." System
Dynamics Review 13(2): 107-129.
Bacharach, S. B. (1989). "Organizational theories: Some criteria for evaluation." The
Academic Management Review 14(4): 496-515.
Barlas, Y. (1992). "Comments on "On the Very Idea of System Dynamics Model of
Kuhnian Science." System Dynamics Review 8(1): 43-47.
Beer, S. (1984). "The Viable Sytem Model: Its Provenance, Development, Methodology
and Pathology." Journal of the Operational Research Society 35(1): 9.
Bergek, A., S. Jacobsson, et al. (2008). "Analyzing the functional dynamics of
technological innovation systems: A scheme of analysis." Research Policy 37:
407-429.
Berkhout, F., A. Smith, et al. (2004). Socio-technological regimes and transition contexts
System Innovation and the Transition to Sustainability. Theory, Evidence and
Policy. B. Elzen, F. W. Geels and K. Greene. Cheltenham, Edward Elgar: 48-75.
Breiter, C. and M. Scardmalia (1993). Surpassing ourselves: An inquiry into the nature
and implications of expertise. Chicago, Open Court.
Carlile, P. R. and C. M. Christensen (2005). "The cycles of theory building in
management research (version 6.0)." Available from http://www. innosight.
com/documents Accessed May 19 2010.
Carlsson, B., S. Jacobsson, et al. (2002). "Innovation systems: analytical, and
methodological issues." Research policy 31: 233-245.
Carlsson, B. and R. Stankiewicz (1991). "On the nature, function, and composition of
technological systems." Journal of Evolutionary Economics 1(2): 93-118.
Christensen, C. M. (1997). The Innovator's dilemma. When New Technologies Cause
Great Firms to Fail, Harvard Business School Press.
Cooke, P., M. G. Uranga, et al. (1997). "Regional systems fo innovation: institutional and
organisational dimensions." Research policy 26: 945-74.
David, P. A. and D. Foray (2002). "An introduction to the economy of the knowledge
society." The International Social Science Journal 171: 9-23.
6/17/2010
33
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
Davidsen, P. I., J. D. Sterman, et al. (1991). "A petroleum life cycle model for the United
States with endogenous technology, exploration, recovery, and demand." System
Dynamics Review 6(1): 66-93.
Dowlatabadi, H. and M. A. Oravetz (2006). "US long-term energy intensity: Backcast
and projection." Energy Policy 34: 3245-3256.
Eden, C., F. Ackermann, et al. (2009). "Integrating modes of policy analysis and strategic
management practice: requisite elements and dilemmas." Journal of the
Operational Research Society 60: 2-13.
Edquist, C. (2004). Systems of innovation: perspectives and challenges. The Oxford
Handbook of Innovation. J. Fagerberg, D. Mowery and R. Nelson. Oxford,
Oxford University Press: 181-208.
Eisenhardt, K. (1989). "Building theories from case study research." Academy of
Management Review 14: 532-550.
Encaova, D. (2009). Nature of the european technology gap: creative destruction or
industrial policy. The new economics of technology policy. D. Foray.
Cheltenham, Elgar: 281-314.
Fagerberg, J. (2009). "Innovation studies - The emerging structure of a new scientific
field." Research policy 38: 218-233.
Ferraro, F., J. Pfeffer, et al. (2009). "How and why theories matter: a comment on Felin
and Foss (2009)." Organization Science 20(3): 669-675.
Fiddaman, T. (2009). "Dynamics of climate policy." System Dynamics Review 23(1):
21-34.
Foray, D. (2009). General introduction. The new economics of technology policy. D.
Foray. Cheltenham, Edward Elgar Publishing Ltd: 1-2.
Foray, D., Ed. (2009). The new economics of technology policy. Cheltenham, Elgar.
Forrester, J., W. (1968). Principles of Systems. Cambridge, MIT Press.
Forrester, J. W. (1961). Industrial Dynamics. Cambridge, MIT Press.
Forrester, J. W. (1969). Urban Dynamics. Massachusetts, MIT Press.
Forrester, J. W. (1971). "Counterintuitive behavior of social systems." Technology
Review 73(3): 52-68.
Forrester, J. W. (2003). Economic Theory for the New Millennium. 21st International
Conference of the System Dynamics Society (July 20-24, 2003), New Y ork City,
The System Dynamics Society.
Freeman, C. (1987). Technology Policy and Economic Performance: Lessons from Japan.
London, Pinter.
Freeman, C. (1988). Japan: a new national system of innovation Technical Change and
Economic Theory. G. Dosi and e. al. London, Francis Pinter: 330-348.
Freeman, C. and L. Soete (2009). "Developing science, technology and innovation
indicators: What we can learn from the past." Research policy 38: 583-589.
Galiana, I. and C. Green (2009). "Let the global technology race begin." Nature 462(3):
570-571.
Gancia, G. and F. Zilibotti (2009). "Technological change and the wealth of nations."
Annual Review of Economics 1: 93-120.
Geels, F. W. (2005). "The dynamics of transitions in socio-technical systems: A multi-
level analysis of the transition pathway from horse-drawn carriages to
6/17/2010
34
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
automobiles (1860 - 1930)." Technological Anlysis & Strategic Management
17(4): 445-476.
Geels, F. W. (2005). "Processes and patterns in transitions and system innovations:
Refining the co-evolutionary multi-level perspective." Technological Forecasting
and Social Change 72: 681-696.
Geels, F. W., M. P. Hekkert, et al. (2008). "The dynamics of sutainable innovation
journeys." Technological Analysis & Strategic Management 20(5): 521-536.
Geels, F. W. and R. P. J. M. Raven (2006). "Non-linearities and expectations in niche-
development trajectories: ups and downs in Dutch biogas development (1973-
2003)." Technological Analysis & Strategic Management 18(3/4): 375-92.
Gerlagh, R. (2007). "Measuring the value of induced technological change." Energy
Policy 35(11): 5287-5297.
Ghaffarzadegan, N., J. Lyneis, et al. (2009). "Why and how small system dynamics
models can help policymakers: A review of two public policy models."
Proceedings of the 26th International Conference of the System Dynamics
Society, July 26-31, Albuquerque, NM USA.
Goulder, L. H. and S. H. Schneider (1999). " Induced technological change and
the attractiveness of CO2 abatement policies. 21, 211-253." Resource and Energy
Economics 21: 211-253.
Hekkert, M. P. and S. O. Negro (2009). "Functions of innovation systems as a framework
to understand sustainable technological change: Empirical evidence for earlier
claims." Technological Forecasting & Social Change 76: 584-594.
Herling, R. W. (2000). "Operational definitions of expertise and competence." Advances
in developing human resources 2(1): 8-21.
Howick, S., F. Ackermann, et al. (2006). "Linking event thinking with structural
thinking: methods to improve client value in projects." System Dynamics Review
22(2): 113-140.
Jacobsson, S. (2008). "The emergence and troubled growth of a ["]biopower' innovation
system in Sweden." Energy Policy 36(4): 1491-1508.
Kemp, R., D. Loorbach, et al. (2007). "Transition management as a model for managing
processes of co-evolution." The Intemational Journal of Sustainable Development
and World Ecology: 78-91.
Kline, S. J. and N. Rosenberg (1986). An overview of innovation. The positive sum
strategy: Harnessing technology for economic growth. R. Landau and N.
Rosenberg. Washington D.C. , National Academy Press.
Kopainsky, B. and L. F. Luna-Reyes (2008). "Closing the loop: Promoting synergies with
other theory building approaches to improve system dynamics practice." Systems
Research and Behavioral Science 25: 471-486.
Krugman, P. (1994). "Competitiveness: a dangerous obsession." Foreign A ffairs 73(2):
28-44.
Lane, D.C. (2001a). "Rerum cognoscere causas: Part I - How do the ideas of system
dynamics relate to traditional social theories and the voluntarism/determinism
debate?" System Dynamics Review 17(2): 97-118.
Lane, D. C. (2007). "The power of the bond between cause and effect: Jay Wright
Forrester and the field of system dynamics." System Dynamics Review 23: 95-
118.
6/17/2010
35
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
Levinthal, D. A. (1998). "The slow pace of rapid technological change: Gradualism and
punctuation in technological change." Industrial and Corporate Change 7(2): 217-
247.
Luna-Reyes, L. F. and D. L. Andersen (2003). "Collecting and analyzing qualitative
data." System Dynamics Review 19(4): 271-296.
Lundvall, V. A. (1992). National Systems of Innovation: Towards a Theory of Innovation
and Interactive Learning. Pinter London.
Mahapatra, K., L. Gustavsson, et al. (2008). "Bioenergy innovations: The case of wood
pellet systems in Sweden." Technological Analysis & Strategic Management
19(1): 99-125.
Maier, F. H. (1987). "New product diffusion models in innovation management - a
system dynamics perspective." System Dynamics Review 14(285-308): 285-308.
Malerba, F. (2002). "Sectoral systems of innovation." Research policy 31(2): 247-264.
Meadows, D. H., D. L. Meadows, et al. (1973). "A response to Sussex." Futures 5: 135-
152.
Meadows, D. H. andJ. M. Robinson (2002). "The electronic oracle: computer models
and social decisons." System Dynamics Review 18(2): 271-308.
Meadows, D. L., D. H. Meadows, et al. (1972). The Limits to Growth. New Y ork,
Universe Books.
Meadows, D. L. andJ. M. Robinson (1985). The electronic oracle. Computer models and
social decision. Chichester, John Wiley and Sons.
Milling, P. (1996). "Modeling innovation processes for decision support and management
simulation." System Dynamics Review 12(3): 211-234.
Milling, P. M. (2002). "Understanding and managing innovation processes." System
Dynamics Review 18(1): 73-86.
Morecroft, J. D. W. (1988). "System dynamics and microworlds for policymakers."
European Journal of Operational Research 35(1988): 301-320.
Morecroft, J. D. W., D. C. Lane, et al. (1991). "Modeling growth strategy in a
biotechnology startup firm." System Dynamics Review 7(2): 93-116.
Morlacchi, P. and B. R. Martin (2009). "Emerging challenges for science, technology and
innovation policy research: A reflexive overview " Research policy 38: 571-582.
Mueller, M., S. Ulli-Beer, et al. (2010). "How do we know whom to include in
collaborative research? Towards a methof for the identification of experts."
European journal of operational research forthcoming.
Negro, S. and M. P. Hekkert (2008). "Explaining the success of emerging technologies
by innovation system functioning: the case of biomass digestion in Germany."
Technological Analysis & Strategic Management 35(2): 465-482.
Nelson, R. R. (1993). National system of innovations: A comparative study. Oxford,
Oxford University Press.
Nemet, G. F. (2009). "Demand-pull, technology-push, and government-led incentives for
non-incremental technical change." Research Policy 38(5): 700-709.
Nordhaus, W. D. (1997). " Modeling induced innovation in climate-change policy.
Working Paper, Y ale
University." Working paper, Y ale University.
OECD (1981). The measurement of scientific and technical activities: "Frascati Manual".
Paris, OECD publishing.
6/17/2010
36
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
OECD (2005). Oslo Manuel: Guidelines for collecting and interpreting innovation data
3rd Edition). Paris, OECD Publishing.
OECD and Eurostat (1997). OSLO Manual: The measurement of scientific and
technological activities (2nd Edition). Paris, OECD pulishing.
Pardue, J. H., T. D.C. Jr, et al. (1999). "Modeling short- and long-term dynamics in the
commercializaton of technical advances in IT producing industries." System
Dynamics Review 15: 97-105.
Pfeffer, J. (1993). "Barriers to the advancement of organizational science: Paradigm
development as a dependent variable." The academic management review 18(4):
599-620.
Pizer, W. A. and D. Popp (2008). "Endogenizing technological change: Matching
empiricl evidence to modeling needs." Energy Economics 30: 2754-2770.
Pool, R. (1992). "The third branch of science debuts." Science 256: 44-47.
Rahmandad, H. and D. M. Weiss (2009). "Dynamics of concurrent software
development." System Dynamics Review 25: 224-249.
Repenning, N. P. (2003). "Selling system dynamics to (other) social scientits." System
Dynamics Review 19(4): 303-327.
Richardson, G. P. (1991). Feedback Thought in Social Science and Systems Theory.
Philadelphia, University of Pennsylvania Press.
Schot, J. and F. W. Geels (2007). "Niches in evolutionary theories of technical change: A
critical survey of the literature." Journal of Evolutionary Economics 17: 605-622.
Schot, J. and F. W. Geels (2008). "Strategic niche management and sustainable
innovation journeys: theory, findings, research agenda, and policy."
Technological Analysis & Strategic Management 20(5): 537-554.
Soete, L. (2009). Research without frontiers. The new economics of technology policy.
D. Foray. Cheltenham UK, Edward Elgar Publishing Ltd.
Sterman, J. (2002). "The 2001 Jay W. Forrester A ward: citation for the winner Peter
Milling." System Dynamics Review 18(1): 71-72.
Sterman, J. D. (1994). "Learning in and about complex systems." System Dynamics
Review 10(2-3): 291-330.
Sterman, J. D. (2000). Business Dynamics. Systems Thinking and Modeling fora
Complex World. Boston, Irwin McGraw-Hill.
Sterman, J. D. (2007). "Exploring the next great frontier: system dynamics at fifty."
System Dynamics Review 23(2/3).
Sterman, J. D. and 1. B. Sweeny (2002). "Cloudy Skies: assessing public understanding of
global warming." System Dynamics Review 18(2): 207-240.
Streatfeild, G. (1973). "World dynamics challenged." Futures 5(4).
Swanson, R. A. (2003). "A disservice to the ideas of theory, research and expertise."
human resource development review 2: 206-2010.
Swanson, R. A. (2007). "Theory framework for applied disciplines: Boundaries,
contributing, core, useful, novel, and irrelevant components." Human resource
development review 6(321-339).
Trajtenberg, M. (2009). The rumblings of a paradigm shift: concluding comments. The
new economics of technology policy. D. Foray. Cheltenham, UK, Edward Elgar
Publishing Ltd.
6/17/2010
37
Paper to be presented at the 28th International System Dynamics Conference, Seoul, Korea -- J uly 25 - 29, 2010.
Ulli-Beer, S., S. Bruppacher, et al. (2006). Introducing an Action Science V enture:
Understanding and accelerating the diffusion process of energy-efficient buildings
Proceedings of the 24th International Conference of the System Dynamics
Society, (23-27 July 2006) Nijmegen NL.
Utterback, J. M. (1996). Mastering The Dynamics of Innovation. Boston, Massachusetts,
Harvard University Business School Press.
Vennix, J. A. M. (1996). Group Model Building. Facilitating Team Learning Using
System Dynamics. Chichester, John Wiley & Sons.
Weil, H. B. (2007). "Application of system dynamics to corporate strategy: an evolution
of issues and frameworks." System Dynamics Review 23(2/3): 137-156.
Woolthuis, R. K., M. Lankhuizen, et al. (2005). "A system failure framework for
innovation policy design." Technovation 25(609-619).
Yin, R. K. (2003). Case study research: Design and methods (3rd ed.). Thousand Oaks,
CA, Sage.
6/17/2010
38