Content uploaded by Siavash Farahbakhsh
Author content
All content in this area was uploaded by Siavash Farahbakhsh on Sep 10, 2019
Content may be subject to copyright.
Modeling
Organizational
and Institutional
Complexity
3–4 maggio 2019 / Room C4.01
Fourth Agent-Based Models of Organizational
Behavior (ABMO4) Workshop
Campus Bozen-Bolzano
Workshop Proceedings
Preface
This short booklet collects the proceedings of work presented at the Fourth Agent-
Based Models of Organizational Behavior (ABMO4) workshop on “Modeling Organi-
zational and Institutional Complexity”. Organizers prepared a Call for Abstracts and
selected thirteen to be included in the workshop, for two days of work and discussions,
on 3-4 May 2019 at the Free University of Bolzano-Bozen.
The purpose of the was to explore the complexity of social systems across multi-
levels of individual, organization and sectors/fields. The use of computational simu-
lations alone and/or in combination with empirical data was one of the elements that
organizers highlighted in the call. To deal with this challenging objective, the workshop
welcomed contributions from any discipline, including but not limited to psychology,
sociology, management, computer science, cognitive science, decision science, lan-
guage science, artificial intelligence, economics, philosophy. This interdisciplinary
approach was complemented with the aid of a distinguished scientific committee to
cover different research areas.
Stephen Cowley, University of Southern Denmark, Denmark
Alan Kirman, EHESS, Aix-Marseille University, France
Daniel A. Levinthal, University of Pennsylvania, USA
John F. Padgett, University of Chicago, USA
Flaminio Squazzoni, University of Milan, Italy
Massimo Warglien, Ca’ Foscari University, Italy
After a very stimulating and engaging two days, the opportunity to share our work
with each other, and to make it available to others outside of the workshop participants
came to the fore. Collecting all the extended abstracts in one short outlet seemed like
a good idea that meets these needs of sharing our work both internally and externally.
Among the thirteen, twelve agreed to have their abstract appear in this booklet. The
two days also included two keynote speakers, Alessandro Lomi, University of Ital-
ian Switzerland & University of Exeter, and Martin Neumann, Johannes-Gutenberg-
University Mainz, Institute of Sociology.
Many thanks to all the participants to ABM04, the scientific committee, and to all
the authors who agreed to make their work available.
Siavash Farahbakhsh
Alessandro Narduzzo
Davide Secchi
i
ii
Contents
Agent-Based Modeling as a Prescriptive Tool: Evidences from an Application in
Urban Planning
1
An Organizational Ecology of Garbage Cans 6
A Simulated Minimum-Distance Method for the Calibration of ABMs 9
Challenges of Designing and Implementing Simulation Models of Peer Review 12
Effects of Heterogeneous Cost of Effort in Organizations: Results of an Agent-Based
Simulation on the Role of Task Complexity
16
How motivational Orientation Determines the Pattern of Advice-Seeking
Relationships within Organizational Contexts?
23
Irrational or Risk Averse? A Sticky Expectations Agent-Based Model for Households’
Unemployment Predictions
28
Is Non-Functional Information Necessary to Organizational Adaptation? Typology
and Simulation
33
Learning across Prisoners’ and Hunters’ Games 37
Messy Dynamic Capabilities: Effects of Disorganization on Sensing, Seizing and
Transforming in Top Management Teams
42
Multilevel Dynamics of Emerging Institutions 47
The Cognitive Affiliation to Institutional Norms 51
Annex: Summary 55
iii
Francesco Orsi ABM04
AGENT-BASED MODELING AS A PRESCRIPTIVE TOOL: EVIDENCES FROM AN
APPLICATION IN URBAN PLANNING
Francesco Orsi
francesco.orsi@unitn.it
Department of Civil, Environmental and Mechanical Engineering
University of Trento, Italy
Introduction
Defining the ideal allocation of resources within a complex system is a very difficult task be-
cause any intervention on the system may reverberate through it in articulated and hardly pre-
dictable ways owing to the many interactions among the system’s various components. Agent-
based models (ABMs) have the capability to handle these interactions and may therefore integrate
or even replace more traditional prescriptive techniques like optimization (Barbati et al., 2012).
This has been done in manufacturing process planning (Sabar et al., 2009), transport and logis-
tics (B¨
ocker et al., 2010), facility location (Uno et al., 2008) and landscape planning (Bone and
Dragi´
cevi´
c, 2010).
A field that may considerably benefit from prescriptive ABMs is urban planning and particu-
larly the siting and design of urban green spaces. In fact, designing a city that enables all of its
inhabitants to enjoy easy access to both services and the natural environment is inherently com-
plex because any attempt at guaranteeing such access to a household affects, and is affected by, the
possibility to give other households access to the same elements.
This paper presents an agent-based modeling approach to the problem, where the “ideal” urban
configuration emerges from repeated interactions among several virtual households, each aiming
to find a satisfactory residential location in terms of distance from the center and from green space
(Orsi, 2019). Embedded in the model is a cooperation mechanism by which households choose
1
Francesco Orsi ABM04
a location considering not just their own benefit, but also the burden they impose on their neighbors.
Method
The proposed ABM considers two entities: cells, that is parcels constituting the landscape, and
households, namely mobile agents looking for a satisfactory residential location based on distance
from the center, distance from green space and density. When one or more households are on
a cell, that cell is considered built-up, whereas if a cell is empty, it is considered green space.
Households are assigned a basic additive utility function including the three variables above, and
three parameters whose values are randomly assigned at the beginning of the simulation to reflect
a unique attitude. However, the actual utility a household seeks to maximize during the simulation
is the summation of the basic utility above (individualistic component) and the change in utility
the arrival of the household induces in its neighbors (altruistic component).
The simulation starts with all households randomly spread across the landscape. At each time
step, households check their current location and the eight adjacent locations and move to the one
maximizing their comprehensive utility (i.e. individualistic plus altruistic). The simulation ends
when little or no movements are observed in the population: that represents an achieved acceptable
equilibrium and hence an ideal urban configuration.
Model outputs were compared to those of an optimization model (partly fed with information
supplied by the ABM) to test for efficiency, and used as the input of a simple rent formation model
to assess their likely affordability.
Results
Multiple simulations run with 1000 households showed that the best configurations, in terms
of the maximization of households individual utilities, are achieved when household behavior is
partly cooperative (i.e. 60% individualistic and 40% altruistic) (Figure 1). Instead when behavior
is totally individualistic or totally altruistic, unstable configurations and stable yet unsatisfactory
configurations are obtained, respectively.
2
Francesco Orsi ABM04
Figure 1: Improvement in the summation of all households’ utility with respect to utility in the
initial random configuration for three levels of cooperation: none (0), medium (0.4) and full (1).
The analysis of the outputs’ spatial characteristics showed that the greatest benefits for house-
holds are achieved when density declines for increasing distances from the center, the boundary
between city and countryside is clear and green areas are spread across the entire settlement (Figure
2).
Figure 2: Urban configurations (in 2D and 3D) obtained from the interactions of households having
medium sensitivity to density (a) and low sensitivity to density (b).
3
Francesco Orsi ABM04
A lower sensitivity to density may be compatible with less yet larger green areas and a rougher
rural-urban fringe (Figure 2). Large green areas are to be preferred in anisotropic environments,
where travel is markedly faster along some directions.
Configurations generated by the ABM are better than those of the optimization model by all
measures, including utility improvement, average accessibility to green space and average density
(Figure 3). They are also largely affordable as the poorest half of the population could choose from
nearly 75% of all locations.
Figure 3: Comparison of the configurations generated by the proposed ABM and an “equivalent”
optimization model considering an anisotropic environment (i.e. travel faster along four direc-
tions).
Discussion and conclusion
This study follows in the footsteps of early work on the use of ABMs as planning support sys-
tems (Ligtenberg et al., 2009), introducing the idea that virtual households rather than traditional
planning actors (e.g. administrators, environmental organizations) interact to define an urban con-
figuration. This is clearly not consistent with what happens in the real-world, but has proven to
supply interesting results.
Simulations have shown that cooperation among agents is key to achieving an equilibrium and
seems to confirm the assumption that in many situations the individualistic behavior may only
bring in temporary benefits (Axelrod, 1984). From a planning perspective, this also suggests that
urban interventions at a given location may improve the well-being of the community only if the
conditions of households around the location are adequately accounted for. The application of the
4
Francesco Orsi ABM04
model has unveiled or reaffirmed important considerations for planning, including the importance
of compactness and the widespread distribution of green areas as well as the conditions under
which large green spaces are suitable.
Given its stand-alone nature compared to the equivalent optimization model (which requires
some input to achieve similar results) and its efficiency, the proposed ABM may inform the design
of urban settlements and their densification.
References
Axelrod, R. (1984). The evolution of cooperation. New York: Basic Books.
Barbati, M., Bruno, G., and Genovese, A. (2012). Applications of agent-based models for
optimization problems: A literature review. Expert Systems with Applications,
39(5):6020–6028.
Bone, C. and Dragi´
cevi´
c, S. (2010). Simulation and validation of a reinforcement learning
agent-based model for multi-stakeholder forest management. Computers, Environment and
Urban Systems, 34(2):162–174.
Ligtenberg, A., Beulens, A., Kettenis, D., Bregt, A. K., and Wachowicz, M. (2009). Simulating
knowledge sharing in spatial planning: an agent-based approach. Environment and Planning B:
Planning and Design, 36(4):644–663.
Orsi, F. (2019). Centrally located yet close to nature: A prescriptive agent-based model for urban
design. Computers, Environment and Urban Systems, 73:157–170.
Sabar, M., Montreuil, B., and Frayret, J.-M. (2009). A multi-agent-based approach for personnel
scheduling in assembly centers. Engineering Applications of Artificial Intelligence,
22(7):1080–1088.
Uno, T., Katagiri, H., and Kato, K. (2008). An evolutionary multi-agent based search method for
stackelberg solutions of bilevel facility location problems. International Journal of Innovative
Computing, Information and Control, 4(5):1033–1042.
5
Guido Fioretti ABM04
AN ORGANIZATIONAL ECOLOGY OF GARBAGE CANS
Guido Fioretti
guido.fioretti@unibo.it
Department of Management
University of Bologna
The Garbage Can Model (GCM) proposed by Cohen et al. (1972) focused on one single organi-
zation, which the GCM represents as a sort of vessel where four sorts of agents move, occasionally
meet and eventually make decisions. These fours sorts of agents are the participants to decision
processes, the opportunities to make choices, the solutions available for decision making, and the
problems that must eventually be solved.
Cohen, March and Olsen’s understanding of organizations is peculiar in many respects. How-
ever, its most notable departure from conventional wisdom is the assumption that solutions exist
independently of decision-makers, detached from their minds though floating within organization
space. Conventional wisdom would assume that rational decision makers carefully evaluate avail-
able options formulating appropriate solutions, but GCM decision-makers do not. They rather pick
up whatever solution they find in their organization and apply it to whatever choice opportunity
they happen to be involved in.
We deem that, far from being a shortcoming, this assumption rather reflects a deep conso-
nance between the GCM and the ecological view of organizations. We maintain that the GCM’s
”solutions” are a crude representation of organizational routines, stable patterns of behaviour that
organizations themselves cannot change - hence the similarity with the genome of living beings.
The GCM was presented in 1972. The ecological view of organizations stems from the second half
of the 1970s. Cohen, March and Olsen could not explicitely refer to an approach that did not yet
exist but, possibly, in the 1970s it was in the air somehow. Since we find that these two streams of
organizational theory are coherent with one another, we set out to establish a link.
We constructed a computational model of an organizational ecology governed by GCM-rules.
6
Guido Fioretti ABM04
We still have a vessel where participants, opportunities, solutions and problems move and interact,
but this time, Cohen, March and Olsen’s framework represents society at large, where organi-
zations eventually emerge. In our model, organizations are founded whenever an independent
participant makes a decision: the participant retains the solution that he employed, which becomes
the organization’s characteristic solution (representing its routines).
In our model, organization members make decisions according to GCM rules just like indepen-
dent participants do. Organization members can make decisions by resolution (where they actually
solve a problem) or by oversight (in order to obtain social legitimacy), just like independent par-
ticipants do. They also resort to ’flights’ by either postponing problems or passing them to other
participants, just like independent participants do. Just like the original GCM, participants are
charaterized by their ’ability’, solutions by their ’efficiency’, problems by their ’difficulty’, and
make decisions according to the same rules.
Since organizations make their solution readily available to their members, they provide clear
advantages in terms of the number of decisions that their members can make. A further advan-
tage is that organization members can more easily resort to flights in order to escape from difficult
problems, both because they can postpone them to any choice opportunity in the organization and
because they can pass the buck to colleagues instead of waiting for randomly walking opportuni-
ties or independent participants to come to their place, respectively. Thus, independent participants
are eager to join organizations. Hence we assumed that whenever independent participants jump
on a position where an organization is, they immediately join that organization. Thus organiza-
tions typically grow once they have been founded. These organizations are characterized by two
hierarchical levels: organization founders at the top and organization associates at the bottom.
Organizations need resources in order to operate. Organizations are initially endowed with
an amount of resources equal to the ability of their founder. Later on, each time an organization
member makes a decision by resolution the organization increases its resource endowment by
an amount equal to the difficulty of the problems that it solved. However, at each time step an
organization’s resources decrease by an amount proportional to its size. Eventually, the resources
of organizations that do not solve sufficiently many problems shrink down to zero, in which case
organizations are dissolved. Thus, we have a rich dynamics where organizations emerge, grow and
eventually die.
7
Guido Fioretti ABM04
Our GCM organizational ecology the following three properties of the basic GCM:
1. Decisions by oversight are much more common than decisions by resolution;
2. Decisions by resolution are mostly made at low hierarchical levels;
3. Flights increase efficiency, understood as the number of decisions that are made.
Furthermore, we establish a link with the literature on the liability of newness and adolescence.
In our artificial environment other forms of liability (the liability of smallness, the liability of
obsolescence and the liability of senescence) do not exist by design. In our controlled experiment
we find that:
4. Organizations that are founded with a decision by resolution have a less skewed size distri-
bution and are characterized by a larger mean size than organizations that are founded with
a decision by oversight;
5. Organizations that are founded with a decision by resolution have a longer mean life than
organizations that are founded with a decision by oversight. Thus, organizations that are
founded with a decision by resolution are subject to the liability of adolescence whereas
organizations that are founded with a decision by oversight are rather subject to the liability
of newness.
These findings are coherent with the available empirical literature, which ascribes the liability
of adolescence to hight-tech firms that start with good management teams. In the GCM, decisions
by resolution are most likely made by participants who have a high ability and who are employing
a solution characterized by a high efficiency. Thus, organizations that are founded with a decision
by resolution are likely to receive a higher initial endowment - which is equal to the founder’s
ability - than organizations that are founded with a decision by oversight. Hence the consonance
with the empirical literature.
References
Cohen, M. D., March, J. G., and Olsen, J. P. (1972). A garbage can model of organizational
choice. Administrative Science Quarterly, 17(1):1–25.
8
Raffaello Seri, Mario Martinoli, Davide Secchi & Samuele Centorrino ABM04
A SIMULATED MINIMUM-DISTANCE METHOD FOR THE CALIBRATION OF ABMS
Raffaello Seri
raffaello.seri@uninsubria.it
Insubria University, Italy
Mario Martinoli
mariomartinoli@alice.it
Insubria University, Italy, and Stony Brook University, USA
Davide Secchi
secchi@sdu.dk
Research Centre for Computational & Organizational Cognition
University of Southern Denmark, Denmark
Samuele Centorrino
samuele.centorrino@stonybrook.edu
Stony Brook University, USA
When agent-based models (ABMs) deal with complex adaptive systems, there is always the
possibility that the simulated data reflects the complexity typical of those environments (Edmonds
and Meyer, 2015). For this reason, modelers should equip themselves with tools that are capable
of dealing effectively with complexity, especially when it appears at multiple interactive (some-
times nested) levels, for example, from institutional to organizational and from organizational to
individual (Secchi and Cowley, 2018). Statistical tools are, in this respect, particularly useful. In
this paper, we consider the situation, often encountered in the study of ABMs, in which a modeller
wants to make sure that the output of an ABM is close enough to a historical or real-world trend
or pattern. As simulated data are created out of a mix of inputs and assumptions on the behaviours
of agents in an environment, there can be a number of parameter values that produce patterns that
are close to the one in the empirical data. The choice of the parameter values is often based on
graphical representations in which the combination of parameter values more closely replicating
9
Raffaello Seri, Mario Martinoli, Davide Secchi & Samuele Centorrino ABM04
the observed pattern is chosen as the true one. In this paper, we present a method of simulated
minimum distance that allows to make better informed judgements on the validity of this prefer-
ence. The method can be seen as an estimation technique that tries to identify which values of the
parameters of an ABM provide simulations that are as similar as possible to a set of real data.
The method works as follows. We suppose to have a pattern of evolution over time and a
distance between time series. The pattern is supposed to be fixed. We suppose to identify K
combinations of parameters among which we want to find the one that more closely replicates
the pattern. For each combination of parameters, we simulate nrealizations of the ABM and
we compute the distances of these realizations with respect to the original time pattern. While
other choices are clearly possible, in this paper we choose the combination of the parameters
yielding the smallest empirical average distance. As the number of possible parameters is finite,
this is a case of discrete parameter estimation (see Choirat et al., 2012). It is well known that
the rate of convergence of the algorithm to the true value is exponential, i.e. the probability of
choosing a combination of parameters that does not yield the smallest average distance converges
exponentially fast to 0.
The main problem of this algorithm is that it is sensitive to outliers and is prone to incorrectly
identify a point whenever two combinations of parameters provide very similar patterns. Therefore
we propose two procedures that investigate the sensitivity of the procedure.
The first one is a derivation of the model confidence set of Hansen et al. (2011). Let us denote
as M∗the set of combinations of parameters that yield the smallest average distance. Given a
level of confidence α, the method identifies a set c
M∗
−αof combinations of parameters. When nis
large, the set c
M∗
−αcontains with probability at least 1−αthe set M∗. Therefore the algorithm
identifies a set of combinations of parameters that approximate fairly well the pattern.
We propose to supplement this analysis with a second one. The rate of decrease of the proba-
bilities not corresponding to the combination of parameters yielding the smallest average distance
can be characterized through so-called large deviations principles (see, e.g., Dembo and Zeitouni,
2010). We show how, on the basis of the data, it is possible to characterize the rate of decrease of
these probabilities. This tool can be used to make sensitivity analyses on some combinations of
parameters.
As the techniques explained above appear quite daunting, we then use an existing validated
10
Raffaello Seri, Mario Martinoli, Davide Secchi & Samuele Centorrino ABM04
model – the inquisitiveness model (see. Bardone and Secchi, 2017) – to apply the methods.
References
Bardone, E. and Secchi, D. (2017). Inquisitiveness: Distributing rational thinking. Team
Performance Management, 23(1/2):66–81.
Choirat, C., Seri, R., et al. (2012). Estimation in discrete parameter models. Statistical Science,
27(2):278–293.
Dembo, A. and Zeitouni, O. (2010). Large deviations techniques and applications. In Stochastic
Modelling and Applied Probability. Springer, second edition edition.
Edmonds, B. and Meyer, R. (2015). Simulating social complexity. Springer, second edition
edition.
Hansen, P. R., Lunde, A., and Nason, J. M. (2011). The model confidence set. Econometrica,
79(2):453–497.
Secchi, D. and Cowley, S. J. (2018). Modeling organizational cognition: The case of impact
factor. Journal of Artificial Societies and Social Simulation, 21(1).
11
Thomas Feliciani, Pablo Lucas, Junwen Luo & Kalpana Shankar ABM04
CHALLENGES OF DESIGNING AND IMPLEMENTING SIMULATION MODELS OF
PEER REVIEW
Thomas Feliciani
thomas.feliciani@ucd.ie
Pablo Lucas
pablo.lucas@ucd.ie
School of Sociology & Geary Institute for Public Policy, Ireland
Junwen Luo
junwen.luo@ucd.ie
Kalpana Shankar
kalpana.shankar@ucd.ie
School of Information and Communication studies &
Geary Institute for Public Policy University College Dublin, Ireland
Abstract
Science relies on peer review. Through this mechanism, manuscripts are selected for publica-
tion and grant proposals for funding. However, the processes of peer review do not operate in a
vacuum; they reflect the priorities, norms, and practices of the institutions in which they are em-
bedded, such as scientific communities, funding agencies, publishers, and scholarly societies, each
with their own perspectives and logics (Bollen et al., 2014; Benner and Sandstr¨
om, 2000). Peer
review is a multi-level system. At the macro level a funding agency sets its priorities and goals
for funding based on national priorities and legal mandates. At the meso level, funding agencies
12
Thomas Feliciani, Pablo Lucas, Junwen Luo & Kalpana Shankar ABM04
use peer review to select which proposals to fund, but also integrate their own strategic objectives
(gender balance, geographical diversity, disciplinary needs for example) into the selection process.
At the micro level, individual reviewers and panels bring their own perspectives to bear on the
review processes. In particular, the dynamics of meso- and micro-level complexity provides an
area of exploration that could benefit from simulation studies for two reasons. Simulation studies
help us understand what features of the peer review process emerge from different norms, relation-
ships, attitudes and behaviors of the actors and organizations involved. These methods also allow
us to develop and test policy recommendations for the improvement of peer review in these same
organizations.
In our own project we started by mapping existing simulation models of peer review and iden-
tified knowledge gaps in the literature, then started developing a simulation model to address these
gaps. We found that numerous researchers had studied peer review systems by means of formal
and computational modeling, such as agent-based models (ABM) (Squazzoni and Tak´
acs, 2011).
We counted 44 papers on simulation models of peer review published since 1969: some were used
to compare the efficiencies of alternative peer review systems (e.g. Kovanis et al., 2017); some
compared different behavioral strategies of authors, editors or reviewers (e.g. Thurner and Hanel,
2011; Squazzoni and Gandelli, 2013); some sought the origin of the issues of peer review, such as
biases, high costs and inefficiencies (e.g. Righi and Tak´
acs, 2017).
Reviewing existing models
To perform a scoping review of the literature (in process), we ran queries on Scopus and Web
of Science to find a comprehensive set of publications on simulation models of peer review. We
integrated the set by reference chaining and with papers from our knowledge. We then classified
the models based on the kind of models, the kind of peer review systems, the prominent model
features, and the research questions explored with the model. Besides proposing a taxonomy of
models of peer review, our scoping review identifies some open issues and knowledge gaps.
13
Thomas Feliciani, Pablo Lucas, Junwen Luo & Kalpana Shankar ABM04
First issue: limited model integration
In our review we found a highly fragmented landscape of models, assumptions, and findings.
None of the papers we reviewed attempted to compare previous models, and only in a few cases
was a model further developed after its initial publication. The lack of integration between models
carries important consequences for the generalizability of their findings. This is the case for some
key model assumptions, which constitute the difference between some models. An example is
the assumption that submissions (e.g. manuscripts or grant proposals) have intrinsic, “true” qual-
ity. Whereas some models are built on this assumption, other models assume no intrinsic quality
of these submissions. The intrinsic quality assumption represents a modeler’s perspective about
the role of reviewers: if submissions do have an intrinsic quality, then it is the reviewer’s task to
estimate the quality as accurately as possible. Conversely, assuming no intrinsic quality implies
that reviews are purely subjective, and disagreement between reviewers does not imply that some
reviewers are wrong. We ask: would a model’s predictions be different if we did (or did not) as-
sume that submissions have an intrinsic quality? As models with these two alternative modeling
assumptions have never been systematically compared, we do not know the extent to which their
findings are robust to the assumption.
Second issue: lack of empirical data
Despite growing interest and calls by the computational modeling community for the empirical
calibration and validation of simulation models (Hedstr¨
om and Manzo, 2015), we found that only
a minority of papers made use of empirical data. Access to peer review data is difficult, as it does
require an appropriate management protocol to ensure both confidentiality and anonymity. When
empirical data was used, few model parameters were calibrated based on that, and few model
predictions were compared to the available empirical data.
14
Thomas Feliciani, Pablo Lucas, Junwen Luo & Kalpana Shankar ABM04
Work in progress
We are developing an ABM of peer review process at one national funding agency that: (1)
integrates features from existing relevant models in literature, and (2) is empirically calibrated and
validated with qualitative and quantitative data, including textual data from policy and organiza-
tional documents and interviews.
The integration of features from previous models (1) is done by ‘aligning’ alternative imple-
mentations found in the literature (see e.g. Axtell et al., 1996). This implies that in the simulation
environment of our ABM we are able to compare alternative assumptions/features from previous
work, all other factors being kept constant. This allows us to test the effects of these assumptions
and the robustness of our findings against them.
The use of diverse data sources for ABM calibration and validation (2) will be done by integrat-
ing insights from both qualitative and quantitative data sources. This presents some methodological
challenges. On the one hand, it is clear that expert interviews can be useful to inform a simulation
model on the functioning of the peer review process. On the other hand, there are no guidelines
or best practices for translating interview transcripts into a formal system. Relatedly, it is not clear
how to handle discrepancies between different data sources, e.g. when micro- and meso-level ac-
tors provide contradictory descriptions of peer review practices, or when their description clashes
with the quantitative evidence. With our ongoing modeling work, we are currently considering
ways to overcome these challenges.
References
Axtell, R., Axelrod, R., Epstein, J. M., and Cohen, M. D. (1996). Aligning simulation models: A
case study and results. Computational & Mathematical Organization Theory, 1(2):123–141.
Benner, M. and Sandstr¨
om, U. (2000). Institutionalizing the triple helix: research funding and
norms in the academic system. Research Policy, 29(2):291–301.
Bollen, J., Crandall, D., Junk, D., Ding, Y., and B¨
orner, K. (2014). From funding agencies to
scientific agency: Collective allocation of science funding as an alternative to peer review.
EMBO Reports, 15(2):131–133.
Hedstr¨
om, P. and Manzo, G. (2015). Recent trends in agent-based computational research: A
brief introduction. Sociological Methods & Research, 44(2):179–185.
15
Friederike Wall ABM04
EFFECTS OF HETEROGENEOUS COST OF EFFORT IN ORGANIZATIONS:
RESULTS OF AN AGENT-BASED SIMULATION ON THE ROLE OF TASK
COMPLEXITY
Friederike Wall
friederike.wall@aau.at
Department for Management Control and Strategic Management
University of Klagenfurt, Austria
Keywords: Agent-based simulation, cost of effort, free riding, heterogeneity, imperfect infor-
mation, NK fitness landscapes, task complexity
The cost of effort a decision-maker has to incur for implementing an option is regarded to
reduce the decision-maker’s net “income” and, hence, is widely reflected in economic modelling
of managerial decision-making. For example, much of the literature in principal-agent theory
captures effort as a (quadratically increasing and additively separable) disutility reflecting personal
costs of that contracting party (agent) being in charge of a task; in consequence, for the other
party (principal) it is the more costly to induce a certain level of effort the higher the agent’s cost
of effort (e.g., Baker, 1992; Feltham and Xie, 1994; Lambert, 2001) (e.g.and Xie 1994). Apart
from its particular behavioural assumptions (for a discussion, e.g., Kirman, 1992; Axtell, 2007),
this perspective predominantly emphasizes the “vertical” effects of cost of effort, and, usually, the
overall task and structure of an organization is boiled down to rather reduced settings even if
multi-agent-/multi-task situations are considered (e.g., Bushman et al., 1995, for a discussion with
further references see. Wall (2016)).
Against this background the research endeavour introduced here, seeks to take a different per-
spective by studying the following research question: How do cost of efforts affect an organiza-
tion’s performance when potentially heterogeneous decision-makers search for superior solutions
16
Friederike Wall ABM04
for an overall organizational task and how do interactions among the decision-makers’ sub-tasks
affect the performance obtained?
For investigating the research question an agent-based simulation is employed. In the model,
the task environment of the organizations is represented by fitness landscapes as introduced to
managerial science by Levinthal (1997) and, in particular, the landscapes are modelled according
to the NK framework (Kauffman and Levin, 1987; Kauffman, 1993). A key feature of the NK
framework is that it allows to easily control for the complexity of the task to be performed (Li
et al., 2006). In the simulations, artificial organizations have to solve an N-dimensional binary
decision problem which is decomposed into Mdisjoint sub-problems of equal size - each of which
exclusively assigned to a department r= 1, ..., M (see examples in Figure 1): according to the
NK-framework, Kgives the overall level of interactions among the Nchoices, each of which
contributing to overall performance (for overviews on NK landscapes, e.g., Altenberg, 1997; Li
et al., 2006; Csaszar, 2018). Additionally, Kex denotes the level of cross-departmental interactions
also affected by task complexity in terms of interdependencies (Liu and Li, 2012). As familiar
in agent-based modelling, the department heads show some form of bounded rationality (Simon,
1955,9): (1) they cannot oversee their particular search space at once and employ a stepwise search
for superior options; (2) they cannot anticipate the fellow departments’ choices; (3) their ex ante-
evaluation of options is afflicted with some noise. Altering the status quo of a sub-problem in
favour of an alternative configuration causes cost of effort for department head rquadratically
increasing in the number of binary choices altered and shaped by cost parameter cr. In order to
study the effects of heterogeneous costs of efforts, in the experiments two of three departments
operate at a moderate level of costs, while a third unit’s cost is subject to a variation across five
levels (see Table 1 for parameter settings).
The results (Figure 2) suggest some interesting effects of increasing cost of effort of the “high
cost department” (HCD) (red lines, circle marks, in Figure 2) in interrelation with task complex-
ity: (1) Not only does the average effort of the HCD decrease with its higher cost of effort, but,
for higher levels of complexity, also the effort of the other departments; (2) with increasing task
complexity, the overall organizational performance (blue line, triangle marks) increases the with
increasing HCD’s cost of efforts; (3) The decrease of the HCD’s performance (red line, circle
marks) with increasing cost of effort is subject to task complexity; for higher complexity, the
17
Friederike Wall ABM04
HCD’s performance declines, if at all, are rather low; (4) for higher levels of complexity, the par-
tial performances of those units operating at a moderate cost level remarkably increase the higher
the HCD’s cost of effort.
In sum, results indicate subtle mutual effects between heterogeneous levels in cost of effort of
units and task complexity: Contrary to what intuition may suggest, for higher levels of complexity,
the overall organizational and the other units’ performances raise when the cost of effort of one
department increases and with higher cost of effort of one unit also the other units’ effort declines.
An explanation may lie in the subtle interference of effort cost with the imperfect information
at the units’ sites in cases of cross-unit interactions: head of department rwhen forming pref-
erences without knowing about the intentions of the fellow units may not only be surprised by
the actual performance obtained in the previous period but also by the other units’ choices which,
due to Kex >0, has affected the “own” performance. Thus, this eventually lets unit r’s head
adapt the “own” choices for the next period and so forth leading to frequent time-delayed mutual
adjustments. However, whether unit head rleaves the status quo is shaped by cost of efforts: in
particular, costs of effort apparently cause that only worthwhile alterations are made, i.e., alter-
ations whose perceived performance gains exceed the costs of effort. Hence, reasonably, with one
“high cost department” mutually reinforcing effects that stabilize the search processes are initiated
(further indicators, like the number of alterations, on the adaptive walks support this explanation).
These findings suggest that (heterogeneously) higher cost of effort which may be associated
with being less efficient also may have their beneficial effects for organizational performance
(in a similar vein, see. K¨
olle, 2015). However, these findings call for extending the model and
experiments presented in order to capture some further aspects of collaborative decision-making.
As such, for example, cost-induced reduction of effort by one agent may be regarded as a kind of
free-riding by the other units (Delton et al., 2012; Albanese and Van Fleet, 1985) which may react
with certain punishments. Moreover, organizations may employ further institutional arrangements
to foster cooperative behaviour in terms of inducing effort.
18
Friederike Wall ABM04
Table 1: Parameter settings for the simulations presented in Figure 2
Parameter Values/Type
Observation time T= 250 periods
Number of choices
(decision problem)
N= 12,d= (d1, d12)
Number of departments M= 3 with
(sub-problems) d1= (d1, d4),d2= (d5, d8),d3= (d9, d12)
Interactions structures a. decomposable: N= 12,K= 3,Kex = 0 (see Figure1.a)
b. near decomposable: N= 12,K= 4,Kex = 1
c. non-decomposable: N= 12,K= 7,Kex = 4 (see Figure1.b)
d. non-decomposable: N= 12,K= 8,Kex = 5
Precision of ex-ante
evaluation for
sub-problems
relative error, Gaussian distribution with mean 0and standard
dev.σr= 0.1for each department r, errors independent of each other
Cost coefficients department 1: c1= 0.0025;
department 2: c2= 0.0025
department 3: c3= (0.0025; 0.005; 0.01; 0.015; 0.02)
Simulation runs per
scenario (given by
interaction and cost
coefficient c3)
2500, with 10 simulations on 250 distinct fitness landscapes
Figure 1: Interaction structure of a task decomposable into 3 sub-tasks (a) and an example for a
non-decomposable task (b)
19
Friederike Wall ABM04
Figure 2: Final performance (left column) and average efforts per period (right column) of departments and overall organization for different
levels of task complexity and for different levels of heterogeneity of departments with respect to cost of effort. Heterogeneity is given as the
difference between cost coefficient c3of department 3 and cost coefficients of department 1 or 2, respectively. The performances for the partial
sub-problems and the overall problem (blue line), are normalized to the respective local or global maximum for the respective fitness landscape,
respectively. For parameter settings see Table 1.
20
Friederike Wall ABM04
References
Albanese, R. and Van Fleet, D. D. (1985). Rational behavior in groups: The free-riding tendency.
Academy of Management review, 10(2):244–255.
Altenberg, L. (1997). B2. 7.2. nk fitness landscapes. In B ¨
ack, T., Fogel, D. B., and Michalewicz,
Z., editors, The handbook of evolutionary computation. Oxford University Press.
Axtell, R. L. (2007). What economic agents do: How cognition and interaction lead to emergence
and complexity. The Review of Austrian Economics, 20(2-3):105–122.
Baker, G. P. (1992). Incentive contracts and performance measurement. Journal of political
Economy, 100(3):598–614.
Bushman, R. M., Indjejikian, R. J., and Smith, A. (1995). Aggregate performance measures in
business unit manager compensation: The role of intrafirm interdependencies. Journal of
Accounting research, 33:101–128.
Csaszar, F. A. (2018). A note on how nk landscapes work. Journal of Organization Design,
7(1):15.
Delton, A. W., Cosmides, L., Guemo, M., Robertson, T. E., and Tooby, J. (2012). The
psychosemantics of free riding: dissecting the architecture of a moral concept. Journal of
personality and social psychology, 102(6):1252.
Feltham, G. A. and Xie, J. (1994). Performance measure congruity and diversity in multi-task
principal/agent relations. Accounting Review, pages 429–453.
Kauffman, S. and Levin, S. (1987). Towards a general theory of adaptive walks on rugged
landscapes. Journal of Theoretical Biology, 128(1):11–45.
Kauffman, S. A. (1993). The origins of order: Self-organization and selection in evolution.
Oxford: Oxford University Press.
Kirman, A. P. (1992). Whom or what does the representative individual represent? Journal of
Economic Perspectives, 6(2):117–136.
K¨
olle, F. (2015). Heterogeneity and cooperation: The role of capability and valuation on public
goods provision. Journal of Economic Behavior & Organization, 109:120–134.
Lambert, R. A. (2001). Contracting theory and accounting. Journal of Accounting and
Economics, 32(1-3):3–87.
Levinthal, D. A. (1997). Adaptation on Rugged Landscapes. Management Science,
43(7):934–950.
Li, R., Emmerich, M. T., Eggermont, J., Bovenkamp, E. G., B¨
ack, T., Dijkstra, J., and Reiber,
J. H. (2006). Mixed-integer nk landscapes. In Parallel Problem Solving from Nature-PPSN IX,
pages 42–51. Springer.
21
Friederike Wall ABM04
Liu, P. and Li, Z. (2012). Task complexity: A review and conceptualization framework.
International Journal of Industrial Ergonomics, 42(6):553–568.
Simon, H. A. (1955). A behavioral model of rational choice. The Quarterly Journal of
Economics, 69(1):99–118.
Simon, H. A. (1959). Theories of decision-making in economics and behavioral science. The
American Economic Review, pages 253–283.
Wall, F. (2016). Agent-based modeling in managerial science: an illustrative survey and study.
Review of Managerial Science, 10(1):135–193.
22
Vojkan Nedkovski, Viktor Stojkoski, & Marco Guerci ABM04
HOW MOTIVATIONAL ORIENTATION DETERMINES THE PATTERN OF
ADVICE-SEEKING RELATIONSHIPS WITHIN ORGANIZATIONAL CONTEXTS?
Vojkan Nedkovski
vojkan.nedkovski@unimi.it
Graduate School of Social and Political Sciences
Universit`
a degli studi di Milano, Italy
Viktor Stojkoski
vstojkoski@manu.edu.mk
Macedonian Academy of Sciences and Arts
Marco Guerci
marco.guerci@unimi.it
Graduate School of Social and Political Sciences
Universit`
a degli studi di Milano, Italy
One of the current debates in social network research concerns how personal characteristics
shape the social networks that individual actors construct around them within organizational set-
tings (Kilduff and Brass, 2010). It has been shown, for example, that individuals with high self-
monitoring personality tend to be more prompt in providing work-related advice and help to others
than those with low self-monitoring personality (Flynn et al., 2006). Similarly, individuals that
score high in openness to experience personality trait are more likely to have open friendship net-
works in which their direct contacts are disconnected from each other (L¨
onnqvist et al., 2014).
However, this line of inquiry has largely overlooked a key personal characteristic, that is, in-
dividuals’ motivational orientation. Motivational orientation is, indeed, a key topic in psychology
and organization studies as it describes the motives behind individuals’ actions, and it is fundamen-
tal concept for understanding organizational behavior (Mitchell and Daniels, 2003). The motives
that drive individuals’ engagement into work-related activates can be intrinsic or extrinsic to the
working process itself (Deci and Ryan, 1985; Ryan and Deci, 2000). Thus, intrinsic motivational
23
Vojkan Nedkovski, Viktor Stojkoski, & Marco Guerci ABM04
orientation refers to the desire for engagement into a working activity because of the interest, cu-
riosity and enjoyment of the work itself (Amabile, 1993; Ryan and Deci, 2000). On the other hand,
the extrinsic motivational orientation refers to the desire for engagement into a working activity for
the sake of factors that are external to the work itself, such as rewards, promotions, and recognitions
(Amabile, 1993). Although the extant literature has shown that intrinsic and extrinsic motivational
orientations are key determinants of several desired organizational outcomes such as knowledge
sharing (Andreeva and Sergeeva, 2016), and creativity (Amabile et al., 1994), it remains unclear if
and how they affect the patters of relationships in a social network.
Following the theoretical model of advice-seeking behaviors (Bamberger, 2009; Borgatti and
Cross, 2003), in this paper, we argue that organizational members perceive and value diversely the
intrinsically and extrinsically motivated individuals; therefore, the probability that the focal actor
creates ties with his/her colleagues depends on their motivational orientations.
Specifically, we develop our hypotheses on the basis of three theoretical arguments, related
to the evaluations that the focal actors develop on regard to his/her possible ties in respect to the
quality of advice, the quality of a relationship, as well as the willingness of a person to provide
advice (Borgatti and Cross, 2003). Those evaluations, indeed, are shaped by the perception of the
focal actor in regard to the motivational orientations of his/her possible contacts, and affect the
probability that the focal actor creates a tie with those colleagues.
Our first theoretical argument is based on the fact that the information seeking model (Borgatti
and Cross, 2003), posits that the probability that an individual will seek advice from another person
depends from the perceptions about another’s person expertise and skills. If an ego perceives an
alter as a person who possess knowledge and expertise that can be useful and applicable within the
ego’s work-domain, then the ego is likely to ask advice from that person (Agneessens and Wittek,
2012; Bonaccio and Dalal, 2006; Porter and Woo, 2015). Because of the interest and curiosity to
learn about the tasks at hand, intrinsically motivated individuals will invest more time and efforts
in work-related activities than the individuals who are less intrinsically motivated (Grant, 2008;
Dysvik and Kuvaas, 2013). Accordingly, we argue that organizational members will hold positive
expectations about the quality of advice that can be obtained in the interaction with intrinsically
motivated individuals. The reverse logic holds for individuals who possess high level of extrinsic
motivation. Unlike the individuals with high level of intrinsic motivation, the extrinsically moti-
24
Vojkan Nedkovski, Viktor Stojkoski, & Marco Guerci ABM04
vated individuals will engage into working activities with less enjoyment, interest and curiosity to
learn about the tasks at hand, which will make extrinsically motivated people less prone to per-
sist while solving complex, and unfamiliar problems that require ‘out of box’ thinking (Amabile,
1997). As the information seeking model suggests (Borgatti and Cross, 2003), this will translate
into fewer advice seeking requests from other organizational members.
Our second theoretical argument is related to the quality of the relationship with the potential
advice providers, which is the second key criterion in the advice seeking model (Borgatti and Cross,
2003). In their experimental study, Wild et al. (1997) developed a theoretical model suggesting
that by perceiving the task engagement of a target person to be driven by intrinsic motivation, the
perceiver will expect that the relationship with the target person will be of a greater quality, and
that performing the same task as the target person will be followed by feelings of pleasure and
enjoyment. The effects are reversed when the participants in the experiment were exposed to an
extrinsically motivated individual. The participants showed lower expectations about the quality of
the relationship and experienced fewer feelings of pleasure and enjoyment for engaging into tasks
that were performed by extrinsically motivated individuals.
Our third theoretical argument regards the accessibility, and willingness of an advice provider
to meet the requests of the advice seeker, which is the third criterion upon which the advice seekers
will decide who to choose as a potential advice provider (Borgatti and Cross, 2003). The research
on knowledge sharing suggests that intrinsic, rather than extrinsic motivation, is a better predictor
of tacit knowledge sharing (Jeon et al., 2011; Ozlati, 2015). Conversely, the individuals who are
moved to act because of promises for material benefits such as rewards and compensation, were
found to withhold their knowledge and expertise (Stenius et al., 2016). According to the informa-
tion seeking model (Borgatti and Cross, 2003), this increases the costs for the other organizational
members to seek advice from people with high level of extrinsic motivation.
Taking into consideration the three above presented theoretical arguments, we advance the
following Hypotheses:
HP 1: Individuals with high level of intrinsic motivation are more likely than individuals with
low level of intrinsic motivation to be sought out for work-related advice.
HP 2: Individuals with low level of extrinsic motivation are more likely than individuals with
high level of extrinsic motivation to be sought out for work-related advice.
25
Vojkan Nedkovski, Viktor Stojkoski, & Marco Guerci ABM04
In keeping with this view, here we develop an agent-based model where agents seek to form
advice relationships based on the motivational orientations of their actors, as their motivational
orientation affects the perceived quality of advice, the quality of the relationship, and willingness
of a third party to share advice. In particular, we assume simple dynamics where in every round
of interaction each agent seeks advice from another randomly chosen agent. The probability with
which he/she chooses a specific agent is proportional to its previous experience with that agent,
whereas the quality of the advice, quality of the relationships, and willingness of the third party
to share advice will be determined by advice provider’s intrinsic and extrinsic motivation. We
test the model predictions on an empirical dataset capturing 1,988 (overall density = 0.11) advice
ties among all the 134 employees in a mid-size consulting company. Within the same dataset, we
also collected data about employees’ work motivation, and the response rate was 88% (N= 118).
We show that the model adequately reproduces the characteristics of the empirical network and it
provides a better depiction on the reality when compared to other network-generation models.
References
Agneessens, F. and Wittek, R. (2012). Where do intra-organizational advice relations come from?
the role of informal status and social capital in social exchange. Social Networks,
34(3):333–345.
Amabile, T. M. (1993). Motivational synergy: Toward new conceptualizations of intrinsic and
extrinsic motivation in the workplace. Human Resource Management Review, 3(3):185–201.
Amabile, T. M. (1997). Motivating creativity in organizations: On doing what you love and
loving what you do. California Management Review, 40(1):39–58.
Amabile, T. M., Hill, K. G., Hennessey, B. A., and Tighe, E. M. (1994). The work preference
inventory: assessing intrinsic and extrinsic motivational orientations. Journal of personality
and social psychology, 66(5):950.
Andreeva, T. and Sergeeva, A. (2016). The more the better or is it? the contradictory effects of hr
practices on knowledge-sharing motivation and behaviour. Human Resource Management
Journal, 26(2):151–171.
Bamberger, P. (2009). Employee help-seeking: Antecedents, consequences and new insights for
future research. In Research in personnel and human resources management, pages 49–98.
Emerald Group Publishing Limited.
Bonaccio, S. and Dalal, R. S. (2006). Advice taking and decision-making: An integrative
literature review, and implications for the organizational sciences. Organizational Behavior
and Human Decision Processes, 101(2):127–151.
26
Vojkan Nedkovski, Viktor Stojkoski, & Marco Guerci ABM04
Borgatti, S. P. and Cross, R. (2003). A relational view of information seeking and learning in
social networks. Management Science, 49(4):432–445.
Deci, E. L. and Ryan, R. M. (1985). The general causality orientations scale: Self-determination
in personality. Journal of Research in Personality, 19(2):109–134.
Dysvik, A. and Kuvaas, B. (2013). Intrinsic and extrinsic motivation as predictors of work effort:
The moderating role of achievement goals. British Journal of Social Psychology,
52(3):412–430.
Flynn, F. J., Reagans, R. E., Amanatullah, E. T., and Ames, D. R. (2006). Helping one’s way to
the top: self-monitors achieve status by helping others and knowing who helps whom. Journal
of Personality and Social Psychology, 91(6):1123.
Grant, A. M. (2008). Does intrinsic motivation fuel the prosocial fire? motivational synergy in
predicting persistence, performance, and productivity. Journal of Applied Psychology,
93(1):48.
Jeon, S., Kim, Y.-G., and Koh, J. (2011). An integrative model for knowledge sharing in
communities-of-practice. Journal of Knowledge Management, 15(2):251–269.
Kilduff, M. and Brass, D. J. (2010). Organizational social network research: Core ideas and key
debates. Academy of Management Annals, 4(1):317–357.
L¨
onnqvist, J.-E., Itkonen, J. V., Verkasalo, M., and Poutvaara, P. (2014). The five-factor model of
personality and degree and transitivity of facebook social networks. Journal of Research in
Personality, 50:98–101.
Mitchell, T. and Daniels, D. (2003). Motivation, volume 12. New York: John Wiley.
Ozlati, S. (2015). The moderating effect of trust on the relationship between autonomy and
knowledge sharing: A national multi-industry survey of knowledge workers. Knowledge and
Process Management, 22(3):191–205.
Porter, C. M. and Woo, S. E. (2015). Untangling the networking phenomenon: A dynamic
psychological perspective on how and why people network. Journal of Management,
41(5):1477–1500.
Ryan, R. M. and Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic definitions and
new directions. Contemporary Educational Psychology, 25(1):54–67.
Stenius, M., Hankonen, N., Ravaja, N., and Haukkala, A. (2016). Why share expertise? a closer
look at the quality of motivation to share or withhold knowledge. Journal of Knowledge
Management, 20(2):181–198.
Wild, T. C., Enzle, M. E., Nix, G., and Deci, E. L. (1997). Perceiving others as intrinsically or
extrinsically motivated: Effects on expectancy formation and task engagement. Personality and
Social Psychology Bulletin, 23(8):837–848.
27
Luca Gerotto & Paolo Pellizzari ABM04
IRRATIONAL OR RISK AVERSE? A STICKY EXPECTATIONS AGENT-BASED
MODEL FOR HOUSEHOLDS’ UNEMPLOYMENT PREDICTIONS
Luca Gerotto∗
Department of Economics, Ca’ Foscari University of Venice, Italy
OCPI, Universit Cattolica del Sacro Cuore, Italy
luca.gerotto@unive.it
Paolo Pellizzari
Department of Economics, Ca’ Foscari University of Venice, Italy
paolop@unive.it
Keywords: Agent-based models; Bounded rationality; Risk aversion; Sticky expectations; Un-
employment expectations.
Empirical evidence suggests that the expectations of households are systematically biased with
respect to the ones of the experts and also with respect to the real state of the world. Models featur-
ing a staggered update of information, like Carroll (2003, 2006), provide reasons for the relatively
slow aggregate changes and the cross-sectional dispersion in households expectations, but still fail
to explain the presence of any systematic over- or under-estimation of a given macroeconomic or
financial variable. For example, Coibion et al. (2018) reports that in the years 2000s inflation ex-
pectations of households were on average 3.5%, much higher than the ones of professionals and
financial market participants (2%). Noteworthy, also top executives surveyed in 2018 had inflation
expectations that were close to the ones of households and much higher than the ones made by
experienced forecasters. Furthermore, a cursory look at unemployment expectations data from the
Consumer Survey run by the European Commission shows that households seldom expected the
unemployment rate to decrease, even though there have been prolonged periods in which unem-
ployment has actually decreased.
∗Corresponding author. Address: Department of Economics, Ca’ Foscari University, Cannaregio 873, Fondamenta
San Giobbe, 30121, Venice, Italy. The views expressed in the paper are those of the authors and do not involve the
responsibility of the institutions they are affiliated with.
28
Luca Gerotto & Paolo Pellizzari ABM04
Part of this evidence could be related to irrationality, stubbornness, or to the different expecta-
tions that households with different demographic characteristics tend to have. As reported in Pe-
saran and Weale (2006), in the literature there are several studies, mostly concerning the Survey of
Consumers run by the University of Michigan, that try to understand whether there are systematic
differences in survey expectations among different groups. Dominitz and Manski (2011) present
summary statistics from the Michigan Survey concerning expectations of a positive nominal equity
return. They find that men are on average more optimistic than women, and that optimism increases
with education and decreases with age. Similar results have been obtained by Bryan et al. (2001),
who study inflation expectations. For UK, Blanchflower and MacCoille (2009) find that infla-
tion expectations rise with age and decrease for the more educated and among home-owners, who
also have more precise one-year expectations. In this regard, Souleles (2004) tests for systematic
demographic components in households’ forecast error in the Michigan Survey. He finds that de-
mographic variables are jointly significant both from a statistical and from an economic point of
view, reporting that “the inflation forecast error is about 0.4 percentage points larger in magnitude
for those without high school education, relative to those with high school education”. The error
tends to decrease with income and age, and to be larger also for females with respect to males and
for members of racial minorities with respect to whites. There are also some common patterns of
heterogeneity in the degree of financial literacy around the world (Lusardi and Mitchell, 2011): the
more educated people are the more informed, women are less financially literate than men, and the
older population tends to be overconfident on his knowledge.
A study that takes heterogeneity into account focusing on Italian data is Easaw et al. (2013).
The authors analyse households inflation expectation from February 2003 to October 2010. They
find that, as reported also in Malgarini (2009), “expected inflation decreases with age and educa-
tion [...] and women expect higher rates of inflation than men”. Moreover, in a sticky information
framework similar to the one of Mankiw and Reis (2002) and Carroll (2003), they also find that
the frequency of update of information is heterogeneous as a function of the demographic charac-
teristics of the individual. The more educated have also the highest absorption rate. Similarly, the
self-employed inform themselves more frequently, while there is mixed evidence on the working
status, which appears to depend on the interaction with the education level.
In this paper we develop a model featuring a staggered update of information together with the
29
Luca Gerotto & Paolo Pellizzari ABM04
novel idea that a risk-averse economic agent may, in un uncertain environment, over(under)-predict
a variable in order to hedge against a costly under(over)-prediction (Capistr´
an and Timmermann,
2009). Put boldly, some agents may (cautiously) introduce biases in their forecasts. In particular,
the idea is that not all individuals have absorbed and processed the most recent information, and
the less an individual is informed the more he may “choose” to bias upwards (or downwards), or be
inclined to bias, his forecast. The presence of these not-fully-informed individuals leads to a sys-
tematic over(under)-estimation in the aggregate households expectation. Whether this adjustment
process is rational or unconscious is open to debate.
In the empirical part, we concentrate on unemployment expectations of Italian households.
Unemployment expectations are a proxy for zero-income probability (Carroll et al., 1992) and it
is reasonable to assume any risk-averse individual would prefer to over-predict the risk of getting
no income rather than under-predicting it; conversely, for other indicators (i.e.: the inflation rate)
there are weaker theoretical reasons to assume that all individuals should optimally bias the forecast
in the same direction. Moreover, unemployment expectations have been shown to be significant
drivers of the consumption/saving behaviour (Carroll et al., 2012, 2014).
Our calibrated agent-based model reveals, firstly, that empirical data on households expecta-
tions can be explained by a combination of infrequent update of information and asymmetric loss
aversion in a way that is consistent with the notion that unemployment expectations are upward
biased for hedging reasons. Secondly, we take into account the expectations of different demo-
graphic groups and test if these differences can be explained by a different frequency of update of
information, as suggested by Easaw et al. (2013), and by a different intensity of the hedging motive
(related to the degree of risk aversion). We find that the inclusion of heterogeneity in the model
allows for a very good fit of the data and there is strong evidence that the main driver of the differ-
ences in expectations between different groups are the different frequencies at which they collect
information. In particular, the frequency of update of information is increasing in the education
level, is larger for men than for women, is larger for working people than for non-working people
and is lower in the Southern regions of Italy than in the rest of the country.
30
Luca Gerotto & Paolo Pellizzari ABM04
References
Blanchflower, D. G. and MacCoille, C. (2009). The formation of inflation expectations: an
empirical analysis for the UK. NBER Working Paper 15388, Cambridge, MA: National Bureau
of Economic Research.
Bryan, M. F., Venkatu, G., et al. (2001). The demographics of inflation opinion surveys. Federal
Reserve Bank of Cleveland, Research Department.
Capistr´
an, C. and Timmermann, A. (2009). Disagreement and biases in inflation expectations.
Journal of Money, Credit and Banking, 41(2-3):365–396.
Carroll, C., Slacalek, J., and Sommer, M. (2012). Dissecting saving dynamics: measuring credit,
wealth and precautionary effects. Available at: http://www. econ2. jhu.
edu/people/ccarroll/papers/cssUSSaving. pdf.
Carroll, C. D. (2003). Macroeconomic expectations of households and professional forecasters.
The Quarterly Journal of Economics, 118(1):269–298.
Carroll, C. D. (2006). The Epidemiology of Macroeconomic Expectations. In Blume, L. E. and
Durlauf, S. N., editors, The Economy as an Evolving Complex System, III: Current
Perspectives and Future Directions, pages 5–29. Oxford University Press, Oxford, New York.
Carroll, C. D., Hall, R. E., and Zeldes, S. P. (1992). The buffer-stock theory of saving: Some
macroeconomic evidence. Brookings papers on economic activity, 1992(2):61–156.
Carroll, C. D., Slacalek, J., and Tokuoka, K. (2014). The distribution of wealth and the mpc:
implications of new european data. American Economic Review, 104(5):107–11.
Coibion, O., Gorodnichenko, Y., Kumar, S., and Pedemonte, M. (2018). Inflation expectations as
a policy tool? Technical report, National Bureau of Economic Research.
Dominitz, J. and Manski, C. F. (2011). Measuring and interpreting expectations of equity returns.
Journal of Applied Econometrics, 26(3):352–370.
Easaw, J., Golinelli, R., and Malgarini, M. (2013). What determines households inflation
expectations? Theory and evidence from a household survey. European Economic Review,
61:1–13.
Lusardi, A. and Mitchell, O. S. (2011). Financial literacy around the world: an overview. Journal
of pension economics & finance, 10(4):497–508.
Malgarini, M. (2009). Quantitative inflation perceptions and expectations of Italian consumers.
Giornale degli Economisti e Annali di Economia, pages 53–80.
Mankiw, N. G. and Reis, R. (2002). Sticky information versus sticky prices: a proposal to replace
the New Keynesian Phillips curve. The Quarterly Journal of Economics, 117(4):1295–1328.
Pesaran, M. H. and Weale, M. (2006). Survey expectations. Handbook of economic forecasting,
1:715–776.
31
Luca Gerotto & Paolo Pellizzari ABM04
Souleles, N. S. (2004). Expectations, heterogeneous forecast errors, and consumption: Micro
evidence from the Michigan consumer sentiment surveys. Journal of Money, Credit, and
Banking, 36(1):39–72.
32
Davide Secchi & Stephen J. Cowley ABM04
IS NON-FUNCTIONAL INFORMATION NECESSARY TO ORGANIZATIONAL
ADAPTATION? TYPOLOGY AND SIMULATION
Davide Secchi
secchi@sdu.dk
Stephen J. Cowley
cowley@sdu.dk
Research Centre for Computational & Organizational Cognition
University of Southern Denmark, Denmark
The study of management and organizations in general has historically been mostly a study of
success. Whether it is about achieving “superior” performance, leading effectively, executing tasks
optimally, or solving problems efficiently, scholars have usually emphasized positive results (e.g.,
Alvesson and Spicer, 2012; Abrahamson and Freedman, 2013). It is, of course, important to accu-
mulate knowledge on the path to success and, as a result of this, there is quite a significant amount
of research on various aspects of information “handling” (from the work of Chester Barnard on-
wards). For example, streams of research have started from how to search for information (e.g.,
Levinthal and March, 1981), how information is processed (e.g., Simon, 2013), the flaws of the
process (e.g., Kahneman, 2003; Gigerenzer and Selten, 2001), and ways to correct them (e.g.,
Thaler and Sunstein, 2008). In other words, there has been a focus on functional information and
its use in various aspects of organizational life.
However, as some have started to point out (Alvesson and Spicer, 2012; Abrahamson and
Freedman, 2013), information is not always functional. Large portions of what is available and
screened while individuals make decisions may end up to be irrelevant. This is a well known
fact in organizational research, and some have referred to it by calling it different names such
as ambiguity (e.g., March, 1978; Einhorn and Hogarth, 1986), uncertainty (e.g., Michel, 2007),
or disorder (e.g., Abrahamson, 2002). From a functional perspective, the smallest proportion of
33
Davide Secchi & Stephen J. Cowley ABM04
information deemed irrelevant is an estimate of the efficiency with which a decision is made. But,
from an experiential point of view, the path towards a decision may reveal to be useful in the
future, even if most of the information may not be. Hence, a question one may ask is how could
information that is not functional be classified, and what would its use be. We propose here three
categories: (a) seemingly functional, (b) dysfunctional, and (c) irrelevant information. The first
type of information describes a situation where information is judged to be relevant to the task,
decision or problem but it ultimately does not fit or cannot be made to fit by those working on
it. The misjudgment could have multiple causes, and be driven by either a superficial analysis, a
lack of comprehension, or a simple mismatch between means and ends. The second type is termed
dysfunctional because it disrupts the process from happening, either aggravating the problem or
sprouting new ones. The third category, named irrelevant, refers to information that pertains to
tasks, problems, and decisions different than the one at hand.
All of the arguments above have been produced by having in mind the relevance/irrelevance as
it refers to, for example, a task, a problem, or a decision. And yet, there are cognitive qualifiers that
shall not be ignored when referring to the exploitation of information (Magnani, 2007). We refer
to these as embedded in the social organizational environment, usually apparent through coopera-
tion, collaboration, coordination, shared objectives, common purpose, and interaction (Secchi and
Cowley, 2018).
The ways in which individuals make sense of their surroundings while they perform activities
is embedded into the so-called social organizing (Secchi and Cowley, 2018). In other words, we
are claiming that the categorization proposed above need to be necessarily reconciled with cogni-
tive processes, and that the differences between the four types (including functional information) is
dynamic and adaptive, if cognition is defined socially (and organizationally). This process makes
the boundaries between the four categories undefined and blurry. This means that a piece of in-
formation cannot be considered in vacuum but it must be always referred to the cognitive process
that qualifies it. Moreover, even when this match is performed, how much the categorization still
applies depends on the systemic convergence of organizational cognitive processes (Cowley et al.,
2017).
In short, by studying these three types of non-functional information, we are arguing that orga-
34
Davide Secchi & Stephen J. Cowley ABM04
nizational adaptation1can be thought in terms of the tolerance that an organization has for them.
As a corollary to this statement, we are also claiming that this tolerance does not rely simply on
organizational processes, procedures or routines, but it must rely on the behavioral nature of so-
cial interactions. These are rooted to what we have defined elsewhere as social organizing, or to
the dynamic and adaptive locus where individuals are able to meaningfully create knowledge i.e.
where cognition happens in organizations (Secchi and Cowley, 2018).
An agent-based computational simulation (Edmonds and Meyer, 2017) explores the extent
to which functional, seemingly functional, dysfunctional, and irrelevant information constitute a
source of adaptation for an organization. As we have done in previous simulations (Secchi and
Gullekson, 2016; Secchi and Cowley, 2018; Bardone and Secchi, 2017), we assume that teams are
organized through systemic e-cognition (as a way to represent social organizing). The simulation
presents external shocks, defined by variants of institutional pressure that changes the general
macro structure governing relationships within the organization. Agents in this simulation are
employees that choose to interact, cooperate, and act to solve one or more tasks. Each agent-
employee works in a team and has two sets of qualifications that are defined as task-related and
social/organization-related. These two sets identify the agent’s bounded/extendable rationality
(Secchi, 2011) and constitute the way in which information is read/perceived, partly by the agent
as individual, partly as member of the team, and partly as a member of the organization as a whole.
The employees as well as the teams and the organization have resilience and keep trace of history,
in the form of discarded non-functional information (as well as functional).
From the simulation, we expect to find that some of the non-functional types may reveal to play
a role in adaptation, if not it in the short, at least in the long run.
References
Abrahamson, E. (2002). Disorganization theory and disorganizational behavior: Towards an
etiology of messes. Research in Organizational Behavior, 24:139–180.
Abrahamson, E. and Freedman, D. H. (2013). A perfect mess: The hidden benefits of disorder.
Hachette UK.
1We intend this not in the biological sense, but more broadly in terms of adjustments to both internal and environ-
ment change.
35
Davide Secchi & Stephen J. Cowley ABM04
Alvesson, M. and Spicer, A. (2012). A stupidity-based theory of organizations. Journal of
management studies, 49(7):1194–1220.
Bardone, E. and Secchi, D. (2017). Inquisitiveness: Distributing rational thinking. Team
Performance Management, 23(1/2):66–81.
Cowley, S. J., Vall´
ee-Tourangeau, F., and Vall´
ee-Tourangeau, F. (2017). Cognition beyond the
brain. Springer.
Edmonds, B. and Meyer, R. (2017). Simulating Social Complexity. A Handbook. Heidelberg:
Springer, second edition.
Einhorn, H. J. and Hogarth, R. M. (1986). Decision making under ambiguity. Journal of
Business, pages S225–S250.
Gigerenzer, G. and Selten, R. (2001). Bounded rationality: The adaptive toolbox. MIT Press.
Kahneman, D. (2003). A perspective on judgment and choice: mapping bounded rationality.
American Psychologist, 58(9):697.
Levinthal, D. and March, J. G. (1981). A model of adaptive organizational search. Journal of
Economic behavior & Organization, 2(4):307–333.
Magnani, L. (2007). Morality in a technological world: Knowledge as duty. Cambridge
University Press.
March, J. G. (1978). Bounded rationality, ambiguity, and the engineering of choice. The Bell
Journal of Economics, pages 587–608.
Michel, A. A. (2007). A distributed cognition perspective on newcomers’ change processes: The
management of cognitive uncertainty in two investment banks. Administrative Science
Quarterly, 52(4):507–557.
Secchi, D. (2011). Extendable Rationality - Understanding Decision Making in Organizations.
Springer.
Secchi, D. and Cowley, S. J. (2018). Modeling organizational cognition: The case of impact
factor. Journal of Artificial Societies and Social Simulation, 21(1).
Secchi, D. and Gullekson, N. (2016). Individual and organizational conditions for the emergence
and evolution of bandwagons. Computational and Mathematical Organization Theory,
22(1):88–133.
Simon, H. A. (1947/2013). Administrative behavior. Simon and Schuster.
Thaler, R. H. and Sunstein, C. R. (2008). Nudge: Improving decisions about health, wealth, and
happiness. Yale University Press.
36
Marco LiCalzi & Roland M¨
uhlenbernd ABM04
LEARNING ACROSS PRISONERS’ AND HUNTERS’ GAMES∗
Marco LiCalzi
licalzi@unive.it
Roland M¨
uhlenbernd
roland.muhlenbernd@unive.it
Department of Management, Universit Ca’ Foscari Venezia, Italy
Introduction. This paper studies agents who face a stream of similar (but not identical) 2×2
games. We are interested in modeling the empirical evidence about which actions agents play and
how their choices change over time across similar games. The key ideas in our approach are four:
1) agents categorize games into two sets, one per each pure action, and play the corresponding
action; 2) the choice of the category is stochastic; 3) the categorizing propensities are proportional
to linear functions of three game features; 4) the functions weights change over time in response to
the observed play. We test the model across a wide set of laboratory experiments from the literature
and find a very good fit for the main stylized facts.
Game descriptors and individual motivations. Given a 2×2game, consider the payoff matrix
for the row player:
C D
C a b
D c d
∗THIS PROJECT HAS RECEIVED FUNDING FROM THE EUROPEAN UNION’S HORIZON 2020 RE-
SEARCH AND INNOVATION PROGRAMME UNDER GRANT AGREEMENT NO. 732942.
37
Marco LiCalzi & Roland M¨
uhlenbernd ABM04
The actions are conventionally labelled C (cooperation) and D (defection). A Prisoners’ Dilemma
(PD) and a Stag Hung are characterized by the preference order c>a>d>band a > c ≥d>b,
respectively.
In a meta-study of 96 laboratory experiments on the Prisoners’ Dilemmas (PD), published or
carried out between 1967 and 2014, Mengel (2017) identifies three key variables for predicting
cooperation in PDs. She calls them temptation (TEMPT), risk (RISK) and efficiency (EFF). Each
of these three descriptors takes values in [0,1]. They pertain to the game under play: we say that
they are extrinsic.
We assume that, when agents evaluate a strategic decision in a PD or SH game, they are influ-
enced by the strength of three individual motivations:
fear: an impulse to avoid the lowest individual payoff of the game, given by b;
greed: an impulse to achieve the highest individual payoff of the game, given by cin PD and ain
SH;
harmony: an impulse to coordinate on the highest common payoff, given by a.
The relative strength of each impulse is denoted by f, g, h, whereby 0≤f , g, h ≤1and f+g+h=
1. The three impulses pertain to the individual agent: we say that they are intrinsic.
The three extrinsic descriptors (RISK, TEMPT, EFF) and the three intrinsic impulses (fear,
greed, harmony) interact in ordered pairs: RISK with fear, TEMPT with greed, and EFF with
harmony. We model their pairwise complementarity using a simple product operator. Each product
generates a disposition in [0,1]:
dR=f·Ris the disposition to avoid the lowest individual payoff;
dT=g·Tis the disposition to achieve the highest individual payoff;
dE=h·Eis the disposition to coordinate on the highest common payoff.
Individual choice over a game. An agent confronting a game has to decide whether to play C
or D. Following Erev and Roth (1998), our descriptive model for the laboratory evidence attempts
to predict the probability of choice for each action, and compares predicted and observed behavior
by computing the mean squared deviation (MSE).
We posit that the agent’s attraction to either choice depends on the three dispositions aroused
by the game. The attraction towards C is increasing in dEand decreasing in dRfor both PD and
38
Marco LiCalzi & Roland M¨
uhlenbernd ABM04
SH. On the other hand, the attraction towards C is decreasing in dTfor PD but increasing for SH.
Then the probability Pi(C)that an agent iwith weights (f, g, h)ichooses C in a game Gwith
descriptors (T, R, E)Gis proportional to dEin a PD and to dE+dTin a SH.
Reinforcement learning. This rule of probabilistic choice applies for an agent facing a game
Gat a given period. Our model let intrinsic motivations change over time in accordance with
observed past play (over the same game or over similar games). We model this updating process
using reinforcement learning.
The update rule is based on three elements, common to both PD and SH games. First, after
a recent history of defecting opponents, the agent becomes more concerned about avoiding the
lowest payoff and thus we increase the weight ffor fear. Second, after a recent history of cooper-
ating opponents, the agent feels more tempted to achieve the highest payoff and thus we increase
the weight gfor greed. Third, after a recent history of inconclusive evidence (some opponents
defected, other cooperated), the agent experiences more strategic uncertainty: because there seems
to be no settled convention yet, the common payoffs on the main diagonal provide some guidance
about the relative benefits of coordinating on one.
Our implementation of these ideas aims for simplicity and robustness, not for accuracy. We
assume that an agent estimates opponents’ behavior using only the last two rounds of interaction
with them. Given a (PD or SH) game Gwith descriptors (T, R, E)Gand the last two observed
opponents’ actions (st−1, st), the vector of impulse strengths (f, g, h)ifor player iis updated to a
new vector (f0, g0, h0)iusing the rule:
(f0, g0, h0)i=
f+γR, g −γR
2, hi−γR
2i
if st−1=st=D
f−γF
2, g +γF, h −γf
2, hi−γR
2i
if st−1=st=C
f−γE
2, g −γE
2, hi+γEi
otherwise
This update rule ensures that the revised weights always sum to 1, but allow single weights to
become negative. Our implementation ensures that this cannot occur, but for the sake of readability
we omit details.
39
Marco LiCalzi & Roland M¨
uhlenbernd ABM04
PD: Experiments and results. We test our learning model for PD against the laboratory evi-
dence collected by others. We use the 28 games collected in Mengel (2018) that are based on
the random matching protocol, where agents play the same game over several periods against ran-
domly chosen opponents. The main stylized fact is that the initial cooperation rate is relatively
high, but it declines over time.
We calibrate the initial weights for the stochastic choice rule to match the available evidence
on the initial cooperation rate, by minimizing the mean squared error (MSE) between observations
and predictions. Then we apply the update rule with a (non-calibrated) learning rate γ= 0.2.
The initial calibration considers the average initial cooperation rate over the 28 PD games.
We compute a triple (f∗, g∗, h∗)that minimizes the mean squared error between the observed
average cooperation rate πGand the cooperation probability PG(C)predicted by the model across
28 games. We find that the optimal initial weights are f∗= 0.36, g∗= 0.28, and h∗
i= 0.36. Using
these weights entails an average squared error of 0.0218, with a Pearson correlation of 0.691 across
the 28 games.
Experiment I concerns the PD games. We first ran 100 simulations for each of the 28 games,
over the same number of periods used in the original experiment. Each simulation started with the
optimal initial weights (f∗, g∗, h∗)for PD games and a learning rate γ= 0.2. We computed the
average cooperation rates for each game across time (averaged across simulations) and we com-
pared it against the average cooperation rate (averaged across players) in the original experiments:
the Pearson correlation is 0.566.
We also computed the trend in weights. Generally speaking, the weight for fear increases over
time, while the weights for greed and harmony decrease. This implies that the cooperation rate
decreases over time, matching the main stylized fact. Intuitively, people learn to focus more on the
motivation associated with the opponent’s defection as players defect more and more.
SH: Experiments and results. We tested our model for SH against a database of 22 games from
experimental studies conducted between 1995 and 2008. The database contains 22 games but only
17 are different with respect to payoffs, because 5 of them were used twice in different studies.
Using the same methodology, we calibrate the initial weights for the stochastic choice rule
over SH: the optimal initial weights are f∗
i= 0.22, g∗
i= 0.13, and h∗
i= 0.65. Using these weights
40
Marco LiCalzi & Roland M¨
uhlenbernd ABM04
entails an average squared error of 0.0234, with a Pearson correlation of 0.597 across the 22 games.
Experiment II concerns the SH games. We ran 100 simulations for each of the 22 games,
over the same number of periods as in the original experiment. Each simulation started with the
optimal initial weights (f∗, g∗, h∗)for SH games and a learning rate γ= 0.2. We computed the
final cooperation rate for each game (averaged across simulations) and we compared it against the
final cooperation rate (average across players) in the original experiments: the Pearson correlation
is 0.835. Generally speaking, the trend in weights is that harmony always decreases while either
fear or greed increases, depending on whether agents learn to cooperate or defect. Intuitively,
people learn to focus less on harmony and more on the motivation associated with the opponent’s
emerging convention.
References
Erev, I. and Roth, A. E. (1998). Predicting how people play games: Reinforcement learning in
experimental games with unique, mixed strategy equilibria. American economic review, pages
848–881.
Mengel, F. (2017). Risk and temptation: A meta-study on prisoner’s dilemma games. The
Economic Journal, 128(616):3182–3209.
41
Dinuka B. Herath ABM04
MESSY DYNAMIC CAPABILITIES: EFFECTS OF DISORGANIZATION ON SENSING,
SEIZING AND TRANSFORMING IN TOP MANAGEMENT TEAMS
Dinuka B. Herath
D.herath@hud.ac.uk
The University of Huddersfield
Dealing with complexity is an everyday occurrence in modern firms (Reeves et al., 2016).
However, many organizations find this exercise tedious and difficult (Frank et al., 2017). A firm’s
inability to deal with complexity can be severely detrimental to its survival (Mouzelis, 2017). In
modern times, market winners and losers are often determined by the agility and adaptability of
firms (Reeves and Deimler, 2009). Firms which can seize and capitalize on opportunities faster
than their competitors have a better chance of success and survival (Peteraf et al., 2013). Given its
importance, the study of how firms constantly improve the ways in which they manage complexity
and adapt both to its indigenous and exogenous pressures has become an important area of inquiry
(Suddaby and Greenwood, 2009; Secchi and Cowley, 2018). In the vein, a multitude of approaches
has been developed to enhance firm adaptability (Padgett and Powell, 2012). Dynamic capabilities
(DC), is one such branch with a rich body of knowledge behind it (Teece et al., 2016). In its most
rudimentary form, dynamic capabilities can be seen as those capabilities that an organization needs
to sustain competitive advantage within complex environments (Eisenhardt and Martin, 2000).
While there is a healthy amount of work covering dynamic capabilities and their importance
for organizations, there still remain some important unanswered questions. Teece (2018) points
outs that there is the need for research which focuses on “specific aspects of dynamic capabilities”
which, includes a focus on “flexibility to illuminate aspects of business model innovation and
implementation” (p.47). In addition, DCs have also been criticized for being rather a black box
when it comes to the precise mechanisms at play when an organization is developing its dynamic
capabilities (Abell et al., 2008). In addressing these gaps, work is being carried out in order to
42
Dinuka B. Herath ABM04
determine how to incubate and enhance the development of dynamic capabilities in firms (Felin
et al., 2012). Recent work in dynamic capabilities, especially in highly volatile environments, have
indicated that routinization and structuring of dynamic capabilities alone do not seem to produce
the desired results (Kindstr¨
om et al., 2013). In fact, when environmental pressures are heightened,
non-routine dynamic capabilities are favored (Kruss et al., 2015). This has prompted us to seek
better ways in which we design organizations in order to foster and harness the strategic power
of dynamic capabilities suited for dealing with modern pressures. Therefore, this study models
meso-level dynamic capability development under a varying set of organizational settings in order
to determine the optimal designs suited for effective dynamic capability development. Taking a
Teecian view towards dynamic capacities (Teece et al., 1997) this study specifically models the
sensing, seizing and transforming processes among top management teams in firms.
Through this process, we examine how the development of dynamic capabilities differs in
traditional organizational settings compared to more modern settings. When considering, the or-
ganizational designs in question, the study has a large emphasis on learning the effects of fluid,
disorganized settings with minimal structural and functional boundaries. This interest is justified
by previous work1(Herath et al., 2016,0) which has demonstrated, that environments which are ac-
tively disorganized, tend to be highly effective for individual and team problem-solving. Therefore,
building on previous agent-based models (ABM) of disorganization, this study plans to determine
if disorganization is better suited for dynamic capability development and if so, to what extent.
The prediction capability of ABM and the ability to demonstrate drastically different situations
with relative ease and ability to model self-organization (emergence) at various scales (Miller and
Page, 2007; Bazghandi, 2012) further justifies the choice of the simulation method.
It should be noted that “disorganization” here has a very precise meaning. Abrahamson (2002)
defines it as the “stochastic accumulation of entities within hierarchically ordered complex human
structures” (p, 139). The concept of disorganization as it is articulated here has very important
and inescapable outcomes. This is that disorganization is a phenomenon which is highly resilient
and cannot be completely eradicated regardless of how well organized a firm is (Secchi, 2019).
Therefore, no matter how many instances of disorganization are re-organized, in time the disorga-
nization will re-emerge (Herath et al., 2019). Now, if an organization working under the traditional
1This work has also been featured in ABMO1 (UK) and ABMO2 (Denmark) workshops
43
Dinuka B. Herath ABM04
assumption that “order is good” encounters disorganization, they will be locked in a never-ending
incessant tussle to keep organizing every instance of disorganization they encounter (Abrahamson
and Freedman, 2013). This means that for n instances of disorganization n+1 efforts to reorganize
should be carried out (i.e. 100 instances of disorganization, require 101 efforts to reorganize).
These efforts to reorganize require allocation of resources (financial and manpower) and grows
rapidly as the stakes get higher (multimillion dollar projects will require comparable amounts of
resources for re-organization purposes). Disorganization research, therefore, has been built with
the motivation to break this cycle of organizing-disorganizing by studying methods in which disor-
ganization (and the complexity it brings) can be embraced rather than feared (Alvesson and Spicer,
2012; Herath et al., 2019). This is not only because disorganization is inevitable, but upon careful
scrutiny, some distinct benefits of disorganization have been postulated (Cohen et al., 1972; Abra-
hamson, 2002). Among these benefits, actively leveraging disorganization (setting conditions for
disorganization to occur structurally and functionally within firms) has shown to enhance creativ-
ity, innovation, task performance and decision making (Harford, 2017). All of these elements are
seen as important factors affecting effective dynamic capability development (Teece et al., 1997).
Therefore, in this study, we explore how disorganized organizational designs affect the develop-
ment of dynamic capabilities in firms.
In this model, the primary parameters which will be modeled are organizational structure,
top management teams (with behaviors for sensing, seizing and transforming), and indigenous
& exogenous opportunities each with differing levels of complexity. Each of the parameters and
the rules which govern them will be directly derived from dynamic capability literature and the
literature on disorganization.
Once completed, this simulation will have the ability to impose or dispose of various organi-
zational designs (ranging from highly structured to fully disorganized) at will demonstrating their
effects on the sensing, seizing and transforming processes of dynamic capability development. The
simulation will provide a clear understanding of levels of optimization needed for better dynamic
capability development while also demonstrating levels of structure or disorganization which are
clearly detrimental for dynamic capability development. Furthermore, the simulation will be sus-
ceptible to future improvement thus making the research effort more suitable for further expansion.
Ultimately, the simulation produced through this research will help explore current theory as well
44
Dinuka B. Herath ABM04
as set off complimentary research efforts in the study of dynamic capability development and dis-
organization.
References
Abell, P., Felin, T., and Foss, N. (2008). Building micro-foundations for the routines, capabilities,
and performance links. Managerial and Decision Economics, 29(6):489–502.
Abrahamson, E. (2002). Disorganization theory and disorganizational behavior: Towards an
etiology of messes. Research in Organizational Behavior, 24:139–180.
Abrahamson, E. and Freedman, D. H. (2013). A perfect mess: The hidden benefits of disorder.
Hachette UK.
Alvesson, M. and Spicer, A. (2012). A stupidity-based theory of organizations. Journal of
management studies, 49(7):1194–1220.
Bazghandi, A. (2012). Techniques, advantages and problems of agent based modeling for traffic
simulation. International Journal of Computer Science, 9(1):115.
Cohen, M. D., March, J. G., and Olsen, J. P. (1972). A garbage can model of organizational
choice. Administrative Science Quarterly, 17(1):1–25.
Eisenhardt, K. M. and Martin, J. A. (2000). Dynamic capabilities: what are they? Strategic
Management Journal, pages 1105–1121.
Felin, T., Foss, N. J., Heimeriks, K. H., and Madsen, T. L. (2012). Microfoundations of routines
and capabilities: Individuals, processes, and structure. Journal of Management Studies,
49(8):1351–1374.
Frank, H., G¨
uttel, W., and Kessler, A. (2017). Environmental dynamism, hostility, and dynamic
capabilities in medium-sized enterprises. The International Journal of Entrepreneurship and
Innovation, 18(3):185–194.
Harford, T. (2017). Messy: The power of disorder to transform our lives. Penguin.
Herath, D., Costello, J., and Homberg, F. (2017). Team problem solving and motivation under
disorganization–an agent-based modeling approach. Team Performance Management: An
International Journal, 23(1/2):46–65.
Herath, D., Secchi, D., and Homberg, F. (2016). The effects of disorganization on goals and
problem solving. In Agent-Based Simulation of Organizational Behavior, pages 63–84.
Springer.
Herath, D., Secchi, D., Homberg, F., and Herath, G. B. (2019). Business plasticity through
disorganization.
45
Dinuka B. Herath ABM04
Kindstr¨
om, D., Kowalkowski, C., and Sandberg, E. (2013). Enabling service innovation: A
dynamic capabilities approach. Journal of Business Research, 66(8):1063–1073.
Kruss, G., McGrath, S., Petersen, I.-h., and Gastrow, M. (2015). Higher education and economic
development: The importance of building technological capabilities. International Journal of
Educational Development, 43:22–31.
Miller, J. H. and Page, S. E. (2007). Complex Adaptive Systems. An Introduction to
Computational Models of Social Life. Princeton, NJ: Princeton University Press.
Mouzelis, N. P. (2017). Organizational pathology: Life and death of organizations. Routledge.
Padgett, J. F. and Powell, W. W. (2012). The emergence of organizations and markets. Princeton
University Press.
Peteraf, M., Di Stefano, G., and Verona, G. (2013). The elephant in the room of dynamic
capabilities: Bringing two diverging conversations together. Strategic Management Journal,
34(12):1389–1410.
Reeves, M. and Deimler, M. S. (2009). Strategies for winning in the current and post-recession
environment. Strategy & Leadership, 37(6):10–17.
Reeves, M., Levin, S. A., and Ueda, D. (2016). The biology of corporate survival. Harvard
Business Review, 2016(January-February).
Secchi, D. (2019). What a mess! In Herath, D., editor, Business plasticity through
disorganization. Bingley, England: Emerald.
Secchi, D. and Cowley, S. J. (2018). Modeling organizational cognition: The case of impact
factor. Journal of Artificial Societies and Social Simulation, 21(1).
Suddaby, R. and Greenwood, R. (2009). Methodological issues in researching institutional
change. In The Sage handbook of organizational research methods, pages 177–195. Sage
London.
Teece, D., Peteraf, M., and Leih, S. (2016). Dynamic capabilities and organizational agility: Risk,
uncertainty, and strategy in the innovation economy. California Management Review,
58(4):13–35.
Teece, D. J. (2018). Business models and dynamic capabilities. Long Range Planning,
51(1):40–49.
Teece, D. J., Pisano, G., and Shuen, A. (1997). Dynamic capabilities and strategic management.
Strategic Management Journal, 18(7):509–533.
46
Siavash Farahbakhsh & Alessandro Narduzzo ABM04
MULTILEVEL DYNAMICS OF EMERGING INSTITUTIONS
Siavash Farahbakhsh
sfarahbakhsh@unibz.it
Alessandro Narduzzo
anarduzzo@unibz.it
Faculty of Economics and Management, Free University of Bozen-Bolzano, Italy
Keywords: Institutionalism, organization studies, dynamics, complexity, multilevel
The main subject of this essay is the complexity in institutions. The term complexity has been
widely used in the institutional theory literature and many studies were conducted to study insti-
tutional change processes acknowledging the complexity of institutions. However, there has been
disperse and different approaches on how to diagnose the complexity in institutional change pro-
cesses. All the different approaches toward the complexity share a common fundamental, which
is heterogeneity as early organizational scholars such as March and Simon (1958) argue. Nev-
ertheless, it remains unclear how to apply the complexity lens on the institutional theory. This
impediment motivates this essay to provide an insightful approach toward the complexity tailored
for the institutional theory.
In order to start a conversation on this subject, two concepts should be clarified as reference
points. First, institutions; they are norms, rules, and practices that are shared among a group of ac-
tors — individuals and organizations — (North, 1990; Scott, 2008; Thornton et al., 2012). Second,
fields; they are a shared space among actors and who are living an institutional life (Greenwood
and Suddaby, 2006; Scott, 2008). Therefore, organizations and individuals with different practices
share a common space, which is an institutional field and is, arguably, socially constructed. In the
institutional theory, the field has been mostly associated with industrial sectors to provide clear
boundaries and explain organizational shared cultures (Boxenbaum and Jonsson, 2017).
47
Siavash Farahbakhsh & Alessandro Narduzzo ABM04
With this context, the complexity can be associated with the existing heterogeneity in the in-
stitutional field. The heterogeneity can be in forms of the differences between organizational and
individual characteristics such as power, resources, predominant behaviors, connectedness, etc.
In this line of argumentation, the institutional field can be viewed as a system of relations be-
tween hierarchical interacting and interdependent actors (Giddens, 1984; Coleman, 1990; Scott,
2008). This view is similar to the theory of complex adaptive systems (CASs), which a CAS con-
sists of heterogeneous interacting units with emerging properties (Miller and Page, 2007; Holland,
2014). Markedly, the state of the social system (field) carries a totality, which may influence actors’
decision-making and pressurize them to change (Giddens, 1984). Similarly, the actors can initiate
a change process, which may influence other connected actors and eventually aggregates in a form
of a collective movement toward a system-level change. In this process, thus, two main levels can
be identified; actor and system 1. Giddens (1984) considers the duality of structure within a social
system, where actors are interacting agents with power, and system of relations who can shape and
be shaped by the structure. For him, the structure is the rules and resources that constrains the ac-
tors. Perhaps, one of his main arguments is that structural change can take place both intentionally
and unintentionally by the actors, which highlights the role of the agency.
With a similar approach, more recent scholars, characterize the macrostructure of the social
system (totality) as institutional logics (Frieland and Alfrod, 1991; Thornton et al., 2012). They
define institutional logics as broader organizational and individual templates giving the actors’
behaviors consistent forms (Pache and Santos, 2013). The actors following such logics may pur-
posefully organize works and practices resulting in a field-level change (Lawrence et al., 2009).
Connecting the institutional logic approach to complexity, the heterogeneity in logics of ac-
tors’ behaviors creates a complex institutional environment. In this sense, institutional complexity
scholars associate the complexity with the multiplicity of co-existing institutional logics with de-
grees of incompatibility within a field (Greenwood et al., 2011). The multiple logics, basically,
frame the structure of the system, and change in their domination represents the institutional and
structural change (Thornton et al., 2005).
Despite the recent attempts to investigate institutional change processes given the complexity of
1Perhaps most of the existing literature related to institutional change stemming from institutional theory identify
these two levels using different terms.
48
Siavash Farahbakhsh & Alessandro Narduzzo ABM04
the institutional environment, still further research is required to unfold the dynamics of complexity
in institutions (Greenwood et al., 2011; Vermeulen et al., 2016). Capturing the dynamics of the
change processes requires careful consideration of institutional contexts together with the usage of
proper methods (Suddaby and Greenwood, 2009; Hwang and Colyvas, 2011). To tackle this gap,
this essay is aimed to provide a multilevel framework on how to study the complexity of institutions.
In doing so, this study, first, identifies the institutional levels with respect to the complexity in
institutions and institutional change processes. Second, it illustrates how the institutional levels
interact and how the global state (totality) of the social system is created and changed. Third, it
continues on investigating the dynamics of change at the actor-level knowing the internal com-
plexity of the actors and the external complexity imposed by the institutional environment. Fourth,
while proposing appropriate methods, it provides a holistic view on the field level complexity.
References
Boxenbaum, E. and Jonsson, S. (2017). Isomorphism, diffusion and decoupling: Concept
evolution and theoretical challenges. In The Sage handbook of organizational institutionalism,
volume 2, pages 79–104. Sage.
Coleman, J. S. (1990). Foundations of Social Theory. Harvard University Press.
Frieland, R. and Alfrod, R. (1991). Bringing society back in: Symbols, practices and institutional
contradictions. In Walter W., P. and DiMaggio, P. J., editors, The New Institutionalism in
Organizational Analysis. University of Chicago Press.
Giddens, A. (1984). The constitution of society: Outline of the theory of structuration. University
of California Press.
Greenwood, R., Raynard, M., Kodeih, F., Micelotta, E. R., and Lounsbury, M. (2011).
Institutional complexity and organizational responses. Academy of Management Annals,
5(1):317–371.
Greenwood, R. and Suddaby, R. (2006). Institutional entrepreneurship in mature fields: The big
five accounting firms. Academy of Management Journal, 49(1):27–48.
Holland, J. H. (2014). Complexity: A very short introduction. Oxford Univeristy Press.
Hwang, H. and Colyvas, J. A. (2011). Problematizing actors and institutions in institutional work.
Journal of Management Inquiry, 20(1):62–66.
Lawrence, T. B., Suddaby, R., and Leca, B. (2009). Institutional work: Actors and agency in
institutional studies of organizations. Cambridge University Press.
49
Siavash Farahbakhsh & Alessandro Narduzzo ABM04
March, J. G. and Simon, H. A. (1958). Organizations. Oxford.
Miller, J. H. and Page, S. E. (2007). Complex Adaptive Systems. An Introduction to
Computational Models of Social Life. Princeton, NJ: Princeton University Press.
North, D. C. (1990). Institutions, Institutional Change and Economic Performance. Cambridge
University Press.
Pache, A.-C. and Santos, F. (2013). Inside the hybrid organization: Selective coupling as a
response to competing institutional logics. Academy of Management Journal, 56(4):972–1001.
Scott, W. R. (2008). Institutions and organizations: Ideas and interests. Sage.
Suddaby, R. and Greenwood, R. (2009). Methodological issues in researching institutional
change. In The Sage handbook of organizational research methods, pages 177–195. Sage
London.
Thornton, P. H., Jones, C., and Kury, K. (2005). Institutional logics and institutional change in
organizations: Transformation in accounting, architecture, and publishing. In Transformation
in Cultural Industries, pages 125–170. Emerald Group Publishing Limited.
Thornton, P. H., Ocasio, W., and Lounsbury, M. (2012). The institutional logics perspective: A
new approach to culture, structure, and process. Oxford University Press on Demand.
Vermeulen, P. A., Zietsma, C., Greenwood, R., and Langley, A. (2016). Strategic responses to
institutional complexity. Strategic Organization, 14(4):277–286.
50
Gayanga B. Herath & Davide Secchi ABM04
THE COGNITIVE AFFILIATION TO INSTITUTIONAL NORMS
Gayanga B. Herath
gayanga@sdu.dk
Davide Secchi
secchi@sdu.dk
Research Centre for Computational & Organizational Cognition
University of Southern Denmark, Denmark
In the recent years institutional theory has become a dominant macro perspective on orga-
nizational research (Suddaby, 2010). In this regard, Teraji (2018) states that institutions are in-
separably connected to agents and that institutional and cognitive processes merge to influence
human decision-making (see also Beckert, 2010). Institutional elements come mostly from out-
side the organization and should be intended as the norms, rules, and practices that emerge in a
society/economy (North, 1990; Powell and DiMaggio, 1991). However, they are brought into the
organization by individuals as agents of institutional pressure. The claim is that that any cogni-
tive process, when associated with institutional elements (e.g., norms, rules, and practices), may
change the dynamic of interactions in the workplace. This research is aimed at studying the extent
to which these institutional elements enter and affect cognition in organizations.
In a recent study, Secchi and Cowley (2016, 2018) state that organizational cognition can be
described as social organizing i.e. the way through which individuals discern what is around them
through the means of others. This notion of social organizing has been intended as a “meso” layer
between macro structural elements and the micro elements, constituted by individual agents. As an
intermediate layer, social organizing is mostly concerned with the interactions between individuals
that operate daily routines, procedures, or processes, where the influence of macro organizational
elements (e.g., values, social norms, rules, history) becomes visible. In other words, social organiz-
ing can be intended as the cognitive foundation of organizations (Secchi and Cowley, 2016, 2018).
51
Gayanga B. Herath & Davide Secchi ABM04
In stating this, cognition in organizations is identified as eminently social, because interaction with
others stays at the core of a large number of organizational behaviors. Given the perspective as out-
lined here, one possibility for the understanding of how institutional (macro) elements enter and
affect cognition in organizations is to look at how/when institutions (e.g., societal norms) affect
the way in which individuals operate that is, while performing their job (e.g., making decisions,
solving problems, performing a task, executing a routine).
Accordingly, due to the broad scope of this research question, it is more feasible to focus our at-
tention solely on a single institutional element (either norms, rules, or practices), as it would require
too vast of a research to fit a single article otherwise. As previously mentioned, cognition is bound
to institutional norms in a way that is very much affecting how it turns into organizational action
(especially through social interconnections). Hence, studying specifically institutional norms may
unveil cognitive mechanisms that more directly affect organizational life. Since action is inher-
ently a core component of cognition (Magnani, 2007; Secchi and Bardone, 2009), our grounds for
focusing the attention on norms is judged to be promising.
Consequently, when looking at the current literature on norms (Saam et al., 1999; Neumann,
2008; Hollander and Wu, 2011; Proietti and Franco, 2018; Mercuur et al., 2019), two issues be-
come apparent. Firstly, it came to our attention that most studies on norms do not actually look at
“institutional” norms, but they rather focus on just norms, independent of their source. Secondly,
most studies on norms analyze the effects on individual behavior and/or cognition (e.g., Armitage
and Conner, 2001). It is unclear whether effects might change if one adopts a social organizing
perspective. In other words, how does institutional norms affect social elements that are considered
constitutive of organizational cognition?
In this article we use an agent-based computational simulation model (ABM). The aim of the
model is that of exploring the impact of institutional norms on organizational cognition. The
ABM includes different small-sized organizations – in between 20 to 50 employees –, and one
hierarchical level (only with the intention of keeping things simple at first). Each employee is
equipped with his/her own competencies, skills, and an associated team in which they work. Most
importantly, each employee has a function that allows them to make decisions using information
coming from others (i.e. is docile Simon, 1993; Secchi, 2016), to some extent. Concurrently,
there are also certain tasks that employees perform either individually or as a team. In order to
52
Gayanga B. Herath & Davide Secchi ABM04
apply the external pressures of institutional norms the organization has an adaptation function that
modifies in relation to pressure coming from the environment. This adaptation is dependent on the
type of institutional pressure to which the organization is exposed to and it could be immediate
(for example, when legislation is passed), delayed in time (e.g., in case of workarounds or best
practices), costly but necessary (e.g., professional standards), or it could just be risky but optional
(e.g., an innovation that one may decide to adopt or not). As a result, these decision dynamics have
consequences on the way in which employees perform their jobs/tasks, since each one of them
either individually or as a team understand the institutional norm in a certain way. Therefore, there
is a function to “filter” what happens at the institutional level, and this “filter” is a function of both
individual attitudes and team dynamics (the social organizing as described above). This way, we
intend to study how institutional change affect the way employees “think” about their work and in
turn experiment if there is an impact on performance.
References
Armitage, C. J. and Conner, M. (2001). Efficacy of the theory of planned behaviour: A
meta-analytic review. British Journal of Social Psychology, 40(4):471–499.
Beckert, J. (2010). How do fields change? the interrelations of institutions, networks, and
cognition in the dynamics of markets. Organization Studies, 31(5):605–627.
Hollander, C. D. and Wu, A. S. (2011). The current state of normative agent-based systems.
Journal of Artificial Societies and Social Simulation, 14(2):6.
Magnani, L. (2007). Animal abduction. In Model-based reasoning in science, technology, and
medicine, pages 3–38. Springer.
Mercuur, R., Dignum, V., and Jonker, C. M. (2019). The value of values and norms in social
simulation. Journal of Artificial Societies & Social Simulation, 22(1).
Neumann, M. (2008). Homo socionicus: a case study of simulation models of norms. Journal of
Artificial Societies and Social Simulation, 11(4):6.
North, D. C. (1990). Institutions, Institutional Change and Economic Performance. Cambridge
University Press.
Powell, W. W. and DiMaggio, P. J. (1991). The New Institutionalism in Organizational analysis.
University of Chicago Press.
Proietti, C. and Franco, A. (2018). Social norms and the dominance of low-doers. Journal of
Artificial Societies and Social Simulation, 21(1).
53
Gayanga B. Herath & Davide Secchi ABM04
Saam, N. J., Harrer, A., et al. (1999). Simulating norms, social inequality, and functional change
in artificial societies. Journal of Artificial Societies and Social Simulation, 2(1):2.
Secchi, D. (2016). Boundary conditions for the emergence of docility in organizations:
Agent-based model and simulation. In Agent-Based Simulation of Organizational Behavior,
pages 175–200. Springer.
Secchi, D. and Bardone, E. (2009). Super-docility in organizations: an evolutionary model.
International Journal of Organization Theory & Behavior, 12(3):339–379.
Secchi, D. and Cowley, S. (2016). Organisational cognition: What it is and how it works. In 16th
Annual Conference of the European Academy of ManagementEuropean Academy of
Management. Proceedings. European Academy of Management.
Secchi, D. and Cowley, S. J. (2018). Modeling organizational cognition: The case of impact
factor. Journal of Artificial Societies and Social Simulation, 21(1).
Simon, H. A. (1993). Altruism and economics. The American Economic Review, 83(2):156–161.
Suddaby, R. (2010). Challenges for institutional theory. Journal of Management Inquiry,
19(1):14–20.
Teraji, S. (2018). The Cognitive Basis of Institutions: A Synthesis of Behavioral and Institutional
Economics. Academic Press.
54
Annex: Summary
In this workshop, 13 scientific works were presented. The presenters and the par-
ticipants were from different universities in Austria, Denmark, England, France, Ger-
many, Ireland, Italy, and Macedonia. The presented topics were from different disci-
plines and this made the workshop a multidisciplinary international event. The research
areas addressed themes in management, organization, human resources, institutions,
environmental economics, game theory, urban policy-making, and sociology research.
Among all the presentations of different disciplines, two elements were shared: com-
plexity, and multi-levels.
Furthermore, two keynote speakers, Alessandro Lomi (Institute of Computational
Science, University of Italian Switzerland & University of Exeter) and Martin (Neu-
mann Johannes-Gutenberg-University Mainz, Institute of Sociology), gave talks around
the topics of institutional complexity and emergence. In this respect, Alessandro Lomi’s
presentation title was “Trajectories of Institutional Complexity: Theoretical Mech-
anisms, Empirical Models, and Computational Validation”. He, with a multilevel
approach, gave a speech on how structural similarities in institutions using different
mechanisms emerge. Martin Neumann’s presentation title was “Emergence Reconsid-
ered: Computation, Computability and the Limits of Knowledge”. He spoke about
different views in conceptualizing the emergence as a topic and he challenged some of
the recent linear approaches toward diagnosing the emergence.
This workshop was sponsored by the Faculty of Economics and Management–
Free University of Bozen-Bolzano (unibz), the European Academy of Management–
Organisational Behaviour (EURAM-OB), and the Research Centre for Computational
& Organisational Cognition (CORG)–University of Southern Denmark (SDU).
Thanks to the sponsors, the workshop’s registration was free, the meals and the
social dinner for all the participants were covered, and the lodging of the two keynote
speakers was paid.
Best regards to the sponsors, the scientific committee, the participants, the presen-
ters, and the administrative staff of unibz.
Siavash Farahbakhsh
Alessandro Narduzzo
Davide Secchi
55
