'HE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986. 851
DRAFT
A SYSTEM-BEHAVIORAL METHODOLOGY
FOR BUSINESS POLICY RESEARCH*
I. Hall
Department of Business Administration
University of Manitoba
Winnipeg, MB. R3T 2N2
Abstract
A process modeling approach is used to describe three major
elements of policy making, namely, the workings of the corporate
_System of a firm, its representation in Managers' Cause Maps, and
the Policy Formation Procedures used by the policy making elite.
System Dynamics provides an expert system to aid the
construction of the Corporate System. Cause Map and Behavioral
Decision Making theory, on the other hand, provides the
artificial intelligence (modeling the collective decision
making behavior of a senior management) that drives the Corporate
System. Potential applications of the methodology are put
forward
* The author wishes to acknowledge the incisive comments and helpful
advice of Dr. Malcolm C. Mmro in editing parts of this paper.
The study was supported by a Strategic Grant from the Social
Sciences and Humanities Research Council of Canada.
ist Draft, May, 1986
852 THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986.
Introduction
The business policy research to be described is based on process
modeling.
According to Mohr (1982):
Process models are little used in organization theory. When
they are used, they are often underdeveloped. There is a
process but to omit the forces that drive the movement from
one stage to another. The latter, however, are essential .
Description (of a process) as theory in the more mature
sciences has targeted the form, matter, and motion of
phenomena, but the kind of description that would seem to
have the greatest potential in social science is description
of processes—how things are done by people and groups.
The processes to be modeled concern the interaction of the policy making
system (describing the beliefs in causality and decision making
processes of the dominant policy elite of the firm) amd the corporate
System (describing the actual operations and activities of a firm). An
analogy canbe drawn with geological exploration. From scraps of
information obtained from test borings and accepted geological theory,
the geologist draws a map of the subterranean configurations of old
river beds, and etc., that are thousands of feet below the surface.
“these maps are used by exploration experts to decide on where to drill
to maximize the chance of, say, finding oil. If, subsequently, oil is
discovered from exploratory drilling, more is learned about the exact
nature of the subterrain, leading to a revision of exploration strategy,
and soon. Similarly, from scraps of information about operations,
IME 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY, SEVILLA, OCTOBER, 1986. 853
customers, suppliers, competitors, government regulations or impending
legislation, senior managers piece together "Cognitive Maps"'--individual
mental maps of their policy domains—-which they then use in their policy
determinations. The results from implementing these decisions advance
the understanding of the firm's operations and environment by the
managers, leading to a revision of policies and so on.
Axelrod (1976) has found the decisions by policy makers to be extremely
rational within the structure of their "cognitive maps". Unfortunately,
these "cognitive maps" often contain gross simplifications of reality
because policy makers have more causal assertions than they can handle
mentally. The human mind, also, seems incapable of handling causal
feedback relations that confound the individual's map.
Group behavior can cause even greater distortions to the accepted "map
of causality"—-group cause maps of the policy domain—on which the
group's decisions are based (Hall, 1981). Furthermore, changes in
customer's tastes, the state of the economy, the nature of the
competition, and etc., add a dynamic element requiring the maps to be
continually updated. But there is a growing concern that senior
executives are making important decisions based on out-of-date maps that
are gross simplifications of reality and deficient of feedback loops of
causality that will create unforeseen deletarious policy side effects
There seems a general lack of formal methods for handling this problem
854 THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986,
in private firms that is similar to the military's 'C3I' approach
(Commmication, Control, Command and Intelligence). A good map is like
an insurance policy: when the environment is benign, almost any policy
based on the most crude assumptions will be successful, but when the
environment “is hostile, then survival may depend on having a good map of
the policy terrain. ‘The formal map building methodology being developed
by the author is based on cause mapping and group behavior theory and
computer simlation. It is envisaged that it would complement and not
replace existing formal methods such as market analysis, economic
forecasting and business policy/strategy fommllation. It -would focus
primarily on (a) helping the organization cope with greater complexity
in its domain, (b) finding policies for stability in the face of
destabilizing events, and (c) training managers by providing a rich map
of their domains that they can use to exploit opportunities and defend
against threats as they emerge in 'real' time.
Methodology
To accomplish the task, the author has developed a framework of analysis
(see Figure 1) with two distinct submethodologies: (1) to model the
corporate system of a firm (e.g., production rates, sales and cash
flows), and (2) to model the management policy making system (e.g., the
“cause map" and organizational behavior processes--the driving forces
used in policy determinations; such as what proportion of available
funds/cash flows to devote to different activities).
THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986, 855
1
For modeling the corporate system, Management System Dynamics (Coyle,
1977a; Forrester, 1968; Roberts et al., 1983; Richardson and Pugh, 1981;
Iyneis, 1980) and its associated computer simulation languages Dynamo
(Pugh, 1983; Pugh and Paton, 1986) provide a ready-made ‘expert
system'. The steps involved in building such a corporate system model
are:
1. From interviews with people in the firm and from industry statistics
and company operating reports, an influence diagram, depicting the
operations of a company, is put together using the directed digraph
method (Axelrod, 1976; Hall, 1978). Note the sign of correlation (+
or -) of each link representing causality in the system.
2. A system flow diagram is developed from the influence diagram to
facilitate programming. This employs control engineering symbols to
categorize the conserved subsystems (e.g., inventories and cash
balances) and the rates that control the flows in the subsystems
The interconnecting webs of information and decision protocols that
determine the activities/rates (the driving forces) in the conserved
subsystems, are added.
3. A computer system simulation model is programmed from the system
flow diagram using the system simulation language of DYNAMO. This
language uses parallel processing (i.e., the order of the computer
cards or records is unimportant) and is tailor made to assist this
kind of study (e.g., documenting, dimensional checking, loop
856 THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986.
analysis and comparative plotting facilities are built in).
4. Computer runs of the model, together with its flow diagrams and
documented assumptions, are shown to the potential users who
criticize the realiem of the results. Changes are made to the model
in light of these criticisms (a trivial task with parallel
processing, but often, a major operation without this facility).
The process is repeated until the opinion of the majority is
expressed that the model represents reality for the purpose at
hand. Finally, the validation methods appropriate toa System
Dynamics model (Bell and Senge, 1980; Forrester and Senge, 1980) can
be applied.
Cause Mapping and Policy Making
Whereas the System Dynamics method helps one to capture the essence of
the actual workings of the corporate system of a company, the "Cause
Mapping" method assists one in modeling the management's collective view
of how the corporate system works. The characteristics of the "cause
maps" of policy makers have been established by Axelrod (1976) (in
particular the ways they differ from the very complex real system they
seek to represent). The author has developed an Artificial Intelligence
‘process model! of management policy making*based on the structure of
the accounting/budgeting framework used by a company, together with
macro organizational aspects derived from the structure of the firm (its
major departments and divisions) and the driving forces of policy
THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986. 857
formation (e.g., 'equivocality reduction’ and group status enhancement;
Hall, 1984). The author has been able to demonstrate that with such a
model one can predict which policy will be adopted and what the
organization will learn from its implementation.
In the next stage of this research, it is proposed to program such a
policy and organizational learning model and connect it to the corporate
system simlation model. Such a complete system model of a firm has not
been attempted before (to the author's knowledge). With the development
of micro-camputer versions of suitable languages, the application of
this technique in situ to firms of all sizes (not just those with access
to large main frame computers) is now possible.
The steps involved in modeling the policy making and controlling system
are:
1. Construct a Management Cause Map by restructuring the influence
diagram using Axelrod's (1976) scheme (i.e., recategorize the
concept variables into Policy Variables, Performance Variables and
Intervening Variables for each major department or division in the
firm and map their interrelations).
2. Conduct an analysis of the Management Cause Map as follows: Trace
all paths from primary policy variables to departmental and overall
performance variables or goals. Sum the ‘negative' signs of
correlation of the links along each path to find the
path-correlation between policy and performance variables (e.g., an
858.
THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986.
uneven number of negative link-correlations means a ‘negative!
overall path-correlation, otherwise the path-correlation is
‘positive'). For each set of paths from a particular policy toa
particular performance variable, note whether the net
policy-performance correlation is indeterminant (N.B., indeterminacy
exists when two or more paths have opposing path-correlations, and
it infers that the effect of the policy on the performance variable
is problematic). Where indeterminacy exists, note the correlation
of the shortest path (with fewest links). The policy suggested by
the shortest path is a strong candidate for adoption by policy
makers to force the issue. If recursive paths (feedback loops) are
found, note their path-correlations (polarity). If the polarity is
negative, this suggests a tendency for self correction should any
variable in the loop change its value. If positive, this suggests
the opposite--any change will be amplified. Positive feedback loops
are sources of growth, decay and potential uncontrollability (i.e.,
the system will tend to have a life of its’ own) and demand special
attention, however, as noted before, feedback loops and their
associated side effects are not usually apparent to policy making
4
groups.
Derive the policies that the management are most likely to adopt
from a set of standard policy making procedures supplied in Table 1.
These hypothesized procedures—described in detail in Hall
(1981)—-draw heavily on the seminal work of Cyert and March (1963),
'HE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986. 859
Lindblom (1968), Axelrod (1976) and other decision school theorists
who have observed the way managers and groups of managers go about
their decision making work. The hypothesized procedures are evoked
by the socio-political driving forces associated with subunit
(departmental /divisional) status enhancement (or defense against
loss of status) (Pettigrew, 1973; Mumford and Pettigrew, 1985;
Salancik and Pfeffer, 1977) and the social-psychological driving
force associated with the reduction of equivocality threatening
confusion and chaos (Weick, 1969; Jung, 1969). The procedures invoke
a search of the Management Cause Map for remedial policies.
Subsequent learning’ from the success (or lack thereof) in
implementing the policies leads to an updating of the Map, that, in
turn will effect the subsequent policies evoked, and so on. The
scheme models the continuous process of learning from experience.
For an example of this kind of analysis, see Hall (1984).
4. The Policies that it is predicted will be used to dispel the
symptoms of problems can now be compared with the actual management
policies chosen in similar circumstances. This will provide a rough
check of the credance of the model. The policies so chosen can be
used to drive the Corporate System simulation model. If the results
are counter-intuitive, further investigation can be undertaken to
find the cause.
Table 1 about here
860 THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986.
Potential Developments and Applications
Corporate system models are used for a different purpose (namely to aid
the management on a journey of discovery into the policy areas of the
organization) and hence, complement other techniques that are oriented
more to prediction, forecasting and strategy analysis. Same of the
potential developments and applications of this technique are as
follows:
Intuitive-Logical Policy Analysis
The insights generated by experimenting with a corporate system
simulation model lead to the identification of the factors causing
unsatisfactory behavior and to the derivation of policies logically
(albeit intuitively) to prevent the deterioration of the system's
perfomance (see for example, Nord, 1963; Packer, 1964; Roberts et.
al., 1968; Hall, 1976). This is one of the more conventional uses of
corporate system simlation. For example, with the aid of a corporate
system model, the change in fortunes in a magazine publishing company
was explained and a policy for survival devised (Hall, 1976). It became
evident that information critical to the survival of a magazine (such as
the turnover of regular readers) was not being supplied by the company's
information system or recognized in policy making. The method could be
used for Critical Success Factor analysis (ll and Munro, 1986) leading
to the formulation of more sensible policies.
THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986. 861
Stability Analysis
The feedback loop structure of a system model will determine its dynamic
stability (or lack thereof): how the system will react to external
disturbances: and its own controls. Coyle (1977a: Ch. 7 and 8) presents
a loop analysis method based on tabulating loop polarity, gain, number
of pure integrations and length of exponential delays. A better
understanding of the causes of instability (e.g., combinations of phase
shift due to delays or integrations, and loop gain) can be derived,
leading to prescriptions (e.g., changes in gain or delays, or "short
circuiting" offending loops and their implications for policy change) to
. remedy the situation.
Day (1982) has shown that the simple feedback structures embodied in
self-organizing systems, such as firms and their markets, when certain
critical values of parameters are approached, can produce wild
fluctuations and chaotic results. Similarly, the unusual and sudden
changes in the basic behavior of a positive feedback loop (also found in
self-organizing systems such as firms and their markets) has been
demonstrated by Rahn (1982). Although the study of chaos is relatively
new, it does not take mich imagination to perceive the potential use of
system modeling to warn organizations when their markets are becoming
chaotic or their own internal policies are leading them into a ‘zone of
chaos.'
These studies of chaos suggest that organizations can suddenly encounter
862 THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986.
periods of great turbulence for reasons that are difficult to
ascertain, The consequent internal political activities set in motion
can compound the situation by favouring the conditions for internecine
warfare and vaccilating strategies from which the organization may not
recover.
Again it would seem that the System Dynamics methodology could come to
the rescue here, since it is a particularly apt technique for modeling
complex interactive feedback systems and analyzing them for stability in
the face of uncontrollable external variability. From such a study it
is usually possible to demonstrate the effects on the system of, say, a
proposed compromise agreement, and devise policies for the organization
that are "robust"--i.e., reduce the destabilizing effects of the
compromise on the system (Sharp, 1977). It offers a way for putting
control back into the system.
Clearing House for Values
Organizations tend to be made up of individuals or groups vying with
each other for status and power over resources (Pettigrew, 1973). The
competition can become very intense and potentially damaging to the
organization as a whole. System models can be used to demonstrate the
effect of unilateral actions by any individual or group on the others.
It can provide a means for clarifying issues and a stimulus to searching
for creative policies that will simltaneously satisfy several
contending forces. For example, Coyle (1977b) was able to show with the
THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986, 863
aid of a simple corporate system model of an international mining
company that the natural policies being pursued by both the parent
company and its more independent subsidiaries were mutually harmful.
Policies simultaneously beneficial to both were generated by the
analysis.
Without a well informed board of directors, who is to supervise the
management? Roos and Hall (1980) using influence diagramming, have
shown that a viable role for the evaluator of an organization is to
uncover the power strategies used by managers to acquire excess
resources.
Crisis Simulation
The Limits to Growth (Forrester, 1971; Meadows et. al., 1972)
simulations of the collapse of the world are supposed to have had a
profound effect on the thinking of statesmen. It has been suggested
that a similar simulation of the collapse of a firm could have the same
effect on its management (Hall, 1979). The use of such a model could
facilitate the changes in values, attitudes and orientation associated
with a crisis (Turner, 1976) before, rather than after, the onset of the
crisis.
Weick (1969) has suggested that the selection and retention processes of
organizational adaption are driven by the need to reduce equivocality
and not necessarily to optimize per se. Inacrisis situation, the
procedures for reducing equivocality tend to become political--the
864 — THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986,
dominant group prescribes a policy most in line with its interests
(Hall, 1981; Pettigrew, 1973).
The lack of attention to both complexity and novel alternatives in the
deliberations of an organization during a threat have been noted by
Staw, Sanderlands and Dutton (1981):
...search for information may change as a threat develops,
from an initial flurry when a threat is recognized, to a low
point as channels become overloaded, and on to a second peak
as decisions are confirmed or implemented. However,
throughout these changes in information search, the number
of genuinely new or novel alternatives considered by the
organization may still be relatively low. Even when search
is increased, information received is likely to be similar
to that of the past, due to heavy reliance on standard
operating procedures, previous ways of understanding, or
commmnication that is low in camplexity...(p. 513).
Alternatively, crude and emotive arguments based on simplistic
assumptions hold sway and complexity and uncertainty are assumed away
(Steinbrunner, 1974). The chance of selecting an inappropriate policy is
obviously increased by such primitive group processes. As Pettigrew
(1974) puts it:
For organizations as for groups and individuals, extreme
situations provide the opportunity for learning which will
only be taken up if the participants have the capacity to
unravel what has been experienced from what has been
learned, and the motivation to do the after-the-fact
reflection and analysis which will disentangle the noise of
the experience from the message of learning (p. 7).
Corporate System Modeling could be invaluable in a crisis situation
(particularly when survival is at stake) by reducing the equivocality
surrounding the problem (an essential step in coping) yet aiding in the
THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986. 865
construction of a rich map of the policy terrain with which to search
for a safe passage. It becomes a part of the organizational process for
learning to cope with an uncertain and threatening situation. Using a
system model in this way as an organizational intervention tool would
seem to provide a fruitful field for future action research and a
potentially important extension of analytical methods in organizational
and policy issues. Hall and Menzies (1983) have reported on the
successful application of such a model in saving a distinguished sports
club from collapsing in its centennial year. Such an analysis leads
naturally to the identification of the vunerabilities and the associated
information that is critical to the survival of the organization.
Efficient Policy Generation
Nelson and Krisberg (1974) have shown that a search algorithm, such as
Bandler's (1971) Razor Search Program, can be used in conjunction with a
system simulation model to generate more complex policies for managing
the system. The policies so generated exhibit not only more policy
variables in tandem but also in sequence (e.g., adopt policy A for so
many months and then switch to policy B). The Razor Search Program has
the capability to search for an optimum value of same objective function
in the kind of discontinuous solution space associated with system
models. Setting the weight to the criteria of an objective function
does, however, pose a problem since managers can rarely agree on the
relative merits of achieving various goals. Experimenting with
different goals weightings can help the management team clarify their
866 THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986.
collective objectives (Kelocharju, 1982). A systematic procedure for
model simplification by removing links in the model that do not
significantly alter its behavior has also been devised (Keloharju and
Iuostarinen, 1982; Keloharju, 1983).
Policy Training Aid
A Corporate System Model can be turned into a game using GAMING-DYNAMO
(Pugh-Roberts, 1984) or reprogrammed in FORTRAN by using the translation
facilities of DYNAMO II/F to produce a FORTRAN module to which
subroutines controlling the game and generating reports can be added
(see, for an example, Hall, 1974; Hall and Iai, 1984). The participants
in the game make decisions or set policies and receive feedback of the
results. In the process, they can gain a better understanding df the
sensitivity (or lack thereof) of the corporate system to changes in
policies, and learn to incorporate more complexity into their policy
determinations. The game can also be used to demonstrate or study
decision making behavior and organizational learning such as the
inability to perceive recursive paths of causality.
Summary
This paper has attempted to develop a methodology for business policy
research based on the notion of Process modeling. The workings of the
Corporate System (turning inputs into outputs), Management Cause Maps
(the management's collective representation of how the Corporate System
works) and the Policy Formation procedures (whereby problems are
THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986. 867
recognized and the Cause Maps searched for solutions) are described in
process form. It is suggested that System Dynamics be used as an expert
system to aid the construction of the Corporate System. Cause Mapping
and Behavioral Decision Making theory, on the other hand, can be used to
provide the artificial intelligence to model of the. way a group of
managers might seek to understand and control the Corporate System. Such
a model, it is suggested, can provide the driving force of action,
learning and adaption in a particular corporate environment. Lastly,
potential applications of the methodology have been put forward.
NOTES
1. A corporate system is defined here as an organizational entity that
is capable of being managed in such.a way that, at the very least,
it is self-regulatory.
2, Standard modules of typical configurations are available to assist
the construction of the System Flow Diagram (see Wolstenholme and
Coyle, 1983).
3. A policy is defined here as an important decision resulting from
group processes within the organization and not imposed from above
or without (as for example, a president or receiver empowered to
make sweeping changes unilaterally). It may or may not be tied toa
strategy or long-term master plan for the organization. In fact the
natural policy making process (e.g., raising prices to offset
short-run profit shortfalls) may systematically subvert a strategy
868
THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY, SEVILLA, OCTOBER, 1986.
(e.g., to produce a low priced product for mass sale). This
interplay of natural policy process and strategy raises same
interesting questions for business policy research.
To examine this phenomenon, students in classes studying decision
making participated in a magazine publishing game (Hall, 1974; Hall
and Lai, 1984). Working in teams and assuming the roles of managers
of the departments of a magazine publishing company, the
participants (over 200) made decisions and received feedback from a
computer simulation model that simulated 20 years of operations
spread over a 10-week period. After instruction in cause mapping
(Axelrod 1973, Hall 1978), they were asked to draw the perceived
relationships in the computer model. Few were able to discern any
of the six feedback loops built into the model. Nor could the
participants interpret the meaning of such loops in causality when
made avare of their presence. This is consistent with the
observations of Axelrod (1976).
Subsequent correspondence and interviews with the presidents of five
leading national and international magazines gave the impression
that the availability of this strategy was not generally
appreciated. All the presidents expressed the belief that the
number of editorial pages was not directly related to the amount of
advertising, although the plots of editorial versus advertising
pages (using data they furnished) cast serious doubt on this
statement. Most used separate companies to handle their
THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986. 869
subscription sales and had little idea about the churning effect of
subscribers described by such statistics as the percentages of trial
and regular subscribers renewing their subscriptions. Yet small
changes in these percentages can have dramatic effects on the long
term success and viability of the magazines. It is perhaps not
surprising that most have since gone out of business.
A project is underway to build a simulation version of the policy
making processes of an organization and use it to drive a corporate
system model. It is intended to examine the budget planning
process, for example, in detail—e.g., how many cycles through the
budget were required before an acceptable decision was found, was it
necessary for daminant coalition to force a decision, what did the
organization leam from the results and how did this effect
subsequent decisions?
REFERENCES:
Axelrod, R., The Structure of Decision: The Cognitive Maps of Political
Elites, Princeton University Press, Princeton, N.J., 1976.
Bandler, J.W., "The Razor Search Program", IEEE Transactions on
Microwave Theory and Techniques, July, 1971, p. 667.
Bell, J.A. and Senge, Peter, M. "Method for Enhancing Refutability in
System Dynamics Modelling" in A.A. Legasto, Jr., J.W. Forrester and
J.M. Iyneis (eds.), TIMS Studies in the Management Sciences:
System Dynamics, Vol. 14, North-Holland, New York, 1980,
Pp.
870 THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986.
67-73.
Coyle, R.G., Management System Dynamics, John Wiley, London, 1977a.
+ ‘Modeling the Future of Mining Groups", Dynamica, Vol. 3,
1977b, pp. 132-171.
Cyert, Richard M. and March, James G. A Behavioral Theory of the Firm,
Prentice-Hall, Englewood Cliffs, N.J., 1963.
Day, Richard H., "Complex Behaviour in System Dynamic Models", Dynamica,
Vol. 8, 1982, pp. 82-89.
Forrester, Jay W., Principles of Systems, Wright-Allen Press, Cambridge,
Mass., 1968.
, World Dynamics, Wright-Allen Press, Cambridge, Mass.,
1971.
, and Senge, Peter M., "Tests for Building Confidence in
System Dynamics Models", in A.A. Legasto, Jr., J.W. Forrester and
J.M. Iyneis (eds.) TIMS Studies in the Management Science System
Dynamics, Vol. 14, North-Holland, New York, 1980, pp. 209-228.
Hall, Roger I., "Managing a Publishing Company: A Decision Making Game"
in Carney 1T., Constructing Instructional . Simulation Games,
University of Manitoba, Winnipeg, Manitoba, 1974, pp. 22-29.
, “A System Pathology of an Organization: The Rise and Fall
THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986, 871
of the Old Saturday Evening Post", Admin. Sci. Quart., Vol. 21,
1976, pp. 185-211.
, "Simple Techniques for Constructing Explanatory Models of
Complex Systems for Policy Analysis", Dynamica, Vol. 3, 1978, pp.
101-144.
, "Simulating a Crisis", in Highland, Spiegel and Shannon
(eds.), Winter Simulation Conference, IEEE, New York, 1979, pp.
195-204.
, “Decision Making in a Complex Organization", in England,
G.W., A. Neghandi and B. Wilpert (eds.) The Rimctioning of Complex
Organizations, Oelgeschlager, Gunn and Hain, Cambridge, Mass.,
1981, Chapter 5, pp. 111-144.
, “The Natural Logic of Management Policy Making: Its
Implications for the Survival of an Organization", Management
Science, Vol. 30, 1984, pp. 905-927.
, and Menzies, William. "A Corporate System Model of a
Sports Club: Using Simulation as an Aid to Policy Making ina
Crisis", Management Science,"Vol. 29, 1983, pp. 52-64.
, and Lai, Daniel. Magazine Publishing Game and Research
Tool Operation Manual, Working Document, Faculty of Administrative
Studies, University of Manitoba, Winnipeg, Canada, 1984.
872 THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986.
, and Malcolm Munro, "Corporate System Modeling as an Aid to
Defining Critical Success Factors." Working Paper, Faculty of
Management, University of Alberta, 1986.
Jung, R. "Systems of Orientation", in D.M. Kochen (ed.) Some Problems in
Information Science, Scarecrow Press, New York, 1965.
Kiloharju, Raimo. "System Dynamics or Super Dynamics", Dynamica, Vol. 4,
1977, pp. 26-43.
, Relativity Dynamics, The Helsinki School of Economics,
Helsinki, Finland, 1983.
, and Inostarinen, Ari. "Achieving Structural Sensitivity by
Automatic Simplification", Proc. 7th Internat. Conf. on System
Dynamics, University Libre de Bruxelles, Il, 1982.
Lyneis, James M., Corporate Planning and Policy Design: A System
Dynamics Approach. Cambridge, Mass.: MIT Press, 1980.
Meadows, Donnella, H., Meadows, Dennis L., Randers, Jorgen and Behrens,
William W. The Limits to Growth, Universe Books, New York, 1972.
Memford, Enid and Pettigrew, Andrew M. Implementing Strategic Decisions,
Longman, London, 1975.
Nelson, Carl W. and Krisberge, Harold M. "A Search Procedure for Policy
Oriented Simulations: Applications to Urban Dynamics", Management
Science, Vol. 20, 1974, pp. 1164-1174.
THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986. 873
Nord, Ole C., Growth of a New Product, M.I.T. Press, Cambridge, Mass.,
1963.
Packer, David W., Resource Acquisition and Corporate Growth, M.I.T.
Press, Cambridge, Mass., 1964.
Pettigrew, Andrew M., The Politics of Organizational Decision Making,
Tavistock Press, London, 1973.
, “Learning from Extreme Situations", International
Institute for Advanced Studies in Management, Brussels, Working
Paper 74-33, 1974.
Pugh, Alexander, L. III, Dynamo II User's Manual, M.I.T. Press,
Cambridge, Mass., 1970.
Pugh, Alexander L. III and Dawna -L. Paton. Professional ‘Dynamot Users!
Guide and Reference Manual. Cambridge, Mass.: Pugh-Roberts
Associates Inc., Five Lee Street 02139, 1986.
Rahn, R. Joel. "Unity-Gain Positives Feedback Systems", Dynamica, Vol.
8, 1982, pp. 96-104.
Richardson, George P. and Alexander L. Pugh III, Introduction to System
Dynamics Modeling with DYNAMO. Cambridge, Mass.: MIT Press, 1981.
Roberts, Nancy, D.F. Anderson, R.M. Deal, M.S. Garet and W.A. Shaffer,
Introduction to Computer Simulation: The System Dynamics Approach.
Addison-Wesley, Reading, MA., 1983...
874 — THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986.
Roos, Leslie, L. and Hall, Roger I. "Influence Diagrams and
Organizational Power", Admin. Sci. Quart., Vol. 25,. 1980, pp.
57-71.
Salancik, G.R. and J. Pfeffer. "Who Gets Power — and How They Hold on
to It: A Strategic-Contingency Model of Power", Organizational
Dynamics, Winter, 1977, pp., 3-21. ‘
Sharp, J.A. "System Dynamics Applications in Industrial and Other
Systems", J. Oper. Res. Soc., Vol. 28, No. 3, 1977, pp. 489-504.
Schrieber, A.N., Corporate Simulation Models, Univ. of Washington,
Seattle, Wash., 1970.
Staw, Barry, M., Sandelands, Lance E., and Dutton, Jane £E.
"Threat-Rigidity Effects in Organizational Behavior: A Multilevel
Analysis", Admin. Sci. Quart., Vo. 26, 1981, pp. 501-524.
Steinbrunner, J.D., The Cybernetic Theory of Decision, Princeton
University Press, Princeton, N.J., 1974.
Turner, Barry A. "The Organizational and Interorganizational Development
of D: ", Admin. Sci. Quart., Vol. 21, 1976, pp. 378-397.
Weick, Karl E. The Social Psychology of Organizing, Addison-Wesley,
Reading, Mass., 1969.
Winter, Sidney G. "Satisficing, Selection and Innovating Remnant",
Institute of Public Policy Studies, University of Michigan,
THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986. 875
discussion paper no. 18, 1970.
Wolstenholme, E.F. and R.G. Coyle. "The Development of System Dynamics
as a Methodology for System Description and Qualitative Analysis",
J. Opl. Res. Soc., No. 34, 1983, pp. 569-581.
Uncontrolled Environmental \—
Disturbances
Policy Changes
Budget Plan Decisions
and
Controlling Actions
S| FRAMEWORK FOR MODELING POLICY MAKING "SYSTEM 7
\
N\
N\
\
FIGURE 1 - Framework of Analysis
using System Dynamics as an Expert System based on
influence diagramming techniques tA
system simulation modeling 7
statistical inference
computer simulation /
CORPORATE SYSTEM
Performance
Flows and accumulations of:
ces
customers/clients, products/services,
expenditures and revenues.
Periodic Reports
and Reviews
POLICY MAKING §& CONTROLLING
SYSTEM
The processes of: Infusion of
forming goal priorities, mapping the + Hew ideas
environment, fomulating plans
recognizing problems, choosing and Comparative
implementing policies. 7 Industry
Performance
Indices
os pe me a
using an Artificial Intelligence approach based on: 1
evolutionary learning theory
“cognitive mappin,
human learning, Judgement and intuition in
politics of organizational decision making
the organization's culture, beliefs and
orientation /
behavioral theory of the firm ,
918
"986} ‘H3GOL90 ‘VTIIASS ‘ALZIOOS SOIWVNIG WALSAS SHL 40 3ONSUS4NOO TWNOLLVNYSLNI 986+ SHI
‘THE 1986 INTERNATIONAL CONFERENCE OF THE SYSTEM DINAMICS SOCIETY. SEVILLA, OCTOBER, 1986.
TABLE 1 ~ Standard Policy Making Procedures based on Organizational Behavior Assumptions
Detving Fore
To FORMULATE BUDGET PLANS
Reduce the uacercataty of producing
Unsatietactory (nancial ceaulte at
year ends
Reduce sabiguicy sbout the asnuaptions
to be used for computing aach Line of
the budget.
‘TO RECOGNIZE PROBLENS
Reduce dleagreenent in recoga! ring
problem
Reduce uncertainty and dtexgre
Aiagnosing the cause of probless
jen Sn
TO CHOOSE REMEDIAL POLICIES
agutnee threata to it.
Mintatze incer-group conflict
Bring closure to the policy process when
Kacer-group conflict cannot be avotd
Bring closure to the policy process vhea
choosing aaong several competing pollcter-
lring cloaure to the polley process then
Indeterminancy exiscer
ring closure to che policy process when
indecerainaney existe under steersful
condietons.
Avotd uncertainty concerning the reaction
Of the organization's environment to 1ts
Bring order to planning and coordinating
‘the proposals of subuntee
TO HARE THE PLAN WORK
Reduce the uncertainty of noc aeering
budgeted target
Reduce che uncerteinty associated with
lover achieving eargete:
Reduce the uncertelaty of oegstive
reactions © policier that aaaipulate
slack,
The procedures evoked by the driving
foxes
Construct « budget plan using the structure
fof the financial accounts as the basis of
che plan.
Estimate each budget tte ust
Srtabliened relationships vetelned fro part
experience. If these relatloaehipe ate on
Glear, use staple forecaste based on extra”
OLS Ene cpeab seaulees
Coupare the cesulte computed by the budget
igh the fatal eation'e expectations Yor tach
Goal. Identity shortfalls and surpluse
Schleving the goales Tete define the erga”
taatton's problems.
Use standard Finatictal procedures to compute
operating ratios and grovth rates of itene in
the proposed budget. Couper vith previous
year's Figures to identity the eymptone of
the problea.
Esch eubuntt evokes preferred policies to
Gtapel the ayaptous of the dlasatlefied goals
ring tee retained wap of caueality-
A search fe made for acceptable policies thet
donot violate eubuatt goals.
Selece policies het meet che goals of the
politically poverfol subunits at che expense
OF the politically weak
‘Choose the policy moat frequently used be-
fore:
Evoke policies thac actend co the éla~
setiatied goals ope at time
Choose the policy based oo the moat staple
and direct argunent offering iamediate cangt~
ble results
Nake only saall sncrenentel changer to the
ahorefalle and surpluses
Repeat the process until all prob
solved of 90 solution can be found
LC che target ta detng under subscribed and
slack reaourees exiat, fovoke « slack reduc-
Elon program (e-p-, cut production coats)
1s oe arget i over subecribed, invoke 4
frearch and development) -
Control tncersal variables only (tes, 40 not
change variablea that affect the envizonuent,
such ae prices if at all posnibie):
877
